% ============================================================================
\section{Results}
\label{sec:results}
% ============================================================================

\subsection{Portfolio sorts: a weak unconditional relation}
Table~\ref{tab:sorts} reports the quintile sorts over the full sample of fifteen formation years.
The relation between AI language and future returns is positive and broadly monotone but modest.
The equal-weighted Q5$-$Q1 spread is $+4.6$ percentage points per year for AI adoption
\tstat{1.83}, $+3.9$ for AI infrastructure \tstat{1.58}, $+5.3$ for AI disruption risk
\tstat{2.27}, and $+4.9$ for the composite \tstat{1.88}. Value-weighting attenuates the adoption
and composite spreads, indicating that the relation is stronger among smaller index members. The
left panel of Figure~\ref{fig:quint} displays the full-sample pattern: returns rise gently from the
lowest to the highest AI quintile. By the standards of \citet{harveyliuzhu2016}, none of these
spreads would qualify as a discovery, and the averages blend two very different sub-periods.

\input{tables/tab_sorts}

\subsection{Does AI language earn risk-adjusted alpha unconditionally? (H1)}
Table~\ref{tab:alpha} reports CAPM and FF6 alphas of the monthly long--short portfolios over the
178 months for which the French factors are available. The composite earns a raw spread of about
$+2.3\%$ per year, of which the FF6 alpha is $+1.9\%$ and statistically insignificant
\tstat{1.24}. The adoption and infrastructure long--shorts are similarly insignificant. The one
exception is the disruption-risk spread, whose FF6 alpha of $+2.8\%$ per year is marginally
significant \tstat{2.10}. The factor loadings show that the long--shorts are roughly
market-neutral, with modest negative profitability (RMW) and investment (CMA) tilts consistent with
a growth-oriented long leg. Panel~C of Table~\ref{tab:postinf}, discussed below, adds the
technology-sector factor: its loading is 0.35 \tstat{11.1} and the full-sample FF6 alpha of the
composite falls to $+0.4\%$ per year \tstat{0.36}. The verdict on H1 is clear: unconditionally, AI
language is at best weakly priced, and what premium there is resembles a technology-sector tilt.

\input{tables/tab_alpha}

\subsection{Panel and Fama--MacBeth estimates}
Table~\ref{tab:panel} juxtaposes the two-way fixed-effects panel of equation~\eqref{eq:panel} with
the Fama--MacBeth estimator of equation~\eqref{eq:fm}. In the Fama--MacBeth regressions the AI
adoption and infrastructure premia are positive---$+1.3$ and $+1.6$ percentage points per standard
deviation when entered separately---with $t$-statistics of 1.5--1.6, a weak cross-sectional
premium on average; when the three collinear signals are entered jointly, only infrastructure
retains a positive coefficient. In the two-way fixed-effects panel every AI coefficient is
indistinguishable from zero. Because firm fixed effects remove permanent cross-firm differences,
the signal lives across firms, not within them: a given firm whose filings resemble the AI anchors
more closely from one year to the next does not earn a higher subsequent return. Throughout, size
carries a negative premium ($t\approx-8$) and volatility a positive one.

\input{tables/tab_panel}

\subsection{The post-ChatGPT change: large on average, concentrated in one cohort (H2)}
\label{subsec:break}
Splitting the sample at the ChatGPT launch, the centre and right panels of Figure~\ref{fig:quint}
show two different pictures. Before 2022 the relation between AI quintile and forward return is
nearly flat; after 2022 the schedule tilts upward, and the composite Q5$-$Q1 spread rises from
$+2.5$ percentage points per year to $+14.6$. Table~\ref{tab:break} quantifies the change for each
signal. In sub-period sorts the adoption spread rises from $+2.2\%$ to $+14.3\%$, the composite
from $+2.5\%$ to $+14.6\%$, and infrastructure from $+1.4\%$ to $+13.6\%$. In the pooled
interaction regressions of equation~\eqref{eq:break} the post-2022 increment is positive for all
three signals, between $+2.0$ and $+2.8$ percentage points per standard deviation, and
statistically significant only for AI infrastructure \tstat{2.63}, the facet most directly tied to
the chip, cloud, data-centre, and power firms that led the 2023--2025 rally.

\begin{figure}[htbp]\centering
\includegraphics[width=\textwidth]{fig1_quintile_returns.png}
\caption{\textbf{Forward returns by AI-language quintile, before and after ChatGPT.} Each panel
plots the mean forward twelve-month return across quintiles of the AI composite: the left panel
pools the full sample, the centre panel the pre-ChatGPT formation years (2010--2021), and the right
panel the post-ChatGPT formation years (2022--2024). The Q5$-$Q1 spread rises from $+2.5$
percentage points per year before ChatGPT to $+14.6$ after; the post-period average rests on three
cohorts and is dominated by the 2024 cohort (Table~\ref{tab:postinf}).}\label{fig:quint}
\end{figure}

