HTML 77.2%
TypeScript 10.5%
Python 9.6%
JavaScript 2.5%
1<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"2 xmlns:content="http://purl.org/rss/1.0/modules/content/"3 xmlns:wfw="http://wellformedweb.org/CommentAPI/"4 xmlns:dc="http://purl.org/dc/elements/1.1/"5 xmlns:atom="http://www.w3.org/2005/Atom"6 xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"7 xmlns:slash="http://purl.org/rss/1.0/modules/slash/"8 xmlns:media="http://search.yahoo.com/mrss/">910<channel>11 <title>NVIDIA Blog</title>12 <atom:link href="https://blogs.nvidia.com/feed/" rel="self" type="application/rss+xml" />13 <link>https://blogs.nvidia.com/</link>14 <description></description>15 <lastBuildDate>Fri, 11 Sep 2026 16:54:30 +0000</lastBuildDate>16 <language>en-US</language>17 <sy:updatePeriod>18 hourly </sy:updatePeriod>19 <sy:updateFrequency>20 1 </sy:updateFrequency>21 <generator>https://wordpress.org/?v=7.0.4</generator>22 <item>23 <title>Skild AI Taps NVIDIA Physical AI to Teach Robots New Tasks From a Single Video</title>24 <link>https://blogs.nvidia.com/blog/skild-ai-s1-physical-ai/</link>25 26 <dc:creator><![CDATA[Sasa Docca]]></dc:creator>27 <pubDate>Thu, 10 Sep 2026 16:30:35 +0000</pubDate>28 <category><![CDATA[Robotics]]></category>29 <category><![CDATA[Artificial Intelligence]]></category>30 <category><![CDATA[Cosmos]]></category>31 <category><![CDATA[Customer Stories]]></category>32 <category><![CDATA[Industrial and Manufacturing]]></category>33 <category><![CDATA[Isaac]]></category>34 <category><![CDATA[NVIDIA Blackwell]]></category>35 <category><![CDATA[Omniverse]]></category>36 <category><![CDATA[Physical AI]]></category>37 <category><![CDATA[Simulation and Design]]></category>38 <category><![CDATA[Synthetic Data Generation]]></category>39 <category><![CDATA[TensorRT]]></category>40 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98131</guid>4142 <description><![CDATA[Manufacturing floors, warehouses and production lines rarely stay fixed — tasks change, layouts shift and new products arrive, and most robots can’t keep up without significant reprogramming. Skild AI’s new S1 robot foundation model helps address this, designed to learn previously unseen, long-horizon tasks from a single video demonstration. The model, launched last week, uses […]]]></description>43 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">Manufacturing floors, warehouses and production lines rarely stay fixed — tasks change, layouts shift and new products arrive, and most robots can’t keep up without significant reprogramming.</span></p>44<p><a target="_blank" href="https://www.nvidia.com/en-us/case-studies/skild-ai/"><span style="font-weight: 400;">Skild AI’s</span></a><span style="font-weight: 400;"> new </span><a target="_blank" href="https://www.skild.ai/blogs/s1"><span style="font-weight: 400;">S1</span></a><span style="font-weight: 400;"> robot foundation model helps address this, designed to learn previously unseen, long-horizon tasks from a single video demonstration. The model, launched last week, uses video as input to understand and execute the task without updating its weights or undergoing task-specific post-training — a technique called in-context learning. </span></p>45<p><span style="font-weight: 400;">Skild built S1 and conducted the research on NVIDIA AI infrastructure, part of a broader collaboration spanning synthetic data generation, model training, simulation and real-world physical AI deployment. The companies are working together to move adaptable robot intelligence from the lab into factories and other dynamic operating environments.</span></p>46<p><span style="font-weight: 400;">“Learning by experience, and not preprogramming, is the step change that has happened in robotics,” said Deepak Pathak, cofounder and CEO of Skild AI. “NVIDIA Isaac Lab and NVIDIA Cosmos technologies help Skild create the scalable, diverse experience its robots need to learn across many scenarios and embodiments.”</span></p>47<p><span style="font-weight: 400;">The launch comes as the company reached a <a target="_blank" href="https://www.skild.ai/blogs/how-to-make-100m">$100 million annual revenue</a> run rate 10 months after its first commercial deployment. In that time, Skild has built more than 60 deployment partnerships with work spanning manufacturing, logistics, inspection, security, food preparation and other applications. </span></p>48<h2><b>Learning New Work From One Video</b></h2>49<p><span style="font-weight: 400;">Most industrial robots are built for fixed jobs, so each new product, process or layout requires more data, retraining and validation.</span></p>50<p><span style="font-weight: 400;">S1 takes a different approach: An operator records a video of the desired task and provides it to the model as a prompt. It interprets the demonstrated intent, objects and sequence, then maps them into actions for the robot in front of it — with no retraining — and often for a task not covered by its pretraining dataset. </span></p>51<p><span style="font-weight: 400;">S1 can perform unfamiliar tasks lasting up to 10 minutes, including plant potting, pancake making, pour-over coffee brewing and kit assembly. These tasks can span dozens of manipulation steps and require the robot to compose skills in sequences it hasn’t previously performed. </span></p>52<div style="width: 1200px;" class="wp-video"><video class="wp-video-shortcode" id="video-98131-1" width="1200" height="675" loop preload="auto" controls="controls"><source type="video/mp4" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotics-promo-skildai-corp-blog-1600x900-one.mp4?_=1" /><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotics-promo-skildai-corp-blog-1600x900-one.mp4">https://blogs.nvidia.com/wp-content/uploads/2026/09/robotics-promo-skildai-corp-blog-1600x900-one.mp4</a></video></div>53<p> </p>54<p><span style="font-weight: 400;">In one plant-potting test, the Skild AI team moved from recording the demonstration to autonomous execution on hardware in just 11 minutes. The model can also adjust when objects move, recover from errors and combine skills in sequences that weren’t explicitly programmed.</span></p>55<p><span style="font-weight: 400;">In Skild’s tests on new, multistep tasks, its S1 robot succeeded about 66% of the time at each step, compared with 9% for a similar AI system — a more than sevenfold improvement. Skild also estimates that showing the robot one short video example can be as useful as giving it roughly 380 hands-on training examples. A person collecting those examples manually could take 50-100 hours.</span></p>56<h2><b>From Research to Factory Work</b></h2>57<p><span style="font-weight: 400;">S1 breaks the cycle of needing to constantly retrain robots for new factors by letting operators demonstrate new tasks directly without requiring a new dataset or training run for every change. Where customer agreements permit, experience from Skild’s commercial deployments can inform the broader model and help accelerate future deployments.</span></p>58<p><span style="font-weight: 400;">That work is already in action on the factory floor.</span> <a target="_blank" href="https://www.skild.ai/blogs/reindustrial-revolution"><span style="font-weight: 400;">Skild, NVIDIA and Foxconn</span></a><span style="font-weight: 400;"> are deploying the Skild Brain on dual-arm manipulators for high-precision assembly of </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/technologies/blackwell-architecture/"><span style="font-weight: 400;">NVIDIA Blackwell</span></a><span style="font-weight: 400;"> systems. In one demonstrated workflow, a robot installs a busbar and limit block, fastens 16 screws and adapts to disturbances across a multistep task. The work requires precise motion, contact-aware control, sequence tracking and recovery when the scene differs from the plan.</span></p>59<div style="width: 1200px;" class="wp-video"><video class="wp-video-shortcode" id="video-98131-2" width="1200" height="675" loop preload="auto" controls="controls"><source type="video/mp4" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-video-2.mp4?_=2" /><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-video-2.mp4">https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-video-2.mp4</a></video></div>60<p> </p>61<h2><b>NVIDIA Technology Across the Development Cycle</b></h2>62<p><span style="font-weight: 400;">NVIDIA accelerated computing gives Skild the scale to train its shared robot brain using simulation, human video, teleoperation and, where permitted, deployment data. </span><a target="_blank" href="https://www.nvidia.com/en-us/ai/cosmos/"><span style="font-weight: 400;">NVIDIA Cosmos</span></a><span style="font-weight: 400;"> open world foundation models help diversify training data and turn video into structured descriptions, while Cosmos Curator helps annotate, filter and organize data at scale.</span></p>63<p><span style="font-weight: 400;">Skild is extensively using NVIDIA’s open simulation frameworks to train and validate its robot brain before real-world deployment. </span><a target="_blank" href="https://www.nvidia.com/en-us/omniverse/"><span style="font-weight: 400;">NVIDIA Omniverse</span></a><span style="font-weight: 400;"> libraries</span><span style="font-weight: 400;"> and the </span><a target="_blank" href="https://developer.nvidia.com/isaac/sim"><span style="font-weight: 400;">NVIDIA Isaac Sim</span></a><span style="font-weight: 400;"> framework provide physically based virtual environments for generating data, testing edge cases and validating behaviors.</span></p>64<p><span style="font-weight: 400;">Skild further strengthens the skills of its brain through reinforcement learning in </span><a target="_blank" href="https://developer.nvidia.com/isaac/lab"><span style="font-weight: 400;">Isaac Lab</span></a><span style="font-weight: 400;">, an open modular robot learning framework. Powered by the </span><a target="_blank" href="https://developer.nvidia.com/newton-physics"><span style="font-weight: 400;">Newton physics engine</span></a><span style="font-weight: 400;">, Isaac Lab helps Skild’s engineers accurately model various physical parameters, such as forces, contact, collision and pressure, and reduce the simulation-to-reality gap.</span></p>65<p><span style="font-weight: 400;">Skild and NVIDIA are also jointly developing new GPU-accelerated simulation solvers that quickly and accurately model how robots physically touch, grip and manipulate solid objects. They’ll soon be made available to all developers as part of Newton. </span></p>66<p><span style="font-weight: 400;">As models move toward production, </span><a target="_blank" href="https://developer.nvidia.com/nsight-systems"><span style="font-weight: 400;">NVIDIA Nsight</span></a><span style="font-weight: 400;"> tools help engineers find performance bottlenecks during training, and the </span><a target="_blank" href="https://developer.nvidia.com/tensorrt"><span style="font-weight: 400;">NVIDIA TensorRT</span></a><span style="font-weight: 400;"> software development kit optimizes inference so robots can respond quickly in the physical world. Together, these technologies connect the data, simulation, training and deployment stages instead of treating them as separate systems.</span></p>67<p><i><span style="font-weight: 400;">Read </span></i><a target="_blank" href="https://www.skild.ai/blogs/s1"><i><span style="font-weight: 400;">Skild AI’s S1 research</span></i></a><i><span style="font-weight: 400;"> and explore the</span></i><a target="_blank" href="https://developer.nvidia.com/isaac"> <i><span style="font-weight: 400;">NVIDIA Isaac robotics platform</span></i></a><i><span style="font-weight: 400;">.</span></i></p>68]]></content:encoded>69 70 71 <enclosure url="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotics-promo-skildai-corp-blog-1600x900-one.mp4" length="13416672" type="video/mp4" />72<enclosure url="https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-video-2.mp4" length="9899884" type="video/mp4" />7374 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-kv-1920x1080-1.jpg" type="image/jpeg" width="1920" height="1080">75 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/skildai-kv-1920x1080-1-842x450.jpg" width="842" height="450" />76 <media:title type="html"><![CDATA[Skild AI Taps NVIDIA Physical AI to Teach Robots New Tasks From a Single Video]]></media:title>77 <media:description type="html"></media:description>78 </media:content>79 </item>80 <item>81 <title>Physical AI Takes the Wheel: How the World’s Robotaxi Leaders Are Building With NVIDIA Technologies</title>82 <link>https://blogs.nvidia.com/blog/robotaxi-leaders-full-stack-open-platform/</link>83 84 <dc:creator><![CDATA[Ali Kani]]></dc:creator>85 <pubDate>Thu, 10 Sep 2026 16:00:04 +0000</pubDate>86 <category><![CDATA[Driving]]></category>87 <category><![CDATA[Artificial Intelligence]]></category>88 <category><![CDATA[Cosmos]]></category>89 <category><![CDATA[Customer Stories]]></category>90 <category><![CDATA[Mobility]]></category>91 <category><![CDATA[NVIDIA Blackwell]]></category>92 <category><![CDATA[NVIDIA DGX]]></category>93 <category><![CDATA[NVIDIA DRIVE]]></category>94 <category><![CDATA[NVIDIA Halos]]></category>95 <category><![CDATA[Omniverse]]></category>96 <category><![CDATA[Physical AI]]></category>97 <category><![CDATA[Simulation and Design]]></category>98 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98141</guid>99100 <description><![CDATA[The global robotaxi market — physical AI’s first commercial breakthrough — is projected to reach $400 billion by 2035, with over 6 million commercial vehicles in operation as driverless fleets are already moving people through some of the world’s busiest and most complex streets. Deploying a driverless vehicle is one challenge. Scaling a fleet is […]]]></description>101 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">The global </span><a target="_blank" href="https://www.nvidia.com/en-us/glossary/robotaxi/"><span style="font-weight: 400;">robotaxi</span></a><span style="font-weight: 400;"> market — </span><a target="_blank" href="https://www.nvidia.com/en-us/glossary/generative-physical-ai/"><span style="font-weight: 400;">physical AI’s</span></a><span style="font-weight: 400;"> first commercial breakthrough — is projected to reach </span><a target="_blank" href="https://www.goldmansachs.com/insights/articles/robotaxis-to-become-a-400-billion-dollar-market-in-2035"><span style="font-weight: 400;">$400 billion by 2035</span></a><span style="font-weight: 400;">, with over 6 million commercial vehicles in operation as driverless fleets are already moving people through some of the world’s busiest and most complex streets.</span></p>102<p><span style="font-weight: 400;">Deploying a driverless vehicle is one challenge. Scaling a fleet is a next-level computing challenge; it means delivering the same safe, reliable performance across thousands of vehicles. </span></p>103<p><span style="font-weight: 400;">Meeting those demands requires enormous amounts of compute across the robotaxi development lifecycle, from preparing and training AI models to simulating and validating driving behavior, as well as real-time processing in the vehicle. </span></p>104<p><span style="font-weight: 400;">NVIDIA provides an open platform for </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/ai-training/"><span style="font-weight: 400;">AI training</span></a><span style="font-weight: 400;">, </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/simulation/"><span style="font-weight: 400;">simulation</span></a><span style="font-weight: 400;"> and </span><a target="_blank" href="https://www.nvidia.com/en-us/ai-trust-center/halos/autonomous-vehicles/"><span style="font-weight: 400;">safety validation</span></a><span style="font-weight: 400;">, with libraries, software development kits, workflows and models that developers can use alongside their own technology stacks.</span></p>105<p><span style="font-weight: 400;">Every major robotaxi program operating at commercial scale today is running on NVIDIA’s modular stack, spanning AI training, simulation, </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/in-vehicle-computing/"><span style="font-weight: 400;">in-vehicle computing</span></a><span style="font-weight: 400;"> — or a combination of the three — to develop and deploy fleets at scale. </span></p>106<h2><b>What Is a Robotaxi Technology Stack?</b></h2>107<p><span style="font-weight: 400;">A robotaxi technology stack is the end-to-end set of technologies used to develop, validate and deploy autonomous vehicles (AVs) — from data and AI model training to simulation, safety validation and real-time in-vehicle computing. </span></p>108<p><span style="font-weight: 400;">NVIDIA’s robotaxi and </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/"><span style="font-weight: 400;">AV platform</span></a><span style="font-weight: 400;"> brings these capabilities together in a three-computer solution: the model training computer, simulation and validation computer, and in-vehicle computer.</span></p>109<h3><b>1. Training Computer: NVIDIA DGX</b></h3>110<p><span style="font-weight: 400;">Robotaxi intelligence advances as programs turn growing volumes of fleet data into increasingly capable models. Driving models can be trained on </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/dgx-platform/"><span style="font-weight: 400;">NVIDIA DGX</span></a><span style="font-weight: 400;"> systems. </span></p>111<p><span style="font-weight: 400;">The </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/alpamayo/"><span style="font-weight: 400;">NVIDIA Alpamayo</span></a><span style="font-weight: 400;"> portfolio of open reasoning vision language action (VLA) models, simulation frameworks and physical AI datasets gives developers building blocks they can adapt to their own data, requirements and technology stacks. Its reasoning models help address long-tail AV challenges by breaking complex driving situations into smaller steps, reasoning through each one and selecting the safest trajectory. </span></p>112<p><span style="font-weight: 400;">NVIDIA also provides physical AI datasets, reinforcement learning blueprints and recipes for post-training and distillation, helping developers optimize models for their target vehicles.</span></p>113<figure id="attachment_98145" aria-describedby="caption-attachment-98145" style="width: 1680px" class="wp-caption alignnone"><img fetchpriority="high" decoding="async" class="wp-image-98145 size-large" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-1680x945.jpeg" alt="" width="1680" height="945" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-1680x945.jpeg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-960x540.jpeg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-1280x720.jpeg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-1536x864.jpeg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-1290x725.jpeg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-630x354.jpeg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-300x169.jpeg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart-400x225.jpeg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-vla-chart.jpeg 1920w" sizes="(max-width: 1680px) 100vw, 1680px" /><figcaption id="caption-attachment-98145" class="wp-caption-text">On a challenging autonomous driving evaluation, adding meta-action and chain-of-thought reasoning data improved a VLA model’s trajectory prediction accuracy, reducing minimum average displacement error — the predicted path’s average deviation from the reference route — by 43%, from 2.08 to 1.18.</figcaption></figure>114<h3><b>2. Simulation and Validation Computer: NVIDIA Omniverse and Cosmos on NVIDIA RTX PRO </b></h3>115<p><span style="font-weight: 400;">Robotaxi programs can’t rely on physical miles alone to capture rare, long-tail driving scenarios. </span><a target="_blank" href="https://docs.nvidia.com/nurec/"><span style="font-weight: 400;">NVIDIA Omniverse NuRec</span></a><span style="font-weight: 400;"> models reconstruct real-world driving scenarios from sensor data, while NVIDIA Cosmos world foundation models generate physically based variations of them, enabling developers to turn thousands of real-world corner cases into millions of combinations of driving behavior, traffic, weather, lighting and sensor conditions.</span></p>116<figure id="attachment_98146" aria-describedby="caption-attachment-98146" style="width: 1680px" class="wp-caption alignnone"><img decoding="async" class="wp-image-98146 size-large" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-1680x945.jpeg" alt="" width="1680" height="945" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-1680x945.jpeg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-960x540.jpeg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-1280x720.jpeg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-1536x864.jpeg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-1290x725.jpeg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-630x354.jpeg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-300x169.jpeg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart-400x225.jpeg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/09/robotaxi-momentum-synthetic-data-chart.jpeg 1920w" sizes="(max-width: 1680px) 100vw, 1680px" /><figcaption id="caption-attachment-98146" class="wp-caption-text">From real-world corner cases to thousands of synthetic permutations spanning behavior and content, NVIDIA Cosmos variations expand AV training data and accelerate model deployment.</figcaption></figure>117<p><span style="font-weight: 400;">Running on </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/products/rtx-pro-server/"><span style="font-weight: 400;">NVIDIA RTX PRO Servers</span></a><span style="font-weight: 400;">, NVIDIA Omniverse and Cosmos support closed-loop simulation and validation. The NVIDIA AlpaSim simulation framework extends the workflow for training and evaluating reasoning-based autonomous-driving models, helping developers identify weaknesses before deployment.</span></p>118<h3><b>3. In-Vehicle Computer and Sensor Architecture: NVIDIA DRIVE Hyperion With DRIVE AGX</b></h3>119<p><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/drive-hyperion/"><span style="font-weight: 400;">NVIDIA DRIVE Hyperion</span></a><span style="font-weight: 400;"> is NVIDIA’s modular in-vehicle</span> <span style="font-weight: 400;">compute and sensor</span> <span style="font-weight: 400;">reference architecture for level-4-ready robotaxis. DRIVE Hyperion 10 pairs dual NVIDIA DRIVE AGX Thor systems-on-a-chip, built on the </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/technologies/blackwell-architecture/"><span style="font-weight: 400;">NVIDIA Blackwell platform</span></a><span style="font-weight: 400;">, with 14 high-definition cameras, nine radars, three lidars and 12 ultrasonics for real-time, 360-degree sensor fusion. Its redundant compute and sensing design supports fail-operational driving if a sensor or compute component fails.</span></p>120<p><span style="font-weight: 400;">The dual </span><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/in-vehicle-computing/#hardware"><span style="font-weight: 400;">DRIVE AGX Thor</span></a><span style="font-weight: 400;"> is designed to run modern AI workloads — including VLA models — for perception, reasoning, path planning and driving actions.</span></p>121<p><img decoding="async" class="alignnone wp-image-98147 size-full" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1.jpg" alt="" width="1920" height="1080" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1.jpg 1920w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hyperion-stack-blog-1920x1080-1-400x225.jpg 400w" sizes="(max-width: 1920px) 100vw, 1920px" /></p>122<p><a target="_blank" href="https://www.nvidia.com/en-sg/ai-trust-center/halos/autonomous-vehicles/"><span style="font-weight: 400;">NVIDIA Halos</span></a><span style="font-weight: 400;"> provides a production-ready safety foundation through Halos OS, and a broader validation and certification framework spanning independent inspection, system validation, large-scale simulation and continuous testing from cloud to car.</span></p>123<p><img loading="lazy" decoding="async" class="alignnone wp-image-98167 size-full" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi.png" alt="" width="1920" height="1080" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi.png 1920w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-960x540.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-1680x945.png 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-1280x720.png 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-1536x864.png 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-1290x725.png 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-630x354.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-300x169.png 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/halos-av-robotaxi-400x225.png 400w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /></p>124<h2><b>Robotaxi Leaders Adopting NVIDIA’s Robotaxi Technology Stack</b></h2>125<p><span style="font-weight: 400;">NVIDIA’s robotaxi ecosystem spans every region where commercial robotaxi services are emerging today: Asia, Europe, the Middle East and North America. Across these markets, mobility providers, AV developers and automakers are adopting NVIDIA’s three-computer architecture to train AI models, simulate and validate driving behavior, and deploy autonomous vehicles at scale.</span></p>126<h3><b>Scaling Robotaxi Services Globally</b></h3>127<ul>128<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://investor.uber.com/news-events/news/press-release-details/2026/NVIDIA-to-Launch-L4-Software-Driven-Robotaxis-on-Uber-Across-28-Cities-by-2028/default.aspx"><b>Uber</b></a> <span style="font-weight: 400;">is scaling its fleet of NVIDIA DRIVE Hyperion, with plans to reach 28 cities by 2028. Uber and NVIDIA are also building a robotaxi AI data factory on NVIDIA Cosmos to curate fleet driving data for rare scenarios. Together, </span><b>Uber</b><span style="font-weight: 400;"> and NVIDIA are collaborating with </span><b>Autobrains, Avride, Lucid, May Mobility, Mercedes-Benz, Momenta, Nissan, Nuro, Pony.ai, Stellantis, Waabi, Wayve, WeRide </b><span style="font-weight: 400;">and </span><b>Zoox</b><span style="font-weight: 400;"> to bring NVIDIA-powered robotaxi services to the Uber platform.</span></li>129<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://investor.nvidia.com/news/press-release-details/2025/NVIDIA-Makes-the-World-Robotaxi-Ready-With-Uber-Partnership-to-Support-Global-Expansion/default.aspx"><b>May Mobility</b></a> <span style="font-weight: 400;">is planning to operate autonomous ride-hailing services through Uber’s network, while developing its software stack on the NVIDIA DRIVE platform.</span></li>130<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://bolt.eu/en/blog/nvidia-bolt-announcement/"><b>Bolt</b></a> <span style="font-weight: 400;">uses NVIDIA technologies to develop and scale AVs across Europe.</span></li>131<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://investor.lyft.com/news-events-presentations/press-releases/detail/193/lyft-to-make-rides-smarter-and-more-efficient-through-agentic-ai-accelerate-av-future-with-nvidia-drive-hyperion"><b>Lyft</b></a> <span style="font-weight: 400;">plans to use NVIDIA DRIVE Hyperion as a reference architecture for future autonomous fleets. </span><b>May Mobility</b> <span style="font-weight: 400;">vehicles are also currently operating on Lyft’s network in Atlanta, powered by NVIDIA DRIVE.</span></li>132<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Through its partnership with</span> <b>Grab</b><span style="font-weight: 400;">,</span> <a target="_blank" href="https://ir.weride.ai/news-releases/news-release-details/weride-showcases-robotaxi-gxr-powered-nvidia-drive-hyperion"><b>WeRide</b></a> <span style="font-weight: 400;">plans to bring its DRIVE Hyperion- and DRIVE AGX Thor-based GXR to key markets across Southeast Asia.</span></li>133<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://waymo.com/blog/2026/08/look-under-our-trunk/"><b>Waymo</b></a><span style="font-weight: 400;"> partners with NVIDIA to help build its autonomous computing system.</span></li>134</ul>135<h3><b>Building Robotaxi Intelligence </b></h3>136<p><span style="font-weight: 400;">Behind these services, AV developers are using NVIDIA accelerated computing, simulation and in-vehicle platforms to build the intelligence that operators and automakers deploy.</span></p>137<ul>138<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://wayve.ai/press/wayve-nissan-robotaxi-gtc/"><b>Wayve</b></a><span style="font-weight: 400;">, </span><b>Nissan</b><span style="font-weight: 400;"> a</span><span style="font-weight: 400;">nd </span><b>Uber</b><span style="font-weight: 400;"> are developing a global robotaxi program using a prototype vehicle that combines Nissan’s vehicle engineering, Wayve’s embodied AI and the NVIDIA DRIVE Hyperion platform.</span></li>139<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://autobrains.ai/autobrains-and-uber-launch-agentic-ai-robotaxi-program-in-munich-built-on-nvidia-drive-hyperion/"><b>Autobrains</b></a><span style="font-weight: 400;"> is developing robotaxi programs with </span><b>Uber</b><span style="font-weight: 400;"> in Munich and </span><a target="_blank" href="https://vinfastauto.us/newsroom/press-release/vinfast-and-autobrains-launch-first-agentic-ai-l4-program-for-southeast-asia"><b>VinFast</b></a> <span style="font-weight: 400;">in Southeast Asia, built on NVIDIA DRIVE Hyperion and enabled by Autobrains’ Agentic AI technology.</span></li>140<li style="font-weight: 400;" aria-level="1"><a href="https://blogs.nvidia.com/blog/nvidia-zoox-autonomous-ride-hailing/"><b>Zoox</b></a> <span style="font-weight: 400;">uses NVIDIA DRIVE for in-vehicle computing and cloud-based training and simulation. </span></li>141<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://nvidianews.nvidia.com/news/nvidia-uber-robotaxi"><b>Momenta</b></a><span style="font-weight: 400;"> is developing its software stack based on NVIDIA DRIVE AGX running on DriveOS.</span><span style="font-weight: 400;"> </span></li>142<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://ir.pony.ai/news-releases/news-release-details/pony-ai-inc-announces-new-generation-autonomous-driving-domain"><b>Pony.ai</b></a> <span style="font-weight: 400;">developed its new-generation autonomous-driving domain controller with NVIDIA DRIVE Hyperion and DRIVE AGX Thor. </span></li>143<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://www.tensor.auto/press/lyft2025"><b>Tensor</b></a> <span style="font-weight: 400;">is developing its level 4 Robocar with eight NVIDIA DRIVE AGX Thor systems-on-a-chip in its in-vehicle supercomputer. </span></li>144<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://waabi.ai/insights/waabi-secures-1-billion-in-new-funding-to-lead-physical-ai-revolution"><b>Waabi</b></a> <span style="font-weight: 400;">expands into the robotaxi market through a deployment collaboration with </span><b>Uber</b><span style="font-weight: 400;">; its Waabi Driver platform is built on NVIDIA DRIVE AGX Thor.</span></li>145<li style="font-weight: 400;" aria-level="1"><b>TIER IV</b><span style="font-weight: 400;"> and </span><a target="_blank" href="https://www.isuzu-global.com/en/newsroom/20260317_1.html"><b>Isuzu</b></a> <span style="font-weight: 400;">are deploying level 4 autonomous buses built on NVIDIA DRIVE Hyperion and DRIVE AGX Thor.</span></li>146<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://news.lenovo.com/pressroom/press-releases/lenovo-works-with-swm-to-develop-next-generation-robotaxi-on-nvidia-drive-agx-thor/"><b>Lenovo</b></a> <span style="font-weight: 400;">is supplying its NVIDIA DRIVE AGX Thor-based AD1 level 4 domain controller for a next-generation robotaxi program with </span><b>SWM</b><span style="font-weight: 400;">.</span></li>147<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://www.prnewswire.com/news-releases/deeprouteai-presents-40b-vision-language-action-foundation-model-at-nvidia-gtc-2026-accelerating-autonomous-driving-at-scale-302716046.html"><b>DeepRoute.ai</b></a> <span style="font-weight: 400;">is developing a new generation of robotaxis built on the NVIDIA DRIVE Hyperion platform with DRIVE AGX Thor.</span></li>148</ul>149<h3><b>Bringing Robotaxis Into Production</b></h3>150<p><span style="font-weight: 400;">As these systems move from development into production, automakers are integrating NVIDIA technology into autonomous and robotaxi-ready vehicle programs.