"Agent vs. US Treasury Bonds" - Who will dominate the US stock market?

The clash between the AI agent wave and rising US Treasury yields is tearing the US stock market into two completely different worlds.

This week, Meta released its Agentic AI model, Muse, resulting in a $220 billion increase in market capitalization in a single week and propelling the Nasdaq 100 to become the best-performing major stock index. However, at the same time, rising US Treasury yields pushed interest rate-sensitive assets to a low point—the Russell 2000 index significantly lagged behind, and the S&P 500 index fluctuated around 7700 points, unable to achieve a significant breakthrough.

The market's internal divergence has reached an extreme level. Excluding AI-related stocks, the S&P 500 actually fell by about 1% this week; the number of stocks hitting new lows on the NYSE has exceeded the number hitting new highs for nine consecutive trading days, marking the longest losing streak since October 2023.

Tony Pasquariello, head of Goldman Sachs' global hedge fund business, characterized the current situation as a "frustrating cat-and-mouse game" between the stock market and the interest rate market, and gave clear advice: if you insist on holding long positions in stocks, you should simultaneously short US Treasuries to hedge against interest rate risk.

AI Agents Trigger Structural Market Trends, Market Capitalization Concentration Reaches New High

Meta's Muse model became the biggest catalyst for the market this week. The launch of this agentic AI product added $220 billion to Meta's market capitalization in a single week, bringing its total market capitalization back above the $2 trillion mark, and boosting the entire AI industry chain—the semiconductor sector led the gains due to the demand for hardware driven by the "inference economy," and agentic AI-related stocks collectively benefited. Goldman Sachs' Agentic AI Basket Index (ticker symbol GSXUAGNT) subsequently rose, with the high-beta momentum basket recording nine consecutive days of gains.

Goldman Sachs senior trader Rich Privorotsky points out that Muse is just the first of many entrants, and competitors from Google and OpenAI will emerge in the coming weeks and months. He believes that as large enterprises take over security, infrastructure, and distribution functions, a large number of economic frictions and intermediaries will be eliminated, which in itself is a productivity boost and structurally has an anti-inflationary effect, resulting in a positive overall impact on the stock market.

It is worth noting that the concentration of market capitalization in the US stock market has climbed to a historical extreme. Pete Callahan, chief technology analyst at Goldman Sachs, pointed out that there are now 10 US TMT companies with a market capitalization of over $1 trillion in the S&P 500, while just three years ago, at the beginning of the AI cycle, this number was only 3.

The market is severely fragmented, and "diffusion trading" has come to an end.

Behind the AI rally lies a sharp deterioration in market breadth. This week, the technology sector stood out, while the energy and financial sectors both lagged behind; the Mag7 (the seven tech giants) saw gains almost three times that of the remaining 493 S&P 500 components. ZeroHedge, citing data, points out that "diffusion trading" is now defunct—the number of stocks hitting new lows on the NYSE has exceeded new highs for nine consecutive days, marking the longest streak in nearly three years.

Meanwhile, the AI market itself has also shown divergence. AI-related power stocks have been weak this week, and their performance is almost perfectly negatively correlated with the probability of the Democrats winning the House of Representatives in the US midterm elections—the market is concerned that if the Democrats control the House, existing and future data center construction projects may be indefinitely postponed, which could potentially suppress investment in US AI infrastructure.

Furthermore, credit spreads for hyperscale cloud computing companies (such as Hyperscaler) widened significantly this week, while their stock valuations remained high. Goldman Sachs noted that the divergence between Hyperscaler's credit risk and Nasdaq implied volatility has expanded to an abnormal level, and this gap warrants attention.

US Treasury yields become the biggest obstacle, with the S&P 500 repeatedly failing to break through the 7700 mark.

The continued rise in US Treasury yields is the key variable suppressing market breakthroughs this week. After consolidating around 7700 points for two months, the S&P 500 index showed signs of breaking upwards on Tuesday, but was immediately met with a strong reaction from the bond market, falling back to its original level that evening.

Pasquariello describes this situation as a "frustrating cat-and-mouse game" between the stock market and the interest rate market: whenever the stock market attempts to break out, interest rates jump, suppressing risk appetite; and the rise in interest rates, in turn, directly suppresses interest rate-sensitive assets by tightening financial conditions. The Russell 2000 index has underperformed the S&P 500 for six consecutive weeks, and Goldman Sachs admits that small-cap stocks have been continuously sold off by long-term funds and wealth management channels, and their technical indicators have also broken down.

In terms of valuation, research from Goldman Sachs' US equity strategy team shows that the market has already priced in the sustainability of earnings growth for AI infrastructure to some extent—the median forward P/E ratio for AI infrastructure stocks has fallen from 32 times in April 2026 to the current 22 times, and some pessimistic expectations have been priced in.

Goldman Sachs: Stock holdings need hedging; a catalyst is still needed for a breakout.

In response to this divisive situation, Pasquariello offered a clear operational framework: to maintain a significant long position in stocks, one must simultaneously establish a simple short position in US Treasuries to hedge. He rated the current speculative long positions in the market at +3 on a standard scale of -10 to +10, believing that the overall position is not crowded.

He also pointed out that share buybacks are slowing, new share issuances are increasing, and US retail investors are adopting a wait-and-see approach, indicating that the current market structure is not perfect. However, he also acknowledged that considering the recent sharp rise in interest rates, it is already quite an achievement that the stock market has managed to hold onto most of its gains.

Pasquariello concludes that if a catalyst for an upward breakout appears in the market, quick money will have to rush in to buy; but before that, the game between the stock market and interest rates will continue, and ZeroHedge adds a final word to this – "yet".

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.