Goldman Derivatives Trader: AI Trade Momentum Fading, SPX May Be Range-Bound
Nashnova编辑部
Goldman's Shawn Tuteja says the AI trade is losing its collective momentum as clients pick among AI sub-themes rather than buying the basket whole; he sees the S&P 500 trading sideways for roughly six weeks, with a year-end target of 8,000 unchanged.
Earnings are beating — so why won't stocks rally?
This quarter's numbers are strong: 64% of S&P 500 companies that reported Q2 results beat EPS consensus by at least one standard deviation, among the highest readings on record. The beat lifted 2027 EPS forecasts by 2%.
Yet the market shrugged. Stocks that beat by more than one standard deviation posted a median next-day excess return of just 33 basis points versus the S&P 500 — far below the 95 bps historical median since 2010.
Tech fared worse: tech stocks that beat actually underperformed the S&P 500 by 99 bps the next day. This means → the market's capacity to reward earnings surprises is saturated; the good news was already in the price.
How did sentiment flip from fear to complacency so fast?
Two weeks ago clients were broadly de-risking. Now net positioning sits at the 67th percentile of its five-year range and gross positioning at the 89th percentile — hard to call that "underweight."
Last Tuesday, S&P 500 call volume hit 4 million contracts, an all-time record. Short-dated put/call skew — a gauge of how much more investors pay for downside protection versus upside bets — posted its largest two-day drop in nearly a decade.
Tuteja frames the shift as moving from "fear of a hike" to a "heads-I-win, tails-I-win" read on the September FOMC meeting. In plain terms = the market has priced in the best outcome as the default script, leaving very little cushion for a surprise.
If AI stalls, what is holding the index up?
Tuteja built a "SPXXAI" gauge — the S&P 500 stripped of AI-linked names. Since semiconductors peaked on June 22, SPXXAI has risen roughly 7%.
The catch-up came from quality software, alternative asset managers, large-cap financials, and metals & mining — a rotation into non-AI pockets. This means → the AI trade's ebb has not dragged the broader market down, but the replacement drivers are scattered and lack a unifying theme.
In plain terms = the index is still rising, but the leader changed from one "AI army" to several "guerrilla squads" — enough collective force to hold the level, but hard to push to new highs.
What is splitting apart inside the AI trade?
Clients no longer treat AI as a single position. They are differentiating sub-themes — compute infrastructure versus AI application software, for instance, follow very different logic.
After the late-July bounce, AI stocks have diverged sharply. Positions went through a washout and implied volatility — the options market's pricing of future swings — has reset lower.
This reflects a shift from "buy first, ask later" to "show me the sub-theme before I commit." Winners and losers within AI are heading in opposite directions.
What should investors watch over the next six weeks?
Tuteja keeps his year-end S&P 500 target at 8,000 but expects the index to trade in a narrow range for roughly the next six weeks, consistent with the path implied volatility is pricing.
Two catalysts to watch: ① whether the upcoming CPI print materially lowers the probability of a September rate hike; ② whether AI capital expenditure continues to deliver in subsequent earnings reports.
In plain terms = for the index to break out of this sideways drift, either inflation data must open the door for the Fed to cut, or real AI spending must prove the story is still alive — without at least one, the range-bound pattern is hard to break.
Content is for reference only, not financial advice.