Goldman Sachs Warns AI Infrastructure Spending Drives Half of S&P 500 Earnings Growth

Goldman Sachs Flags AI-Driven Earnings Bubble Risk

Peter Oppenheimer, Goldman Sachs' chief global equity strategist, has escalated his warnings about a potential AI-driven earnings bubble, backing his August concerns with hard data in a new report titled "Competition for Capital."1

1

The strategist argues that AI infrastructure spending and government borrowing are directly competing for the same pool of capital, driving up the global cost of capital and creating systemic risk in technology stocks.

Oppenheimer's analysis reveals that capital expenditure growth among AA-rated technology issuers surged 65% year-over-year in the second quarter, marking the tenth consecutive quarter of aggregate capex growth exceeding 35%.1

1

U.S. convertible bond issuance reached $135 billion year-to-date, with AI-related borrowers accounting for 44% of total volume. Goldman's credit team raised its full-year U.S. investment-grade issuance forecast by $200 billion to a record $2.3 trillion, with AI-related issuers now representing a quarter of all supply.

Hyperscaler Spending Cycles Dominate S&P 500 Earnings

Ben Snider, Goldman's chief U.S. equity strategist, confirmed that AI infrastructure spending is driving roughly half of all S&P 500 earnings growth in 2026.2

2

Hyperscale technology companies are expected to pour $754 billion into capital expenditure in 2026, an 83% jump from 2025. Analyst consensus estimates project hyperscaler spending will reach $905 billion in 2027, though Goldman Sachs believes this figure remains too conservative.

The concentration risk becomes stark when examining NVIDIA's performance. The chipmaker reported $96.2 billion in revenue for its second quarter of fiscal 2027, a 106% increase year-over-year, with data center sales accounting for $89.0 billion—more than 92% of quarterly revenue.2

2

NVIDIA guided third-quarter revenue to $108 billion, reflecting commitments already locked in by buyers who have announced full-year capital budgets running into hundreds of billions each.

Cost of Capital Competition Intensifies

Torsten Slok, chief economist at Apollo Global Management, independently echoed Oppenheimer's framing five days before Goldman's report, arguing that a two-decade "savings glut" has transformed into a "savings shortage."1

1

Slok explained that when projects are abundant and capital is scarce, capital competes by demanding higher returns. He noted that spreads on hyperscalers' longest-dated bonds have widened, with most paper issued in 2026 trading wider than where it priced.

Slok offered a precise explanation for why long-term yields specifically have moved more than short-term ones: "Data centers, power generation, transmission and government deficits are all long-duration claims on savings. So the competition for capital concentrates at the long end of the curve, which is why long rates have moved more than short rates."1

1

Goldman Raises S&P 500 Forecast Despite Warning Signs

Despite the concentration risk warnings, Goldman Sachs raised its 2026 earnings-per-share (EPS) forecast for the S&P 500 to $340, projecting 24% annual profit growth.2

2

This upgraded estimate anchors a year-end index target of 8,000, up from the bank's earlier call of 7,600. The S&P 500 climbed toward new highs in 2026 while its valuation multiple fell, indicating that profits rather than multiple expansion are driving the rally.

Aggregate S&P 500 earnings grew 18% year-over-year in the first quarter, with the median company posting its strongest quarterly growth rate in a decade outside the 2018 tax-cut surge and the post-pandemic reopening.2

2

However, capital spending remains discretionary, and a pullback by two or three major hyperscalers would significantly thin out the profit growth powering the index.

AI Productivity Boost Remains Largely Unreported

Goldman's forecasts embed a 0.4 percentage point AI productivity boost to S&P 500 earnings this year, rising to 1.5 percentage points in 2027.2

2

Snider noted that enterprise adoption of AI tools is broadening beyond raw infrastructure into productivity and enterprise applications. However, most S&P 500 companies still exclude AI-driven earnings gains from their reported figures, meaning the productivity dividend remains largely invisible in official results.

Jensen Huang, NVIDIA's founder and CEO, argued that compute capacity has become a revenue-generating asset rather than a back-office cost: "AI has reached its inflection point. It's doing useful work. Its tokens are productive and profitable. Now, compute is revenue."2

2

Positioning Indicators Signal Caution Ahead

Goldman's proprietary positioning indicator, which tracks hedge funds, mutual funds, and other institutional investors, has dropped to its lowest reading since March 2026.2

2

Light positioning cushions potential selloffs because less crowded exposure leaves less to unwind during downturns. The analysis flagged seasonal weakness in September 2026, amplified by midterm-year patterns, and a Volatility Index drifting above 16.

Oppenheimer's report references historical parallels including the 2008 financial crisis, the dot-com bubble of the late 1990s, and Japan's bubble in the late 1980s as prior instances where earnings rather than valuations collapsed first.1

1

For investors holding broad S&P 500 exposure, quarterly capital expenditure guidance from Microsoft, Alphabet, Amazon, and Meta deserves close attention as a leading indicator for index-level earnings, given that these hyperscalers are generating a substantial portion of the index's total earnings expansion.