• Goldman Sachs (GS)' preferred breadth measure has fallen to its weakest level since the dot-com era, with the median S&P 500 stock trading 16% below its 52-week high.
  • Investor positioning has retrenched to around its market-low level earlier this year, indicating headline index strength is not matched by broad conviction.
  • Goldman sees potential for lagging stocks to catch up if macro uncertainty eases, but warns of fragility if AI leadership disappoints.

AI Dominance Distorts Market Signals

The S&P 500's near-record level is masking unusually weak participation, according to Goldman Sachs. A small group of AI-linked megacaps is carrying the benchmark while the typical constituent lags materially, creating a two-sided outlook: fragility if AI leadership disappoints, but potential for a broad "catch-up" rally if macro uncertainty recedes and capital rotates into laggards.

Goldman's preferred breadth measure has fallen to its weakest reading since the dot-com era.

The S&P 500 was roughly 1% below its August record high, but the median constituent was 16% below its own 52-week high.

Investor positioning has retrenched to around its market-low level earlier this year, according to Goldman. That means headline index strength is not being matched by broad investor conviction.

Technical indicators underscore the divergence: fewer than half of S&P 500 stocks were above their 200-day moving averages, approximately 70% were below their 50-day averages, and new 52-week lows had exceeded new highs for nine straight sessions.

Market Structure and Concentration Risks

The immediate catalyst is not one company-specific event. It is a market-structure development tied to the concentration of gains in AI beneficiaries—semiconductors, cloud platforms, data-center infrastructure, and large technology firms—rather than a synchronized rise across the 500 companies.

Goldman previously estimated that the largest technology names generated 53% of the S&P 500's 2025 return, and it projected AI infrastructure investment would account for about half of the forecast earnings growth in 2026. The same concentration that lifts reported earnings and index returns also leaves the market more exposed to disappointment in a relatively small set of firms.

The concentration is especially important because the S&P 500 is capitalization weighted: a handful of the biggest companies can lift the index even if most constituents decline. One analysis citing Goldman's work put AI-linked stocks at nearly 45% of S&P 500 market value, up from about 25% around ChatGPT's late-2022 launch.

| Area | What it includes | Why it matters | |---|---|---| | AI semiconductors | Chip designers, memory firms, foundries, networking equipment | Their revenue and earnings expectations are closely tied to data-center capital expenditure | | Cloud and platforms | Hyperscale cloud providers, software platforms, AI-model developers | They fund much of the computing buildout and may monetize AI through cloud usage, software pricing, and advertising | | Power and data-center infrastructure | Utilities, equipment suppliers, electrical systems, real estate/data centers | AI expansion requires substantial power generation, grid capacity, cooling, and physical computing space | | Broader S&P 500 | Financials, consumer companies, industrials, health care, small-growth firms | These groups have not participated evenly, leaving the cap-weighted index less representative of the typical stock |

Macro Backdrop Less Accommodating

The macro backdrop is currently less accommodating than a simple "AI boom" narrative suggests. On September 16, the Federal Reserve raised the federal-funds target range by 25 basis points to 3.75%–4.00%, saying economic activity was solid but inflation remained elevated. Higher rates raise the discount rate applied to future profits—particularly relevant to long-duration growth and AI stocks—and can restrain credit-sensitive sectors that might otherwise lead a broadening rally.

For AI leaders, persistent capital investment supports earnings, but high valuations make these shares more sensitive to slowing growth, cost overruns, energy constraints, or higher bond yields. For lagging sectors, if inflation eases and the Fed can stop tightening, banks, industrials, consumer cyclicals, health care, and smaller companies could benefit from lower financing pressure and a rotation away from crowded AI trades.

Politically, no new law is the direct subject of the Goldman note. Still, government policy matters through monetary policy, AI and chip regulation, export restrictions, energy permitting, grid investment, trade policy, and fiscal incentives for domestic semiconductor and data-center capacity. The Fed's stated focus remains bringing inflation back toward 2%, so inflation data and policy expectations are central to whether market breadth improves.

Dot-Com Parallels and Outlook

The analogy is about breadth and concentration, not proof that today's market is identical to 2000. Goldman has emphasized a key difference: the current AI leaders are supported by substantial earnings and capital spending, whereas the late-1990s episode included many firms whose valuations rested on more speculative business expectations. Even so, Goldman has warned that some individual AI beneficiaries have valuation characteristics resembling dot-com-era winners, where a miss in expected growth can trigger a sharp valuation reset.

Goldman's strategic case is cautiously constructive: light positioning and depressed breadth leave room for underperforming stocks to catch up if uncertainty fades. Its broader 2026 research had forecast S&P 500 earnings per share of $340, up 24% year over year, and a year-end index target of 8,000, while explicitly flagging narrow breadth and momentum as caution signals.

| Scenario | What could drive it | Likely market effect | |---|---|---| | Broadening / catch-up rally | Easing macro uncertainty, declining inflation pressure, stable or lower yields, improving earnings revisions outside AI | Laggard sectors and equal-weight indices outperform; the overall index can advance with healthier breadth | | Continued narrow advance | AI earnings and investment remain exceptional while other sectors' profit outlooks stay subdued | The S&P 500 may remain resilient, but concentration and drawdown risk stay elevated | | AI-led correction | Disappointing AI monetization, weaker data-center spending, tighter financial conditions, or rising yields | A small group of heavyweight stocks could pull the index lower disproportionately | | Broad risk-off decline | Inflation, policy tightening, geopolitical disruption, recession fears, or credit stress | Weak breadth becomes more consequential because there are fewer rising stocks to cushion the index |

The most relevant related development is the policy-rate increase: with the Fed now at 3.75%–4.00% and stating that inflation remains elevated, a rotation beyond AI likely requires evidence that price pressures are moderating without a major deterioration in growth.

Bottom line: The headline is a warning about index composition, not a definitive crash signal. It argues for distinguishing "the S&P 500" from "the average S&P 500 stock," monitoring AI earnings and capital-expenditure guidance, and watching whether breadth measures improve as macro conditions evolve.