The trade is showing up in options prices before it shows up in most portfolios. The cost of hedging a decline in the Invesco QQQ ETF, the exchange-traded fund that tracks the Nasdaq-100, has climbed far above the equivalent protection on small-cap funds, a gap that Goldman Sachs derivatives strategists say reflects a broad retreat from America’s largest technology companies.
Brian Garret, Goldman’s head of derivative strategy for the region, laid out the case in a note Thursday: investors are underweighting U.S. technology stocks, especially the group of megacaps known as the Magnificent Seven, and shifting money toward companies they see as direct beneficiaries of artificial intelligence, above all the semiconductor industry. “Underweighting large-cap stocks appears to be the prevailing strategy,” Garret said.
The caution stems from performance. The Magnificent Seven, the cluster of megacap names that drove the bull market’s early phase, have lagged the broader market in recent months, and investors who bought them for growth have watched their returns trail the index they were meant to beat. The options market’s pricing of downside protection on the QQQ reflects that disappointment, the note said.
The numbers tell the story. The Magnificent Seven, a group that includes the largest platform, software and hardware companies in American markets, carried the indexes through 2023 and 2024, then gave back their leadership as investors questioned whether growth could keep up with valuations. The group’s aggregate performance has trailed the S&P 500 over the recent stretch, and the underperformance has fed the rotation rather than reversing it.
The money is not leaving the market; it is rotating. Investors are ignoring what Garret described as the blue-chip sector broadly, while favoring AI supply-chain names whose earnings are tied to the physical build-out of data centers. Semiconductor companies have absorbed much of that demand, with investors betting that whoever wins the race to sell AI models, the chipmakers get paid first.
The strategy has a caveat attached. Goldman said investors may stay cautious on the largest technology companies unless hyperscalers, the cloud providers that buy chips at enormous scale, show stronger earnings growth. The banks’ clients are asking whether the computing capacity being installed is generating revenue fast enough to justify the spending, and the answer, for now, is not obvious to everyone.
The rotation marks a change in how the AI trade is being expressed. In the early phase of the boom, investors bought the companies writing the checks: the platform giants with cloud businesses and massive capital budgets. The new phase favors the companies receiving the checks, the chip designers and equipment makers whose order books are full regardless of which software products succeed.
Semiconductor names have become the market’s chosen expression of the AI trade for a straightforward reason: their revenue is booked before software products prove themselves. Chip orders are placed on capacity plans and long lead times, so the earnings are visible months in advance. Platform companies, by contrast, must show that AI features translate into subscriptions and advertising dollars, a connection investors are still waiting to see in the numbers.
Garret’s note landed in a week when the divergence was visible in the tape. Semiconductor names have held up better than the broader technology complex in recent sessions, while megacap software and platform stocks have drifted, a pattern consistent with the underweighting he described.
The hedging data adds texture to the shift. Demand for protection on the QQQ has been strong relative to demand for protection on small-cap indexes, an unusual posture because small caps are normally seen as riskier. The inversion says investors fear a drawdown in megacaps more than they fear a stumble in the rest of the market, and they are paying accordingly.
Garret’s own guidance is hedged. The note does not argue that megacap technology is overvalued in absolute terms, only that the relative trade has shifted, and it lists the conditions under which the rotation would reverse, including a pickup in hyperscaler earnings and a stabilization in AI spending plans. The derivatives market, he said, is simply pricing the most likely path from here, and that path currently runs through chips.
Analysts who follow fund flows said the positioning could persist. Institutional investors have spent months trimming megacap holdings and rotating into industrial and semiconductor names, and nothing in the earnings calendar so far has forced a reversal. The second-quarter reporting season, due to begin in the coming weeks, will test whether the rotation is durable or whether the megacaps reclaim their place once numbers land.
The sector’s earnings season will provide the next data points. Chipmakers report in the coming weeks alongside the platform giants, and the contrast between the two sets of numbers will likely determine whether the underweighting of megacaps becomes a longer-term allocation change or a passing phase in a market that has already rotated several times.
Goldman’s note is one view in a crowded debate. Other banks have argued that the megacaps’ balance sheets remain strong enough to support their valuations, and that a rotation of this kind historically runs for a quarter or two before reversing. The options market, for now, is pricing the bearish version, and the semiconductor trade, for now, is where the money is going.