\input{tables/tab_break}

The Newey--West $t$-statistics of Table~\ref{tab:break} are computed on three post-period
observations and carry little information. Table~\ref{tab:postinf} asks directly how much evidence
the post-period contains. Panel~A lists the three post-launch cohorts: the 2022 cohort (held from
July 2023 to June 2024) earned a composite spread of $+9.3$ points, the 2023 cohort $-4.0$, and the
2024 cohort (held from July 2025 to June 2026) $+38.4$. Two pre-launch cohorts, 2012 and 2015,
earned more than the 2022 cohort, and nine of the twelve pre-launch spreads are positive. The
post-period average therefore owes its size to a single cohort: dropping 2024 leaves a post-period
mean of $+2.6$ points, essentially the pre-period value of $+2.4$, whereas dropping either of the
other two cohorts raises the mean. Figure~\ref{fig:spread}, which plots the annual spreads, makes
the same point visually; the first version of this paper described the post-2022 cohorts as
``uniformly and substantially positive,'' which the figure does not support.

\begin{figure}[htbp]\centering
\includegraphics[width=0.9\textwidth]{fig3_annual_spread.png}
\caption{\textbf{Annual AI-language return spreads by cohort.} The figure plots the annual
high-minus-low (Q5$-$Q1) forward-return spread on the AI composite by formation year (the filing
year of the signal). Spreads are of mixed sign through 2021; among the post-ChatGPT cohorts (in
red) the 2022 spread is positive, the 2023 spread negative, and the 2024 spread, whose holding
period is July 2025--June 2026, far larger than any other. The post-period average of
Table~\ref{tab:break} is driven by the 2024 cohort.}
\label{fig:spread}
\end{figure}

Panel~B of Table~\ref{tab:postinf} pools the firm-years in the top and bottom quintiles of the three
post-launch cohorts and regresses the forward return on a top-quintile indicator with year dummies
and standard errors clustered by firm. The post-period spread is $+14.7$ points \tstat{2.18} on raw
returns but $+7.2$ points \tstat{2.21} on returns winsorized at the first and ninety-ninth
percentiles: half of the raw spread comes from a handful of extreme winners in the top quintile of
the 2024 cohort, among them Western Digital ($+969\%$), Micron ($+730\%$), and Seagate ($+624\%$),
the memory and storage manufacturers whose prices rose with AI-related demand in 2025--2026.
Restricting the pooled test to the 2022 and 2023 cohorts gives $+2.6$ points \tstat{0.46}, and the
difference between the post- and pre-period spreads in the pooled sample is $+12.1$ points
\tstat{1.79}. Panel~C turns to the monthly long--short series, which has 36 post-period
observations. The composite portfolio earned 7.5\% per year after July 2023 with an annualized
Sharpe ratio of 1.32 and a HAC $t$-statistic of 2.20, against 1.1\% per year and a Sharpe ratio
of 0.22 before; its FF6 alpha is $+5.2\%$ per year \tstat{1.72}. The monthly mean is about half
the annual spread because the monthly portfolio rebalances equal weights every month, which caps
the contribution of stocks that multiply in value within the year, whereas the annual buy-and-hold
spread lets them run. Adding the technology-sector factor, on which the post-period loading is 0.47
\tstat{6.53}, reduces the alpha to $+1.9\%$ per year \tstat{0.86}. The monthly evidence thus
supports a post-2022 increase in the return to AI language that is statistically detectable, but
it is largely spanned by the technology sector's own outperformance over the same months, which
averaged 15.8\% per year in our equal-weighted sector factor.

\input{tables/tab_postinference}

Two further lenses show the same regime change and the same concentration. Figure~\ref{fig:cum}
plots the cumulative value of one dollar in the equal-weighted, annually rebalanced long--short
portfolio: it drifts sideways for a decade and rises after the dashed ChatGPT line, with most of
the rise in the last year. Figure~\ref{fig:fmyear} plots the year-by-year Fama--MacBeth slope on
the composite with controls: it fluctuates around zero through 2021, is positive in 2022
($+2.9$ points per standard deviation), negative in 2023 ($-1.1$), and large in 2024 ($+14.3$).
With sector dummies (Table~\ref{tab:reviewrobust}, Panel~D) the 2024 slope falls to $+10.2$ and the
post-period average from $+5.5$ \tstat{2.00} to $+2.9$ \tstat{1.36}.

