AI Is Driving Earnings, but Read the Footnotes: A $98 Billion Gain Inside the Q2 Numbers

The headline of the second-quarter earnings season is that profits for S&P 500 companies grew by about half. The footnote is that two of the stars of the quarter did not earn it the usual way.

• As of August 7, FactSet showed 50.4 percent blended Q2 earnings growth for the S&P 500, but 32.0 percent excluding Alphabet and Amazon ^[1].
• Alphabet reported EPS of $9.11 against a $2.88 estimate, helped by gains on equity securities; Amazon reported $5.75 against $1.82, helped by gains on its Anthropic investment ^[1].
• Even without the two, growth was the second consecutive quarter above 25 percent and the seventh straight quarter of double-digit growth ^[1].
• Energy, Communication Services and Consumer Discretionary led sector growth ^[1].

The headline

FactSet’s August 7 update, with 88 percent of S&P 500 companies reported, showed blended earnings growth of 50.4 percent for the second quarter, up from 47.4 percent the previous week and 23.1 percent at quarter-end. Companies collectively beat estimates by 29.2 percent, the highest surprise since FactSet began tracking in 2008 ^[1].

The footnote

Two companies account for much of that. Alphabet reported earnings per share of $9.11 against an estimate of $2.88, and Amazon reported $5.75 against $1.82. Both benefited from substantial investment gains, Alphabet from equity securities and Amazon from its investment in Anthropic ^[1]. Without them, the growth rate would still have been 32.0 percent ^[1].

Why it matters

An investment gain is real money on the income statement, but it is not the same as selling more products. It depends on the market value of a stake and may not recur. Analysts therefore look at operating results too, and on that measure the quarter is still strong: it is the second straight quarter above 25 percent growth and the seventh straight with double-digit growth, with revenue growth of 15.0 percent ^[1].

Doing the subtraction

Simple subtraction shows how much the two companies mattered. Blended growth was 50.4 percent as of August 7 and 32.0 percent excluding Alphabet and Amazon, so the two lifted the figure by 18.4 percentage points (my arithmetic) ^[1]. At the July 24 snapshot, when only Alphabet had reported among the two, growth was 37.9 percent and 25.9 percent excluding Alphabet, a 12.0-point effect ^[2]. That implies Amazon added roughly another 6 points, although the mix of reporting companies also changed between the dates, so that split is approximate.

Two companies are about 0.4 percent of the 500 constituents, so the effect shows how much a weighted average depends on a few large items. The underlying picture is still strong on the FactSet figures: even without them growth was above 25 percent for the second straight quarter and double-digit for the seventh ^[1]. The footnote is not that the quarter was weak, but that headline growth of 50 percent describes the mix of operating results and investment gains, and the two should not be read as one.

Reading the next report

• Check whether the EPS beat came from operations or from one-time items.
• Compare the headline with the figure excluding outliers; FactSet publishes both.
• Look at revenue growth, which is harder to flatter than EPS.

The lesson is not that the earnings are fake. It is that the average has outliers, and in an economy where a few companies hold large investment stakes in the AI labs, those outliers can be very large. This is analysis, not investment advice.

Questions readers ask

What was S&P 500 earnings growth in Q2 2026?

FactSet’s August 7 blended figure was 50.4 percent, or 32.0 percent excluding Alphabet and Amazon ^[1].

Why was Alphabet's EPS so far above estimates?

It reported $9.11 versus a $2.88 estimate, boosted by gains on equity securities ^[1].