</span></p>151<ul>152<li style="font-weight: 400;" aria-level="1"><b>Tesla</b> <span style="font-weight: 400;">trains its autonomous-driving neural networks on NVIDIA supercomputers. </span></li>153<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://media.mbusa.com/releases/release-cfaf7728c8957661f23433449e08d179-mercedes-benz-accelerates-future-robotaxi-ecosystem-and-collaborates-with-industry-leading-partners"><b>Mercedes-Benz</b></a> <span style="font-weight: 400;">and NVIDIA are collaborating with </span><b>Uber</b><span style="font-weight: 400;"> to develop a robotaxi ecosystem based on the new S-Class, built on the NVIDIA DRIVE Hyperion architecture, full-stack NVIDIA DRIVE AV L4 software, and NVIDIA Alpamayo open AI models, simulation tools and datasets to support reasoning-based, safety-first autonomy.</span></li>154<li style="font-weight: 400;" aria-level="1"><b>Stellantis, Wayve </b><span style="font-weight: 400;">and </span><b>Uber</b> <span style="font-weight: 400;">are collaborating to develop and deploy L4 driverless mobility services, leveraging NVIDIA DRIVE Hyperion and AI computing technologies.</span></li>155<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://ir.lucidmotors.com/news-releases/news-release-details/lucid-nuro-and-uber-unveil-global-robotaxi-ces-announce?utm_source=chatgpt.com"><b>Lucid</b></a><span style="font-weight: 400;">, </span><b>Nuro</b> <span style="font-weight: 400;">and </span><b>Uber </b><span style="font-weight: 400;">are developing a global robotaxi service using NVIDIA DRIVE AGX Thor, part of the DRIVE Hyperion platform.</span></li>156<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://nvidianews.nvidia.com/news/hyundai-motor-kia-autonomous-driving"><b>Hyundai Motor</b></a><span style="font-weight: 400;"> and </span><b>Kia</b> <span style="font-weight: 400;">are expanding their collaboration with NVIDIA to develop data-driven autonomous-driving systems built on NVIDIA DRIVE Hyperion. NVIDIA will also explore expanded collaboration with Hyundai Motor Group’s joint venture, </span><b>Motional</b><span style="font-weight: 400;">,</span><span style="font-weight: 400;"> to advance level 4 robotaxi services.</span></li>157<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://www.globenewswire.com/news-release/2026/03/18/3258289/0/en/geely-expands-strategic-partnership-with-nvidia-across-physical-enterprise-and-industrial-ai.html"><b>Geely</b></a><span style="font-weight: 400;">, </span><span style="font-weight: 400;">alongside its ecosystem partners, plans to develop and commercialize robotaxis using DRIVE Hyperion.</span></li>158<li style="font-weight: 400;" aria-level="1"><a target="_blank" href="https://www.prnewswire.com/news-releases/zeekr-announces-january-2025-delivery-update-302365781.html"><b>Zeekr</b></a><span style="font-weight: 400;">, a</span><span style="font-weight: 400;"> Geely Auto Group brand, has adopted DRIVE AGX Thor for a centralized domain controller.</span></li>159</ul>160<p><span style="font-weight: 400;">From cloud to car, nearly every layer of the robotaxi platform is being developed on NVIDIA accelerated computing. </span></p>161<p><i><span style="font-weight: 400;">Explore NVIDIA’s complete </span></i><a target="_blank" href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/"><i><span style="font-weight: 400;">platform for robotaxi development</span></i></a><i><span style="font-weight: 400;">.</span></i></p>162]]></content:encoded>163 164 165 166 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hero-kv-blog-r7-1920x1080-1.jpg" type="image/jpeg" width="1920" height="1080">167 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/auto-robotaxi-hero-kv-blog-r7-1920x1080-1-842x450.jpg" width="842" height="450" />168 <media:title type="html"><![CDATA[Physical AI Takes the Wheel: How the World’s Robotaxi Leaders Are Building With NVIDIA Technologies]]></media:title>169 <media:description type="html"></media:description>170 </media:content>171 </item>172 <item>173 <title>d-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment</title>174 <link>https://blogs.nvidia.com/blog/d-matrix-nvlink-fusion/</link>175 176 <dc:creator><![CDATA[Jesse Clayton]]></dc:creator>177 <pubDate>Thu, 10 Sep 2026 13:00:21 +0000</pubDate>178 <category><![CDATA[Corporate]]></category>179 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98153</guid>180181 <description><![CDATA[AI inference chipmaker d-Matrix today announced it will use NVLink Fusion to connect its next-generation Raptor XPUs to NVIDIA’s AI infrastructure platform — joining a growing roster of ecosystem partners. By connecting Raptor to NVIDIA NVLink scale-up and Spectrum-X scale-out networking, the NVIDIA MGX rack architecture and the broader NVIDIA AI platform, NVLink Fusion gives […]]]></description>182 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p>AI inference chipmaker d-Matrix today <span draggable="true"><a href="https://www.d-matrix.ai/announcements/d-matrix-rackscale-nvidia/" target="_blank" rel="noopener noreferrer"><u>announced</u> </a></span>it will use NVLink Fusion to connect its next-generation Raptor XPUs to NVIDIA’s AI infrastructure platform — joining a growing roster of ecosystem partners.</p>183<p><span style="font-weight: 400;">By connecting Raptor to NVIDIA NVLink scale-up and Spectrum-X scale-out networking, the NVIDIA MGX rack architecture and the broader NVIDIA AI platform, NVLink Fusion gives d-Matrix an accelerated, lower-risk path from custom silicon to large-scale deployment. </span></p>184<p><span style="font-weight: 400;">“Demand for inference is soaring, but capital, time and energy remain finite,” said Sid Sheth, cofounder and CEO of d-Matrix during a press briefing yesterday. “With NVLink Fusion and MGX, we can integrate our Raptor XPUs into a broadly deployed, liquid-cooled architecture, giving customers a faster, lower-risk path to deploy and scale ultralow-latency inference.” </span><span style="font-weight: 400;"><br />185</span><span style="font-weight: 400;"><br />186</span><span style="font-weight: 400;">The NVIDIA AI platform is vertically integrated and horizontally open. NVLink Fusion extends this openness to XPUs and CPUs, allowing silicon companies to focus on their processor innovations while using NVIDIA infrastructure to deploy them at AI factory scale.</span></p>187<h2><b>From Custom Silicon to Rack-Scale Deployment</b></h2>188<p>Building an XPU is only the first step. Deploying it at AI factory scale requires a complete platform spanning networking, rack architecture, power, cooling, software and a proven supply chain. Each of the steps — sourcing chips, integrating high-speed interfaces, validating a scale-up networking solution, designing and certifying rack architecture — adds time, costs and risks.</p>189<p><span style="font-weight: 400;">NVLink Fusion lets silicon innovators </span><a href="https://blogs.nvidia.com/blog/nvlink-fusion-xpu-ai-factory/"><span style="font-weight: 400;">connect directly into NVIDIA’s proven platform</span></a><span style="font-weight: 400;">. It’s the high-bandwidth, low-latency technology that connects custom XPUs and CPUs to the NVIDIA stack.</span></p>190<p><span style="font-weight: 400;">By adopting NVLink Fusion, d-Matrix can tap into the NVIDIA MGX ecosystem’s mature, validated rack designs, supply chain, power and cooling infrastructure. By standardizing on a common rack, data centers can be built once and support GPUs, CPUs and XPUs — without requiring a separate rack architecture for each processor type.</span></p>191<h2>NVLink Fusion Integrates d-Matrix XPUs Into AI Factories</h2>192<p><span style="font-weight: 400;">Using NVIDIA NVLink, d-Matrix plans to connect its XPUs in a single high-bandwidth, low-latency scale-up domain. Its racks can also work alongside NVIDIA GPU-based systems like NVIDIA Vera Rubin NVL72 for disaggregated inference. </span></p>193<p><span style="font-weight: 400;">d-Matrix also plans to integrate NVIDIA Vera CPUs, NVIDIA ConnectX-9 SuperNICs, NVIDIA BlueField-4 DPUs and NVIDIA Spectrum-X Ethernet networking. Together with NVLink and MGX, these technologies give d-Matrix a proven foundation to deploy specialized inference alongside NVIDIA systems within flexible, unified AI factories.</span></p>194<h2>NVLink Fusion Opens NVIDIA AI Factories to Specialized XPU Architectures</h2>195<p><span style="font-weight: 400;">The NVIDIA full-stack AI factory platform includes NVIDIA Vera Rubin NVL72, Groq 3 LPX, the Vera CPU rack, Vera BlueField-4 STX storage and Spectrum-6 SPX Ethernet networking. It’s designed to be completely fungible — running every AI workload, model and model architecture — with the best performance per watt and lowest cost per token. </span></p>196<p><span style="font-weight: 400;">NVLink Fusion gives customers the flexibility to match the right compute to each workload within a common AI factory platform, opening access to NVIDIA networking, systems, software and global supply chain. Silicon innovators like d-Matrix can use NVLink Fusion to increase performance, accelerate time to market and reduce the risk of deploying semi-custom AI factories.</span></p>197<p> </p>198]]></content:encoded>199 200 201 202 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/gpu-architecture-corp-blog-dmatrix-nvlink-fusion-1280x680-5692448-r1.png" type="image/png" width="1280" height="680">203 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/gpu-architecture-corp-blog-dmatrix-nvlink-fusion-1280x680-5692448-r1-842x450.png" width="842" height="450" />204 <media:title type="html"><![CDATA[d-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment]]></media:title>205 <media:description type="html"></media:description>206 </media:content>207 </item>208 <item>209 <title>Boots on the Ground: ‘WARDOGS’ Goes All Out on GeForce NOW at Early-Access Launch</title>210 <link>https://blogs.nvidia.com/blog/geforce-now-thursday-wardogs/</link>211 212 <dc:creator><![CDATA[GeForce NOW Community]]></dc:creator>213 <pubDate>Thu, 10 Sep 2026 13:00:18 +0000</pubDate>214 <category><![CDATA[Gaming]]></category>215 <category><![CDATA[Cloud Gaming]]></category>216 <category><![CDATA[GeForce NOW]]></category>217 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98103</guid>218219 <description><![CDATA[Gear up: The latest PC games and major updates are ready to play on GeForce NOW this week. WARDOGS drops onto the cloud at early-access launch, alongside the Valheim 1.0 Deep North update and Bus Simulator 27 — part of nine new titles joining the cloud. The newest PC releases can demand serious hardware, storage […]]]></description>220 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400">Gear up: The latest PC games and major updates are ready to play on </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=24109411186&gbraid=0AAAAAD4XAoHFeYjF4UfakOH41Ho0uO9IY&gclid=Cj0KCQjwkt_UBhDMARIsALpnOAw2w5dF_WNG3Y9tF0N7hBh4z3LTF2AErnlCr8_N89SM6JQbquAmT5oaAgiJEALw_wcB"><span style="font-weight: 400">GeForce NOW</span></a><span style="font-weight: 400"> this week. </span><i><span style="font-weight: 400">WARDOGS </span></i><span style="font-weight: 400">drops onto the cloud at early-access launch, alongside the </span><i><span style="font-weight: 400">Valheim </span></i><span style="font-weight: 400">1.0 Deep North update and </span><i><span style="font-weight: 400">Bus Simulator 27</span></i><span style="font-weight: 400"> — part of nine new titles joining the cloud.</span></p>221<p><span style="font-weight: 400">The newest PC releases can demand serious hardware, storage space and upgrade budgets. GeForce NOW puts GeForce RTX-powered performance in the cloud, so members can jump into the newest titles across </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/system-reqs/#windows-pc"><span style="font-weight: 400">supported devices</span></a><span style="font-weight: 400"> without needing expensive PC upgrades to keep up.</span></p>222<p><span style="font-weight: 400">From a 100-player tactical firefight to a frozen Viking frontier and a growing bus empire, every adventure is ready when members are — no downloads or storage management required.</span></p>223<h2><b>Who Let the ‘WARDOGS’ Out?</b></h2>224<p><iframe loading="lazy" title="WARDOGS | Gameplay Trailer" width="1200" height="675" src="https://www.youtube.com/embed/v1ISv_h_qXY?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>225<p> </p>226<p><span style="font-weight: 400">Take on the Control Zone: </span><i><span style="font-weight: 400">WARDOGS </span></i><span style="font-weight: 400">launches on GeForce NOW today, bringing BULKHEAD’s large-scale tactical first-person shooter to the cloud at early-access launch.</span></p>227<p><span style="font-weight: 400">Up to 100 players across three teams battle to control randomized zones in a large-scale military sandbox, where tactical gunplay, combined-arms combat, building and destruction can change the course of a match.</span></p>228<p><span style="font-weight: 400">There’s no single route to victory. Outfit a loadout, coordinate a squad, fortify a strategic position or take the fight to the enemy in vehicles. Every team play action earns cash, giving players more ways to invest in gear and vehicles that can turn the tide.</span></p>229<p><span style="font-weight: 400">Ultimate members can stream the fight with GeForce RTX 5080-class performance in the cloud for a tactical advantage. Pick up the battle across supported devices without waiting on a huge game install or clearing local storage first.</span></p>230<p><b>Ice, Ice in the Deep North, Baby</b></p>231<p><iframe loading="lazy" title="Valheim: 1.0 & Deep North – Release Date Reveal Trailer" width="1200" height="675" src="https://www.youtube.com/embed/eWULkdcaR5o?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>232<p><span style="font-weight: 400">Sharpen the axes and raise the shields – </span><i><span style="font-weight: 400">Valheim</span></i><span style="font-weight: 400"> has left early access with its 1.0 update, letting members stream the final biome in Iron Gate’s Viking survival saga through </span><a target="_blank" href="https://store.steampowered.com/app/892970?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400"> and </span><a target="_blank" href="https://www.xbox.com/games/store/valheim-game-preview/9NCBL78CG9N7?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox Store</span></a><span style="font-weight: 400">. </span></p>233<p><span style="font-weight: 400">Set sail for the Deep North, a frozen frontier filled with new enemies to face, weapons to craft and strongholds to build before the cold closes in. The 1.0 launch also arrives alongside new platform versions with full crossplay support, making it easier to rally the clan.</span></p>234<p><span style="font-weight: 400">Whether returning to a long-running world on a PC or Mac or beginning a fresh voyage on a Steam Deck or Amazon Fire TV Stick, members can head to the Deep North across supported devices without waiting for downloads.</span></p>235<p><b>The Wheels on the Bus Go ‘Round in the Cloud</b></p>236<figure id="attachment_98106" aria-describedby="caption-attachment-98106" style="width: 1200px" class="wp-caption aligncenter"><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-scaled.jpg"><img loading="lazy" decoding="async" class="size-large wp-image-98106" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-1680x945.jpg" alt="Bus Simulator 27 on GeForce NOW" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-scaled.jpg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/BS27-GFN-2-400x225.jpg 400w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a><figcaption id="caption-attachment-98106" class="wp-caption-text">The route to the top starts here.</figcaption></figure>237<p><span style="font-weight: 400">Hop into the driver’s seat. </span><i><span style="font-weight: 400">Bus Simulator 27</span></i><span style="font-weight: 400"> is now streaming from GeForce NOW. </span></p>238<p><span style="font-weight: 400">Drive a fleet of more than 45 officially licensed buses from 13 manufacturers through Felicia Bay, a sunny region with two cities, 20 districts and a full day-and-night cycle. Build routes, manage timetables and grow a successful bus company in Story, Career or Sandbox mode.</span></p>239<p><span style="font-weight: 400">And there’s even more to stream from the </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/games/"><span style="font-weight: 400">GeForce NOW library</span></a><span style="font-weight: 400"> this week:</span></p>240<ul>241<li style="font-weight: 400"><i><span style="font-weight: 400">Bus Simulator 27 </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2397320/Bus_Simulator_27/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>242<li style="font-weight: 400"><i><span style="font-weight: 400">Halloween: The Game </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3219630/Halloween_The_Game/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>243<li style="font-weight: 400"><i><span style="font-weight: 400">Honeycomb: The World Beyond </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/1510440/Honeycomb_The_World_Beyond/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>244<li style="font-weight: 400"><i><span style="font-weight: 400">Wanderburg </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3624140/Wanderburg/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>245<li style="font-weight: 400"><i><span style="font-weight: 400">Shroom and Gloom </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3271280/Shroom_and_Gloom/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>246<li style="font-weight: 400"><i><span style="font-weight: 400">Train Sim World 7 – Advanced Access</span></i><span style="font-weight: 400"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/4678800/?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>247<li style="font-weight: 400"><i><span style="font-weight: 400">WARDOGS </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/1867240/WARDOGS/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>248<li style="font-weight: 400"><i><span style="font-weight: 400">Welcome to Elderfield </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3195440/Welcome_to_Elderfield/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>249<li style="font-weight: 400"><i><span style="font-weight: 400">Death Stranding Director’s Cut</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://nvidia.custhelp.com/app/answers/detail/a_id/5377/"><span style="font-weight: 400">Epic Games Store</span></a><span style="font-weight: 400">)</span></li>250</ul>251<p><span style="font-weight: 400">Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after release and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.</span></p>252<p><a target="_blank" href="https://www.reddit.com/r/GeForceNOW/comments/1w4tblj/im_back_at_gfn_after_6_years_ultimate_tier_and/"><span style="font-weight: 400">One GFN community member</span></a><span style="font-weight: 400"> recently returned after six years, saying </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-september-2026-games-list/"><span style="font-weight: 400">NVIDIA DLSS 5</span></a><span style="font-weight: 400"> technology renewed their interest in GeForce NOW cloud gaming. Ready to try it out, too? Start with a </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/"><span style="font-weight: 400">day pass</span></a><span style="font-weight: 400"> to test premium cloud gaming before committing to a membership. For those who decide to keep playing, the cost can be applied toward their first monthly membership.</span></p>253<p><span style="font-weight: 400">What are you planning to play this weekend? Let us know on </span><a target="_blank" href="https://www.twitter.com/nvidiagfn"><span style="font-weight: 400">X</span></a><span style="font-weight: 400"> or in the comments below.</span></p>254]]></content:encoded>255 256 257 258 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_September_10.jpg" type="image/jpeg" width="1920" height="1080">259 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_September_10-842x450.jpg" width="842" height="450" />260 <media:title type="html"><![CDATA[Boots on the Ground: ‘WARDOGS’ Goes All Out on GeForce NOW at Early-Access Launch]]></media:title>261 <media:description type="html"></media:description>262 </media:content>263 </item>264 <item>265 <title>NVIDIA Brings Real-Time AI to Broadcast, Sports and Global Streaming at IBC</title>266 <link>https://blogs.nvidia.com/blog/ibc-news-2026/</link>267 268 <dc:creator><![CDATA[NVIDIA Writers]]></dc:creator>269 <pubDate>Wed, 09 Sep 2026 16:00:42 +0000</pubDate>270 <category><![CDATA[AI]]></category>271 <category><![CDATA[Pro Graphics]]></category>272 <category><![CDATA[Artificial Intelligence]]></category>273 <category><![CDATA[Events]]></category>274 <category><![CDATA[Holoscan for Media]]></category>275 <category><![CDATA[Media and Entertainment]]></category>276 <category><![CDATA[NVIDIA NIM]]></category>277 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98109</guid>278279 <description><![CDATA[At the IBC conference, running Sept. 11-14 in Amsterdam, the creative, technology and business communities are coming together to turn ideas into action and discuss innovations across the media and entertainment industries. More than 44,000 attendees from 170+ countries are gathering to explore 1,300+ exhibitions in 14+ halls and outdoor spaces, with over 600 speakers […]]]></description>280 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p>At the IBC conference, running Sept. 11-14 in Amsterdam, the creative, technology and business communities are coming together to turn ideas into action and discuss innovations across the media and entertainment industries. More than 44,000 attendees from 170+ countries are gathering to explore 1,300+ exhibitions in 14+ halls and outdoor spaces, with over 600 speakers delivering insights.</p>281<p>Read on to learn more about what NVIDIA’s highlighting at the show.</p>282<hr />283<h2 id="ai-for-media"><b>NVIDIA AI for Media Brings Real-Time Intelligence to Broadcast, Sports and Production Workflows <a href="https://blogs.nvidia.com/blog/ibc-news-2026/#ai-for-media"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></h2>284<p><span style="font-weight: 400;">Media companies are increasingly integrating AI into live production, sports, news and streaming workflows to unlock richer performance insights, verify video authenticity, and enhance and localize content — all without disrupting trusted broadcast environments.</span></p>285<p><span style="font-weight: 400;">At IBC 2026 in Amsterdam, NVIDIA announced a major expansion to NVIDIA AI for Media — a collection of GPU-accelerated software development kits (SDKs), NVIDIA NIM microservices, playbooks, and blueprints that enhance audio, video and augmented-reality effects for media and entertainment workflows — to unlock new ways to understand motion, verify and enhance video, localize programming and build AI-powered media applications. </span></p>286<p><span style="font-weight: 400;">The </span><a target="_blank" href="https://build.nvidia.com/nvidia/synthetic-video-detector"><b>NVIDIA Synthetic Video Detector</b></a><b> (SVD)</b><span style="font-weight: 400;"> NIM microservice, announced earlier this year at </span><a href="https://blogs.nvidia.com/blog/siggraph-news-2026/#synthetic-video"><span style="font-weight: 400;">SIGGRAPH</span></a><span style="font-weight: 400;">,</span> <span style="font-weight: 400;">helps organizations assess the probability of whether footage is authentic or AI-generated, giving editorial, content-authentication, digital-forensics and media-integrity teams another point of analysis in their review process.</span></p>287<p><span style="font-weight: 400;">Since its initial release, SVD’s accuracy has reached 99.3% for text-to-video content and 97.7% for image-to-video content, with especially large gains on difficult image-to-video cases. </span></p>288<p><span style="font-weight: 400;">Dalet is integrating SVD into a secure, cloud-hosted verification workflow for news organizations. This allows editorial teams to submit footage through SVD, inspect and review the resulting scores and metadata within a Dalet interface. </span></p>289<p><span style="font-weight: 400;">TwelveLabs announced the general availability of Compliance by TwelveLabs, its first application built on the company’s video intelligence platform, helping media and broadcast teams rapidly screen content against regional and custom compliance standards. The solution integrates SVD to add frame-level authenticity signals and confidence scores, enabling media teams to identify potentially synthetic media within the same compliance workflow.</span></p>290<p><span style="font-weight: 400;">Wowza</span><i><span style="font-weight: 400;">, </span></i><span style="font-weight: 400;">whose</span> <a target="_blank" href="https://www.wowza.com/streaming-engine"><span style="font-weight: 400;">Wowza Streaming Engine</span></a><span style="font-weight: 400;"> media server technology powers more than 35,000 video deployments across over 170 countries, will distribute SVD through the</span> <a target="_blank" href="https://www.wowza.com/video-intelligence-framework"><span style="font-weight: 400;">Wowza Video Intelligence Framework</span></a><span style="font-weight: 400;">. The solution, powered by NVIDIA-accelerated infrastructure, will enable broadcasters, streaming providers and other organizations to analyze live video feeds and extract data around detected objects, scenes and signs of AI generation in real time. It can be deployed and run on premises, at the edge, in the cloud, across hybrid deployments or fully air-gapped, giving organizations greater control over critical media workflows.</span></p>291<p><b>NVIDIA 3D Body Pose </b><span style="font-weight: 400;">estimates 2D and 3D human joint locations and angles from video captured by a single camera, helping turn motion into structured data without marker-based capture systems.</span></p>292<p><span style="font-weight: 400;">For sports organizations, that data can support player and athlete movement tracking, biomechanics and performance analysis, replay enhancement, officiating and adjudication workflows, player-safety applications, and virtual interaction and immersive experiences. </span></p>293<p><span style="font-weight: 400;">The technology can also provide structured human-motion data for content-creation workflows. When mapped to a compatible character rig, joint and motion data can serve as input for animation blocking, digital doubles, character retargeting and virtual-production experiences.</span></p>294<p><span style="font-weight: 400;">Vizrt is using Body Pose technology in live virtual-studio environments, with tracked body movement driving real-time 3D lighting effects such as reflections, shadows and environmental rendering.</span></p>295<p><b>Video Frame Generation (VFG) </b><span style="font-weight: 400;">makes video motion appear smoother by using generative AI to create new frames between the original frames of a video. It can increase frame rates by 2x or 4x while preserving visual quality and temporal consistency, enabling more fluid sports, slow-motion replays, live media and other high-motion video experiences. VFG also supports frame-rate conversion and frame boosting for generative AI video workflows.</span></p>296<p><span style="font-weight: 400;">Ross Video is integrating VFG into its Rio Replay platform to create AI-assisted slow-motion video for sports production.</span></p>297<p><span style="font-weight: 400;">The work supports 6x slow-motion generation for sports replay. Development is underway toward 8x interpolation, meaning generated intermediate frames can give replay teams smoother motion without requiring every frame to be captured by an ultrahigh-frame-rate source camera. </span></p>298<p><b>NVIDIA Video Super Resolution</b><span style="font-weight: 400;"> (VSR) uses AI to upscale video while reducing noise, blur and compression artifacts. New streaming modes let developers choose between real-time performance and higher image quality, while adjustable controls help achieve the desired level of enhancement. VSR also adds 10-bit video support and improves overall performance and quality. VSR is available through the NVIDIA Video Effects SDK and a NIM microservice for use in streaming, broadcast, conferencing, video playback and content-creation applications. </span></p>299<p><span style="font-weight: 400;">The technology can support video players, conferencing applications, creator tools, streaming services, transcoders and broadcast systems through a common interface.</span></p>300<p><b>NVIDIA TrueHDR</b><span style="font-weight: 400;"> converts standard-dynamic-range video into high-dynamic-range output in real time, reaching up to approximately 2,000 nits while preserving local contrast and adapting brightness to the content.</span></p>301<p><span style="font-weight: 400;">VSR, VFG and TrueHDR can be combined within a single video-effects pipeline — helping media companies enhance existing content libraries for streaming, transcoding, gaming and creator workflows. </span></p>302<p><span style="font-weight: 400;">The </span><a target="_blank" href="https://build.nvidia.com/nvidia/lipsync"><b>NVIDIA LipSync</b></a><span style="font-weight: 400;"> and </span><a target="_blank" href="https://build.nvidia.com/nvidia/active-speaker-detection"><b>Active Speaker Detection</b></a><span style="font-weight: 400;"> NIM microservices help developers build localization systems for interviews, news, sports, entertainment and other programming where multiple people may appear on screen.</span></p>303<p><span style="font-weight: 400;">LipSync transforms mouth movement in an input video to match a target audio track while preserving natural head pose, blinking and body movement. The new release improves facial occlusion handling and better preserves teeth, lip and facial textures.</span></p>304<p><span style="font-weight: 400;">The new Active Speaker Detection NIM microservice no longer requires speaker diarization for multiple audio tracks, adds voice activity detection and expands NIM microservice deployment support through a gRPC interface and broader GPU compatibility.</span></p>305<p><a target="_blank" href="https://ndi.video/stories/press/ndi-nvidia-ai-multilingual-content/"><span style="font-weight: 400;">NDI</span></a> <span style="font-weight: 400;">is using NVIDIA AI for Media, including the NVIDIA LipSync NIM microservice, to enable real-time translation, lip-synced dubbing and regional language adaptation within existing broadcast workflows. By generating multiple language experiences from a common media stream, the approach can help broadcasters reach global audiences while reducing the bandwidth, infrastructure and production complexity traditionally required for multilingual distribution.</span></p>306<p><b>Studio Voice </b><span style="font-weight: 400;">includes new Microphone Profiles built on NVIDIA Studio Voice NIM microservices, giving users more control over the tonal character of enhanced speech.</span></p>307<p><span style="font-weight: 400;">The capability is designed to suppress background noise, reduce room reverberation and improve speech clarity, then shape the enhanced output into a selected microphone profile for more polished live communications, streaming, podcasting and content creation.</span></p>308<p><i><span style="font-weight: 400;">Try </span></i><a target="_blank" href="https://build.nvidia.com/models?label=nvidia+ai+for+media"><i><span style="font-weight: 400;">NVIDIA AI for Media NIM microservices</span></i></a><i><span style="font-weight: 400;">. See the latest NVIDIA and partner workflows at </span></i><a target="_blank" href="https://www.nvidia.com/en-us/events/ibc/"><i><span style="font-weight: 400;">IBC 2026</span></i></a><i><span style="font-weight: 400;">. </span></i></p>309<hr />310<h2 id="mxl"><b>NVIDIA Holoscan for Media Provides Open Media Exchange Layer to Build and Connect Live Media Applications </b><a href="https://blogs.nvidia.com/blog/ibc-news-2026/#mxl"><b><em><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></em></b></a><b></b><b></b></h2>311<p><img loading="lazy" decoding="async" class="alignnone size-large wp-image-98113" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-1680x945.jpg" alt="" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/09/holoscan-for-media-1920x1080-1.jpg 1920w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></p>312<p><span style="font-weight: 400;">As broadcasters, streaming services and sports organizations adopt software and AI, the infrastructure behind live content is becoming more flexible, more connected and increasingly built on shared accelerated computing.