The US earnings season for the second quarter of fiscal 2026 is reshaping market expectations for corporate profitability. S&P 500 companies reported a year-over-year EPS growth rate of 45%, significantly exceeding the roughly 22% forecast at the start of the quarter, with the proportion of earnings beats also at historic highs.

However, behind these impressive figures, the composition of earnings is shifting. A Goldman Sachs analysis reveals that approximately 19 percentage points of the S&P 500's Q2 EPS growth came from "other income" at Alphabet and Amazon, primarily driven by unrealized gains on equity investments rather than growth in core operating profits.

Excluding these non-operating gains, the S&P 500's year-over-year EPS growth for Q2 still stands at roughly 26%, indicating that the underlying business fundamentals are not distorted. Yet, US earnings are becoming increasingly reliant on a handful of AI beneficiaries, with mega-cap tech companies persistently expanding capital expenditures and turning to debt and equity financing to support their AI infrastructure investments.

In other words, the core issue of this earnings season is not whether earnings are growing, but where the growth is coming from and whether it is sustainable.

Nearly Half of 45% EPS Growth Comes from Investment Income

The most closely watched data point this quarter was the S&P 500's 45% year-over-year EPS growth. While this pace far exceeded market expectations, Goldman Sachs noted that a substantial portion came from "other income" line items.

Data shows that Alphabet and Amazon together contributed roughly 19 percentage points to Q2 EPS growth, mainly from unrealized gains on equity holdings. Microsoft also contributed about $3 billion in other income. Specifically, Alphabet posted approximately $98 billion in quarterly other income, while Amazon reported around $53 billion, largely stemming from the increased value of private company equity investments.

After removing these factors, the S&P 500's year-over-year EPS growth for Q2 is about 26%. While clearly lower than the headline 45% figure, it still represents one of the fastest growth rates since 2021.

This suggests that US earnings are not entirely reliant on "window dressing" from accounting gains, but the quality of earnings is changing: investment income is becoming a significant component of large tech companies' profits. Data indicates that for mega-cap tech companies, "other income" as a share of GAAP profits rose to 61% in Q2 2026, a noticeable increase from levels seen in recent years.

Stronger Earnings, Weaker Market Rewards

While the earnings numbers remain robust, market reactions have shifted.

As of July 31, 61% of S&P 500 companies had reported Q2 results, covering approximately 66% of the index's market capitalization. Among them, about 64% of companies beat EPS estimates by at least one standard deviation, a proportion near historical highs. However, the stock price gains following earnings beats have been significantly weaker than in the past.

Goldman Sachs data shows that historically, S&P 500 companies that beat EPS estimates saw their stocks outperform the index by an average of 95 basis points on the next trading day. In this quarter, however, TMT companies that exceeded earnings expectations actually underperformed the S&P 500 by an average of about 192 basis points the following day.

In contrast, non-TMT companies that beat forecasts recorded an average positive excess return of roughly 75 basis points on the next day. The market is sending a clear signal: for AI leaders, investors have already priced in high growth expectations, and simply beating earnings estimates is no longer sufficient to drive stock prices higher.

AI is Reshaping the US Earnings Landscape

Current US earnings growth is increasingly concentrated in the AI supply chain. Goldman Sachs data indicates that AI infrastructure-related companies contributed roughly one-third of the S&P 500's Q2 EPS growth. Looking ahead to the second half of 2026 and into 2027, the sector is expected to account for over half of the total earnings increment.

By company contribution, Alphabet contributed about 28% of the S&P 500's Q2 EPS growth, Amazon contributed roughly 16%, Micron contributed about 10%, and Nvidia is estimated to contribute about 9%. The top ten contributors together accounted for nearly 79% of earnings growth. This means the overall earnings performance of the S&P 500 is increasingly dependent on a few beneficiaries of AI infrastructure.

For index investors, earnings growth remains strong. But for market structure, rising earnings concentration implies greater risk concentration.

AI Investment Soars, Cash Flow Comes Under Pressure

Another notable change this earnings season is the pressure from capital expenditures at mega-cap tech companies. Alphabet, Amazon, and Microsoft reported Q2 cloud revenue growth of 48% year-over-year, accelerating from 39% growth in Q1, as AI demand is translating into cloud computing growth.

However, to secure an edge in AI infrastructure, these companies' capital investment is also expanding rapidly. The combined capital expenditure of these mega-cap tech companies reached $182 billion in Q2, while their free cash flow was only about $5 billion. This funding gap is forcing companies to rely more heavily on external financing. During the quarter, these companies collectively issued about $51 billion in bonds and completed approximately $50 billion in equity financing.

Goldman Sachs forecasts that mega-cap tech companies' capital expenditure will exceed $1 trillion in 2027, a year-over-year increase of about 33%, a significant upward revision from earlier this year. As the AI investment cycle continues, tech giants may need to consistently tap into financing markets. Goldman Sachs' credit team expects these companies could issue around $400 billion in investment-grade bonds in 2027.

AI-driven earnings growth is ongoing, but whether capital expenditures can translate into sufficient returns will become a key focus for the market going forward.