Meta — meta-datacenters, meta-news
meta-datacenters (Meta-owned facilities)
| Source | https://datacenters.atmeta.com/all-locations/ (WordPress, server-rendered <article class="card"> per site) |
| Kind | cloud_provider, kind facility, parser cloud_meta_datacenters_v1 (operator Meta, hyperscale, isHyperscale) |
| Fetch | L1 direct, 0 credits (robots.txt: allow all) |
| Schedule | weekly |
| License / attribution | Publicly published operator pages; facts only — "Meta — datacenters.atmeta.com" |
| Verified live | 2026-09-11: 33 sites, 33 valid, 0 rejected |
Card → eyebrow = state/country ("oregon", "DENMARK", "canada"), title = town ("Prineville"), excerpt = "$2 billion+
investment | 2010 break ground | 1,000 skilled trade workers … | 350+ operational jobs …". Extracted:
name "Meta Data Center", city/state/country (curated coordinates, precision city; counties/parishes such as
Sarpy, Henrico, Richland Parish, Stanton Springs, Sturgeon County → approximate), constructionStartedOn = break-ground
year, externalIds.published_investment (+ parsed amount/currency when parseMoney understands it), info-sheet PDF
and Facebook page links. Same-name towns are guarded by the state (Lebanon IN, Kansas City MO, Montgomery AL).
Status is a documented heuristic because the page never states it: break-ground year > current year → announced;
≥ current year − 1 → under_construction (5 sites: Sturgeon County, Lebanon, Tulsa, El Paso, Beaver Dam…); older →
unknown (28 sites — Prineville 2010, Altoona 2013, … are left for reconciliation with other sources rather than
assumed operational). The description carries the published excerpt verbatim. Info sheets are 1–2 MB PDFs, not fetched.
Location pages exist (/richland-parish-data-center/, /ohio-new-albany/ in page-sitemap.xml) and could later
provide status text.
meta-news (newsroom)
Feed about.fb.com/feed/ (10 items, mostly product news, filtered on title/summary before fetch) + sitemap
datacenters.atmeta.com/post-sitemap.xml (~68 dated posts: "Kansas City, we are online!", "Hello, Bowling Green!",
grants, cooling explainers); discovery.include restricts to dated article URLs so static pages are not turned into
news. Live 2026-09-11: 4 news_events (construction_started, investment_announced, …), 0 rejected, 0 credits. Daily.