</span></p>313<p><span style="font-weight: 400;">The integration of Media Exchange Layer (MXL) with </span><a target="_blank" href="https://developer.nvidia.com/holoscan-for-media"><span style="font-weight: 400;">NVIDIA Holoscan for Media</span></a><span style="font-weight: 400;"> accelerates that transition — providing the common exchange layer that helps media applications connect and operate together. </span></p>314<p><span style="font-weight: 400;">Holoscan for Media is an open reference architecture and developer toolkit for building AI-powered media functions and applications for software-defined live production. MXL adds an open way for those software-based media functions to exchange live video, audio and data across a distributed environment. </span></p>315<p><span style="font-weight: 400;">As production functions move into software, developers can build applications that share accelerated infrastructure, connect dynamically and evolve independently. That can help media companies use infrastructure more efficiently, introduce new capabilities faster and reduce the amount of custom integration required between applications.</span></p>316<p><span style="font-weight: 400;">The integration also creates a stronger foundation for AI in live media. AI processing, video applications and traditional media functions can increasingly operate on the same accelerated infrastructure and in the same software-defined environment.</span></p>317<p><span style="font-weight: 400;">For technology vendors, this expands the opportunity to build applications that can work across broader, multi-vendor ecosystems. For media companies, it creates a path toward infrastructure that can adapt as formats, applications and AI capabilities change.</span></p>318<p><i><span style="font-weight: 400;">See the demo at IBC in </span></i><a target="_blank" href="https://directory.ibc.org/8_0/floorplan/?hallID=F&level=1&st=exhibitor&selectedBooth=booth%7E10.D21"><i><span style="font-weight: 400;">EBU Stand 10.D21</span></i></a><i><span style="font-weight: 400;">. Learn more about </span></i><a target="_blank" href="https://developer.nvidia.com/holoscan-for-media"><i><span style="font-weight: 400;">Holoscan for Media</span></i></a><i><span style="font-weight: 400;">.</span></i></p>319<hr />320<h2 id="sports-intelligence-playbooks"><b>NVIDIA Sports Intelligence Playbooks Chart a Path to Multimodal AI for Sports </b><a href="https://blogs.nvidia.com/blog/ibc-news-2026/#sports-intelligence-playbooks"><b><em><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></em></b></a><b></b><b></b></h2>321<p><img loading="lazy" decoding="async" class="alignnone size-large wp-image-98118" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-1680x945.jpg" alt="" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/09/sports-intelligence-playbooks-1920x1080-1.jpg 1920w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></p>322<p><span style="font-weight: 400;">Sports is becoming a proving ground for a broader shift in AI: from general-purpose models toward fine-tuned open models built on proprietary data.</span></p>323<p><span style="font-weight: 400;"><a target="_blank" href="https://nvidia.github.io/sports-intelligence-playbooks/latest/">NVIDIA Sports Intelligence Playbooks</a> are designed to accelerate that transition. They give leagues, media companies and technology providers structured frameworks to fine-tune NVIDIA open models on their own sports footage and annotations, creating multimodal AI that can understand the rules, players, scoring, strategy and context unique to a sport.</span></p>324<p><span style="font-weight: 400;">Sports organizations hold large volumes of proprietary video, metadata and performance information that are difficult for competitors to replicate. The playbooks provide a practical blueprint for converting those assets into AI capabilities that can underpin new analytics products, media experiences, automation tools and revenue streams.</span></p>325<p><span style="font-weight: 400;">The playbooks span the AI lifecycle, including data preparation, fine-tuning, inference, evaluation, optimization and deployment, and bring together NVIDIA technologies including </span><a target="_blank" href="https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/"><span style="font-weight: 400;">Nemotron</span></a><span style="font-weight: 400;">, </span><a target="_blank" href="https://docs.nvidia.com/nemo/automodel"><span style="font-weight: 400;">NeMo AutoModel</span></a><span style="font-weight: 400;">, </span><a target="_blank" href="https://docs.nvidia.com/nemo/megatron-bridge/latest/"><span style="font-weight: 400;">Megatron Bridge</span></a><span style="font-weight: 400;">, </span><a target="_blank" href="https://www.nvidia.com/en-us/ai-data-science/products/nim-microservices/"><span style="font-weight: 400;">NIM microservices</span></a><span style="font-weight: 400;"> and </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/solutions/accelerated-computing/"><span style="font-weight: 400;">NVIDIA accelerated computing</span></a><span style="font-weight: 400;">. </span></p>326<p><span style="font-weight: 400;">By providing an integrated path from model customization to production, Sports Intelligence Playbooks can reduce the cost and complexity of building specialized sports AI while increasing demand across its compute, software and inference stack.</span></p>327<p><span style="font-weight: 400;">Early testing demonstrates the potential of domain specialization. When evaluated on previously unseen footage using question formats similar to those used in training, multiple-choice accuracy increased from approximately 53% to 94% and open-ended evaluation from approximately 5.7% to 66%.</span></p>328<p><span style="font-weight: 400;">Machina Sports is integrating Sports Intelligence Playbooks with its sports-native data, evaluation and agent infrastructure, enabling rights holders to turn proprietary media and expertise into private, deployable intelligence for live production, content and fan experiences.</span></p>329<p><span style="font-weight: 400;">The opportunity also expands as agentic AI becomes increasingly adopted. With the </span><a target="_blank" href="https://build.nvidia.com/nvidia/aiq"><span style="font-weight: 400;">NVIDIA AI-Q Blueprint</span></a><span style="font-weight: 400;">, organizations can use their domain-specific sports models as expert intelligence within agents that reason across video, enterprise data and software systems, extending the playbook from sports understanding into decision-making and automation.</span></p>330<p><span style="font-weight: 400;">Wowza is integrating vision language models, including NVIDIA Cosmos 3 and Nemotron, into the Wowza Video Intelligence Framework, fine-tuned through NVIDIA Sports Intelligence Playbooks to detect sports-specific moments in live streams and reduce time to action. </span></p>331<p><i><span style="font-weight: 400;">Explore </span></i><a target="_blank" href="https://github.com/NVIDIA/sports-intelligence-playbooks"><i><span style="font-weight: 400;">NVIDIA Sports Intelligence Playbooks</span></i></a><i><span style="font-weight: 400;">.</span></i></p>332<hr />333<h2 id="content-localization"><b>NVIDIA Brings Multilingual Content Localization to Live Broadcast </b><a href="https://blogs.nvidia.com/blog/ibc-news-2026/#content-localization"><b><em><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></em></b></a></h2>334<p><span style="font-weight: 400;">Reaching global audiences with live programming requires more than translating words. Language nuances, voice, timing, facial movement, captions and onscreen graphics must work together in real time, while preserving the editorial intent and production quality of the original program.</span></p>335<p><span style="font-weight: 400;">To help broadcasters, sports leagues, rights holders and streaming services bring these elements into a unified, software-defined, real-time localization workflow, NVIDIA is bringing its Content Localization technologies to the NVIDIA Holoscan for Media developer toolkit. Designed for broadcast and streaming developers, the reference workflow enables captions, translated audio, dubbing, synchronized video and localized graphics.</span></p>336<p><span style="font-weight: 400;">Content Localization with Holoscan for Media provides a reference for how localization technologies can work together in software-defined broadcast applications. Developers can select the capabilities needed for each program, market or distribution channel rather than deploying separate infrastructure for every localized version.</span></p>337<p><span style="font-weight: 400;">Content Localization with Holoscan for Media incorporates the latest advancements from NVIDIA AI for Media, including improved LipSync when faces are partially obscured and enhanced Active Speaker Detection to help applications identify who’s speaking in multi-person scenes.</span></p>338<h3><b>Expanding the Reach of Live Programming</b></h3>339<p><span style="font-weight: 400;">Localization can transform the reach and economics of live programming. A shared, composable workflow can help media companies introduce regional coverage faster, serve more audiences and tailor experiences for individual markets — while preserving the timing, visual context and editorial control required for live production.</span></p>340<p><span style="font-weight: 400;">Technologies from </span><span style="font-weight: 400;">AI-Media</span><span style="font-weight: 400;">, </span><span style="font-weight: 400;">CAMB.AI</span><span style="font-weight: 400;">, Chyron and </span><span style="font-weight: 400;">Panjaya</span><span style="font-weight: 400;"> each address a specific part of content localization with Holoscan for Media, from adapting voice and onscreen delivery to creating multilingual captions and translated audio, localizing graphics, and preserving expression and identity across live and on-demand content.</span></p>341<p><span style="font-weight: 400;">The Content Localization technologies also support file-based, streaming and post-production applications. Developers can use application programming interfaces for on-demand workflows and the Holoscan for Media reference workflow when localization must run as part of a live media environment. Together, they provide a consistent foundation for building multilingual media services across production and distribution.</span></p>342<p><i><span style="font-weight: 400;">Learn more about NVIDIA </span></i><a target="_blank" href="https://developer.nvidia.com/holoscan-for-media"><i><span style="font-weight: 400;">Holoscan for Media</span></i></a><i><span style="font-weight: 400;"> and </span></i><a target="_blank" href="https://developer.nvidia.com/topics/ai/generative-ai/ai-for-media"><i><span style="font-weight: 400;">AI for Media</span></i></a><i><span style="font-weight: 400;">. </span></i></p>343]]></content:encoded>344 345 346 347 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/me-ai-for-media-kv-1920x1080-5262591.jpeg" type="image/jpeg" width="1920" height="1080">348 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/me-ai-for-media-kv-1920x1080-5262591-842x450.jpeg" width="842" height="450" />349 <media:title type="html"><![CDATA[NVIDIA Brings Real-Time AI to Broadcast, Sports and Global Streaming at IBC]]></media:title>350 <media:description type="html"></media:description>351 </media:content>352 </item>353 <item>354 <title>Sparks Fly: NVIDIA Accelerates Local AI at IFA 2026</title>355 <link>https://blogs.nvidia.com/blog/local-ai-ifa-next-gen-agents-nv-pair-rtx-spark/</link>356 357 <dc:creator><![CDATA[Gerardo Delgado]]></dc:creator>358 <pubDate>Thu, 03 Sep 2026 16:00:59 +0000</pubDate>359 <category><![CDATA[AI]]></category>360 <category><![CDATA[Agentic AI]]></category>361 <category><![CDATA[Artificial Intelligence]]></category>362 <category><![CDATA[DGX Spark]]></category>363 <category><![CDATA[Local AI]]></category>364 <category><![CDATA[NVIDIA RTX]]></category>365 <category><![CDATA[Open Source]]></category>366 <category><![CDATA[RTX AI Garage]]></category>367 <category><![CDATA[RTX Spark]]></category>368 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98045</guid>369370 <description><![CDATA[Frontier intelligence is going local. At IFA 2026, NVIDIA, Microsoft and its partners are teaming up to provide faster inference and new tools that make agents easier to set up and run locally on NVIDIA hardware. New compact NVIDIA RTX Spark Windows PCs are also coming in October to give AI enthusiasts, developers and creators […]]]></description>371 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">Frontier intelligence is going local. At IFA 2026, NVIDIA, Microsoft and its partners are teaming up to provide faster inference and new tools that make agents easier to set up and run locally on NVIDIA hardware. New compact NVIDIA RTX Spark Windows PCs are also coming in October to give AI enthusiasts, developers and creators more ways to run capable agents locally and securely. </span></p>372<p><span style="font-weight: 400;">Today’s announcements include:</span></p>373<ul>374<li><span style="font-weight: 400;">Simplified local AI support for NVIDIA GPUs is coming in Hermes Agent, OpenClaw and Perplexity Portable Computer.</span></li>375<li><span style="font-weight: 400;">Up to 1.9x faster local inference — new llama.cpp and vLLM optimizations are available now directly and through LM Studio and Ollama. </span></li>376<li><a target="_blank" href="https://www.nvidia.com/en-us/ai-on-rtx/personal-ai-router/"><span style="font-weight: 400;">NVIDIA PAIR</span></a><span style="font-weight: 400;"> — a Personal AI Router tool that intelligently distributes AI inference across the PCs on a user’s local network.</span></li>377<li><span style="font-weight: 400;">NVIDIA RTX Spark arrives in October — with new Windows PCs from Lenovo and Acer. </span><a href="https://blogs.nvidia.com/blog/gamescom-rtx-spark-pc-games-technology/"><span style="font-weight: 400;">Electronic Arts, Embark and Ubisoft</span></a><span style="font-weight: 400;"> are among the latest game publishers and developers bringing their blockbuster titles to NVIDIA RTX Spark.</span></li>378</ul>379<p><span style="font-weight: 400;">Also, August was a busy month for </span><a href="https://blogs.nvidia.com/blog/local-ai-open-source-models-agents-nemotron/#spark-perplexity"><span style="font-weight: 400;">local AI</span></a><span style="font-weight: 400;">:</span></p>380<ul>381<li><b>Nemotron 3.5 Lightning</b><span style="font-weight: 400;"> — which can run on NVIDIA RTX PCs, RTX PRO Workstations, DGX Spark and Jetson — is a 30-billion parameter model that has been launched. Get started with </span><a href="https://blogs.nvidia.com/blog/nemotron-lightning-switchyard-rtx-dgx"><span style="font-weight: 400;">Nemotron 3.5 Lightning</span></a><span style="font-weight: 400;"> today. </span></li>382<li><b>Z.ai’s </b><a target="_blank" href="https://z.ai/blog/glm-5.3-flash"><span style="font-weight: 400;">GLM-5.3-Flash</span></a><span style="font-weight: 400;"> is a multimodal mixture-of-experts (MoE) model that’s bringing agentic AI to DGX Station.</span></li>383<li><b>Qwen</b><span style="font-weight: 400;"> has released </span><a target="_blank" href="https://qwen.ai/blog?id=qwen3.8-flash-next"><span style="font-weight: 400;">Qwen3.8-Flash-Next</span></a><span style="font-weight: 400;">, an open weight multimodal MoE model, which can run locally on DGX Spark and DGX Station, along with </span><a target="_blank" href="https://huggingface.co/Qwen/Qwen3.8-27B"><span style="font-weight: 400;">Qwen3.8-27B</span></a><span style="font-weight: 400;">, a 27-billion-parameter open model optimized for local agentic and coding workloads on NVIDIA GPUs. </span></li>384<li><b>LTX’s </b><a target="_blank" href="https://ltx.io/model/ltx-2-5"><b>LTX 2.5</b></a><span style="font-weight: 400;"> is an open-world video generation model optimized for NVIDIA RTX GPUs, DGX Spark and DGX Station, with new NVFP4, FastVideo and ComfyUI enhancements for faster, more memory-efficient local generation.</span></li>385<li><a target="_blank" href="https://huggingface.co/MiniMaxAI/MiniMax-H3"><b>MiniMax-H3</b></a> <span style="font-weight: 400;">is an open-weight video generation model with synchronized audio that can run locally on NVIDIA GPUs through ComfyUI. FastVideo teamed up with NVIDIA researchers to improve this further by releasing FastH3 — an open-weight, four-step distilled version that improves performance by 7x. Optimized FastVideo recipes for NVIDIA RTX GPUs and DGX Spark are coming soon.</span></li>386<li><b>Meta’s </b><a target="_blank" href="https://developer.nvidia.com/blog/?p=121045"><b>Muse Glimmer</b></a><span style="font-weight: 400;"> is a 30-billion-parameter open-weight model for coding and agentic workloads that can run locally on GeForce RTX PCs, DGX Spark, DGX Station and Jetson. NVIDIA has also released </span><a target="_blank" href="https://huggingface.co/nvidia/Muse-Glimmer-30B-NVFP4"><span style="font-weight: 400;">NVFP4 quantization</span></a><span style="font-weight: 400;"> with DGX Spark support for more memory-efficient local deployment.</span></li>387<li><a target="_blank" href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731"><b>DeepSeek v4 Flash</b></a><span style="font-weight: 400;"> is a 284-billion-parameter MoE model with 13 billion active parameters that can run locally on 2x DGX Spark cluster and DGX Station.</span></li>388</ul>389<h2><b>A Simpler Start for Local Agents</b></h2>390<p><span style="font-weight: 400;">Getting a local agent up and running with local models required some effort — choosing a model, finding a compatible inference server, dialing in quantization settings and keeping everything updated. That friction is disappearing on RTX and DGX systems.</span></p>391<p><span style="font-weight: 400;">Three of the most widely used agent apps will offer simplified local model setup on Windows, each built on llama.cpp and incorporating NVIDIA’s latest inference optimizations. The new setup experiences are designed to reduce manual configuration and make it easier to get local agents up and running. </span></p>392<p><span style="font-weight: 400;">Last month, Perplexity introduced its </span><a target="_blank" href="https://www.perplexity.ai/hub/blog/introducing-portable-computer-for-local-first-ai"><span style="font-weight: 400;">Portable Computer agent</span></a><span style="font-weight: 400;">, giving users a simple way to run Perplexity locally on Linux systems like </span><a target="_blank" href="https://www.nvidia.com/en-us/products/workstations/dgx-spark/"><span style="font-weight: 400;">NVIDIA DGX Spark</span></a><span style="font-weight: 400;"> with the models, orchestration and tools packaged into a single app experience.</span></p>393<p><span style="font-weight: 400;"><b>Perplexity Portable Computer</b> is available on NVIDIA RTX GPUs with at least 24GB VRAM running on Linux, with support on Windows coming soon, bringing that same streamlined setup to a broader group of PC users. Users can run complete workflows locally without consuming credits, while selectively escalating parts of a task to one of 15+ frontier models in the cloud when additional research or reasoning is needed. Portable Computer asks for permission before sending content to the cloud, helping users keep sensitive information on their device. Here’s some example use-cases:</span></p>394<ul>395<li><b>Engineering: </b><span style="font-weight: 400;">Review open PRs in a connected GitHub repo and sort them into ready, blocked, stale, and needs review, each tagged with the next step. Docs that fell out of sync with the latest merge get caught and fixed, with a PR opened for the changes.</span></li>396<li><b>Finance:</b><span style="font-weight: 400;"> Point the agent at two years of brokerage summaries, consolidated 1099s and tax returns and have it trace the recurring holdings creating the most avoidable fees and tax drag, with every figure cited to the exact file and page — all without a document ever reaching a chatbot.</span></li>397<li><b>Startups:</b><span style="font-weight: 400;"> Ask why activation went flat, and the agent analyzes the funnel export locally to find where new signups drop off between install and first completed task, then posts the top insights straight to the team’s Slack channel.</span></li>398</ul>399<p><span style="font-weight: 400;">Try </span><a target="_blank" href="https://www.perplexity.ai/hub/products/portable-computer"><span style="font-weight: 400;">Portable Computer</span></a><span style="font-weight: 400;"> today.</span></p>400<p><b>Hermes Agent — developed by </b><a target="_blank" href="https://nousresearch.com/"><b>Nous Research</b></a><b> — </b><span style="font-weight: 400;">is a general-purpose agent used by millions that excels at reliability and self-improvement. Model- and provider-agnostic, Hermes is built to run all day on local systems, making RTX PCs, RTX PRO workstations and DGX Spark a natural fit.</span></p>401<p><span style="font-weight: 400;">Configuring a local model in Hermes will provide users with one-click setup across RTX and DGX systems on Windows. The agent will automatically detect the NVIDIA GPU, select an appropriate model and configuration, and run it through integrated llama.cpp with NVIDIA inference optimizations already in place, eliminating manual model downloads and tuning. Support for Linux is coming soon.</span></p>402<p><span style="font-weight: 400;">Once it is running, Hermes works the way it does anywhere else. It uses tools, maintains context across tasks, remembers information between sessions and creates reusable skills over time, allowing the agent to become more capable with continued use. Running the model locally on a GPU keeps performance fast while keeping data on the system.</span></p>403<p><iframe loading="lazy" title="The Easiest Way to Run Hermes Locally | @Nousresearch" width="563" height="1000" src="https://www.youtube.com/embed/TaqNUvMCRBs?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>404<p><span style="font-weight: 400;">One-click local model setup is available now on Windows, with support coming soon to Linux. Learn more about </span><a target="_blank" href="https://hermes-agent.nousresearch.com/"><span style="font-weight: 400;">Hermes Agent</span></a><span style="font-weight: 400;">.</span></p>405<p><a target="_blank" href="https://openclaw.ai/"><b>OpenClaw </b></a><span style="font-weight: 400;">has become one of the defining projects of the open-agent movement — the largest AI project on GitHub, with more than 380K stars and a fast-growing community that’s building tools and skills across research, engineering, project management and everyday productivity.</span></p>406<p><span style="font-weight: 400;">NVIDIA, Microsoft and OpenClaw have been working together to make that experience easier to set up on Windows PCs. To reduce onboarding friction, the OpenClaw Windows App simplifies the process of setting up an optimized local model on any RTX GPU with at least 24GB of VRAM.</span></p>407<p><span style="font-weight: 400;">Learn more in the <a target="_blank" href="https://openclaw.ai/blog/macos-installer-windows-local-ai">OpenClaw blog</a></span><span style="font-weight: 400;">.</span></p>408<h2><b>Faster Inference Gives Local Agents a Boost</b></h2>409<p><span style="font-weight: 400;">Inference performance is critical to keeping local agents responsive. NVIDIA is continuing to collaborate with the open-source llama.cpp and vLLM communities to accelerate agentic workloads across local NVIDIA platforms.</span></p>410<p><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/local-ai-ifa-blog-body.jpg"><img loading="lazy" decoding="async" class="aligncenter size-full wp-image-98047" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/local-ai-ifa-blog-body.jpg" alt="" width="1280" height="680" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/local-ai-ifa-blog-body.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/local-ai-ifa-blog-body-960x510.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/local-ai-ifa-blog-body-630x335.jpg 630w" sizes="auto, (max-width: 1280px) 100vw, 1280px" /></a></p>411<p><span style="font-weight: 400;">llama.cpp delivers up to 1.9x higher throughput through kernel optimizations on a GeForce RTX 5090, enhanced speculative decoding techniques and faster prefill. </span></p>412<p><span style="font-weight: 400;">vLLM delivers 1.2x on </span><a target="_blank" href="https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-6000/"><span style="font-weight: 400;">RTX PRO 6000 Blackwell Workstation Edition</span></a><span style="font-weight: 400;"> and up to 1.4x on two DGX Spark clusters. New XQA attention kernels in FlashInfer and backend optimizations help to accelerate inference across both platforms.</span></p>413<p><span style="font-weight: 400;">These gains are available on the llama.cpp and vLLM inferencing backends. </span></p>414<p><span style="font-weight: 400;">Users can also experience these via the LM Studio and Ollama applications.</span></p>415<h2><b>Tap Idle PCs for More Local AI Compute With NVIDIA PAIR</b></h2>416<p><span style="font-weight: 400;">More than half of U.S. households have two or more PCs, and much of that computing power sits idle throughout the day. </span><a target="_blank" href="https://www.nvidia.com/en-us/ai-on-rtx/personal-ai-router/"><span style="font-weight: 400;">NVIDIA Personal AI Router (PAIR)</span></a><span style="font-weight: 400;"> is a free, open source software tool that puts those systems to work together for local AI.</span></p>417<p><iframe loading="lazy" title="NVIDIA PAIR: Hermes 5-Subagent Demo" width="1200" height="675" src="https://www.youtube.com/embed/GjGM-ZKQMa0?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>418<p><span style="font-weight: 400;">Agentic workflows often break complex tasks into smaller jobs that can run in parallel, but performance can slow when every request is competing for the same GPU. PAIR automatically discovers compatible PCs on a local network and routes independent inference requests to whichever system has capacity. It works with Ollama and LM Studio and can adapt as devices join or leave the network.</span></p>419<p><span style="font-weight: 400;">For example, a user could ask Hermes to create a “Sunday Reset” plan by sorting through a cluttered inbox and prioritizing what needs attention now, what can wait and what can be skipped. Hermes can split that work across multiple subagents, while PAIR distributes those jobs across available PCs instead of having them all wait on a single GPU.</span></p>420<p><span style="font-weight: 400;">The result is more compute for local agents, with more tasks running in parallel and the flexibility to move AI workloads to another PC while the main system is being used for gaming, creating or other work.</span></p>421<p><span style="font-weight: 400;">The NVIDIA PAIR beta is available for Windows, macOS and Linux through both graphical and terminal interfaces, supporting NVIDIA GeForce RTX 20 Series GPUs and newer, </span><a target="_blank" href="https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/"><span style="font-weight: 400;">NVIDIA RTX PRO workstation GPUs</span></a><span style="font-weight: 400;"> (Turing architecture and newer), NVIDIA DGX Spark and Apple M4 or newer silicon.</span></p>422<p><span style="font-weight: 400;">Check out the </span><a target="_blank" href="http://developer.nvidia.com/blog/nvidia-pair-virtual-inference-router-expands-available-compute-on-your-local-network"><span style="font-weight: 400;">NVIDIA tech blog</span></a><span style="font-weight: 400;"> to get started with NVIDIA PAIR. </span></p>423<h2><b>Powerful On Device Photo Editing With Cyberlink PhotoDirector AI PC Mode on RTX Spark</b></h2>424<p><span style="font-weight: 400;">Open image and video models enable artists to experiment with Creative AI models on PCs. This enables artists to iterate and explore concepts and ideas, without the dreaded token anxiety and keep more of their creative work private and on-device.</span></p>425<p><span style="font-weight: 400;">CyberLink’s new PhotoDirector AI PC Mode is one of the first applications to integrate these diffusion models directly into a creative software, and turn them into a creative tool at the finger tips of the artists. Coming to PhotoDirector 365 and optimized for NVIDIA RTX Spark when it launches, AI PC Mode users are getting AI-powered editing tools for generative editing, image enhancement, object and distraction removal, background removal and replacement, portrait refinement and the creation of entirely new visuals — with the flexibility to choose between local or cloud processing, depending on the task.</span></p>426<p><span style="font-weight: 400;">On NVIDIA GPUs, PhotoDirector uses TensorRT-RTX and FP8 to accelerate local AI.</span></p>427<p><span style="font-weight: 400;">Start using </span><a target="_blank" href="https://www.cyberlink.com/products/photodirector-photo-editing-software-365/overview_en_US.html"><span style="font-weight: 400;">Cyberlink’s PhotoDirector 365 photo editing software</span></a><span style="font-weight: 400;"> and learn more about PhotoDirector AI PC Mode, launching with RTX Spark in October.</span></p>428<h2><b>NVIDIA RTX Spark Windows PCs Arrive October 2026</b></h2>429<p><a target="_blank" href="https://www.nvidia.com/en-us/products/rtx-spark/"><span style="font-weight: 400;">NVIDIA RTX Spark</span></a><span style="font-weight: 400;"> is coming this October— and at IFA 2026, partners are showing off their hardware. At IFA, newly announced designs join the existing six OEMs shipping in October. Acer showed its compact desktop RTX Spark concept, and Lenovo announced its Yoga Pro 9n and Yoga 9n 2-in-1. </span></p>430<p><span style="font-weight: 400;">RTX Spark is a new beginning for Windows PCs. One PC built for creators, gamers and AI agents. With a powerful 1 Petaflop RTX Blackwell GPU, up to 128GB of unified memory and a highly efficient 20-core Grace CPU, RTX Spark delivers incredible performance and efficiency. This superchip enables high performance thin laptops with all day battery life and compact desktops to power always-on agents. Paired with the new Windows Agent framework, it enables agents that run safely in the background under OS level control.</span></p>431<p><span style="font-weight: 400;">Last week at Gamescom, Electronic Arts, Embark and Ubisoft were among the latest game publishers and developers bringing their blockbuster titles to NVIDIA RTX Spark Windows PCs. They join the publishers that announced RTX Spark support at COMPUTEX in May, including KRAFTON, NetEase, Riot Games and XBOX. </span><a href="https://blogs.nvidia.com/blog/gamescom-rtx-spark-pc-games-technology/"><span style="font-weight: 400;">Read more</span></a><span style="font-weight: 400;">.</span></p>432<p><a target="_blank" href="https://www.nvidia.com/en-us/products/rtx-spark/#notify-me"><span style="font-weight: 400;">Sign up</span></a><span style="font-weight: 400;"> to be notified when RTX Spark laptops and desktops are available.</span></p>433<h2><b>#ICYMI: More Updates From NVIDIA Local AI </b></h2>434<p><span style="font-weight: 400;"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f3ae.png" alt="🎮" class="wp-smiley" style="height: 1em; max-height: 1em;" /></span><b> NVIDIA Brings New RTX Tech and Games to Gamescom </b><span style="font-weight: 400;">— NVIDIA released </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/news/gamescom-2026-nvidia-geforce-rtx-dlss-4-5-announcements/?utm_source=chatgpt.com"><span style="font-weight: 400;">DLSS 4.5 Ray Reconstruction</span></a><span style="font-weight: 400;">, featuring a new second-generation transformer model for improved image quality in ray-traced and path-traced games. Gamescom also brought </span><a href="https://blogs.nvidia.com/blog/gamescom-rtx-spark-pc-games-technology/"><span style="font-weight: 400;">new RTX announcements </span></a><span style="font-weight: 400;">for titles including 007 First Light, CONTROL Resonant and Gears of War: E-Day, plus expanded game support for the upcoming NVIDIA RTX Spark.