The divergence between the US stock market and credit market is approaching a critical tipping point.
Goldman Sachs' latest report shows that credit spreads for AI hyperscalers have widened to their highest level of the year, while corresponding stock prices have yet to fully reflect this pressure.
Analysts state bluntly: either the bond market is "overly pessimistic," or the stock market still has approximately 15% of downside that has not been digested.
Goldman Sachs strategist Louis Miller noted in the latest weekly report that the Goldman Sachs AI Hyperscaler Credit Spread Index (GSUCHS30) has widened to its widest level of the year, and the last time spreads approached current levels, related stocks were priced about 15% lower than now.
Meanwhile, Goldman Sachs' broader AI beneficiary basket has fallen about 25% from its June high relative to the S&P 500 excluding AI components.
Credit market flashes warning ahead of stocks
Goldman Sachs warns that the AI Hyperscaler Credit Spread Index (blue line, shown inverted) has risen to its widest level of the year, while the corresponding stock basket (black line) has only just begun to pull back.
The last time credit spreads widened to near current levels in July, related stock prices were about 15% lower than now.
Credit market stress is not limited to large cloud computing companies.
AI hyperscaler credit default swaps (CDS) have just touched an all-time high.
Further down the capital structure, the situation is even more severe: of 20 junk-rated AI data center bonds, 7 rely on a single tenant CoreWeave for repayment, and CoreWeave's own bonds have fallen to distressed levels.
Miller further pointed out that virtually no part of the AI sector was spared this week.
Goldman Sachs' AI theme dashboard shows that NeoClouds fell 24.8% over five days, optical networking dropped 21.4%, and memory chips declined 16.6%.
The only theme posting a notable gain was "AI-at-Risk" (traditional industries facing AI disruption), which rose 3.5%.
AI capex growth peaks, earnings contribution to gradually turn negative
Goldman Sachs' core explanation for current pressure lies in the inflection point of capital expenditure growth.
Miller warns: "The widening of spreads across the entire sector forces investors to confront a question: what happens when AI spending growth slows. AI capital expenditure has likely already peaked, and financing sustainability and return on capital have replaced near-term earnings estimate revisions as the core of market debate."
According to Goldman Sachs Investment Research data, hyperscaler capital expenditure year-over-year growth peaked at 116% in the third quarter of 2026, and the market generally expects it to slow quarter by quarter thereafter, dropping to about 20% by the end of 2027.
The impact of this trend on S&P 500 earnings has already been clearly calculated.
Goldman Sachs estimates that hyperscaler capital expenditure (net after depreciation) contributed about 11 percentage points to 2026 S&P 500 earnings per share growth, dropping to 7 percentage points in 2027, and by 2028, as massive data center depreciation and amortization begin to take their toll, the contribution will turn to -1 percentage point.
In other words, capital expenditure that once boosted earnings will, once growth slows, become a drag instead.
Internal market indicators simultaneously flash warnings
The pressure on the AI sector is not an isolated phenomenon, and Goldman Sachs' "alarming charts" also show stress signals at a broader market level.
On the sentiment side, Friday's put option volume reached "the second highest of the year, second only to June."
On the risk appetite side, Goldman Sachs' Beta Factor Hedge Portfolio (GSP1BETA) has fallen 15% from its recent high, and high-beta cyclical stocks have broken below their upward trend channel.
On the growth expectations side, the economic growth expectations implied by Goldman Sachs' Cyclicals vs. Defensives Ratio (GSP1CYDE, excluding commodities) have dropped to about 2%, indicating that the market's pessimism about growth has exceeded what fundamental data supports.
On the leverage and positioning side, hedge funds' Beta exposure is higher than what leverage levels would imply 鈥?Beta-adjusted net exposure sits at the 67th percentile of the past five years, while net leverage is only at the 17th percentile.
Miller also noted that financial conditions have "shifted from a positive driver supporting growth to a negative drag," with higher interest rates and wider spreads directly hitting debt-dependent stocks, and Goldman Sachs' Debt-Sensitive Stock Basket (GSX1DEBT) continues to underperform the broader market.
Goldman Sachs offers five response strategies
Despite issuing the above warnings, Miller does not predict a full-scale market collapse.
He believes that "despite rising correlations and increasingly crowded positioning, the market is still digesting pressure through rotation rather than a sharp index-level decline," and proposes five trading strategies.
First, go long the "de-AI-fied" broad market. The S&P 500 excluding AI components (SPXXAI) has outperformed the S&P 500 by 586 basis points since October 6, and the correlation between the AI sector and the rest of the market has turned negative. Miller believes that nominal growth and employment data remain resilient, the effect of interest rates suppressing valuations has been partially released, and the logic supporting rotation rather than systemic decline continues.
Second, go long quality factor, but be mindful of technical risks. Goldman Sachs' Quality Factor Hedge Portfolio (GSP1QUAL) has risen 20% in three months, with RSI exceeding 80. Miller warns of significant reversal risk once inflation and interest rate expectations ease. The European Quality Factor (GSPEQUAL) has risen 7.5% over the past month, showing clear catch-up momentum.
Third, medium-term healthcare trade. Miller favors ACA beneficiary stocks (GSHLCACA), betting that Democrats may restore expired ACA premium tax credit policies to improve patient flow for healthcare providers currently mired in bad debt and self-pay patient pressure.
Fourth, go long uncorrelated opportunities in Brazil. Current valuations are at mid-cycle levels, foreign participation is below the peak of the previous election cycle, and interest-rate-sensitive domestic stocks (GSBZRATE) are lightly positioned. Miller sees upside around Brazil's October 25 runoff election, with the price-to-earnings ratio potentially re-rating from the mid-teens to the high teens.
Fifth, hedge China Taiwan semiconductor exposure. Miller noted that TSMC's stock fell 2.3% even after raising full-year guidance and accelerating capacity expansion, "fully demonstrating the market's current willingness to buy into the AI strong demand narrative." Hedge fund positioning in China Taiwan has reached an all-time high, while performance has "failed to meet extremely high expectations." Goldman Sachs' Asia team recommends shorting the China Taiwan Semiconductor Basket (GSXATWSM) and the China Taiwan Tech (excluding TSMC) Basket (GSXATTXT).
Comments