Big tech stocks could turn more attractive in the artificial intelligence trade as chipmakers show continued volatility, according to Goldman Sachs Group Inc. strategist Christian Mueller-Glissmann.
While and AI capex beneficiaries have led markets higher, rather than hyperscalers, Mueller-Glissmann noted these stocks represent a particularly volatile part of the AI spectrum with significant positioning and leverage through vehicles like exchange-traded funds and options.
“If you continue to see the momentum in AI to be positive, you want to diversify toward the hyperscalers, and maybe diversify away from the semiconducters. Because they are the really volatile part of the AI capex spectrum,” Mueller-Glissmann, the head of asset allocation research at Goldman, said in a Bloomberg Television interview.
Semiconductors have been the favored over the past few months, while hyperscalers — including tech giants Amazon.com Inc., Oracle Corp., Microsoft Corp., Alphabet Inc. and Meta Platforms Inc. — have lagged as investors questioned their excessive capital spending on data centers.
The Philadelphia Stock Exchange Semiconductor Index, or SOX, has meanwhile surged roughly 150% in the past year. Speaking on the recent chip surge, Mueller-Glissmann said “to some extent, it could make sense to diversify away from them.”
Speaking more broadly, Mueller-Glissmann noted ’ risk appetite indicator has ticked higher, driven initially by AI capital expenditure boosting earnings but more recently by the reopening of the Strait of Hormuz.
This has created what Mueller-Glissmann described as a “Goldilocks spectrum” of falling inflation expectations coupled with relatively good earnings growth. But bullish investor sentiment doesn’t automatically signal it’s time to turn bearish on markets even as positioning indicators show risk-on appetite has increased, he said.
“When everybody is bullish it does not automatically mean you want to be bearish, but you have the risk of setbacks and corrections if the market doubts these drivers,” Mueller-Glissmann added.