</span></p>435<p><b><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f40b.png" alt="🐋" class="wp-smiley" style="height: 1em; max-height: 1em;" />Introducing DeepSeek Harness — </b><span style="font-weight: 400;">DeepSeek’s new </span><a target="_blank" href="https://github.com/deepseek-ai/deepseek-harness"><span style="font-weight: 400;">open source harness</span></a><span style="font-weight: 400;"> pairs with </span><a target="_blank" href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731"><span style="font-weight: 400;">DeepSeek-V4-Flash</span></a><span style="font-weight: 400;"> to power local agentic coding workflows on NVIDIA DGX Station and multi-DGX Spark setups.</span></p>436<p><span style="font-weight: 400;"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f4ca.png" alt="📊" class="wp-smiley" style="height: 1em; max-height: 1em;" /></span><b>MLPerf Client v2.0 Expands AI PC Benchmarking</b><span style="font-weight: 400;"> — MLCommons released </span><a target="_blank" href="https://mlcommons.org/2026/08/mlperf-client-v2-0/"><span style="font-weight: 400;">MLPerf Client v2.0</span></a><span style="font-weight: 400;">, developed in collaboration with NVIDIA and other industry leaders. The update adds new benchmarks for agentic AI and image generation, alongside expanded LLM testing for real-world local AI workloads.</span></p>437<p><i><span style="font-weight: 400;">Follow NVIDIA RTX Spark on </span></i><a target="_blank" href="https://x.com/NVIDIARTXSpark"><i><span style="font-weight: 400;">X</span></i></a><i><span style="font-weight: 400;">, </span></i><a target="_blank" href="https://www.instagram.com/nvidiartxspark/"><i><span style="font-weight: 400;">Instagram</span></i></a><i><span style="font-weight: 400;">, </span></i><a target="_blank" href="https://www.tiktok.com/@nvidiartxspark"><i><span style="font-weight: 400;">TikTok</span></i></a><i><span style="font-weight: 400;"> and </span></i><a target="_blank" href="https://www.facebook.com/NVIDIARTXSpark"><i><span style="font-weight: 400;">Facebook</span></i></a><i><span style="font-weight: 400;"> — and stay informed by subscribing to the </span></i><a target="_blank" href="https://www.nvidia.com/en-us/ai-on-rtx/?modal=subscribe-ai"><i><span style="font-weight: 400;">NVIDIA Local AI newsletter</span></i></a><i><span style="font-weight: 400;">. Follow NVIDIA Workstation on </span></i><a target="_blank" href="https://www.linkedin.com/showcase/3761136/"><i><span style="font-weight: 400;">LinkedIn</span></i></a><i><span style="font-weight: 400;"> and</span></i><a target="_blank" href="https://x.com/NVIDIAworkstatn"><i><span style="font-weight: 400;"> X</span></i></a><i><span style="font-weight: 400;">. </span></i></p>438<p><i><span style="font-weight: 400;">See</span></i><a target="_blank" href="https://www.nvidia.com/en-eu/about-nvidia/terms-of-service/"> <i><span style="font-weight: 400;">notice</span></i></a><i><span style="font-weight: 400;"> regarding software product information.</span></i></p>439]]></content:encoded>440 441 442 443 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/nv-blog-1280x680-1.jpg" type="image/jpeg" width="1280" height="680">444 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/nv-blog-1280x680-1-842x450.jpg" width="842" height="450" />445 <media:title type="html"><![CDATA[Sparks Fly: NVIDIA Accelerates Local AI at IFA 2026]]></media:title>446 <media:description type="html"></media:description>447 </media:content>448 </item>449 <item>450 <title>‘NBA 2K27’ With NVIDIA DLSS 5 Leads 28 New Games Coming to GeForce NOW</title>451 <link>https://blogs.nvidia.com/blog/geforce-now-thursday-september-2026-games-list/</link>452 453 <dc:creator><![CDATA[GeForce NOW Community]]></dc:creator>454 <pubDate>Thu, 03 Sep 2026 13:00:35 +0000</pubDate>455 <category><![CDATA[Gaming]]></category>456 <category><![CDATA[Cloud Gaming]]></category>457 <category><![CDATA[GeForce NOW]]></category>458 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98028</guid>459460 <description><![CDATA[September is here with 28 more games streaming on GeForce NOW this month, led by a slam dunk: NBA 2K27 with the NVIDIA DLSS 5 3D-Guided Neural Rendering feature. Through NVIDIA’s close collaboration with Visual Concepts and 2K, DLSS 5 brings a new level of lifelike lighting and material detail to the court — tuned […]]]></description>461 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400">September is here with 28 more games streaming on </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=24109411186&gbraid=0AAAAAD4XAoGPVGKakiCteqkRtoF1Q-KSI&gclid=CjwKCAjws_DTBhB_EiwAXZknGWWMZ3DFuAXLiR7iw_87oxG2j9kfjeNAchFry7mWxbg7xcErpKfKNhoCH6cQAvD_BwE"><span style="font-weight: 400">GeForce NOW</span></a><span style="font-weight: 400"> this month, led by a slam dunk: </span><i><span style="font-weight: 400">NBA 2K27</span></i><span style="font-weight: 400"> with the </span><a target="_blank" href="https://nvidianews.nvidia.com/news/nvidia-dlss-5-delivers-ai-powered-breakthrough-in-visual-fidelity-for-games"><span style="font-weight: 400">NVIDIA DLSS 5</span></a><span style="font-weight: 400"> 3D-Guided Neural Rendering feature.</span></p>462<p><span style="font-weight: 400">Through </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/news/dlss-5-3d-guided-neural-rendering/"><span style="font-weight: 400">NVIDIA’s close collaboration</span></a><span style="font-weight: 400"> with Visual Concepts and 2K, DLSS 5 brings a new level of lifelike lighting and material detail to the court — tuned by developers to support the game’s authentic broadcast-style presentation.</span></p>463<p><span style="font-weight: 400">Members can also walk the line between human and vampire in Rebel Wolves’ </span><i><span style="font-weight: 400">The Blood of Dawnwalker</span></i><span style="font-weight: 400"> and step into genma-haunted Kyoto in Capcom’s </span><i><span style="font-weight: 400">Onimusha: Way of the Sword</span></i><span style="font-weight: 400"> — part of the 10 games joining the </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/games/"><span style="font-weight: 400">GeForce NOW library</span></a><span style="font-weight: 400"> this week.</span></p>464<h2><b>Take the Court With DLSS 5</b></h2>465<p><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-scaled.jpg"><img loading="lazy" decoding="async" class="aligncenter size-large wp-image-98065" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-1680x945.jpg" alt="NVIDIA GeForce NOW NBA2K DLSS 5" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-scaled.jpg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-off-400x225.jpg 400w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a></p>466<figure id="attachment_98066" aria-describedby="caption-attachment-98066" style="width: 1200px" class="wp-caption aligncenter"><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-scaled.jpg"><img loading="lazy" decoding="async" class="size-large wp-image-98066" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-1680x945.jpg" alt="GeForce NOW NBA2K DLSS5" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-scaled.jpg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/09/nba-2k27-nvidia-dlss-5-comparison-001-on-400x225.jpg 400w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a><figcaption id="caption-attachment-98066" class="wp-caption-text">Showcase your moves.</figcaption></figure>467<p><span style="font-weight: 400">NVIDIA,</span> <span style="font-weight: 400">Visual Concepts and 2K have brought DLSS 5 3D-Guided Neural Rendering to </span><i><span style="font-weight: 400">NBA 2K27</span></i><span style="font-weight: 400">, delivering a new level of sports realism for players streaming from the cloud at launch. </span></p>468<p><span style="font-weight: 400">DLSS 5 enhances and tunes lighting and materials to make the action feel even more lifelike — from how light catches hair and skin to the detail preserved in the facial geometry of real-world athletes — achieving new levels of authenticity. </span></p>469<p><span style="font-weight: 400">GeForce NOW Ultimate members in NVIDIA-operated regions can stream </span><i><span style="font-weight: 400">NBA 2K27</span></i><span style="font-weight: 400"> across PCs, Macs, handhelds, mobile devices, TVs and more. DLSS 5 is available when streaming from a GeForce RTX 5080-powered rig in the cloud.</span></p>470<p><span style="font-weight: 400">Tip off without waiting for downloads or managing storage, and experience </span><i><span style="font-weight: 400">NBA 2K27</span></i><span style="font-weight: 400"> from the cloud with the visual fidelity of DLSS 5.</span></p>471<h2><b>Walk the Line Between Day and Night</b></h2>472<figure id="attachment_98067" aria-describedby="caption-attachment-98067" style="width: 1200px" class="wp-caption aligncenter"><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker.jpg"><img loading="lazy" decoding="async" class="size-large wp-image-98067" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-1680x840.jpg" alt="The Blood of Dawnwalker on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_The_Blood_of_Dawnwalker.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a><figcaption id="caption-attachment-98067" class="wp-caption-text">Human by day. Vampire by night.</figcaption></figure>473<p><span style="font-weight: 400">Rebel Wolves’ newly launched </span><i><span style="font-weight: 400">The Blood of Dawnwalker</span></i><span style="font-weight: 400"> is streaming from GeForce NOW. The open-world, dark-fantasy action role-playing game transports players to 14th-century Europe, where war, plague and rising vampire powers have reshaped history.</span></p>474<p><span style="font-weight: 400">Play as Coen, a young Dawnwalker fighting to save his family while caught between his humanity and cursed vampiric strength. His abilities — and the choices available to him — change between day and night, shaping how players explore, fight and uncover the secrets of the world around them.</span></p>475<p><span style="font-weight: 400">Sink teeth into the fresh adventure from nearly any supported device, with Ultimate members tapping into GeForce RTX 5080-class performance in the cloud. Skip the installs and let GeForce NOW handle the heavy lifting, leaving more time to take a bite out of everything Coen’s new world has to offer.</span></p>476<h2><b>Skip the Downloads and Sharpen the Sword</b></h2>477<figure id="attachment_98068" aria-describedby="caption-attachment-98068" style="width: 1200px" class="wp-caption aligncenter"><a href="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword.jpg"><img loading="lazy" decoding="async" class="size-large wp-image-98068" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-1680x840.jpg" alt="Onimusha: Way of the Sword on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/09/GFN_Thursday_Onimusha_Way_of_the_Sword.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a><figcaption id="caption-attachment-98068" class="wp-caption-text">Pick up the Oni Gauntlet.</figcaption></figure>478<p><span style="font-weight: 400">Capcom’s </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-onimusha-coming/"><i><span style="font-weight: 400">Onimusha: Way of the Sword</span></i></a> <span style="font-weight: 400">arrives on GeForce NOW at launch, giving members a way to jump straight into the series’ return instantly from the cloud.</span></p>479<p><span style="font-weight: 400">Fight across a twisted vision of Edo-period Kyoto as a lone samurai wielding the mystical Oni Gauntlet. Master precise swordplay, unleash powerful abilities and absorb the souls of fallen enemies while battling relentless Genma beneath mysterious clouds of Malice.</span></p>480<p><span style="font-weight: 400">Ultimate members can stream every clash with GeForce RTX 5080-class performance in the cloud across PCs, Macs, handhelds, mobile devices, TVs, the newly supported Firefox browser and more. Cut to the chase by skipping any downloads and storage management and step into battle.</span></p>481<p><span style="font-weight: 400">In addition, members can look for the following titles streaming this week:</span></p>482<ul>483<li><i><span style="font-weight: 400">Breathedge 2</span></i><span style="font-weight: 400"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/2412960?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Aug. 31)</span></li>484<li><i><span style="font-weight: 400">Serious Sam: Shatterverse </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2067210?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Aug. 31)</span></li>485<li><i><span style="font-weight: 400">Crimson Moon </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/4317690?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 1)</span></li>486<li><i><span style="font-weight: 400">The Blood of Dawnwalker </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3751260?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 2)</span></li>487<li><i><span style="font-weight: 400">NBA 2K27</span></i><span style="font-weight: 400"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/4356430/NBA_2K27/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 3)</span></li>488<li><i><span style="font-weight: 400">SpeedRunners 2: King of Speed </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3183760?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400"> and </span><a target="_blank" href="https://www.xbox.com/games/store/speedrunners-2-king-of-speed/9ppkr9n31bw6?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox, available on Game Pass</span></a><span style="font-weight: 400">, Sept. 3)</span></li>489<li><i><span style="font-weight: 400">Onimusha: Way of the Sword </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2638890?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 4)</span></li>490<li><i><span style="font-weight: 400">How to Fish </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/4001890?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>491<li><i><span style="font-weight: 400">Shelves and Sorcery: Tidy Up the Enchanted Shop </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/3614130?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>492<li><i><span style="font-weight: 400">VHOLUME </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/4131730?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>493</ul>494<p><span style="font-weight: 400">And look forward to the games coming throughout the month:</span></p>495<ul>496<li style="font-weight: 400"><i><span style="font-weight: 400">Bus Simulator 27 </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2397320/Bus_Simulator_27/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>497<li style="font-weight: 400"><i><span style="font-weight: 400">Halloween: The Game </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3219630/Halloween_The_Game/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>498<li style="font-weight: 400"><i><span style="font-weight: 400">Honeycomb: The World Beyond </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/1510440/Honeycomb_The_World_Beyond/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>499<li style="font-weight: 400"><i><span style="font-weight: 400">Wanderburg </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3624140/Wanderburg/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 8)</span></li>500<li style="font-weight: 400"><i><span style="font-weight: 400">Shroom and Gloom </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3271280/Shroom_and_Gloom/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>501<li style="font-weight: 400"><i><span style="font-weight: 400">WARDOGS </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/1867240/WARDOGS/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>502<li style="font-weight: 400"><i><span style="font-weight: 400">Welcome to Elderfield </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3195440/Welcome_to_Elderfield/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 10)</span></li>503<li style="font-weight: 400"><i><span style="font-weight: 400">Active Matter </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2887580/Active_Matter/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 15)</span></li>504<li style="font-weight: 400"><i><span style="font-weight: 400">Aniimo </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/4126040/Aniimo/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 15)</span></li>505<li style="font-weight: 400"><i><span style="font-weight: 400">Train Sim World 7 </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/4678800/Train_Sim_World_7/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 15)</span></li>506<li style="font-weight: 400"><i><span style="font-weight: 400">The Guild 1 Remake: Europa 1410 </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2977260/The_Guild_1_Remake_Europa_1410/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 17)</span></li>507<li style="font-weight: 400"><i><span style="font-weight: 400">CONTROL Resonant </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3669870/CONTROL_Resonant/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 24)</span></li>508<li style="font-weight: 400"><i><span style="font-weight: 400">Nivalis Nights </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/1488490/Nivalis_Nights/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Sept. 29)</span></li>509<li style="font-weight: 400"><i><span style="font-weight: 400">Mixtape</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/2582320/Mixtape/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">) </span></li>510<li style="font-weight: 400"><i><span style="font-weight: 400">Neon White</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/1533420/Neon_White/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">) </span></li>511<li style="font-weight: 400"><i><span style="font-weight: 400">Outer Wilds</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/753640/Outer_Wilds/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">) </span></li>512<li style="font-weight: 400"><i><span style="font-weight: 400">Stray </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/1332010/Stray/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>513<li style="font-weight: 400"><i><span style="font-weight: 400">What Remains of Edith Finch</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/501300/What_Remains_of_Edith_Finch/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>514</ul>515<h2><b>All the Extras From August</b></h2>516<p><span style="font-weight: 400">In addition to the 26 games </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-august-2026-games-list/"><span style="font-weight: 400">announced</span></a><span style="font-weight: 400"> last month, 18 more joined the GeForce NOW library. </span></p>517<ul>518<li style="font-weight: 400"><i><span style="font-weight: 400">Cat Mail Co. </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/4380490?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>519<li style="font-weight: 400"><i><span style="font-weight: 400">Monster Hunter Wilds Prologue Demo</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>520<li style="font-weight: 400"><i><span style="font-weight: 400">Stars Reach</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/1925650?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>521<li style="font-weight: 400"><i><span style="font-weight: 400">DIVE or DIE – Children of Rain</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/3590290?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>522<li style="font-weight: 400"><i><span style="font-weight: 400">Escape the Backroom</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.xbox.com/games/store/escape-the-backrooms/9mw15l8frt9p?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox</span></a><span style="font-weight: 400">, available on Game Pass)</span></li>523<li style="font-weight: 400"><i><span style="font-weight: 400">Hell Is Us</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.xbox.com/games/store/hell-is-us/9n9097g0xzmb?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox</span></a><span style="font-weight: 400">, available on Game Pass)</span></li>524<li style="font-weight: 400"><i><span style="font-weight: 400">Parcel Simulator</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://store.steampowered.com/app/2424010?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>525<li style="font-weight: 400"><i><span style="font-weight: 400">The Thaumaturge</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.xbox.com/games/store/the-thaumaturge/9mzzzchz26nk?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox</span></a><span style="font-weight: 400">, available on Game Pass)</span></li>526<li style="font-weight: 400"><i><span style="font-weight: 400">Tomb Raider IV-VI Remastered</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.epicgames.com/store/p/tomb-raider-ivvi-remastered-0be789?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Epic Games Store</span></a><span style="font-weight: 400">)</span></li>527<li style="font-weight: 400"><i><span style="font-weight: 400">STAR WARS Zero Company</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.ea.com/games/starwars/zero-company?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">EA app</span></a><span style="font-weight: 400"> and </span><a target="_blank" href="https://store.steampowered.com/app/2075800?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>528<li style="font-weight: 400"><i><span style="font-weight: 400">Baldur’s Gate 3 </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/baldurs_gate_iii?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>529<li style="font-weight: 400"><i><span style="font-weight: 400">Clair Obscur: Expedition 33</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.gog.com/game/clair_obscur_expedition_33?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>530<li style="font-weight: 400"><i><span style="font-weight: 400">Cold Fear </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/cold_fear?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>531<li style="font-weight: 400"><i><span style="font-weight: 400">IRON NEST: Heavy Turret Simulator </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/2950790?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>532<li style="font-weight: 400"><i><span style="font-weight: 400">Kingdom Come: Deliverance II </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/kingdom_come_deliverance_ii?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>533<li style="font-weight: 400"><i><span style="font-weight: 400">Mount & Blade: Bannerlord II </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/mount_blade_ii_bannerlord?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>534<li style="font-weight: 400"><i><span style="font-weight: 400">No Man’s Sky</span></i><span style="font-weight: 400"> (</span><a target="_blank" href="https://www.gog.com/game/no_mans_sky?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>535<li style="font-weight: 400"><i><span style="font-weight: 400">The Witcher 3: Wild Hunt </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/the_witcher_3_wild_hunt?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>536</ul>537<p><span style="font-weight: 400">Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after release and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.</span></p>538<p><span style="font-weight: 400">New to GeForce NOW? Start with a </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=24109411186&gbraid=0AAAAAD4XAoGPVGKakiCteqkRtoF1Q-KSI&gclid=CjwKCAjws_DTBhB_EiwAXZknGUBjXpM5vNQvnwLXUGilFsIHBSkbKPfllg7RGpFFuQZv2GFityJC6BoCq8MQAvD_BwE"><span style="font-weight: 400">day pass</span></a><span style="font-weight: 400"> to try premium cloud gaming before committing to a membership. Better yet, the cost of the day pass can be applied toward a first monthly membership — making it easy to try GeForce RTX-powered cloud gaming with the latest releases, then level up for more.</span></p>539<p><span style="font-weight: 400">What are you planning to play this weekend? Let us know on </span><a target="_blank" href="https://www.twitter.com/nvidiagfn"><span style="font-weight: 400">X</span></a><span style="font-weight: 400"> or in the comments below.</span></p>540]]></content:encoded>541 542 543 544 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/gfn-thursday-9-3-blog-1920x1080-logo.jpg" type="image/jpeg" width="1920" height="1080">545 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/gfn-thursday-9-3-blog-1920x1080-logo-842x450.jpg" width="842" height="450" />546 <media:title type="html"><![CDATA[‘NBA 2K27’ With NVIDIA DLSS 5 Leads 28 New Games Coming to GeForce NOW]]></media:title>547 <media:description type="html"></media:description>548 </media:content>549 </item>550 <item>551 <title>NVIDIA to Acquire Hugging Face</title>552 <link>https://blogs.nvidia.com/blog/nvidia-to-acquire-hugging-face/</link>553 554 <dc:creator><![CDATA[黄仁勋]]></dc:creator>555 <pubDate>Thu, 03 Sep 2026 11:56:49 +0000</pubDate>556 <category><![CDATA[AI]]></category>557 <category><![CDATA[Corporate]]></category>558 <category><![CDATA[Software]]></category>559 <category><![CDATA[Artificial Intelligence]]></category>560 <category><![CDATA[High-Performance Computing]]></category>561 <category><![CDATA[Open Source]]></category>562 <category><![CDATA[Supercomputing]]></category>563 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98049</guid>564565 <description><![CDATA[I’m excited to announce that NVIDIA has agreed to acquire Hugging Face for $12,930,300,000. Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide. Over the past decade, Clem, Julien, Thomas and the team at Hugging Face have built something remarkable: a vibrant home for […]]]></description>566 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p>I’m excited to announce that NVIDIA has agreed to acquire Hugging Face for $12,930,300,000. Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide.</p>567<p>Over the past decade, Clem, Julien, Thomas and the team at Hugging Face have built something remarkable: a vibrant home for the <a target="_blank" href="https://www.nvidia.com/en-us/glossary/open-models/">open model</a> developer community.</p>568<p>More than 18 million developers, researchers and creators use Hugging Face to share more than 3 million models, 500,000 datasets and 1 million applications. More than 200,000 companies use the platform to discover, evaluate, customize and deploy AI.</p>569<p>Hugging Face will remain an open platform for the entire AI ecosystem. Developers will choose the models they want, the frameworks they want, the clouds and inference service providers they want and the computing platforms they want. NVIDIA compute will not be required to build on or deploy through Hugging Face.</p>570<p>Hugging Face will continue to support open source and open weight models from across the ecosystem, from every model builder. It will continue to support multi-cloud and multi-accelerator development and deployment, so builders can use the hardware and infrastructure that best fit their work.</p>571<p>Recently, I coauthored an open letter on the importance of open weights to the AI economy. Joined by leaders from across the industry, we made a simple point: open weights broaden access to AI and help ensure that AI leadership is distributed across companies, institutions and communities.</p>572<p>Open models let startups, businesses, universities and public institutions build on advanced capabilities without training every model from scratch. They enable organizations to match the right model to the right job. That is how AI can advance safely, strengthen cybersecurity and sovereignty, accelerate innovation, and reach factories, hospitals, farms, classrooms and Main Street businesses around the world.</p>573<p>AI advances faster when people can build together.</p>574<p>NVIDIA has been committed to open weight models for years, demonstrated by multiyear investments and major contributions to open source platforms, including Hugging Face. NVIDIA has said that open models, data and tools broaden access to AI, and it has contributed hundreds of open models and datasets to Hugging Face as part of that effort.</p>575<ul>576<li>NVIDIA is the largest contributor of open models and data to Hugging Face, and our contributions continue to grow.</li>577<li>NVIDIA has released more than 500 models on Hugging Face and more than 250 open datasets.</li>578<li>We build our own models, libraries and tools in the open so developers everywhere can use them, modify them and build on top of them.</li>579</ul>580<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-98073" src="https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-scaled.png" alt="" width="2048" height="1423" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-scaled.png 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-960x667.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-1680x1167.png 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-1280x889.png 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-1536x1067.png 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/09/light_highres_source-630x438.png 630w" sizes="auto, (max-width: 2048px) 100vw, 2048px" /></p>581<p><span style="font-weight: 400;">As the opportunity for open models accelerates, Hugging Face can serve the global AI community at unprecedented scale. NVIDIA’s infrastructure, engineering and global reach can help improve platform reliability, safety, model evaluation, inference and deployment capabilities, while preserving the open ecosystem that made Hugging Face foundational.</span></p>582<p><span style="font-weight: 400;">I am honored that Clem came to me as he considered the next chapter of Hugging Face and believed NVIDIA would be a great home for the company, its community and the future of open models. We share this vision, and the Hugging Face team will now bring their passion and expertise to a much larger canvas, with their same iconic <img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f917.png" alt="🤗" class="wp-smiley" style="height: 1em; max-height: 1em;" /> brand.</span></p>583<p>To the millions of builders on Hugging Face: thank you for pushing the boundaries of what is possible. We can’t wait to build the future together with you. Together, we will make AI more open, more capable and more accessible to people and institutions around the world.</p>584]]></content:encoded>585 586 587 588 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/hf-nvidia-partner_hf-nvidia-partner-press-1920x1080-2.png" type="image/png" width="1920" height="1080">589 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/hf-nvidia-partner_hf-nvidia-partner-press-1920x1080-2-842x450.png" width="842" height="450" />590 <media:title type="html"><![CDATA[NVIDIA to Acquire Hugging Face]]></media:title>591 <media:description type="html"></media:description>592 </media:content>593 </item>594 <item>595 <title>NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier</title>596 <link>https://blogs.nvidia.com/blog/nvidia-crowdstrike-fal-con-2026/</link>597 598 <dc:creator><![CDATA[Brian Caulfield]]></dc:creator>599 <pubDate>Tue, 01 Sep 2026 21:19:20 +0000</pubDate>600 <category><![CDATA[AI]]></category>601 <category><![CDATA[Corporate]]></category>602 <category><![CDATA[Software]]></category>603 <category><![CDATA[Cybersecurity]]></category>604 <category><![CDATA[Nemotron]]></category>605 <category><![CDATA[Open Source]]></category>606 <guid isPermaLink="false">https://blogs.nvidia.com/?p=98014</guid>607608 <description><![CDATA[“We’re at an inflection point in cybersecurity,” Jensen Huang told a sold-out crowd at CrowdStrike’s Fal.Con 2026 in Las Vegas Tuesday. Attacks are now automated. Defense has to be, too. The NVIDIA founder and CEO joined CrowdStrike CEO and founder George Kurtz to announce CrowdStrike SafeMind, its agentic cybersecurity system developed by the CrowdStrike Cyber […]]]></description>609 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">“We’re at an inflection point in cybersecurity,” Jensen Huang told a sold-out crowd at CrowdStrike’s Fal.Con 2026 in Las Vegas Tuesday. Attacks are now automated. Defense has to be, too. </span></p>610<p><span style="font-weight: 400;">The NVIDIA founder and CEO joined CrowdStrike CEO and founder George Kurtz to announce CrowdStrike </span><a target="_blank" href="https://www.crowdstrike.com/en-us/press-releases/crowdstrike-launches-frontier-models-for-cybersecurity-with-nvidia/"><span style="font-weight: 400;">SafeMind</span></a><span style="font-weight: 400;">, its agentic cybersecurity system developed by the CrowdStrike Cyber Superintelligence Lab. </span><span style="font-weight: 400;"><br />611</span><span style="font-weight: 400;"><br />612</span><span style="font-weight: 400;">“This is the beginning of a new age of cybersecurity,” Huang told the crowd of 10,000 security professionals. “On the one hand, the adversaries are going to be more armed than ever. On the other hand, all of you are going to be more armed than ever.” </span></p>613<p><span style="font-weight: 400;">SafeMind combines CrowdStrike’s purpose-built, beyond frontier-capable models and customized agentic harnesses, with defensive models built on </span><a target="_blank" href="https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/"><span style="font-weight: 400;">NVIDIA Nemotron</span></a><span style="font-weight: 400;">, in a continuous coevolution loop where offense and defense repeatedly challenge and improve each other. </span></p>614<p><span style="font-weight: 400;">CrowdStrike also announced CrowdStrike Falcon IQ to operationalize Project QuiltWorks through agentic workload automation and expanded its CrowdStrike Guardian AI safety solution.