There have been a few emerging narratives in recent weeks that suggest the AI trade is looking shaky, but Goldman Sachs’ chief US equity strategist isn’t worried.
In an interview with Business Insider, Ben Snider dismissed bubble concerns and said he’s continuing to lean into some of the major themes in the AI trade.
One fear rippling through markets is that hyperscalers could suddenly pull back on their massive spending efforts. If one firm decides its investment isn’t delivering a satisfactory return, it could slash its spending projections. If investors reward such a move, others could follow suit, the thinking goes.
To Snider, however, the risk of such a pivot is small. Growing spending on the AI buildout should continue to fuel the trade, he said.
Another worry is that valuations in the AI and tech space are fairly cheap despite strong earnings growth, a signal that investors may be wary of paying a higher premium for these stocks because they are skeptical about how long the earnings boom can last.
For Snider, it’s a counterintuitive argument.
“It’s interesting to hear investors fear both elevated valuations as a warning sign and discounted valuations as a warning sign,” he said.
“I’m always comforted when I hear bearish arguments,” he continued. “It’s an affirmation to me that there is still some equity risk premium in the market. The day when everyone agrees the outlook is good and there’s nothing to be afraid of, that’s the day there’s a clear signal that the market is overvalued.”
Another concern is valuation. The Shiller CAPE ratio is near all-time highs, floating around peak levels seen during the dot-com era and in 2021. Combine that with the fact that the S&P 500 has rallied more than 20% in the last year, and one starts to wonder if the upside is limited.
But other valuation metrics tell a different story, like the 12-month forward PE. Snider put it succinctly: Despite the S&P 500’s big gain over the last 12 months, its PE multiple is lower than it was a year ago.
It’s a sign that earnings growth, not multiple expansion, continues to drive the market. And Snider says some of the most attractive parts of the market remain those that have outperformed in recent months and years.
3 AI themes to invest in
Snider highlighted three subsets of the AI trade that he’d lean into most right now.
First, he’d stay with the AI infrastructure names as they’re trading at relatively low valuations. This encompasses businesses producing products like semiconductors, servers, and AI networking firms.
“For much of the semiconductor complex, including memory, there’s been very little, if any, expansion in valuation multiples that implies a continued skepticism,” he said.
Second, he said the power infrastructure theme is “extremely attractive” both due to the ongoing AI spend and because of the Iran war’s impact on power and energy. Energy infrastructure has been a pillar of the “HALO trade” in 2026, with the infrastructure buildout key to fueling the AI industry’s insatiable demand for power.
And third, Snider likes the hyperscaler firms that doing all of the AI spending on data centers for their cloud platforms. Think Amazon, Microsoft, Meta, Alphabet, Oracle, and IBM.
These stocks have outperformed since late 2022, but many have struggled this year as investors have rotated into stocks whose earnings are benefitting from hyperscaler spending.
“They’ve just had one of their worst months on record. And as a result, on an earnings basis, they are trading at very discounted valuation multiples,” Snider said.
“On a PE basis, they are trading at the very bottom of their range over the last decade, not dissimilar to how they traded in March 2020 or at the end of the bear market in 2022,” he added.
Examples of funds that offer exposure to these themes include the Tortoise AI Infrastructure ETF (TCAI), the Defiance AI & Power Infrastructure ETF (AIPO), the iShares U.S. Power Infrastructure ETF (POWR), and the Roundhill Magnificent Seven ETF (MAGS).