\begin{figure}[htbp]\centering
\includegraphics[width=0.86\textwidth]{fig2_cumulative_ls.png}
\caption{\textbf{Cumulative performance of the AI long--short portfolio.} The figure plots the
cumulative value of \$1 invested in the equal-weighted, annually rebalanced, self-financing
high-minus-low portfolio, for the AI composite and the AI disruption-risk signal. Both curves drift
sideways for a decade and rise after the dashed ChatGPT line, with most of the increase during the
holding period of the 2024 cohort.}
\label{fig:cum}
\end{figure}

\begin{figure}[htbp]\centering
\includegraphics[width=0.9\textwidth]{fig6_fm_byyear.png}
\caption{\textbf{Year-by-year Fama--MacBeth slopes on the AI composite.} The figure plots the
annual cross-sectional slope on the composite---the forward return per cross-sectional standard
deviation---controlling for size, momentum, and volatility, with error bars of $\pm1$ standard
error (heteroskedasticity-robust, within year). The slope fluctuates around zero through 2021 and
is large only in the 2024 cohort.}\label{fig:fmyear}
\end{figure}

Figure~\ref{fig:bin} shows that, within the post-period, the relation is not confined to the
extreme quintiles: the binned relation between the composite and forward returns is flat before
ChatGPT and upward-sloping afterward across the distribution of the signal. The binned means are
averages over the three post-period cohorts, so the slope inherits the weight of the 2024 cohort.

\begin{figure}[htbp]\centering
\includegraphics[width=\textwidth]{fig5_binscatter.png}
\caption{\textbf{Binned relation between AI language and returns, by sub-period.} Binned scatter of
the forward twelve-month return against the AI composite, separately for the pre- and post-ChatGPT
formation years. Each point is the mean within a bin of the signal and the line is an OLS fit. The
relation is flat before ChatGPT and upward-sloping afterward across the distribution of the signal;
the post-period bins average the 2022--2024 cohorts.}\label{fig:bin}
\end{figure}

\subsection{The anatomy of the post-period spread: size and sector}
Where in the cross-section does the post-2022 spread concentrate? Table~\ref{tab:doublesort} and
Figure~\ref{fig:double} report an AI~$\times$~size double sort. Before 2022 the AI gradient is flat
within both size halves. After 2022 it is steep, and markedly so among smaller firms: small,
high-AI firms earned a mean forward return of 34.6\%, against 19.5\% for small, low-AI firms, a
15-point gap, while the gap among large firms is 7 points. The post-period spread is thus not a
mega-cap phenomenon, a point reinforced in Section~\ref{sec:robustness}; it is strongest among the
smaller, less-followed index members, which is consistent with an attention-based mechanism but
also with the fact that the memory and storage manufacturers of the 2024 cohort were mid-sized
index members at formation.

\input{tables/tab_doublesort}

\begin{figure}[htbp]\centering
\includegraphics[width=\textwidth]{fig9_double_sort.png}
\caption{\textbf{AI language $\times$ firm-size double sort.} Mean forward twelve-month return by AI
tercile within small and large firms, separately for the pre- and post-ChatGPT formation years. The
post-2022 AI gradient is steeper among small firms, where high-AI firms outperform low-AI firms by
about 15 percentage points; the post-period cells average three cohorts.}\label{fig:double}
\end{figure}

\subsection{Economic significance}
The Sharpe ratio of the equal-weighted composite long--short portfolio rises from 0.22 in the
pre-ChatGPT period to 1.32 afterward (Table~\ref{tab:postinf}, Panel~C); for the disruption-risk
spread it rises from 0.41 to 1.09. The broad U.S.\ equity market has historically delivered a
Sharpe ratio in the neighbourhood of 0.4, so a value above one for a single-characteristic
strategy is large. Three qualifications keep the magnitude in proportion. First, the ratio is
estimated on 36 months and its standard error is roughly $1/\sqrt{3}\approx0.58$ in annual units,
so it is distinguishable from the pre-period value but not precisely measured. Second, the
portfolio's post-period return is largely explained by its technology-sector loading, so the
Sharpe ratio is in part that of a sector bet during a sector boom. Third, the annual spreads that
produce the ``$+14.6$ points per year'' headline are dominated by one cohort and by extreme
winners within it; the winsorized pooled spread of $+7.2$ points is the better summary of what a
typical top-quintile firm earned over a typical bottom-quintile firm. The first version of this
paper described the post-2022 re-rating as ``among the largest single-theme cross-sectional effects
in the recent asset-pricing record''; the evidence supports a weaker statement, namely that AI
language, which was unrelated to returns for a decade, became strongly related to them in the
2024 cohort and modestly related in the 2022 cohort.