</span><span style="font-weight: 400;"><br />615</span><span style="font-weight: 400;"><br />616</span><span style="font-weight: 400;">“We have asymmetric advantages because we have a large community of cybersecurity experts who want to work with each other and keep the world safe,” Huang told the crowd. </span></p>617<p><span style="font-weight: 400;">CrowdStrike’s annual conference drew security leaders from financial services, healthcare, the public sector and critical infrastructure.</span></p>618<p><span style="font-weight: 400;">“The real gap that I saw was that the attackers had frontier AI, and the defenders didn’t,” Kurtz told them. “And that changes now.”</span></p>619<h2><strong>SafeMind</strong></h2>620<p><span style="font-weight: 400;">CrowdStrike built SafeMind’s defensive model using NVIDIA Nemotron open models, post-trained with CrowdStrike’s cyber experience and threat data. The SafeMind models are paired with proprietary cybersecurity harnesses optimized to work as an agentic stack.</span></p>621<p><span style="font-weight: 400;">The result ships natively in the CrowdStrike Falcon platform as SafeMind, CrowdStrike’s agentic cybersecurity system. SafeMind brings offensive and defensive AI together in a continuous coevolution loop, where each side adapts to and strengthens the other. This process continuously hardens the security of the customer environment until attacks are unsuccessful.</span></p>622<p><span style="font-weight: 400;">“Your decade and a half of security data that we can train on — we can take a frontier model and make it essentially a super AGI that is incredibly good at cybersecurity,” Huang told Kurtz.</span><span style="font-weight: 400;"><br />623</span><span style="font-weight: 400;"><br />624</span><span style="font-weight: 400;">“Together with NVIDIA, we built cybersecurity’s first complete agentic system for cybersecurity, including the first frontier models and harness purpose-built for defenders,” Kurtz said. “This isn’t a copilot baked into someone else’s intelligence. It’s not a chatbot with a security skin. It is a frontier-class model built and trained by CrowdStrike on our data in partnership with NVIDIA.”</span></p>625<p><span style="font-weight: 400;">NVIDIA Nemotron 3 Ultra orchestrates the defensive agent harness. A fine-tuned Nemotron 3 Super powers SafeMind’s rule-generation sub-agent. </span></p>626<p><span style="font-weight: 400;">By post-training Nemotron with CrowdStrike data, CrowdStrike internal evaluations showed that the Blue Solano model — based on Nemotron 3 Super — delivered higher accuracy rates than leading frontier models at 99% lower cost. </span></p>627<p><span style="font-weight: 400;">While SafeMind can operate as a complete system, the models can be used independently to empower defenders to stay ahead of the adversary. Security experts can also pair their own models with CrowdStrike’s custom harnesses, giving customers the flexibility to use the right models and capabilities for their environment.</span></p>628<p><span style="font-weight: 400;">“The harness is essentially the exoskeleton of the large language model,” Huang said. “The large language model is the brain. The exoskeleton turns it into an agent — and this exoskeleton doesn’t have to be the same shape and capability for every domain.”</span></p>629<h2><strong>When AI Is the Defense</strong></h2>630<p><span style="font-weight: 400;">AI-enabled attacks rose 89% in the past year, and the fastest eCrime breakout time has reached 27 seconds, </span><a target="_blank" href="https://www.crowdstrike.com/en-us/global-threat-report/"><span style="font-weight: 400;">according to CrowdStrike</span></a><span style="font-weight: 400;">. Human-speed response isn’t defense. It’s documentation.</span></p>631<p><span style="font-weight: 400;">“There are many applications in the world where you must have the ability to fine-tune, to post-train — to create an AI that is super good at a particular domain,” Huang said. “Nemotron was created for precisely that. Completely free. Incredibly fast. You have the ability to have an asymmetric advantage against whatever comes your way.”</span></p>632<p><span style="font-weight: 400;">With open Nemotron as the base, CrowdStrike’s security teams post-trained on their own threat data without sending it to an outside provider, and customized the AI to their environment. </span></p>633<p><span style="font-weight: 400;">That’s not possible with a closed frontier model, and in security, the ability to inspect what’s defending matters. </span></p>634<h2><strong>Red vs. Blue</strong></h2>635<p><a target="_blank" href="https://developer.nvidia.com/blog/building-an-adaptive-agentic-cybersecurity-system-with-nvidia-nemotron/"><span style="font-weight: 400;">NVIDIA announced its work testing the CrowdStrike SafeMind models and harnesses</span></a><span style="font-weight: 400;"> in a high-fidelity cyber agent environment running as a simulation of the NVIDIA network. </span></p>636<p><span style="font-weight: 400;">The testing runs SafeMind in an offensive-defensive loop for adversarial coevolution. An offensive red-team agent finds the exploit, a blue-team defensive agent closes it and the findings become actionable detections to block attacks. </span></p>637<p><span style="font-weight: 400;">The red-agent harness runs Recon, Assault and Compromise sub-agents executing attack paths inside the cyber agent environment. The blue-agent harness monitors via Falcon sensors, generates detection candidates, validates them and promotes them. </span></p>638<p><span style="font-weight: 400;">CrowdStrike built the test environment with NVIDIA: a digital twin of NVIDIA’s own accelerated computing infrastructure, validated against NVIDIA’s real threat landscape.</span></p>639<p><span style="font-weight: 400;">“The basic framework of SafeMind — an adversarial model acting on a digital twin of the environment, with a defender model in a continuous cat-and-mouse loop, eventually learning how to secure itself — this basic framework applies to robotics, edge computing, enterprise computing and just about everything,” Huang said.</span></p>640<p><span style="font-weight: 400;">CrowdStrike also announced </span><a target="_blank" href="https://www.crowdstrike.com/en-us/press-releases/crowdstrike-falcon-iq-operationalizes-project-quiltworks-at-machine-speed/"><span style="font-weight: 400;">Falcon IQ</span></a><span style="font-weight: 400;">. NVIDIA Nemotron models help to power the agentic engine at the heart of Charlotte AI AgentWorks, CrowdStrike’s no-code agent development platform where Falcon IQ runs. </span></p>641<p><span style="font-weight: 400;">Falcon IQ uses more than 50 agents working together as a unified agentic workforce to automate the most time-intensive workflows in assessment, prioritization and remediation. </span></p>642<p><span style="font-weight: 400;">Partners use Falcon IQ to deliver customized findings, recommendations and executive outputs to customers. Charlotte AI AgentWorks enables every Falcon user to build their own agentic security workforce.</span></p>643<h2><strong>The Full Stack</strong></h2>644<p><span style="font-weight: 400;">CrowdStrike has thousands of customer organizations generating trillions of daily security events. </span></p>645<p><span style="font-weight: 400;">With NVIDIA’s full-stack accelerated computing platform, the collaboration runs from the chips up through the models to the harnesses acting on what those models find. For Kurtz, that’s the point. </span></p>646<p><span style="font-weight: 400;">“The crowd in CrowdStrike,” Kurtz added, “is the asymmetry that puts the defenders in a unique position to defeat the adversary.”</span></p>647]]></content:encoded>648 649 650 651 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/09/crowdstrike-nvidia-stage-scaled.jpg" type="image/jpeg" width="2048" height="1366">652 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/09/crowdstrike-nvidia-stage-842x450.jpg" width="842" height="450" />653 <media:title type="html"><![CDATA[NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier]]></media:title>654 <media:description type="html"></media:description>655 </media:content>656 </item>657 <item>658 <title>GeForce NOW Gives Gamers More Ways to Play at Gamescom 2026</title>659 <link>https://blogs.nvidia.com/blog/geforce-now-thursday-gamescom-2026/</link>660 661 <dc:creator><![CDATA[GeForce NOW Community]]></dc:creator>662 <pubDate>Thu, 27 Aug 2026 13:00:33 +0000</pubDate>663 <category><![CDATA[Gaming]]></category>664 <category><![CDATA[Cloud Gaming]]></category>665 <category><![CDATA[GeForce NOW]]></category>666 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97971</guid>667668 <description><![CDATA[NVIDIA’s Gamescom announcements are revealing what’s next for GeForce NOW, with new ways to play, more supported devices and platforms, and even more big PC games headed to the cloud. New NVIDIA DLSS 4.5 technology controls give members more ways to fine-tune gameplay, while expanded support for new Steam devices, GOG single sign-on, Firefox browser […]]]></description>669 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400">NVIDIA’s </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/news/gamescom-2026-nvidia-geforce-rtx-dlss-4-5-announcements/"><span style="font-weight: 400">Gamescom announcements</span></a><span style="font-weight: 400"> are revealing what’s next for </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/"><span style="font-weight: 400">GeForce NOW</span></a><span style="font-weight: 400">, with new ways to play, more supported devices and platforms, and even more big PC games headed to the cloud.</span></p>670<p><span style="font-weight: 400">New NVIDIA DLSS 4.5 technology controls give members more ways to fine-tune gameplay, while expanded support for new Steam devices, GOG single sign-on, </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-firefox"><span style="font-weight: 400">Firefox</span></a><span style="font-weight: 400"> browser support and more Amazon Fire TV devices continue broadening where and how GeForce RTX-powered PC gaming can be enjoyed.</span></p>671<p><span style="font-weight: 400">Gamescom is also bringing a blockbuster lineup of PC games to GeForce NOW. </span><i><span style="font-weight: 400">CONTROL Resonant</span></i><span style="font-weight: 400">, </span><i><span style="font-weight: 400">Gears of War: E-Day</span></i><span style="font-weight: 400">, </span><i><span style="font-weight: 400">Onimusha: Way of the Sword</span></i><span style="font-weight: 400">,</span><i><span style="font-weight: 400"> The Blood of Dawnwalker</span></i><span style="font-weight: 400"> and </span><i><span style="font-weight: 400">STAR WARS Zero Company</span></i><span style="font-weight: 400"> are all coming to the cloud at launch. </span></p>672<p><span style="font-weight: 400">For a limited time, members can also get </span><i><span style="font-weight: 400">CONTROL Resonant</span></i><span style="font-weight: 400"> with the purchase of a 12-month GeForce NOW </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=24109411186&gbraid=0AAAAAD4XAoFI7Qv3JuyVS3JfsbTyEsT23&gclid=CjwKCAjwqJXUBhBNEiwA8BgG7tncwPMGvgxfIqAMVMDu9z52vqLYF50jDggeNGVpU_B29eyVk2YCKBoC_YMQAvD_BwE"><span style="font-weight: 400">Ultimate membership</span></a><span style="font-weight: 400">.</span></p>673<p><span style="font-weight: 400">The Gamescom extras don’t stop there. To celebrate, GeForce NOW is giving away more than 40 prizes in our </span><a target="_blank" href="https://www.reddit.com/r/GeForceNOW/comments/1vy33q2/geforce_now_gamescom_giveaway_win_ultimate_gaming/"><span style="font-weight: 400">Community Giveaway</span></a><span style="font-weight: 400"> — including gaming hardware and GeForce NOW Ultimate memberships.</span></p>674<p><span style="font-weight: 400">There’s even more fun to come. Check out the list of </span><span style="font-weight: 400">13</span><span style="font-weight: 400"> games joining the </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/games/"><span style="font-weight: 400">GeForce NOW library</span></a><span style="font-weight: 400"> this week.</span></p>675<h2><b>Next-Level Cloud Gaming</b></h2>676<figure id="attachment_97975" aria-describedby="caption-attachment-97975" style="width: 1024px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-full wp-image-97975" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games.png" alt="DLSS 4.5 games on GeForce NOW" width="1024" height="576" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games.png 1024w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games-960x540.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games-630x354.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games-300x169.png 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-DLSS_4_5_games-400x225.png 400w" sizes="auto, (max-width: 1024px) 100vw, 1024px" /><figcaption id="caption-attachment-97975" class="wp-caption-text"><em>More ways to choose how to play.</em></figcaption></figure>677<p><span style="font-weight: 400">This fall, GeForce NOW is introducing new </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/news/dlss-4-5-dynamic-multi-frame-gen-6x-2nd-gen-transformer-super-res/"><span style="font-weight: 400">NVIDIA DLSS 4.5</span></a><span style="font-weight: 400"> controls in the cloud, giving Ultimate members more choice over how supported games look and perform, even without the latest hardware. Powered by NVIDIA RTX graphics, each control offers something different: DLSS Super Resolution sharpens image quality, while Dynamic Frame Generation delivers lower latency and stutter for smoother streaming, and Ray Reconstruction enhances ray-traced visuals. Together, these technologies make it easier to tune supported games for image quality, responsiveness or a preferred balance between the two. </span></p>678<p><span style="font-weight: 400">Stay tuned to GFN Thursdays for details about best practices for enabling DLSS 4.5 overrides and how Dynamic Frame Generation can help Ultimate members enable more graphic settings while still getting smooth streaming.</span></p>679<figure id="attachment_97977" aria-describedby="caption-attachment-97977" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97977" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-1680x945.jpg" alt="GeForce NOW and Steam hardware" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Steam_Hardware.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97977" class="wp-caption-text"><em>Stream full steam ahead with the Steam ecosystem.</em></figcaption></figure>680<p><span style="font-weight: 400">GeForce NOW is also adding community-requested support for Steam Machine with Steam Controller later this fall, and later for Windows and macOS, delivering more ways to play and making it even easier for longtime Steam players to jump into the cloud using familiar hardware. The additions build on existing </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-steam-deck-geforce-now/"><span style="font-weight: 400">Steam Deck support</span></a><span style="font-weight: 400"> and expand compatibility across the Steam ecosystem.</span></p>681<figure id="attachment_97978" aria-describedby="caption-attachment-97978" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97978" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-1680x945.jpg" alt="GOG SSO on GeForce NOW" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-GOG_Games.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97978" class="wp-caption-text"><em>Get your GOG on with one click into the library.</em></figcaption></figure>682<p><span style="font-weight: 400">Launching GOG games just got even easier. With </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-gdc-2026/"><span style="font-weight: 400">GOG single sign-on</span></a><span style="font-weight: 400"> now available on GeForce NOW, members can securely link their account once, automatically sync supported games and launch them without having to sign in to GOG again — all with a single click.</span></p>683<p><span style="font-weight: 400">With fewer steps between opening the app, skipping any downloads and jumping into gameplay, it’s another way GeForce NOW continues streamlining access to PC gaming libraries, making it easier to play the GOG library of games on TV platforms since gamers are automatically logged into their GOG accounts.</span></p>684<p><span style="font-weight: 400">Find popular titles, like </span><i><span style="font-weight: 400">Cyberpunk 2077</span></i><span style="font-weight: 400"> and </span><i><span style="font-weight: 400">The Witcher 3: The Wild Hunt</span></i><span style="font-weight: 400">, in the app in the dedicated GOG row. Or search and filter for GOG games, then jump into supported titles from almost anywhere — all streamed from the cloud. Check back on GFN Thursdays for the latest games joining the GeForce NOW library, including newly supported games streaming from GOG.</span></p>685<figure id="attachment_97979" aria-describedby="caption-attachment-97979" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97979" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-1680x945.jpg" alt="GeForce NOW and Firefox" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-1.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97979" class="wp-caption-text"><em>Outfox nearly any device with GeForce RTX gaming.</em></figcaption></figure>686<p><span style="font-weight: 400">Plus, GeForce NOW is now officially </span><a target="_blank" href="https://blog.mozilla.org/en/firefox/firefox-nvidia-geforce-now-partnership/"><span style="font-weight: 400">supported on Firefox</span></a><span style="font-weight: 400">, expanding browser access across all major Windows browsers. Members can jump into supported PC games directly from Firefox without downloading a dedicated app. Ultimate members can enjoy GeForce RTX-powered gaming at up to 1440p and 120 frames per second.</span></p>687<p><span style="font-weight: 400">Simply download the </span><a target="_blank" href="https://www.firefox.com/en-US/landing/geforce-now/?utm_source=nvidia&utm_medium=referral&utm_campaign=firefox-geforce-now"><span style="font-weight: 400">latest version of Firefox</span></a><span style="font-weight: 400"> and visit </span><a target="_blank" href="https://play.geforcenow.com"><span style="font-weight: 400">play.geforcenow.com</span></a><span style="font-weight: 400"> to get started. The cloud handles downloads, installs and updates, keeping local storage free while games stay ready to play in the cloud. </span></p>688<figure id="attachment_97976" aria-describedby="caption-attachment-97976" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97976" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-1680x945.jpg" alt="GeForce NOW and Fire TV" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Amazon_FireTV.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97976" class="wp-caption-text"><em>Bring the fire to living rooms with GeForce NOW.</em></figcaption></figure>689<p><span style="font-weight: 400">GeForce NOW is expanding support to additional </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-amazon-fire-tv-app/"><span style="font-weight: 400">Amazon Fire TV devices</span></a><span style="font-weight: 400">. GeForce NOW first launched on Fire TV devices earlier this year, and will soon expand to the Fire TV Stick 4K Select, a 4K streaming stick that’s ready to game right out of the box. Streaming from the cloud means supported PC games are ready to play using a compatible controller, making it easy to enjoy GeForce NOW on the biggest screen in the house with even more options to play from the comfort of the couch. </span></p>690<p><span style="font-weight: 400">And that’s not all the awesomeness from Amazon. Later this year, Fire TV users will be able to purchase GeForce NOW memberships directly through Amazon, creating a more seamless path from device setup to gameplay.</span></p>691<p><span style="font-weight: 400">These new features will arrive alongside some of the biggest games coming to the cloud throughout the year.</span></p>692<h2><b>Get Ready for Launch</b></h2>693<p><span style="font-weight: 400">Five newly announced and highly anticipated games are coming to GeForce NOW at launch: </span><i><span style="font-weight: 400">CONTROL Resonant</span></i><span style="font-weight: 400">, </span><i><span style="font-weight: 400">Gears of War: E-Day</span></i><span style="font-weight: 400">, </span><i><span style="font-weight: 400">Onimusha: Way of the Sword</span></i><span style="font-weight: 400">, </span><i><span style="font-weight: 400">The Blood of Dawnwalker</span></i><span style="font-weight: 400"> and </span><i><span style="font-weight: 400">STAR WARS Zero Company</span></i><span style="font-weight: 400">. </span></p>694<p><span style="font-weight: 400">Better yet, members can lock in one of them early: For a limited time, purchase a 12-month GeForce NOW Ultimate membership and receive </span><i><span style="font-weight: 400">CONTROL Resonant</span></i><span style="font-weight: 400"> at launch at no additional cost, and be first to experience Faden’s next chapter. The offer is available Tuesday, Aug. 25, through Sunday, Sept. 27.</span></p>695<p><span style="font-weight: 400">Ultimate members can stream the full lineup with GeForce RTX 5080-class performance in the cloud, including </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/technologies/dlss/"><span style="font-weight: 400">NVIDIA DLSS 4</span></a><span style="font-weight: 400">, </span><a target="_blank" href="https://developer.nvidia.com/discover/ray-tracing"><span style="font-weight: 400">ray tracing</span></a><span style="font-weight: 400">, </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/technologies/reflex/"><span style="font-weight: 400">NVIDIA Reflex</span></a><span style="font-weight: 400"> and Cinematic Quality Streaming technologies, with up to 5K high dynamic range on supported devices. </span></p>696<p><span style="font-weight: 400">Whether diving into new worlds or returning to iconic franchises, GeForce NOW lets members skip expensive hardware upgrades, massive installs and storage management so they can jump straight into the action the moment each game launches. Here’s a closer look at what’s coming:</span></p>697<figure id="attachment_97980" aria-describedby="caption-attachment-97980" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97980" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-1680x945.jpg" alt="Control Resonant bundle on GeForce NOW" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-CONTROL_Resonant_Bundle.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97980" class="wp-caption-text"><em>Contain the chaos.</em></figcaption></figure>698<p><span style="font-weight: 400">Explore a warped Manhattan on the brink of paranatural annihilation in Remedy’s </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-gdc-2026/"><i><span style="font-weight: 400">CONTROL Resonant</span></i></a><span style="font-weight: 400">. Harness Dylan Faden’s extraordinary powers to battle the Hiss, the Mold and other reality-bending threats while searching for his sister, Federal Bureau of CONTROL Director Jesse Faden.</span></p>699<figure id="attachment_97981" aria-describedby="caption-attachment-97981" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97981" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-1680x840.jpg" alt="Gears of Wars E-day launch to be on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Gears_of_War_E-Day.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97981" class="wp-caption-text"><em>Emergence begins.</em></figcaption></figure>700<p><span style="font-weight: 400">Experience the horror and brutality of Emergence Day in the Coalition’s </span><i><span style="font-weight: 400">Gears of War: E-Day</span></i><span style="font-weight: 400">. Fourteen years before the original </span><i><span style="font-weight: 400">Gears of War</span></i><span style="font-weight: 400">, join Marcus Fenix and Dominic Santiago in a gripping origin-story campaign as the Locust Horde first erupt from below, igniting a desperate fight for survival. Rally squads and fight on in multiplayer with a reimagination of Gears’ iconic PvE mode, Horde Siege, or go head to head with a refined versus PvP.</span></p>701<figure id="attachment_97982" aria-describedby="caption-attachment-97982" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97982" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-1680x840.jpg" alt="Onimusha Way of the Sword launch on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Onimusha_Way_of_the_Sword.jpg 1900w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97982" class="wp-caption-text"><em>Pick up the Oni Gauntlet.</em></figcaption></figure>702<p><span style="font-weight: 400">Fight across a twisted vision of Edo-period Kyoto in Capcom’s </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-onimusha-coming/"><i><span style="font-weight: 400">Onimusha: Way of the Sword</span></i></a><span style="font-weight: 400">. Wield the mystical Oni Gauntlet, master intense swordplay and battle relentless demonic Genma beneath mysterious clouds of Malice.</span></p>703<figure id="attachment_97983" aria-describedby="caption-attachment-97983" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97983" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-1680x840.jpg" alt="Blood of the Dawnwalker launch to be on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-The_Blood_of_Dawnwalker.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97983" class="wp-caption-text"><em>Walk the line between day and night.</em></figcaption></figure>704<p><span style="font-weight: 400">Step into 14th-century Europe in Bandai Namco’s </span><i><span style="font-weight: 400">The Blood of Dawnwalker</span></i><span style="font-weight: 400">, where war, plague and rising vampire powers have reshaped history. Play as Coen, a young Dawnwalker torn between humanity and cursed strength, and shape his path while fighting to save his family.</span></p>705<figure id="attachment_97984" aria-describedby="caption-attachment-97984" style="width: 1200px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="size-large wp-image-97984" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-1680x840.jpg" alt="STAR WARS Zero Company launch to be on GeForce NOW" width="1200" height="600" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-1680x840.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-960x480.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-1280x640.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-1536x768.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company-630x315.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-STAR_WARS_Zero_Company.jpg 2048w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97984" class="wp-caption-text"><em>Command the galaxy’s finest.</em></figcaption></figure>706<p><span style="font-weight: 400">Lead an elite squad behind enemy lines in </span><i><span style="font-weight: 400">STAR WARS Zero Company</span></i><span style="font-weight: 400">, a single-player turn-based game set during the twilight of the Clone Wars. Command former Republic officer Hawks and Zero Company through tactical battles and high-stakes missions where strategy, squad composition and player choices shape the fate of the galaxy.</span></p>707<h2><b>More Ways to Win</b></h2>708<figure id="attachment_97974" aria-describedby="caption-attachment-97974" style="width: 1280px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="wp-image-97974 size-full" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo.jpg" alt="Gamescom news for GeForce NOW" width="1280" height="680" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo-960x510.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo-630x335.jpg 630w" sizes="auto, (max-width: 1280px) 100vw, 1280px" /><figcaption id="caption-attachment-97974" class="wp-caption-text"><em>Gaming gear and Ultimate memberships are up for grabs.</em></figcaption></figure>709<p><span style="font-weight: 400">GeForce NOW is marking Gamescom with a community giveaway featuring more than 40 prizes. Gamers will have a chance to score a MacBook Neo, Steam Deck or Chromebook — each paired with a one-year GeForce NOW Ultimate membership — along with Amazon Fire TV Sticks and GeForce NOW Ultimate memberships ranging from one month to one year.</span></p>710<p><span style="font-weight: 400">Head over to the GeForce NOW community on Reddit for all the giveaway details and how to enter.</span></p>711<h2><b>The Hunt Is On</b></h2>712<p><span style="font-weight: 400">Squad up and fight for survival in </span><i><span style="font-weight: 400">Aliens: Fireteam Elite 2</span></i><span style="font-weight: 400">, streaming on GeForce NOW this week at launch. Assemble a four-player fireteam of Colonial Marines and battle through swarms of Xenomorphs, deadly new threats and increasingly desperate encounters. With deeper squad mechanics, new and improved classes, smarter enemies and expanded customization, every mission puts teamwork — and nerves — to the test.</span></p>713<p><span style="font-weight: 400">Members can also look for the following games joining the cloud this week:</span><i></i></p>714<ul>715<li><i><span style="font-weight: 400">Aliens: Fireteam Elite 2 </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/3448650?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Aug. 25)</span></li>716<li><i><span style="font-weight: 400">Resonance: A Plague Tale Legacy </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://store.steampowered.com/app/2713000?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400"> and </span><a target="_blank" href="https://www.xbox.com/games/store/resonance-a-plague-tale-legacy/9ppbncxp0096?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox</span></a><span style="font-weight: 400">, Aug. 27, available on Game Pass)</span></li>717<li><i><span style="font-weight: 400">STAR WARS Zero Company </span></i><span style="font-weight: 400">(New release on </span><a target="_blank" href="https://www.ea.com/games/starwars/zero-company?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">EA app</span></a><span style="font-weight: 400"> and </span><a target="_blank" href="https://store.steampowered.com/app/2075800?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">, Aug. 27)</span></li>718<li><i><span style="font-weight: 400">Baldur’s Gate 3 </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/baldurs_gate_iii?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>719<li><i><span style="font-weight: 400">Call of Duty: Modern Warfare 4 – Open Beta </span></i><span style="font-weight: 400">Weekend (</span><a target="_blank" href="http://battle.net"><span style="font-weight: 400">Battle.net, </span></a><a target="_blank" href="https://store.steampowered.com/app/4783780?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam </span></a><span style="font-weight: 400">and </span><a target="_blank" href="https://www.xbox.com/games/store/call-of-duty-modern-warfare-4-vault-edition-windows/9pk3g146xxvf?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Xbox</span></a><span style="font-weight: 400"> available from Aug. 28 10am PDT to Sep. 1 at 10am PDT. For more information please visit the </span><a target="_blank" href="https://www.callofduty.com/modernwarfare4/beta"><span style="font-weight: 400">Call of Duty website</span></a><span style="font-weight: 400">)</span><i><span style="font-weight: 400">.