The S&P 500's 32% Earnings Surge Is Real — Just Don't Mistake It for a Broad Boom

Wall Street just raised its full-year 2026 forecast for S&P 500 earnings growth to roughly 32%, up from 24% a few weeks earlier, and called the jump the direct work of the AI build-out. It is the kind of number that lands a headline and then disappears. For an ordinary index investor it deserves a second look, because the growth is not spread across the market the way the round figure implies. More of it than you might think sits in a handful of AI-heavy names — and part of it is a one-time accounting event.

Start with the honest part: the underlying profits are real. Second-quarter S&P 500 earnings rose about 50% year over year, the biggest jump since 2021, and 86% of reporting companies beat analyst estimates, the most since the same year. This is not a paper rally built on valuation alone; the money is showing up on income statements. The question is whose.

Now subtract the distortions. Two of the quarter's biggest beats did not come from operations. Alphabet's earnings per share of $9.11 versus a $2.88 estimate included a $98 billion unrealized gain on equity holdings. Amazon's $5.75 per share versus a $1.82 estimate was inflated by a $53.4 billion gain on its stake in Anthropic. Strip those two out and the quarter's blended growth falls from roughly 50% to 32%, and the aggregate earnings surprise drops from 29.2% to 10.9% — still healthy, still historically above average, but a different story than the one the round numbers tell.

The concentration runs deeper than one quarter. Information Technology alone has driven about half of the S&P 500's 2026 earnings growth, and adding Communication Services pushes two AI-tied sectors to roughly 63% of the total. The ten largest stocks now generate about a third of the index's net income, up from a 15–17% share a decade ago. Inside the sectors, the pattern repeats: technology's earnings growth of 54% collapses to about 30% when Nvidia and Micron are removed, and Communication Services flips from a 49% surge to a decline once Alphabet and Meta are excluded. Strip away the few builders of the AI story and the index's "boom" thins out rapidly.

The underlying driver is capital, and it is real and committed. Goldman Sachs expects hyperscale tech companies to spend about $754 billion on capital expenditures in 2026, an 83% jump from 2025, and it attributes roughly half of all S&P 500 earnings growth in 2026 and 2027 to AI-infrastructure beneficiaries. That is the flow to trace: a small number of companies are spending enormous sums on chips, data centers, and energy, and the earnings of everyone upstream — memory makers, industrial suppliers, utilities, the chip designers themselves — are growing off that spending.

This is where the index's exposure and the reader's risk converge. Technology is now about 37% of the S&P 500's weight, above the 34% peak reached in the 2000 dot-com bubble, and Nvidia alone sits at nearly 8% of the index — the largest single-stock weight in the index's history. The S&P 500 trades at about 21 times forward earnings, in the 88th percentile of its 40-year range. So the market is paying a premium multiple for earnings that depend on a narrow set of AI names continuing to convert massive spending into profit.

That conversion is the live question, and the first signs of friction are showing. Nvidia has warned that surging memory costs are narrowing its margins even as demand booms; Apple's sales outlook disappointed on prices and supply; Meta's shares fell after it reported sharply lower free cash flow even as it reaffirmed its capital-spending plans. None of these is a collapse — this is not 2000, when the boom was debt-funded against speculative revenue. The AI build-out is cash-flow funded, with data-center capacity largely pre-committed. The risk is not that the earnings are fake; it is that the market has concentrated its bet on a handful of names converting a giant investment cycle into durable, recurring profit — and if that conversion slows, the index's weight has nowhere diversified to hide.

An S&P 500 index fund was long a reasonable shorthand for "the U.S. economy." Today it is, to a degree few realize, a leveraged-when-it-works bet on whether AI infrastructure spending becomes AI revenue. Wall Street's 32% is real money and real growth. Read it as a description of what a few dozen companies are doing — not as confirmation that the market beneath them is broad.