Author: Li Jia
Source: Wall Street News
The U.S. stock market is currently displaying a rare split: on one side, consumer confidence has fallen to historic lows and correlations between macro assets have become completely distorted; on the other side, major stock indices continue to set new all-time highs driven by the AI and semiconductor rally. What the market is truly worried about is no longer whether the uptrend can continue, but how long this highly concentrated AI-driven rally can withstand the shocks from oil prices, interest rates, and crowded positioning.
Spurred by rising expectations for U.S.-Iran talks and a surge in the semiconductor sector, U.S. stocks hit new record highs again on Tuesday. The Nasdaq 100 Index surpassed 30,000 points for the first time, and the S&P 500 rose about 0.5%; meanwhile, a decline in oil prices led to lower U.S. Treasury yields, while a rebound in the dollar pressured the performance of gold and Bitcoin.
Semiconductors remain the core driver of this rally. After UBS Group significantly raised its target price for Micron Technology, trading sentiment for memory chips heated up rapidly, with the semiconductor sector rising 14% over the past five days. However, Nvidia has begun to underperform the overall semiconductor index, indicating that funds are shifting from the leader towards sectors with higher elasticity.
Nelson Armbrust, a Goldman Sachs trader, warned that the current correlations between the S&P 500 and interest rates, gold, the VIX, and oil prices have all deviated from their 20-year historical averages and entered extreme ranges. "One side of the market has to give," he said. "The stock indices are not reflecting the full reality of the market."
AI Theme Remains, But Driving Forces Rotate
The strongest driving force in the stock market currently comes from semiconductors, especially the memory chip segment. Goldman Sachs trader Pete Callahan pointed out that the semiconductor index has outperformed Nvidia by about 16.5 percentage points over the past five trading days, marking the largest five-day advantage for the SOX relative to Nvidia since 2018.
Notably, this rally is not being driven by a broad-based surge in mega-cap tech stocks. Large-cap tech stocks overall showed relative weakness, with Nvidia slightly lagging, indicating funds are rotating from the AI leader towards more elastic trading plays within the semiconductor sector. Memory chip-related stocks opened sharply higher, with DRAM ETF nominal trading volume reaching about $30 billion for the day, and Goldman Sachs' Meme-Stocks basket also rose noticeably.
AI remains the market's main theme. AI semiconductors, Agentic AI, and AI data center stocks led the market gains. Goldman Sachs noted that the strength in the momentum factor that day was almost entirely driven by long positions, with the best-performing stocks over the past 12 months continuing to significantly outperform.
Meanwhile, increasingly extreme structural signals are also appearing in the options market. SpotGamma data shows aggressive negative Delta flows in 0-DTE options, primarily driven by selling call options; and the combination of "rising volatility and rising spot prices" persists. This means the current rally is not a typical low-volatility expansion of risk appetite but is increasingly being driven by positioning squeezes and options trading structures.
Consumer Confidence Bottoms, But Behavior and Sentiment Diverge
Goldman Sachs trader Chris Hussey pointed out that many are puzzled: why are U.S. stocks still hitting new record highs when consumer confidence indicators have fallen to historic lows? His explanation is that what consumers "feel" and what they "actually do" are inconsistent. In other words, sentiment is pessimistic, but consumer behavior has not deteriorated in sync.
Meanwhile, fiscal stimulus is still supporting household cash flow. Chris Hussey mentioned that tax relief from the budget bill last July is improving household balance sheets, partially offsetting the pressure from rising gasoline prices.
U.S. macro data also shows clear divergence. The Chicago Fed National Activity Index rebounded sharply, the Conference Board Consumer Confidence Index beat expectations, and the Dallas Fed Manufacturing Index held steady; while the S&P CoreLogic Case-Shiller Home Price Index weakened and the Philadelphia Fed Manufacturing Index missed expectations. Overall, U.S. economic data maintains a state of "slightly stronger than expected."
This also explains the market's current contradiction: consumer sentiment is extremely weak, but economic data has not significantly deteriorated yet, and actual consumer behavior has not shown a synchronous decline. However, the AAII Bull-Bear Spread remains negative, indicating investor sentiment has not truly turned optimistic alongside the new stock market highs.
Macro Correlations Fail, Negative Gamma Intensifies: Goldman Sachs Warns of Structural Cracks in U.S. Stocks
Goldman Sachs trader Nelson Armbrust warned that current U.S. stock indices do not reflect the full reality of the market. Correlations between the S&P 500 and key macro assets have comprehensively deviated from long-term averages: correlation with interest rates is at a decade low, correlation with gold is at a decade high, correlation with the VIX is at a two-year high, and correlation with oil prices is at a decade low.
These are extremely rare levels even when viewed across a 20-year historical dimension. In other words, while U.S. stocks are still rising, their traditional联动 relationships with interest rates, volatility, commodities, and safe-haven assets are breaking down. For investors relying on historical correlations for asset allocation, hedging, and risk budgeting, this means model stability is declining.
Meanwhile, Gamma has turned negative. In a negative Gamma environment, the market becomes more sensitive to price fluctuations, and the state of "spot prices rising while volatility simultaneously increases" also means the current market action is not a typical low-volatility one-sided bull market, but increasingly driven by positioning and options structures.
Goldman Sachs HF Trend Monitor data shows hedge funds' current allocation to the momentum factor has risen to the 90th percentile, semiconductor positioning has reached a record 10%, while software positioning has fallen to its lowest since 2019. Highly crowded positioning implies that the rally may still continue upward in the short term driven by chasing funds, but once a reversal occurs, the pullback could also be more severe.
How Far Can U.S. Stocks Go? It Depends on Three Constraints
The first constraint comes from oil prices. Diplomatic progress can quickly depress geopolitical risk premiums, but cannot immediately repair the buffer capacity of shipping, insurance, refineries, and real supply chains. As long as uncertainty remains around the Strait of Hormuz situation and the prospects for a U.S.-Iran ceasefire, oil prices may fluctuate repeatedly between optimistic expectations and tail risks.
The second constraint comes from semiconductor positioning. The current U.S. stock rally relies increasingly on AI and semiconductors, especially memory chips and momentum long trades. If funds continue chasing this direction, indices may maintain strength; but the more crowded the positioning, the higher the market's sensitivity to earnings, guidance, or fund flow changes—any slight disappointment could be rapidly amplified.
The third constraint comes from correlation breakdowns. The simultaneous deviation of the S&P 500's relationship with interest rates, gold, the VIX, and oil prices from long-term averages means the current rally does not equate to a comprehensive easing of macro risks. More accurately, it is the result of a combination of declining geopolitical risk premiums, falling U.S. Treasury yields, AI/semiconductor momentum, and positioning squeezes.
Therefore, the U.S. stock market's "solo party" may continue, but the stability of the rally is declining. What truly matters is not whether the indices can set new highs, but which variables will force a repricing of this trading logic—whether oil prices will rise again, whether interest rates will re-ascend, whether semiconductor momentum will cool, and when those distorted correlations will revert.