</span></i></li>720<li><i><span style="font-weight: 400">Clair Obscur: Expedition 33 </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/clair_obscur_expedition_33?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>721<li><i><span style="font-weight: 400">Cold Fear </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/cold_fear?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>722<li><i><span style="font-weight: 400">IRON NEST: Heavy Turret Simulator </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/2950790?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>723<li><i><span style="font-weight: 400">Kingdom Come: Deliverance II </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/kingdom_come_deliverance_ii?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>724<li><i><span style="font-weight: 400">Mount & Blade: Bannerlord II </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/mount_blade_ii_bannerlord?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>725<li><i><span style="font-weight: 400">No Man’s Sky </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/no_mans_sky?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>726<li><i><span style="font-weight: 400">Paradise Killer </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://store.steampowered.com/app/1160220?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">Steam</span></a><span style="font-weight: 400">)</span></li>727<li><i><span style="font-weight: 400">The Witcher 3: Wild Hunt </span></i><span style="font-weight: 400">(</span><a target="_blank" href="https://www.gog.com/game/the_witcher_3_wild_hunt?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400">GOG</span></a><span style="font-weight: 400">)</span></li>728</ul>729<p><span style="font-weight: 400">Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after release and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.</span></p>730<p><span style="font-weight: 400">To experience GeForce NOW, gamers can start with a GeForce NOW </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=23934057100&gbraid=0AAAAAD4XAoGnE9lpVE2P7s2w402M9Z88O&gclid=CjwKCAjwmozTBhAeEiwAkEGZzpLVIofm6Fb8NkgPbRUdjDh1f44tfDuIuNq5ZpkT2XfbB6EbuKByaxoCmzsQAvD_BwE"><span style="font-weight: 400">day pass</span></a><span style="font-weight: 400"> and try premium cloud gaming before committing to a membership. Better yet, the cost of a day pass can be applied toward a first membership purchase, making it even easier to level up when ready.</span></p>731<p><span style="font-weight: 400">Check back every GFN Thursday for new games, new features and even more ways to play on GeForce NOW. Which Gamescom announcement has you most excited? Let us know on </span><a target="_blank" href="https://www.twitter.com/nvidiagfn"><span style="font-weight: 400">X</span></a><span style="font-weight: 400"> or in the comments below. </span></p>732<blockquote class="twitter-tweet" data-width="550" data-dnt="true">733<p lang="en" dir="ltr">The future of PC gaming is on display with GeForce NOW. <img src="https://s.w.org/images/core/emoji/17.0.2/72x72/2601.png" alt="☁" class="wp-smiley" style="height: 1em; max-height: 1em;" /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/26a1.png" alt="⚡" class="wp-smiley" style="height: 1em; max-height: 1em;" /></p>734<p><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f5a5.png" alt="🖥" class="wp-smiley" style="height: 1em; max-height: 1em;" />DLSS 4.5 controls<br /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f4f1.png" alt="📱" class="wp-smiley" style="height: 1em; max-height: 1em;" />Expanded device support<br /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/2728.png" alt="✨" class="wp-smiley" style="height: 1em; max-height: 1em;" /> Upcoming AAA games<br /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f3ae.png" alt="🎮" class="wp-smiley" style="height: 1em; max-height: 1em;" /> More ways to play<br /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f381.png" alt="🎁" class="wp-smiley" style="height: 1em; max-height: 1em;" /> Community giveaway<br /><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f98a.png" alt="🦊" class="wp-smiley" style="height: 1em; max-height: 1em;" /> Firefox browser support<br />… and more! <a target="_blank" href="https://x.com/hashtag/gamescom?src=hash&ref_src=twsrc%5Etfw">#gamescom</a></p>735<p>Details <img src="https://s.w.org/images/core/emoji/17.0.2/72x72/27a1.png" alt="➡" class="wp-smiley" style="height: 1em; max-height: 1em;" /> <a target="_blank" href="https://t.co/Rz6RnX0Pt4">https://t.co/Rz6RnX0Pt4</a> <a target="_blank" href="https://t.co/sShKRhzrD3">pic.twitter.com/sShKRhzrD3</a></p>736<p>— <img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f329.png" alt="🌩" class="wp-smiley" style="height: 1em; max-height: 1em;" /> NVIDIA GeForce NOW (@NVIDIAGFN) <a target="_blank" href="https://x.com/NVIDIAGFN/status/2092273299057201244?ref_src=twsrc%5Etfw">August 25, 2026</a></p></blockquote>737<p><script async src="https://platform.x.com/widgets.js" charset="utf-8"></script></p>738]]></content:encoded>739 740 741 742 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo.jpg" type="image/jpeg" width="1280" height="680">743 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-thursday-8-27-nv-blog-1280x680-logo-842x450.jpg" width="842" height="450" />744 <media:title type="html"><![CDATA[GeForce NOW Gives Gamers More Ways to Play at Gamescom 2026]]></media:title>745 <media:description type="html"></media:description>746 </media:content>747 </item>748 <item>749 <title>Delivering Vera: NVIDIA’s First CPU Built for Agents Is Shipping Now</title>750 <link>https://blogs.nvidia.com/blog/vera-cpu-delivery/</link>751 752 <dc:creator><![CDATA[Ian Finder]]></dc:creator>753 <pubDate>Thu, 27 Aug 2026 13:00:17 +0000</pubDate>754 <category><![CDATA[AI Infrastructure]]></category>755 <category><![CDATA[Corporate]]></category>756 <category><![CDATA[Agentic AI]]></category>757 <category><![CDATA[AI Training]]></category>758 <category><![CDATA[NVIDIA Vera]]></category>759 <guid isPermaLink="false">https://blogs.nvidia.com/?p=92992</guid>760761 <description><![CDATA[NVIDIA Vice President of Hyperscale and HPC Ian Buck hand-delivers Vera CPU systems across the AI ecosystem as Vera begins shipping at scale.]]></description>762 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div>]]></content:encoded>763 764 765 766 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/05/vera-key-visual-1920x1080-1.jpg" type="image/jpeg" width="1920" height="1080">767 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/05/vera-key-visual-1920x1080-1-842x450.jpg" width="842" height="450" />768 <media:title type="html"><![CDATA[Delivering Vera: NVIDIA’s First CPU Built for Agents Is Shipping Now]]></media:title>769 <media:description type="html"></media:description>770 </media:content>771 </item>772 <item>773 <title>NVIDIA NVLink Fusion Expands With NVHBM Custom High-Bandwidth Memory</title>774 <link>https://blogs.nvidia.com/blog/nvlink-fusion-nvhbm-custom-high-bandwidth-memory/</link>775 776 <dc:creator><![CDATA[Jesse Clayton]]></dc:creator>777 <pubDate>Wed, 26 Aug 2026 21:05:30 +0000</pubDate>778 <category><![CDATA[AI Infrastructure]]></category>779 <category><![CDATA[Hardware]]></category>780 <category><![CDATA[Networking]]></category>781 <category><![CDATA[NVLink]]></category>782 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97957</guid>783784 <description><![CDATA[The next wave of AI is placing new demands on infrastructure. As AI agents and trillion-parameter workloads become mainstream, the performance of AI infrastructure depends not only on compute, but on how compute, memory, storage, networking and software are designed together as a unified system. To help hyperscalers and AI innovators build the next generation […]]]></description>785 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">The next wave of AI is placing new demands on infrastructure. </span></p>786<p><span style="font-weight: 400;">As AI agents and trillion-parameter workloads become mainstream, the performance of AI infrastructure depends not only on compute, but on how compute, memory, storage, networking and software are designed together as a unified system.</span></p>787<p><span style="font-weight: 400;">To help hyperscalers and AI innovators build the next generation of semi-custom AI infrastructure, NVIDIA today expanded </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/nvlink-fusion/"><span style="font-weight: 400;">NVIDIA NVLink Fusion</span></a> <span style="font-weight: 400;">with </span><a target="_blank" href="http://developer.nvidia.com/blog/nvidia-nvlink-fusion-brings-nvhbm-to-next-generation-ai-infrastructure"><span style="font-weight: 400;">NVIDIA NVHBM</span></a><span style="font-weight: 400;">, a next-generation high-bandwidth memory technology that brings higher memory performance and efficiency to XPUs. It will be validated and offered by leading memory partners, extending this advanced memory capability to NVLink Fusion customers.</span></p>788<p><span style="font-weight: 400;">Traditional HBM architectures place the memory controller on the XPU die, consuming valuable silicon area that could otherwise be dedicated to compute. NVHBM, built on the same technology that NVIDIA will use for future GPUs, integrates NVIDIA’s custom memory controller into the HBM base die. </span></p>789<p><span style="font-weight: 400;">By integrating the memory controller into the 3D HBM stack instead of the XPU, NVHBM delivers </span><span style="font-weight: 400;">up to 30%</span><span style="font-weight: 400;"> greater memory bandwidth and </span><span style="font-weight: 400;">15%</span><span style="font-weight: 400;"> lower HBM power consumption, and </span><span style="font-weight: 400;">frees up to 25%</span><span style="font-weight: 400;"> more area on XPU compute die compared with standard HBM4E.</span></p>790<p><span style="font-weight: 400;">NVIDIA is establishing a standard NVHBM implementation, available from multiple memory providers. This reduces the engineering effort required to integrate and qualify memory across multiple suppliers — giving NVLink Fusion customers a faster path for bringing custom AI chips to market.</span><span style="font-weight: 400;"> </span></p>791<p><a target="_blank" href="https://nvidianews.nvidia.com/news/aws-and-nvidia-to-deliver-2-million-additional-gpus-and-next-generation-infrastructure-for-agentic-and-physical-ai"><span style="font-weight: 400;">Amazon’s Annapurna Labs</span></a><span style="font-weight: 400;"> will be the first to work on NVHBM as part of its broader collaboration with NVIDIA around NVLink Fusion.</span></p>792<h2><b>AWS and NVIDIA Continue NVLink Fusion Collaboration</b></h2>793<p><span style="font-weight: 400;">Amazon’s Annapurna Labs </span><span style="font-weight: 400;">will work with NVIDIA on NVHBM technology and the NVLink scale-up architecture to enhance performance and efficiency for AI workloads.</span></p>794<p><span style="font-weight: 400;">This builds on AWS’s </span><a href="https://blogs.nvidia.com/blog/aws-partnership-expansion-reinvent/"><span style="font-weight: 400;">previously announced</span></a><span style="font-weight: 400;"> support for NVLink Fusion. Annapurna Labs will support NVLink Fusion with its next-generation Trainium chips starting with Trainium4, which would allow Amazon chips and NVIDIA GPUs to work together with common rack-scale architecture.</span><span style="font-weight: 400;"><br />795</span></p>796<p><span style="font-weight: 400;">“NVHBM represents a new architectural approach to advancing high-bandwidth memory performance and efficiency,” said </span><span style="font-weight: 400;">Nafea Bshara, vice president of Annapurna Labs at Amazon</span><span style="font-weight: 400;">. “We look forward to this technology collaboration to benefit future AWS infrastructure designs.” </span></p>797<h2><b>Vertically Integrated and Horizontally Open</b></h2>798<p><span style="font-weight: 400;">NVLink Fusion enables partners to connect custom XPUs and CPUs to NVIDIA’s rack-scale platform.</span></p>799<p><span style="font-weight: 400;">Partners can access NVIDIA NVLink chiplets, NVLink-C2C, NVLink Switches and NVIDIA MGX systems and racks, as well as a broad ecosystem of CPU partners, ASIC designers, system manufacturers and technology providers. </span></p>800<p><span style="font-weight: 400;">Offered with each generation of NVIDIA’s rack-scale system architecture, NVLink Fusion allows hyperscalers and AI-native companies to focus engineering resources on XPU innovation while using a proven technology stack for scale-up and scale-out networking, rack-scale systems and software — creating a faster, lower-risk path to deploying semi-custom AI infrastructure.</span></p>801<p><i><span style="font-weight: 400;">Learn more about </span></i><a target="_blank" href="https://www.nvidia.com/en-us/data-center/nvlink/"><i><span style="font-weight: 400;">NVLink</span></i></a><i><span style="font-weight: 400;"> and </span></i><a target="_blank" href="https://www.nvidia.com/en-us/data-center/nvlink-fusion/"><i><span style="font-weight: 400;">NVLink Fusion</span></i></a><i><span style="font-weight: 400;">.</span></i></p>802]]></content:encoded>803 804 805 806 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/gpu-architecture-corp-blog-nvhbm-1280x680-5547600.jpg" type="image/jpeg" width="1280" height="680">807 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/gpu-architecture-corp-blog-nvhbm-1280x680-5547600-842x450.jpg" width="842" height="450" />808 <media:title type="html"><![CDATA[NVIDIA NVLink Fusion Expands With NVHBM Custom High-Bandwidth Memory]]></media:title>809 <media:description type="html"></media:description>810 </media:content>811 </item>812 <item>813 <title>Leading Publishers Bring Blockbuster PC Games and Technology to NVIDIA RTX Spark</title>814 <link>https://blogs.nvidia.com/blog/gamescom-rtx-spark-pc-games-technology/</link>815 816 <dc:creator><![CDATA[Alexander Mejia]]></dc:creator>817 <pubDate>Tue, 25 Aug 2026 15:30:48 +0000</pubDate>818 <category><![CDATA[Gaming]]></category>819 <category><![CDATA[Game Development]]></category>820 <category><![CDATA[Gamescom]]></category>821 <category><![CDATA[GeForce]]></category>822 <category><![CDATA[GeForce NOW]]></category>823 <category><![CDATA[NVIDIA RTX]]></category>824 <category><![CDATA[Ray Tracing]]></category>825 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97918</guid>826827 <description><![CDATA[NVIDIA is bringing the next wave of RTX gaming to the Gamescom conference running this week in Cologne, Germany, with support for new games, anti-cheat technologies and increased visual quality. Electronic Arts, Embark and Ubisoft are among the latest game publishers and developers bringing their blockbuster titles to NVIDIA RTX Spark ahead of its launch […]]]></description>828 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">NVIDIA is bringing the next wave of RTX gaming to the Gamescom conference running this week in Cologne, Germany, with support for new games, anti-cheat technologies and increased visual quality. </span></p>829<p><span style="font-weight: 400;">Electronic Arts, Embark and Ubisoft are among the latest game publishers and developers bringing their blockbuster titles to </span><a target="_blank" href="https://www.nvidia.com/en-us/products/rtx-spark/"><span style="font-weight: 400;">NVIDIA RTX Spark</span></a><span style="font-weight: 400;"> ahead of its launch this fall. They join the publishers that announced RTX Spark support at COMPUTEX in May, including KRAFTON, NetEase, Riot Games and XBOX. </span></p>830<p><span style="font-weight: 400;">NVIDIA is working closely with game developers and ecosystem partners to ensure today’s top PC games — including EA’s </span><i><span style="font-weight: 400;">EA SPORTS</span></i> <i><span style="font-weight: 400;">F1 25</span></i><span style="font-weight: 400;"> and </span><i><span style="font-weight: 400;">Apex Legends</span></i><span style="font-weight: 400;">, Ubisoft’s </span><i><span style="font-weight: 400;">Anno 117: Pax Romana</span></i><span style="font-weight: 400;"> and Embark Studios’ </span><i><span style="font-weight: 400;">ARC Raiders </span></i><span style="font-weight: 400;">and </span><i><span style="font-weight: 400;">THE FINALS</span></i><span style="font-weight: 400;"> — run seamlessly on RTX Spark devices.</span></p>831<p><i><span style="font-weight: 400;">“</span></i><span style="font-weight: 400;">We’re excited to bring </span><i><span style="font-weight: 400;">Anno 117: Pax Romana</span></i><span style="font-weight: 400;"> to a new generation of RTX-powered PCs,” said Stéphane Jankowski, Ubisoft executive producer of </span><i><span style="font-weight: 400;">Anno 117 : Pax Romana. </span></i><span style="font-weight: 400;">“Our teams build the game to be seen wide: full districts, long sightlines, thousands of details at once. Seeing RTX Spark hold that level of fidelity from a single chip is genuinely impressive.”</span></p>832<p><span style="font-weight: 400;">RTX Spark reinvents personal computing, bringing personal AI agents, advanced content creation and high-performance gaming together in a platform engineered from the ground up for the next wave of Windows PCs. For gamers, RTX Spark combines Windows devices with NVIDIA RTX technologies, including </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/technologies/dlss/"><span style="font-weight: 400;">DLSS</span></a><span style="font-weight: 400;"> and </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce/technologies/reflex/"><span style="font-weight: 400;">Reflex</span></a><span style="font-weight: 400;">. </span></p>833<p><img loading="lazy" decoding="async" class="alignnone size-large wp-image-97946" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-1680x945.jpeg" alt="" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-1680x945.jpeg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-960x540.jpeg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-1280x720.jpeg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-1536x864.jpeg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-1290x725.jpeg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-630x354.jpeg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-300x169.jpeg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1-400x225.jpeg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Body-1.jpeg 1920w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></p>834<h2><b>Bringing Anti-Cheat to RTX Spark</b></h2>835<p><span style="font-weight: 400;">Supporting modern PC gaming goes beyond running the game itself. Many of today’s biggest online games rely on sophisticated anti-cheat technologies, to enable fair play across the multiple platforms these games run on. </span></p>836<p><span style="font-weight: 400;">NVIDIA has been working with EA to bring native EA Javelin Anticheat support to RTX Spark, including for EA’s biggest multiplayer games, such as </span><i><span style="font-weight: 400;">EA SPORTS F1 25</span></i><span style="font-weight: 400;">.</span><span style="font-weight: 400;"> RTX Spark launches this fall. </span></p>837<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97962" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-scaled.jpg" alt="" width="2048" height="1152" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-scaled.jpg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/RTX-Spark-EA-F1-Update-400x225.jpg 400w" sizes="auto, (max-width: 2048px) 100vw, 2048px" /></p>838<h2><b>RTX Delivers Next-Generation Gaming</b></h2>839<p><span style="font-weight: 400;">RTX Spark is just one part of NVIDIA’s lineup of gaming announcements at Gamescom, including:</span></p>840<ul>841<li style="font-weight: 400;" aria-level="1"><b>DLSS 4.5 Ray Reconstruction</b><span style="font-weight: 400;"> is available now, featuring NVIDIA’s latest second-generation transformer model to deliver even higher image quality in ray- and path-traced content. It replaces traditional denoisers with an NVIDIA supercomputer-trained AI network to reconstruct higher-quality pixels in the parts of a ray-traced frame where rays were not sampled.</span></li>842<li style="font-weight: 400;" aria-level="1"><b>Path tracing</b><span style="font-weight: 400;"> comes to </span><i><span style="font-weight: 400;">CONTROL Resonant</span></i><span style="font-weight: 400;"> and </span><i><span style="font-weight: 400;">007 First Light</span></i><span style="font-weight: 400;"> to deliver new levels of visual fidelity. </span></li>843<li style="font-weight: 400;" aria-level="1"><b>NVIDIA ACE</b><span style="font-weight: 400;"> technologies are coming to </span><i><span style="font-weight: 400;">Aniimo</span></i><span style="font-weight: 400;"> in early 2027, letting players naturally interact with their creatures.</span></li>844<li style="font-weight: 400;" aria-level="1"><i><span style="font-weight: 400;">Gears of War: E-Day </span></i><span style="font-weight: 400;">includes </span><b>RTX Mega Geometry</b><span style="font-weight: 400;"> integration, advancing image quality on GeForce RTX GPUs, alongside ray tracing and DLSS 4.5.</span></li>845<li style="font-weight: 400;" aria-level="1"><b>RTX Remix </b><span style="font-weight: 400;">keeps remastering PC classics, with a major </span><i><span style="font-weight: 400;">Painkiller RTX </span></i><span style="font-weight: 400;">mod update available now and </span><i><span style="font-weight: 400;">The Elder Scrolls III: Morrowind RTX </span></i><span style="font-weight: 400;">coming soon.</span></li>846<li style="font-weight: 400;" aria-level="1"><b>GeForce NOW </b><span style="font-weight: 400;">gives members more ways to play with new DLSS 4.5 controls, expanded support for Steam Machine and other devices and platforms, plus blockbuster PC games including </span><i><span style="font-weight: 400;">CONTROL Resonant</span></i><span style="font-weight: 400;">, </span><i><span style="font-weight: 400;">Gears of War: E-Day</span></i><span style="font-weight: 400;"> and more coming to the cloud at launch.</span></li>847</ul>848<p><iframe loading="lazy" title="NVIDIA RTX Spark | RTX On Gaming" width="1200" height="675" src="https://www.youtube.com/embed/3VeAFFry80I?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>849<p><i><span style="font-weight: 400;">Find more on NVIDIA’s announcements at Gamescom in </span></i><a target="_blank" href="https://www.nvidia.com/en-us/geforce/news/gamescom-2026-nvidia-geforce-rtx-dlss-4-5-announcements"><i><span style="font-weight: 400;">this GeForce article</span></i></a><i><span style="font-weight: 400;">. </span></i></p>850]]></content:encoded>851 852 853 854 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Key-Visual.jpg" type="image/jpeg" width="1920" height="1080">855 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/Gamescom-RTX-Spark-Key-Visual-842x450.jpg" width="842" height="450" />856 <media:title type="html"><![CDATA[Leading Publishers Bring Blockbuster PC Games and Technology to NVIDIA RTX Spark]]></media:title>857 <media:description type="html"></media:description>858 </media:content>859 </item>860 <item>861 <title>How XPUs Meet a World-Class AI Factory</title>862 <link>https://blogs.nvidia.com/blog/nvlink-fusion-xpu-ai-factory/</link>863 864 <dc:creator><![CDATA[Jesse Clayton]]></dc:creator>865 <pubDate>Mon, 24 Aug 2026 15:00:54 +0000</pubDate>866 <category><![CDATA[AI Infrastructure]]></category>867 <category><![CDATA[Hardware]]></category>868 <category><![CDATA[Networking]]></category>869 <category><![CDATA[Software]]></category>870 <category><![CDATA[AI Factory]]></category>871 <category><![CDATA[NVIDIA DSX]]></category>872 <category><![CDATA[NVLink]]></category>873 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97869</guid>874875 <description><![CDATA[To generate intelligence at scale, AI factories run continuously, and their economics are defined by delivered output: tokens per second, tokens per watt, cost per token, utilization and uptime. That requires AI infrastructure designed and built as a full factory, not a collection of individual accelerators. Hyperscalers and AI-native companies building custom XPUs must consider […]]]></description>876 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">To generate intelligence at scale, AI factories run continuously, and their economics are defined by delivered output: tokens per second, tokens per watt, cost per token, utilization and uptime. </span></p>877<p><span style="font-weight: 400;">That requires AI infrastructure designed and built as a full factory, not a collection of individual accelerators.</span></p>878<p><span style="font-weight: 400;">Hyperscalers and AI-native companies building custom XPUs must consider not just XPU design, but the design and development of the entire AI platform, including scale-up and scale-out networking, rack-scale architecture, production factory software and a robust supplier ecosystem. </span></p>879<p><span style="font-weight: 400;">At AI factory scale, this path is complex and costly, and represents a fundamental obstacle to getting XPUs to market quickly. </span></p>880<p><span style="font-weight: 400;">Breaking the constraint means combining custom XPUs with proven, mature infrastructure — allowing builders to focus innovation where it matters most while harnessing established technology for the rest. </span></p>881<p><span style="font-weight: 400;">NVLink Fusion delivers on that need, connecting XPUs to NVIDIA’s world-leading AI infrastructure to increase performance, accelerate time to market and mitigate risk for semi-custom AI factories.</span></p>882<h2><strong>Unlock XPU Performance With Fast Scale-Up</strong></h2>883<p><span style="font-weight: 400;">For modern workloads such as running trillion-parameter models, </span><a target="_blank" href="https://www.nvidia.com/en-us/glossary/mixture-of-experts/"><span style="font-weight: 400;">mixture-of-experts</span></a><span style="font-weight: 400;"> architectures and agentic AI, if the scale-up fabric cannot keep up, utilization drops and cost per token rises.</span></p>884<p><span style="font-weight: 400;">A scale-up networking solution must excel on three dimensions: </span></p>885<ul>886<li><b>Delivered performance:</b><span style="font-weight: 400;"> End-to-end network performance, in-network compute and mature software integration.</span></li>887<li><b>Factory resiliency:</b><span style="font-weight: 400;"> Uptime, continuous health monitoring and telemetry, and component-level serviceability while the factory keeps running.</span></li>888<li><b>Platform maturity:</b><span style="font-weight: 400;"> Reduced operational risk by using a mature technology stack with a demonstrated track record of large-scale deployments and realized return on investment. </span></li>889</ul>890<p><span style="font-weight: 400;">As an example, NVLink Fusion brings XPUs into the </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/nvlink/"><span style="font-weight: 400;">NVIDIA NVLink</span></a><span style="font-weight: 400;"> scale-up domain. Sixth-generation NVLink provides leading high-bandwidth, low-latency networking across a 72-XPU domain. The end-to-end latency for XPU-to-XPU transfers is 3x lower than alternative solutions based on off-the-shelf Ethernet, and the packet rate is 10x higher.</span></p>891<p><span style="font-weight: 400;">For end-to-end performance, NVIDIA GB300 NVL72 systems help deliver significantly higher throughput and better interactivity compared with configurations that don’t use NVL72, and future NVLink roadmap configurations include domains of up to 1,152 accelerators and co-packaged optics.</span></p>892<figure id="attachment_97888" aria-describedby="caption-attachment-97888" style="width: 1200px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="size-large wp-image-97888" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-1680x945.jpg" alt="A Pareto chart comparing GB300 NVL72 with B300 inference throughput performance in tokens per second per GPU on DeepSeek-V4-Pro at ISL=1K and OSL=1K sampled at various interactivity points in tokens per second per user. GB300 NVL72 is more than 10x the throughput of B300 in the middle of the Pareto between 70 and 100 tokens per second per user." width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/nvlink-thought-leadership-graph-1920x1080-1.jpg 1920w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /><figcaption id="caption-attachment-97888" class="wp-caption-text">The 72-GPU NVLink scale-up domain enables GB300 NVL72 to deliver higher per-GPU throughput and interactivity compared with NVIDIA B300. Results from NVIDIA’s AI Inference Performance Benchmarks page.</figcaption></figure>893<p><span style="font-weight: 400;">NVLink Fusion also includes NVIDIA NVLink-C2C for connecting XPUs to </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/vera-cpu/"><span style="font-weight: 400;">NVIDIA Vera CPUs</span></a><span style="font-weight: 400;"> or other ecosystem CPUs, delivering up to 6x the energy efficiency of a PCIe interface — helping remove barriers between control and compute for agentic systems.</span></p>894<h2><b>A Proven Stack and Ecosystem for Development and Deployment</b></h2>895<p><span style="font-weight: 400;">Teams developing custom XPUs often underestimate the effort and complexity of turning XPU innovation into data center deployment. This includes:</span></p>896<ul>897<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Integrating high-speed CPU and scale-up interfaces</span></li>898<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Sourcing and validating a scale-up network solution</span></li>899<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Designing compute and switch trays</span></li>900<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Designing and validating a rack architecture, including cooling and power</span></li>901<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Integrating security and storage</span></li>902<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Managing a complex supplier ecosystem</span></li>903</ul>904<p><span style="font-weight: 400;">The ideal platform provides all of this, allowing teams to focus on targeted innovation while using proven solutions for the rest.</span></p>905<p><span style="font-weight: 400;">NVLink Fusion is supported by an ecosystem designed for rapid development, integration and deployment, spanning ASIC design, CPU, and IP and optical interconnect partners.</span></p>906<p><em>“NVLink Fusion gives customers the ability to choose the CPU architecture, the performance level, the software capabilities that best meet their needs for the workloads that they care about,” said Tim Wilson, vice president and general manager of data center silicon engineering at Intel.</em></p>907<p><span style="font-weight: 400;">NVLink Fusion adopters can also use the NVIDIA MGX rack-scale architecture and the same supply chain used for MGX-based systems such as </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/vera-rubin-nvl72/"><span style="font-weight: 400;">NVIDIA Vera Rubin NVL72</span></a><span style="font-weight: 400;">. Manufacturing partners manage design and integration, while MGX suppliers provide the building blocks for rack, cooling, power and emerging </span><a href="https://blogs.nvidia.com/blog/800-vdc-power-architecture-ai-factory/"><span style="font-weight: 400;">800 VDC designs</span></a><span style="font-weight: 400;">.</span></p>908<p><em>“With Vera Rubin [NVL72], we are looking at almost 100% automation of system builds in the manufacturing line,” said Jack Luoh, head of product and solution at QCT and Quanta Computer. “Most of those investments can be leveraged if the XPU leverages NVLink Fusion.”</em></p>909<p><span style="font-weight: 400;">The NVIDIA AI infrastructure platform is vertically integrated and horizontally open. NVLink Fusion adopters can optionally incorporate NVIDIA Rubin GPUs, Vera CPUs, co-packaged optics switches, ConnectX SuperNICs, BlueField DPUs, Mission Control software and full-rack solutions including NVIDIA Vera Rubin NVL72, Vera CPU Rack, LPX, STX and SPX.</span></p>910<h2><b>Managing Risk With Infrastructure Standardization</b></h2>911<p><span style="font-weight: 400;">AI factory planning doesn’t wait for silicon. Power procurement, facility design, cooling, rack layout and network architecture begin long before the final accelerator mix is available. A data center locked to one chip can become a schedule risk.</span></p>912<p><span style="font-weight: 400;">Different workloads may favor different accelerators, including XPUs, GPUs, CPUs and LPUs. GPU systems may work alongside semi-custom systems for training, post-training, reasoning, retrieval and serving.</span></p>913<p><em>“The value of the NVLink Fusion program is … [customers] can deploy their rack-level solution with the NVIDIA GPU, and then they can decouple the development of their XPU and put it at a different pace,” said Vince Hu, corporate senior vice president and general manager of the data center and computing business group at MediaTek.</em></p>914<p><span style="font-weight: 400;">NVLink Fusion addresses these challenges through a unified architecture. XPU- and GPU-based systems such as Vera Rubin NVL72 can share rack footprints, networking, cooling, power delivery and management systems. Operators can move forward with buildout while deferring the precise silicon mix, then reprovision capacity as workload demand, silicon supply and business priorities change. </span></p>915<p><em><span style="font-weight: 400;">“NVLink</span><span style="font-weight: 400;"> Fusion allows the hyperscalers or the custom ASIC designers to integrate their own custom CPU or XPU and bridges the NVIDIA technology with a third-party process to create a unified rack-scale architecture,” said Lie-Szu Juang, chair and chief strategy officer at GUC.</span></em></p>916<h2><b>Designed, Validated and Operated as a Factory</b></h2>917<p><span style="font-weight: 400;">Factory buildout is expensive, and mistakes can require costly rework. Infrastructure must be validated before construction begins. NVLink Fusion aligns with the </span><a target="_blank" href="https://www.nvidia.com/en-us/data-center/products/dsx/"><span style="font-weight: 400;">NVIDIA DSX</span></a><span style="font-weight: 400;"> reference architecture for AI factories: codesigning buildings, power, cooling, compute and networking. The NVIDIA Omniverse DSX AI Factory Blueprint provides a digital twin and open reference design for gigawatt-scale AI factories, enabling partners to model facilities and technology together before deployment.</span></p>918<p><span style="font-weight: 400;">At the rack level, serviceability is part of performance. Reference compute trays feature 100% liquid cooling with no fans, cables or hoses, and allow trays to be removed while the rest of the rack remains operational. NVLink Switch trays are also liquid cooled and support continued operation during service.</span></p>919<p><em>“With NVLink Fusion we can use proven NVL72 rack design to have time-to-market, and we can have access to multiple suppliers to help us to deliver more into the hands of our customers,” said CC Lee, senior hardware development manager at Annapurna Labs, an Amazon company.</em></p>920<p><span style="font-weight: 400;">Software completes the factory. NVIDIA NCCL for distributed workloads, NVIDIA Dynamo and NIXL for disaggregation and NVIDIA Mission Control for cluster management, telemetry and debugging help operators run mixed AI infrastructure as a coordinated system.</span></p>921<p><span style="font-weight: 400;">With NVLink Fusion, XPUs can now meet a world-class AI platform, enabling hyperscalers and AI-native companies to build unified, semi-custom AI factories that combine the strengths of many builders into infrastructure no one company could build alone.</span></p>922<p><i><span style="font-weight: 400;">Learn more about </span></i><a target="_blank" href="https://www.nvidia.com/en-us/data-center/nvlink-fusion/"><i><span style="font-weight: 400;">NVLink Fusion</span></i></a><i><span style="font-weight: 400;">.</span></i></p>923]]></content:encoded>924 925 926 927 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/nvidia-dsx-maxlps-fg-i4of6-d-scaled.jpg" type="image/jpeg" width="2048" height="1153">928 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/nvidia-dsx-maxlps-fg-i4of6-d-842x450.jpg" width="842" height="450" />929 <media:title type="html"><![CDATA[How XPUs Meet a World-Class AI Factory]]></media:title>930 <media:description type="html"></media:description>931 </media:content>932 </item>933 <item>934 <title>With Groq 3 LPX in Full Production, NVIDIA Extends Vera Rubin Inference for Agents</title>935 <link>https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/</link>936 937 <dc:creator><![CDATA[NVIDIA Writers]]></dc:creator>938 <pubDate>Mon, 24 Aug 2026 15:00:41 +0000</pubDate>939 <category><![CDATA[AI Infrastructure]]></category>940 <category><![CDATA[Bluefi]]></category>941 <category><![CDATA[CPX]]></category>942 <category><![CDATA[LPX]]></category>943 <category><![CDATA[NVIDIA Spectrum-X Ethernet]]></category>944 <category><![CDATA[NVIDIA Vera Rubin]]></category>945 <category><![CDATA[NVLink]]></category>946 <category><![CDATA[Vera CPU]]></category>947 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97768</guid>948949 <description><![CDATA[The next era of AI inference won’t be defined by a single breakthrough chip, network or system. It’ll be defined by how every layer of the AI factory works together. That’s why NVIDIA is extending Vera Rubin NVL72 with fast token generation for agentic systems. Announced today, the NVIDIA Vera Rubin rack-scale system NVIDIA Groq […]]]></description>950 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400">The next era of AI inference won’t be defined by a single breakthrough chip, network or system. It’ll be defined by how every layer of the AI factory works together. That’s why NVIDIA is extending Vera Rubin NVL72 with fast token generation for agentic systems.</span></p>951<p><a target="_blank" href="https://nvidianews.nvidia.com/news/nvidia-groq-3-lpx-now-in-full-production-with-world-class-speed-for-agentic-ai"><span style="font-weight: 400">Announced today</span></a><span style="font-weight: 400">, the NVIDIA Vera Rubin rack-scale system NVIDIA Groq 3 LPX is in full production. In an Artificial Analysis benchmark running Gemma 4 31B, an open source agentic model, it delivered 3,400 output tokens per second for 100,000-token long-context use cases critical to agentic systems, 4x faster than the nearest alternative platform. </span></p>952<p><span style="font-weight: 400">Industry partners worldwide are adopting Vera Rubin platform solutions. SpaceXAI announced that NVIDIA Vera CPUs will power its next generation of agentic AI. CoreWeave has deployed into production Spectrum-X Multiplane, which connects NVIDIA Vera Rubin racks using multiple parallel switches to provide high-bandwidth, flat and lossless AI networks. Nebius is the first AI cloud to adopt NVIDIA Groq 3 LPX.</span></p>953<p><span style="font-weight: 400">As AI shifts from training to reasoning and agentic, inference has become the new frontier. Agentic AI systems are generating more tokens, processing dramatically larger context windows and increasingly collaborating with other AI systems to solve complex problems. </span></p>954<div style="width: 1200px;" class="wp-video"><video class="wp-video-shortcode" id="video-97768-3" width="1200" height="675" autoplay preload="metadata" controls="controls"><source type="video/mp4" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_DionHarris_v5_28MB.mp4?_=3" /><a href="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_DionHarris_v5_28MB.mp4">https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_DionHarris_v5_28MB.mp4</a></video></div>955<p><span style="font-weight: 400"></p>956<p>These workloads demand a new class of infrastructure optimized not just for performance but for throughput, responsiveness and economics at unprecedented scale.</span><a target="_blank" href="https://nvidia-my.sharepoint.com/personal/scmartin_nvidia_com/Documents/Microsoft%20Copilot%20Chat%20Files/Rubin%20CPX%20Product%20Aug%202026.pdf"><span style="font-weight: 400"> </span></a></p>957<p><span style="font-weight: 400">At the Hot Chips conference this week in Palo Alto, California, NVIDIA is showcasing how extreme codesign is reshaping the AI factory from end to end. By architecting compute, networking and inference acceleration as a unified system, NVIDIA is helping customers build infrastructure purpose-built for the emerging demands of long-context inference and multi-agent systems.</span></p>958<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97932" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-scaled.jpeg" alt="" width="2048" height="1151" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-scaled.jpeg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-960x540.jpeg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-1680x944.jpeg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-1280x720.jpeg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-1536x863.jpeg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-1290x725.jpeg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-630x354.jpeg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-300x169.jpeg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9931-2-400x225.jpeg 400w" sizes="auto, (max-width: 2048px) 100vw, 2048px" /></p>959<h3><b>Extreme Codesign Optimizes for Performance</b></h3>960<p><span style="font-weight: 400">Extreme codesign is the guiding principle behind NVIDIA platforms. Vera Rubin is engineered to accelerate inference as agents reason over increasingly long sequences. </span></p>961<p><a target="_blank" href="https://www.nvidia.com/en-us/networking/spectrumx/"><span style="font-weight: 400">NVIDIA Spectrum-X Ethernet</span></a><span style="font-weight: 400"> moves those massive data flows efficiently across AI factories, and NVIDIA Groq 3 LPX is built to generate tokens at ultrafast speeds. Together, they show how NVIDIA is optimizing every stage of the AI pipeline, from context and communication to generation, as part of a single, integrated AI factory architecture.</span></p>962<p><span style="font-weight: 400">NVIDIA Groq 3 LPX brings a new low-latency inference architecture designed to work alongside Vera Rubin NLV72, the most versatile AI factory platform, helping enterprises and cloud providers deliver the low latency, extreme throughput and scalable economics required for agentic applications.</span></p>963<p><span style="font-weight: 400">Breakthrough performance comes not from optimizing individual components in isolation, but from codesigning every layer of the stack. From networking and context processing to large-scale inference, NVIDIA’s full-stack platform turns AI factories into integrated engines for intelligence, built to turn ever-growing volumes of tokens into revenue.</span></p>964<div style="width: 1200px;" class="wp-video"><video class="wp-video-shortcode" id="video-97768-4" width="1200" height="675" autoplay preload="metadata" controls="controls"><source type="video/mp4" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips2026_Sizzle_final_30MB.mp4?_=4" /><a href="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips2026_Sizzle_final_30MB.mp4">https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips2026_Sizzle_final_30MB.mp4</a></video></div>965<hr />966<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#nvidia-partners"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>967<h2 id="nvidia-partners" class="wp-block-heading">NVIDIA Partners Adopt Vera Rubin for Lowest Token Costs</h2>968<p><span style="font-weight: 400">Nebius, a leading AI cloud, is first to adopt NVIDIA Groq 3 LPX, giving developers access to leading token generation speeds for highly responsive agentic AI applications. </span></p>969<p><span style="font-weight: 400">Adding NVIDIA Groq 3 LPX to NVIDIA Vera Rubin NVL72 in Nebius Token Factory will boost inference performance so developers can build highly interactive agents, coding systems and other real-time AI experiences at scale.</span></p>970<p><span style="font-weight: 400">Connecting NVIDIA Vera Rubin racks, </span><span style="font-weight: 400">CoreWeave</span><span style="font-weight: 400"> is deploying Spectrum-X Multiplane in production, unlocking advances for its AI cloud infrastructure. </span></p>971<hr />972<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#spacexai-vera-cpus"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>973<h2 id="spacexai-vera-cpus" class="wp-block-heading"><b>SpaceXAI Adopts NVIDIA Vera CPUs for Agentic AI</b></h2>974<p><span style="font-weight: 400">SpaceXAI </span><span style="font-weight: 400">plans to build and scale its future AI architecture around NVIDIA Vera Rubin, from data centers on Earth to orbital satellites. The company plans to deploy NVIDIA Vera CPUs to accelerate the CPU-intensive work behind agentic AI, including orchestration, tool use, code execution, data processing and simulation. </span></p>975<p><img loading="lazy" decoding="async" class="alignnone size-medium wp-image-97839" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-960x540.jpg" alt="" width="960" height="540" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750-400x225.jpg 400w, https://blogs.nvidia.com/wp-content/uploads/2026/08/cpu-press-hot-chips-vera-cpu-1920x1080-5562750.jpg 1920w" sizes="auto, (max-width: 960px) 100vw, 960px" /></p>976<p><span style="font-weight: 400">The SpaceXAI </span><span style="font-weight: 400"><a target="_blank" href="https://nvidianews.nvidia.com/news/spacexai-adopts-nvidia-vera-cpu-to-accelerate-agentic-ai-at-massive-scale">partnership</a> extends NVIDIA’s full-stack AI platform to SpaceXAI, bringing together Vera CPUs, NVIDIA accelerated computing, networking and software to advance AI at unprecedented scale.</span></p>977<p><span style="font-weight: 400">Designed for the agentic era, Vera Rubin provides leading per-core performance, exceptional memory bandwidth and predictable performance under load, helping agents complete tasks faster and keeping valuable GPU infrastructure fully utilized. </span></p>978<hr />979<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#groq-3-lpx"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>980<h2 id="groq-3-lpx" class="wp-block-heading"><b>NVIDIA Groq 3 LPX: The Interactive AI Inference Accelerator</b></h2>981<p><span style="font-weight: 400">Codesigned with the Vera Rubin NVL72 platform, NVIDIA Groq 3 LPX is helping AI factories deliver tokens at the lowest latency for agentic workloads.</span></p>982<p><span style="font-weight: 400">Agentic AI is creating a new performance challenge: decode latency. As AI agents reason, use tools and interact with other systems, they generate responses one token at a time, causing even tiny delays to multiply across complex chains of work. To keep agents operating at the pace users expect, NVIDIA Groq 3 LPX extends the Vera Rubin NVL72 platform with specialized acceleration for token generation. </span></p>983<p><span style="font-weight: 400">NVIDIA Rubin GPUs handle large-scale context processing while LPX accelerates latency-sensitive decode workloads. The result is faster, more predictable token generation that helps AI factories deliver responsive reasoning, smoother agent interactions and greater infrastructure efficiency. </span></p>984<p><img loading="lazy" decoding="async" class="alignnone size-medium wp-image-97900" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-960x540.png" alt="" width="960" height="540" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-960x540.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-1680x944.png 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-1280x720.png 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-1536x864.png 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-scaled.png 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-1290x725.png 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-630x354.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-300x169.png 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/LPX-Chart-400x225.png 400w" sizes="auto, (max-width: 960px) 100vw, 960px" /></p>985<p><span style="font-weight: 400">Together, Rubin GPUs and LPUs are designed to eliminate the traditional tradeoff between speed and throughput, helping AI providers deliver responsive, large-scale inference for the next generation of agentic AI applications.</span></p>986<h3><b>Building the Token Factory</b></h3>987<p><span style="font-weight: 400">As the industry shifts from model training to serving intelligence at scale, infrastructure must evolve into what NVIDIA describes as a “token factory” capable of delivering performance, throughput, intelligence integrity and economic efficiency simultaneously. Agentic AI systems increasingly communicate with other AI systems, access multiple data sources and maintain large amounts of context, creating unprecedented demand for fast inference.</span></p>988<p><span style="font-weight: 400">NVIDIA Groq 3 LPX was designed for exactly these workloads. As an extension of the Vera Rubin NVL72, it enables ultrafast responsiveness even across massive context windows while helping service providers maximize throughput and infrastructure utilization.</span><a target="_blank" href="https://nvidia-my.sharepoint.com/personal/scmartin_nvidia_com/Documents/Microsoft%20Copilot%20Chat%20Files/LPXDecks.pdf"><span style="font-weight: 400"> </span></a></p>989<h3><b>Extreme Codesign for Inference</b></h3>990<p><span style="font-weight: 400">Unlike standalone accelerators, NVIDIA Groq 3 LPX combines the strengths of GPUs and LPUs through extreme codesign. Rubin GPUs and LPUs jointly compute every layer of an AI model, enabling new levels of inference performance for agentic workloads. </span></p>991<p><span style="font-weight: 400">At scale, fleets of LPUs operate as a giant processor optimized for deterministic inference. A rack-scale NVIDIA Groq 3 LPX deployment can include 256 LP30 accelerators connected through direct chip-to-chip links, creating a highly efficient inference engine built for modern AI factories.</span></p>992<h3><b>Designed for the Agentic AI Era</b></h3>993<p><span style="font-weight: 400">As reasoning models grow and agentic workflows generate ever more tokens, the infrastructure required to serve them must evolve. NVIDIA Groq 3 LPX extends the Vera Rubin NVL72 platform with a purpose-built inference architecture designed to maximize responsiveness, throughput and efficiency, helping power the next generation of AI factories.</span></p>994<p><span style="font-weight: 400">And this is only the beginning, more optimizations, more models, more performance when paired with Vera Rubin NVL72 — new levels of throughput and interactivity are coming. Stay tuned. </span></p>995<hr />996<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#spectrum-x-multiplane"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>997<h2 id="spectrum-x-multiplane" class="wp-block-heading">NVIDIA Spectrum-X Multiplane Enables Massive AI Factory Scale on a Flatter, More Resilient Network</h2>998<p><span style="font-weight: 400">As AI factories grow massive, the network has become a critical engine of performance. At Hot Chips, NVIDIA is spotlighting </span><a target="_blank" href="https://www.nvidia.com/en-us/networking/spectrumx/"><span style="font-weight: 400">Spectrum-X Multiplane</span></a><span style="font-weight: 400"> — the latest in the hardware-accelerated Spectrum-X Ethernet architecture that lets Ethernet scale to unprecedented size while avoiding the latency, jitter and cost of adding another network tier.</span></p>999<p><img loading="lazy" decoding="async" class="alignnone size-medium wp-image-97841" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-960x540.png" alt="" width="960" height="540" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-960x540.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-1680x945.png 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-1280x720.png 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-1536x864.png 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-scaled.png 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-1290x725.png 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-630x354.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-300x169.png 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/SpectrumX-Multiplanetech-blog-3840x2160-5205700-1-400x225.png 400w" sizes="auto, (max-width: 960px) 100vw, 960px" /></p>1000<p><span style="font-weight: 400">NVIDIA Spectrum-X Ethernet is designed as an end-to-end, AI-optimized Ethernet platform, combining NVIDIA Spectrum-X Ethernet switches, SuperNICs and software to improve the performance and efficiency of Ethernet-based AI infrastructure for AI factories and clouds. </span><span style="font-weight: 400">The platform is designed to deliver 1.6x better AI networking performance compared with off-the-shelf Ethernet, while providing consistent, predictable performance in multi-tenant environments.</span></p>1001<h3><b>Multiplane Unlocks Scale Without the Tradeoffs of a New Tier</b></h3>1002<p><span style="font-weight: 400">Scaling an AI factory beyond today’s largest clusters traditionally means adding a third network tier, which adds latency, slows things down unpredictably and drives up the cost of cabling, optics and power. Spectrum-X Multiplane takes a simpler approach: It splits each server’s network connection into several independent paths, or “planes,” each running its own lightweight two-tier network. </span><span style="font-weight: 400">The result is a flat, simple network that scales to 512,000 GPUs, without the added cost and complexity of a third tier.</span></p>1003<p><span style="font-weight: 400">This all happens automatically. A dedicated hardware engine inside the NVIDIA ConnectX SuperNIC manages traffic across the planes and instantly reroutes around any failure, so applications and software simply see one fast, reliable connection. In an eight-plane topology, if one plane fails, the network still maintains about 90% of its total bandwidth, with hardware recovery that’s 11x faster than software-based multiplane load balancing. This translates to 1.6x higher AI factory output.</span></p>1004<h3><b>Built Through Extreme Codesign</b></h3>1005<p><span style="font-weight: 400">That reliability comes from extreme codesign of Vera Rubin NVL72, spanning switch silicon, SuperNICs and software. Spectrum-X SN6000 series switches, based on the 102.4Tb/s Spectrum-6 Ethernet ASIC and ConnectX-9 SuperNICs, supporting up to 1,600Gb/s per GPU, are purpose-built for Vera Rubin NVL72 AI factories. Spectrum-XGS Ethernet extends that same codesign across data centers, letting multiple facilities function as a single AI super-factory and accelerating multi-site NCCL collectives</span><span style="font-weight: 400"> by 1.9x.</span></p>1006<hr />1007<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#scale-in"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>1008<h2 id="scale-in" class="wp-block-heading">NVIDIA Introduces Scale-In Infrastructure for Agentic AI Factories, Powered by BlueField-4, DOCA</h2>1009<p><span style="font-weight: 400">NVIDIA is introducing NVIDIA Scale-In, the fifth pillar of NVIDIA AI networking and a new class of accelerated network infrastructure for agentic AI factories. Scale-In extends purpose-built acceleration to the infrastructure services that secure, manage and operate the AI factory.</span></p>1010<p><img loading="lazy" decoding="async" class="alignnone size-medium wp-image-97843" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/bluefield-corp-blog-bluefield-4-1280x680-4468150-3-960x510.png" alt="" width="960" height="510" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/bluefield-corp-blog-bluefield-4-1280x680-4468150-3-960x510.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/bluefield-corp-blog-bluefield-4-1280x680-4468150-3-630x335.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/bluefield-corp-blog-bluefield-4-1280x680-4468150-3.png 1280w" sizes="auto, (max-width: 960px) 100vw, 960px" /></p>1011<p><span style="font-weight: 400">Powered by the NVIDIA BlueField-4 processor and NVIDIA DOCA software platform and connected over NVIDIA Spectrum-X Ethernet, NVIDIA Scale-In transforms the traditional north-south access network into a unified, accelerated infrastructure domain.</span></p>1012<p><span style="font-weight: 400">Cloud computing brought software-defined networking, composability and elasticity to the data center, enabling users, applications, data and services to scale dynamically. </span></p>1013<p><span style="font-weight: 400">Agentic AI represents the next platform shift. AI factories bring together massive accelerated compute with growing numbers of users, applications and autonomous agents, all continuously interacting with data, storage and services. This transforms the demands on the infrastructure that brings AI to life. Networking, storage, cybersecurity and operations must now be accelerated alongside AI compute, combining software-defined flexibility with purpose-built hardware acceleration and full-stack codesign. </span></p>1014<p><span style="font-weight: 400">NVIDIA Scale-In delivers multi-tenant networking, high-performance storage access, in-silicon security, elastic provisioning and real-time observability, while keeping infrastructure processing independent of host compute resources. By accelerating and codesigning these services as part of the AI factory, Scale-In helps security, data access and operations scale alongside AI compute. The result is secure, efficient and manageable shared infrastructure for deploying and operating agentic AI at massive scale.</span></p>1015<hr />1016<p><em>Tuesday, Aug. 24, 8:00 a.m. PT <b><a href="https://blogs.nvidia.com/blog/vera-rubin-lpx-spectrum-x-nvlink-fusion/#xpus"><img src="https://s.w.org/images/core/emoji/17.0.2/72x72/1f517.png" alt="🔗" class="wp-smiley" style="height: 1em; max-height: 1em;" /></a></b></em></p>1017<h2 id="xpus" class="wp-block-heading"><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97931" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-scaled.jpeg" alt="" width="2048" height="1151" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-scaled.jpeg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-960x540.jpeg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-1680x944.jpeg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-1280x720.jpeg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-1536x863.jpeg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-1290x725.jpeg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-630x354.jpeg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-300x169.jpeg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/DSCF9828-2-400x225.jpeg 400w" sizes="auto, (max-width: 2048px) 100vw, 2048px" />NVIDIA NVLink Fusion Connects XPUs to NVIDIA’s Leading AI Platform</h2>1018<p><span style="font-weight: 400"><a href="https://blogs.nvidia.com/blog/nvlink-fusion-xpu-ai-factory/">NVIDIA NVLink Fusion</a> brings custom silicon into NVIDIA’s world-leading AI infrastructure platform, enabling hyperscalers and AI-native companies to build semi-custom AI factories with greater performance, flexibility and speed.</span></p>1019<p><span style="font-weight: 400">As AI models grow in size and complexity, raw compute alone is not enough. AI factories require high-bandwidth, low-latency scale-up networking, proven rack-scale architectures and a full ecosystem spanning power, cooling, management software and supply chain. NVLink Fusion addresses these challenges by connecting custom XPUs and CPUs to NVIDIA’s scale-up and scale-out technology stack.</span></p>1020<p><span style="font-weight: 400">The platform includes sixth-generation </span><a target="_blank" href="https://developer.nvidia.com/blog/nvidia-nvlink-the-scale-up-network-for-ai-factories/"><span style="font-weight: 400">NVIDIA NVLink</span></a><span style="font-weight: 400"> and NVLink Switch purpose-built scale-up networking, as well as NVLink-C2C for energy-efficient connectivity between XPUs and CPUs. Through the NVIDIA MGX ecosystem, adopters can also use production-proven rack designs, components, manufacturing partner solutions and open, extensible software for distributed computing, disaggregated workloads and cluster management.</span></p>1021<p><span style="font-weight: 400">By standardizing GPU- and XPU-based systems on a unified architecture, NVLink Fusion helps decouple data center buildout from silicon readiness. Operators can share rack footprints, networking, cooling, power delivery and management systems, then adjust the mix of GPUs and XPUs as supply and workload requirements evolve.</span></p>1022<p><span style="font-weight: 400">NVLink Fusion extends the NVIDIA AI platform’s vertically integrated, horizontally open approach to custom silicon. It gives partners the freedom to innovate where they differentiate while drawing on NVIDIA technologies across compute, networking, infrastructure and software — creating a single, flexible AI factory that no one company could build alone.</span></p>1023]]></content:encoded>1024 1025 1026 <enclosure url="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_DionHarris_v5_28MB.mp4" length="27590807" type="video/mp4" />1027<enclosure url="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips2026_Sizzle_final_30MB.mp4" length="26022533" type="video/mp4" />10281029 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_2026_MKT_KV_Rubin_Rack_Family_Front_comp_4k_V004-2-1-scaled.png" type="image/png" width="2048" height="1152">1030 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/HotChips_2026_MKT_KV_Rubin_Rack_Family_Front_comp_4k_V004-2-1-842x450.png" width="842" height="450" />1031 <media:title type="html"><![CDATA[With Groq 3 LPX in Full Production, NVIDIA Extends Vera Rubin Inference for Agents]]></media:title>1032 <media:description type="html"></media:description>1033 </media:content>1034 </item>1035 <item>1036 <title>Up to 30x More Work Per Watt: NVIDIA Vera Rubin NVL72 Sets a New Efficiency Standard for AI Agents</title>1037 <link>https://blogs.nvidia.com/blog/vera-rubin-nvl72-efficiency-ai-agents/</link>1038 1039 <dc:creator><![CDATA[Shruti Koparkar]]></dc:creator>1040 <pubDate>Mon, 24 Aug 2026 15:00:19 +0000</pubDate>1041 <category><![CDATA[AI Infrastructure]]></category>1042 <category><![CDATA[Hardware]]></category>1043 <category><![CDATA[Networking]]></category>1044 <category><![CDATA[Software]]></category>1045 <category><![CDATA[Agentic AI]]></category>1046 <category><![CDATA[Inference]]></category>1047 <category><![CDATA[NVIDIA Vera Rubin]]></category>1048 <category><![CDATA[Think SMART]]></category>1049 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97826</guid>10501051 <description><![CDATA[According to OpenRouter data, agentic AI workloads consume 15x more tokens than a simple chat request. Why? Consider what happens when an AI agent researches a company for an investment decision. The agent queries financial databases, searches news and filings, invokes a sub-agent to run peer comparisons and model valuations, then synthesizes everything into a […]]]></description>1052 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">According to </span><a target="_blank" href="https://openrouter.ai/blog/insights/deepseek-v4-adoption/"><span style="font-weight: 400;">OpenRouter data</span></a><span style="font-weight: 400;">, agentic AI workloads consume 15x more tokens than a simple chat request. Why? </span></p>1053<p><span style="font-weight: 400;">Consider what happens when an AI agent researches a company for an investment decision. The agent queries financial databases, searches news and filings, invokes a sub-agent to run peer comparisons and model valuations, then synthesizes everything into a recommendation. Agents and sub-agents keep reasoning until the task is done, driving increased token demand. With every step, the accumulated tokens become the input to the next, making long-context handling central to agentic AI performance.</span></p>1054<p><img loading="lazy" decoding="async" class="alignnone wp-image-97855 size-full" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-corp-blog-diagram-agentx-activation-1280x680-5565900.gif" alt="" width="1280" height="680" /></p>1055<p><span style="font-weight: 400;">The same pattern plays out across every agentic use case, from software development to customer service to deep research. </span></p>1056<p><span style="font-weight: 400;">As agentic AI moves into production across industries, the infrastructure running it needs to meet that token demand efficiently. </span></p>1057<p><span style="font-weight: 400;">New measured performance data shows NVIDIA Vera Rubin NVL72 systems deliver up to 30x higher throughput per megawatt than NVIDIA GB300 NVL72 on agentic workloads. </span><span style="font-weight: 400;">NVIDIA measured this inference throughput data using the SemiAnalysis AgentX workload, consisting of recorded real-world agentic coding sessions, with actual context growth, tool calls and sub-agent spawning preserved. For power-constrained AI factories, that translates directly into 30x more agentic work for the same energy footprint.</span></p>1058<p><span style="font-weight: 400;">These early results for Vera Rubin NVL72 demonstrate NVIDIA’s accelerated pace of innovation. With continuous software optimizations, performance across both Vera Rubin NVL72 and GB300 NVL72 will continue to improve.</span></p>1059<h2><b>Vera Rubin NVL72: 30x Higher Throughput per Megawatt and 35x Lower Token Cost</b></h2>1060<p><span style="font-weight: 400;">Agentic workloads look fundamentally different from chat or document summarization, where input and output sequences typically range from 1K to 8K tokens. In agentic sessions, context accumulates across steps and can reach hundreds of thousands of input tokens, with wide variability in both input and output lengths across requests. Performance measurement must evolve to capture the full agent workflow rather than a single inference request.</span></p>1061<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97856" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-scaled.png" alt="" width="2048" height="1152" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-scaled.png 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-960x540.png 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-1680x945.png 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-1280x720.png 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-1536x864.png 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-1290x725.png 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-630x354.png 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-300x169.png 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/InferenceNeedsAgenticBenchmarks-400x225.png 400w" sizes="auto, (max-width: 2048px) 100vw, 2048px" /></p>1062<p><span style="font-weight: 400;">The results below reflect performance measured on real-world agentic coding trajectories.</span></p>1063<p><span style="font-weight: 400;">In </span><a target="_blank" href="https://newsletter.semianalysis.com/p/agentx-inferencexv3-does-cuda-moat"><span style="font-weight: 400;">SemiAnalysis AgentX</span></a><span style="font-weight: 400;">, the NVIDIA Blackwell platform delivers leading performance across multiple agentic models including Kimi K3, MiniMax M3, GLM5.3, Qwen3.5 and DeepSeek V4 Pro. </span></p>1064<p><span style="font-weight: 400;">For example, <a target="_blank" href="https://developer.nvidia.com/blog/nvidia-vera-rubin-and-blackwell-set-a-new-standard-for-agentic-ai-performance-per-watt/?ncid=so-inst-151306">GB300 NVL72 delivers up to 15x</a> better throughput per megawatt than the NVIDIA Hopper architecture on the DeepSeek V4 Pro model, giving customers a high-performance foundation to run agentic workloads. This leap reflects the advantage of a larger scale-up GPU domain and codesigned software in delivering significantly better inference efficiency.</span></p>1065<p><span style="font-weight: 400;">Vera Rubin extends that advantage, lifting the performance across the entire Pareto curve, to deliver as much as 30x higher throughput per megawatt than GB300 NVL72 on the DeepSeek V4 Pro model. These early results, measured using the SemiAnalysis AgentX workload and currently pending SemiAnalysis review, don’t yet reflect Vera CPU performance for tool calling. </span></p>1066<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97890" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1.jpg" alt="" width="1920" height="1080" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1.jpg 1920w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s1-1920x1080-1-400x225.jpg 400w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /></p>1067<p><span style="font-weight: 400;">NVIDIA DSX MaxLPS technologies manages power across the GPU, rack and workload levels to provision up to 40% more GPUs within the same megawatt budget, pushing throughput per megawatt further at AI factory scale. </span></p>1068<p><span style="font-weight: 400;">Throughput per megawatt also directly impacts the cost of every token produced. At up to 35x lower cost per million tokens than GB300 NVL72, Vera Rubin NVL72 can run agents continuously, at scale, across the full breadth of customers’ workloads. </span></p>1069<p><img loading="lazy" decoding="async" class="alignnone size-full wp-image-97891" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1.jpg" alt="" width="1920" height="1080" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1.jpg 1920w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-social-chart1-agentx-activation-s2-1920x1080-1-400x225.jpg 400w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /></p>1070<p><span style="font-weight: 400;">For power-constrained AI factories, throughput per megawatt determines </span><a target="_blank" href="https://www.cio.com/article/4209776/5-critical-questions-that-define-ai-factory-economics.html"><span style="font-weight: 400;">AI factory revenue</span></a><span style="font-weight: 400;"> and cost per million tokens determines the profit margin on that revenue.</span></p>1071<h2><b>Extreme Codesign for Agentic Scale</b></h2>1072<p><span style="font-weight: 400;">Modern inference optimization spans a range of techniques that are especially critical for agentic AI. NVIDIA Vera Rubin NVL72 enables all of these and more through extreme codesign across every layer of the platform to deliver multifold performance gains.</span></p>1073<ul>1074<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Disaggregated serving separates context processing (prefill) from response generation (decode) so each scales independently.</span></li>1075<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Rate matching synchronizes the speeds at which prefill GPUs and decode GPUs produce tokens to maximize efficiency.</span></li>1076<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Large-scale expert parallelism distributes expert sub-networks in </span><a href="https://blogs.nvidia.com/blog/mixture-of-experts-frontier-models/"><span style="font-weight: 400;">mixture-of-experts models</span></a><span style="font-weight: 400;"> across the scale-up GPU domain. </span></li>1077<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Distributed KV-caching extends memory across the scale-up GPU domain, while KV-cache offloading tiers less-active context to host and storage, keeping previously processed context accessible without recomputation.</span></li>1078<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">KV-aware routing directs incoming requests to the GPUs that already hold the relevant cached context, reducing redundant computation across long sessions. </span></li>1079<li style="font-weight: 400;" aria-level="1"><span style="font-weight: 400;">Fused CUDA kernels like MegaMoE combine many computation and inter-GPU communication operations into a single execution pass, keeping GPUs active rather than waiting for data.</span></li>1080</ul>1081<p><span style="font-weight: 400;">NVIDIA Rubin GPUs’ enhanced fifth-generation</span> <a target="_blank" href="https://www.nvidia.com/en-us/data-center/tensor-cores/"><span style="font-weight: 400;">Tensor Cores</span></a><span style="font-weight: 400;"> and the third-generation</span> <a target="_blank" href="https://www.nvidia.com/en-us/data-center/tensor-cores/"><span style="font-weight: 400;">Transformer Engine</span></a><span style="font-weight: 400;"> accelerate both prefill and decode stages of inference. NVFP4 quantization compresses model weights to 4-bit precision, reducing memory footprint and increasing throughput without sacrificing output quality. </span></p>1082<p><span style="font-weight: 400;">The NVL72 scale-up domain, a defining architecture across Vera Rubin and Grace Blackwell, enables the high-bandwidth and low-latency inter-GPU communication essential for techniques such as large-scale expert parallelism and distributed KV-caching. Purpose-built to power this scale-up domain, NVIDIA NVLink interconnect technology and NVLink Switches, now in their sixth generation, deliver 10x higher packet rates and 3x lower latency than off-the-shelf Ethernet alternatives.</span></p>1083<p><span style="font-weight: 400;">Spanning optimized CUDA kernels, inference runtimes like NVIDIA TensorRT LLM and serving frameworks like NVIDIA Dynamo, NVIDIA’s </span><a href="https://blogs.nvidia.com/blog/inference-software-lowest-token-cost/"><span style="font-weight: 400;">software stack</span></a><span style="font-weight: 400;"> is codesigned with the hardware to enable inference optimizations.</span></p>1084<p><span style="font-weight: 400;">While the results above reflect current Vera Rubin NVL72 performance, the full platform is a seven-chip architecture that also includes the NVIDIA Vera CPU, <a target="_blank" href="https://developer.nvidia.com/blog/how-nvidia-groq-3-lpx-unlocks-ultrafast-interactivity-at-long-context-on-nvidia-vera-rubin/">Groq 3 LPU</a>, NVLink 6 Switch, BlueField-4 DPU, Spectrum-6 SPX and ConnectX-9 SuperNIC, all purpose-built for AI factories deploying agents at scale.</span></p>1085<p><span style="font-weight: 400;">Extreme codesign also extends to NVIDIA’s co-engineering with its partners. Vera Rubin is in full production and is scaling across the ecosystem. </span></p>1086<p><i><span style="font-weight: 400;">Learn more about the </span></i><a href="https://blogs.nvidia.com/blog/vera-rubin/"><i><span style="font-weight: 400;">NVIDIA Vera Rubin platform</span></i></a><i>.</i></p>1087]]></content:encoded>1088 1089 1090 1091 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-tech-blog-agentic-perf-kv-1920x1080-5555700.jpg" type="image/jpeg" width="1920" height="1080">1092 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/end-to-end-tech-blog-agentic-perf-kv-1920x1080-5555700-842x450.jpg" width="842" height="450" />1093 <media:title type="html"><![CDATA[Up to 30x More Work Per Watt: NVIDIA Vera Rubin NVL72 Sets a New Efficiency Standard for AI Agents]]></media:title>1094 <media:description type="html"></media:description>1095 </media:content>1096 </item>1097 <item>1098 <title>Bring the Fire: Play Games on GeForce NOW With New Firefox Browser Support</title>1099 <link>https://blogs.nvidia.com/blog/geforce-now-thursday-firefox/</link>1100 1101 <dc:creator><![CDATA[GeForce NOW Community]]></dc:creator>1102 <pubDate>Thu, 20 Aug 2026 13:00:24 +0000</pubDate>1103 <category><![CDATA[Gaming]]></category>1104 <category><![CDATA[Cloud Gaming]]></category>1105 <category><![CDATA[GeForce NOW]]></category>1106 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97798</guid>11071108 <description><![CDATA[It’s a new way into the cloud. GeForce NOW welcomes Firefox support to the cloud, opening up another way to jump into high-performance PC gaming straight from the browser, starting today. Whether on a school laptop or everyday PC, it’s now even easier to play supported PC games without downloading a dedicated app. Plus, discover […]]]></description>1109 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p><span style="font-weight: 400;">It’s a new way into the cloud. </span></p>1110<p><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/"><span style="font-weight: 400;">GeForce NOW</span></a><span style="font-weight: 400;"> welcomes Firefox support to the cloud, opening up another way to jump into high-performance PC gaming straight from the browser, starting today. Whether on a </span><a href="https://blogs.nvidia.com/blog/geforce-now-thursday-back-to-school-2026/"><span style="font-weight: 400;">school laptop</span></a><span style="font-weight: 400;"> or everyday PC, it’s now even easier to play supported PC games without downloading a dedicated app.</span></p>1111<p><span style="font-weight: 400;">Plus, discover 12 new games streaming from the </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/games/"><span style="font-weight: 400;">GeForce NOW library</span></a><span style="font-weight: 400;"> this week, led by </span><i><span style="font-weight: 400;">Gallipoli</span></i><span style="font-weight: 400;">. </span></p>1112<h2><b>No Downloads. Just Firefox.</b></h2>1113<figure id="attachment_97801" aria-describedby="caption-attachment-97801" style="width: 1200px" class="wp-caption aligncenter"><a href="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-scaled.jpg"><img loading="lazy" decoding="async" class="size-large wp-image-97801" src="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-1680x945.jpg" alt="GeForce NOW on Firefox" width="1200" height="675" srcset="https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-1680x945.jpg 1680w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-960x540.jpg 960w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-1280x720.jpg 1280w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-1536x864.jpg 1536w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-scaled.jpg 2048w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-1290x725.jpg 1290w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-630x354.jpg 630w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-300x169.jpg 300w, https://blogs.nvidia.com/wp-content/uploads/2026/08/gfn-firefox-browser-visual-kv-3840x2160-no-logo-1-400x225.jpg 400w" sizes="auto, (max-width: 1200px) 100vw, 1200px" /></a><figcaption id="caption-attachment-97801" class="wp-caption-text">This cloud is on fire with new ways to play on Firefox.</figcaption></figure>1114<p><span style="font-weight: 400;">GeForce NOW is now </span><a target="_blank" href="https://blog.mozilla.org/en/firefox/firefox-nvidia-geforce-now-partnership/"><span style="font-weight: 400;">supported on Firefox</span></a><span style="font-weight: 400;">, expanding access across even more devices. </span></p>1115<p><span style="font-weight: 400;">Members can outfox big downloads and jump into high-performance PC gaming directly from the browser — no waiting on lengthy game installs or updates. The experience delivers GeForce RTX-powered performance at up to 1440p and 120 frames per second for GeForce NOW Ultimate members. </span></p>1116<p><span style="font-weight: 400;">Simply download the </span><a target="_blank" href="https://www.firefox.com/en-US/landing/geforce-now/?utm_source=nvidia&utm_medium=referral&utm_campaign=firefox-geforce-now"><span style="font-weight: 400;">latest version of Firefox</span></a><span style="font-weight: 400;">, head to </span><a target="_blank" href="https://play.geforcenow.com"><span style="font-weight: 400;">play.geforcenow.com</span></a><span style="font-weight: 400;"> and start playing. The cloud handles the heavy lifting, so games are ready to play without taking up valuable storage space on the device. Firefox joins Chrome, Edge and Opera in supporting GeForce NOW on Windows browsers.</span></p>1117<h2><b>Deploy the Games</b></h2>1118<p><iframe loading="lazy" title="Gallipoli - Release Date Re-Reveal Trailer | Steam, PS5, Xbox Series X|S & Epic" width="1200" height="675" src="https://www.youtube.com/embed/hlnSz-JPiBk?feature=oembed" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe></p>1119<p><span style="font-weight: 400;">Land on the beaches of the Gallipoli Peninsula in BlackMill Games’ </span><i><span style="font-weight: 400;">Gallipoli</span></i><span style="font-weight: 400;">, the latest entry in the WW1 Game Series. Fight as soldiers of the British and Ottoman Empires in intense 25v25 objective-based battles across coastal shores, deserts and urban battlefields, with 10 historical classes and over 50 authentic weapons and equipment options.</span></p>1120<p><span style="font-weight: 400;">Teamwork is key in its grounded, one-shot-one-kill combat. Take command as an Officer, provide cover as a Heavy Machine Gunner or Sniper, keep the squad going as a Stretcher Bearer and more while navigating challenging campaigns.</span></p>1121<p><span style="font-weight: 400;">Head to the frontlines from the cloud with no lengthy game downloads or installs taking up local storage space. Play directly in the newly supported Firefox browser, or take the action across PCs, Macs, Chromebooks, handhelds and more with GeForce NOW cloud saves.</span></p>1122<p><span style="font-weight: 400;">In addition, members can look for the following games streaming throughout the week:</span></p>1123<ul>1124<li style="font-weight: 400;"><i><span style="font-weight: 400;">Stars Reach</span></i><span style="font-weight: 400;"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/1925650?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">, Aug. 18)</span></li>1125<li style="font-weight: 400;"><i><span style="font-weight: 400;">The Sinking City 2</span></i><span style="font-weight: 400;"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/2825860?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">, Aug. 18)</span></li>1126<li style="font-weight: 400;"><i><span style="font-weight: 400;">Gallipoli</span></i><span style="font-weight: 400;"> (New release on </span><a target="_blank" href="https://store.steampowered.com/app/3065940?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">, Aug. 20)</span></li>1127<li style="font-weight: 400;"><i><span style="font-weight: 400;">DIVE or DIE – Children of Rain</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://store.steampowered.com/app/3590290?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">)</span></li>1128<li style="font-weight: 400;"><i><span style="font-weight: 400;">Escape the Backroom</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://www.xbox.com/games/store/escape-the-backrooms/9mw15l8frt9p?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Xbox</span></a><span style="font-weight: 400;">, available on Game Pass)</span></li>1129<li style="font-weight: 400;"><i><span style="font-weight: 400;">Hell Is Us</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://www.xbox.com/games/store/hell-is-us/9n9097g0xzmb?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Xbox</span></a><span style="font-weight: 400;">, available on Game Pass)</span></li>1130<li style="font-weight: 400;"><i><span style="font-weight: 400;">High on Life</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://store.steampowered.com/app/1583230?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">)</span></li>1131<li style="font-weight: 400;"><i><span style="font-weight: 400;">High on Life 2</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://store.steampowered.com/app/2069250?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;"> and </span><a target="_blank" href="https://www.xbox.com/games/store/high-on-life-2/9nf6xpsbttgb?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Xbox</span></a><span style="font-weight: 400;">, available on Game Pass)</span></li>1132<li style="font-weight: 400;"><i><span style="font-weight: 400;">Parcel Simulator</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://store.steampowered.com/app/2424010?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">)</span></li>1133<li style="font-weight: 400;"><i><span style="font-weight: 400;">Starvester</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://store.steampowered.com/app/4194800?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Steam</span></a><span style="font-weight: 400;">)</span></li>1134<li style="font-weight: 400;"><i><span style="font-weight: 400;">The Thaumaturge</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://www.xbox.com/games/store/the-thaumaturge/9mzzzchz26nk?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Xbox</span></a><span style="font-weight: 400;">, available on Game Pass)</span></li>1135<li style="font-weight: 400;"><i><span style="font-weight: 400;">Tomb Raider IV-VI Remastered</span></i><span style="font-weight: 400;"> (</span><a target="_blank" href="https://www.epicgames.com/store/p/tomb-raider-ivvi-remastered-0be789?utm_source=nvidia&utm_campaign=geforce_now"><span style="font-weight: 400;">Epic Games Store</span></a><span style="font-weight: 400;">)</span></li>1136</ul>1137<p><span style="font-weight: 400;">Dates listed above reflect when games are released on their respective stores. GeForce NOW availability may vary, as games are onboarded after release and added throughout the week. Keep an eye on GeForce NOW channels and GFN Thursdays for availability updates on announced titles.</span></p>1138<p><span style="font-weight: 400;">To try GeForce NOW in Firefox, start with a GeForce NOW </span><a target="_blank" href="https://www.nvidia.com/en-us/geforce-now/?ncid=pa-srch-goog-673352&gad_source=1&gad_campaignid=23934057100&gbraid=0AAAAAD4XAoGnE9lpVE2P7s2w402M9Z88O&gclid=CjwKCAjwmozTBhAeEiwAkEGZzpLVIofm6Fb8NkgPbRUdjDh1f44tfDuIuNq5ZpkT2XfbB6EbuKByaxoCmzsQAvD_BwE"><span style="font-weight: 400;">day pass</span></a><span style="font-weight: 400;"> and experience premium cloud gaming directly from the browser. If members later upgrade to a Performance or Ultimate membership, the cost of the day pass can be applied toward the first membership purchase.</span></p>1139<p><span style="font-weight: 400;">What are you planning to play this weekend? Let us know on </span><a target="_blank" href="https://www.twitter.com/nvidiagfn"><span style="font-weight: 400;">X</span></a><span style="font-weight: 400;"> or in the comments below.</span></p>1140]]></content:encoded>1141 1142 1143 1144 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox.jpg" type="image/jpeg" width="1920" height="1080">1145 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/GFN_Thursday-Firefox-842x450.jpg" width="842" height="450" />1146 <media:title type="html"><![CDATA[Bring the Fire: Play Games on GeForce NOW With New Firefox Browser Support]]></media:title>1147 <media:description type="html"></media:description>1148 </media:content>1149 </item>1150 <item>1151 <title>Securing the Infrastructure of Intelligence</title>1152 <link>https://blogs.nvidia.com/blog/securing-the-infrastructure-of-intelligence/</link>1153 1154 <dc:creator><![CDATA[黄仁勋]]></dc:creator>1155 <pubDate>Mon, 17 Aug 2026 12:34:51 +0000</pubDate>1156 <category><![CDATA[AI Infrastructure]]></category>1157 <category><![CDATA[Corporate]]></category>1158 <category><![CDATA[AI Factory]]></category>1159 <guid isPermaLink="false">https://blogs.nvidia.com/?p=97752</guid>11601161 <description><![CDATA[AI factories are the defining infrastructure of the AI era — where compute transforms energy and data into intelligence that powers every business, industry and country. In the AI economy, compute is revenue. AI factories require a full stack of critical resources: advanced chips, packaging, memory and networking — as well as land, power and […]]]></description>1162 <content:encoded><![CDATA[<div id="bsf_rt_marker"></div><p>AI factories are the defining infrastructure of the AI era — where compute transforms energy and data into intelligence that powers every business, industry and country.</p>1163<p>In the AI economy, compute is revenue.</p>1164<p>AI factories require a full stack of critical resources: advanced chips, packaging, memory and networking — as well as land, power and shell.</p>1165<p>Just as NVIDIA has used its scale, long-term visibility and supply-chain partnerships to secure critical semiconductor resources, we are now applying that same discipline to secure LPS capacity exclusively for NVIDIA AI factories.</p>1166<p>Today, we are partnering with SB Energy to secure LPS capacity at the exceptional PORTS-Pike Technology Campus in Portsmouth, Ohio, to host NVIDIA compute. OpenAI will be the tenant.</p>1167<h2><strong>LPS: The Next Strategic Resource</strong></h2>1168<p>For the vast majority of NVIDIA customers, securing LPS has long been a part of their infrastructure strategy.</p>1169<p>The world’s largest cloud service providers and investment-grade enterprises have balance sheets, infrastructure expertise and long-term contracts to secure LPS independently. They build and operate AI factories using NVIDIA accelerated computing, networking, systems and software.</p>1170<p>This model will continue to represent most of NVIDIA’s business.</p>1171<p>But frontier AI labs are different.</p>1172<p>Frontier AI labs have extraordinary demand for training and inference compute, but many are growing faster than their balance sheets and long-term credit profiles can support. They may have strong customer demand and rapidly growing revenue yet still lack the decades-long infrastructure contracts and investment-grade financing capacity needed to secure the AI factory infrastructure independently.</p>1173<p>Their growth is increasingly constrained not by algorithms or customer demand, but by the availability of compute.</p>1174<p>For these companies, more compute means more intelligence, more products, more users and more revenue. NVIDIA is helping provide the infrastructure that powers this flywheel.</p>1175<h2><strong>PORTS-Pike: A Site for Generations of NVIDIA Compute</strong></h2>1176<p><span data-contrast="auto">OpenAI will build and operate a world-class AI factory at PORTS-Pike. The AI factory will use NVIDIA’s full-stack DSX AI factory platform, including GPUs, CPUs, networking and infrastructure software. </span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1177<p><span data-contrast="auto">The initial deployment is expected to provide 4.25 gigawatts of AI factory capacity. Each generation of NVIDIA AI factory systems deployed at PORTS-Pike could represent approximately 1.5 million NVIDIA GPUs, or approximately $150 billion to $200 billion in NVIDIA revenue. Over 20 years, the site can support multiple upgrade cycles.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1178<p><span data-contrast="auto">This is the essential economic point: the LPS commitment secures a long-lived AI factory site, while the NVIDIA compute inside can be upgraded repeatedly. Each new generation can deliver greater production, more intelligence and better economics.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1179<p><span data-contrast="auto">NVIDIA may also choose to extend the arrangement at PORTS-Pike beyond the initial 4.25 gigawatts to secure the remaining capacity of 3.75 gigawatts.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1180<h2><strong>OpenAI and NVIDIA Expanding Compute Opportunity</strong></h2>1181<p><span data-contrast="auto">More broadly, OpenAI has committed to substantial deployments of NVIDIA AI infrastructure through 2030. OpenAI’s existing and planned commitments represent approximately 12 gigawatts of NVIDIA compute, with an opportunity to expand to approximately 16 gigawatts if NVIDIA extends the PORTS-Pike arrangement beyond the initial 4.25 gigawatts.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1182<p><span data-contrast="auto">At these levels, the opportunity represents roughly $600 billion of NVIDIA compute through 2030.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1183<h2><strong>The Important Questions</strong></h2>1184<p><strong>What is NVIDIA guaranteeing, and for how long?</strong></p>1185<p><span data-contrast="auto">NVIDIA is supporting the LPS infrastructure at PORTS-Pike for approximately 4 gigawatts over a 20-year term, securing a site on which NVIDIA compute will be exclusively deployed. </span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1186<p><span data-contrast="auto">Our support is limited to defined portions of lease and power payments, along with a specified residual-value commitment — not the full cost of the site or all of the tenant’s obligations. </span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1187<p><span data-contrast="auto">The guarantee will become effective in phases as data centers are placed in service between 2028 and 2030. As OpenAI makes lease payments and capacity comes online, NVIDIA’s remaining exposure declines.</span></p>1188<p><strong>Why is NVIDIA guaranteeing PORTS-Pike?</strong></p>1189<p><span data-contrast="auto">LPS has become a critical constraint on AI factory deployment. NVIDIA is selectively securing exceptional sites where we can host multiple generations of NVIDIA compute and serve durable customer demand.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1190<p><span data-contrast="auto">The productive life of the site extends through multiple generations of NVIDIA systems, each capable of producing more intelligence and more revenue than the generation before.</span></p>1191<p><strong>Is this circular financing?</strong></p>1192<p><span data-contrast="auto">No. OpenAI will pay the lease.</span></p>1193<p><span data-contrast="auto">NVIDIA uses its scale and long-term visibility to secure PORTS-Pike to host NVIDIA compute. This is the same discipline we apply to supply-chain management: we secure critical inputs when we have visibility into customer demand and when doing so enables long-term productive capacity.</span></p>1194<p><strong>What happens to PORTS-Pike if OpenAI does not use the site in the future?</strong></p>1195<p><span data-contrast="auto">NVIDIA compute is versatile, fungible and broadly adopted. The capacity can be resold to another qualified tenant across NVIDIA’s global ecosystem of cloud service providers, enterprises, AI labs and startups.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1196<p><span data-contrast="auto">CUDA makes NVIDIA compute more than hardware. It gives developers and NVIDIA engineers a common platform to continually improve installed systems.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1197<p><span data-contrast="auto">CUDA makes NVIDIA compute versatile. </span><span data-contrast="auto">Versatility makes it fungible. Fungibility drives utilization and durability — making NVIDIA compute a productive asset: rentable and financeable.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1198<p><span data-contrast="auto">The value of an exceptional site, like PORTS-Pike, is not limited to one customer or one generation of compute. NVIDIA’s standardized platform, broad developer ecosystem and large market of potential users support the ability to redeploy productive capacity over time.</span></p>1199<p><strong>How much LPS will NVIDIA secure?</strong></p>1200<p><span data-contrast="auto">It will be strategic and disciplined.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1201<p><span data-contrast="auto">Most NVIDIA customers will continue to secure their own LPS. The vast majority of LPS hosting NVIDIA compute will continue to be secured directly by CSPs, enterprises, sovereign AI builders and other customers.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1202<p><span data-contrast="auto">NVIDIA will focus selectively on exceptional sites where visible, durable demand can support multiple generations of NVIDIA compute.</span></p>1203<h2><strong>The Infrastructure of Intelligence</strong></h2>1204<p><span data-contrast="auto">PORTS-Pike represents the next step in NVIDIA’s journey.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1205<p><span data-contrast="auto">We began by building accelerated computing chips. We then expanded to systems, networking, CUDA and full-stack AI factories. Today, we are helping secure the critical infrastructure required to build these factories.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1206<p><span data-contrast="auto">NVIDIA is the full-stack AI infrastructure platform.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1207<p><span data-contrast="auto">We are investing in the long-lived foundations of AI factories so our customers can deploy the most productive compute platform in the world, generation after generation.</span><span data-ccp-props="{"201341983":0,"335559740":276}"> </span></p>1208<p><span data-contrast="auto">By securing the critical resources needed to host NVIDIA compute, we can help the world’s most innovative companies build the AI factories that will power the age of intelligence.</span></p>1209]]></content:encoded>1210 1211 1212 1213 <media:content url="https://blogs.nvidia.com/wp-content/uploads/2026/08/partner-promo-pack-logo-lockup-openai-sbenergy-5590550_-press-1920x1080-1.png" type="image/png" width="1920" height="1080">1214 <media:thumbnail url="https://blogs.nvidia.com/wp-content/uploads/2026/08/partner-promo-pack-logo-lockup-openai-sbenergy-5590550_-press-1920x1080-1-842x450.png" width="842" height="450" />1215 <media:title type="html"><![CDATA[Securing the Infrastructure of Intelligence]]></media:title>1216 <media:description type="html"></media:description>1217 </media:content>1218 </item>1219 </channel>1220</rss>1221