The AI trade is going full throttle, a positive energy that is driving up the broader stock market. But it will slow down, Goldman Sachs believes, and Wall Street must be ready when that time inevitably comes.
Goldman analysts sounded the alarm in a note, warning that the AI trade will trail off when Big Tech’s spending for artificial intelligence isn’t increasing at the breakneck speed it is now.
Over the past four quarters, Alphabet, Amazon, Meta Platforms, and other tech companies that operate global networks of data centers—so-called data hyperscalers—have invested more than $300 billion in AI capital expenditures.
That huge sum has been a boon to the earnings of companies that support the data center buildout—power suppliers and the makers of electrical equipment andsemiconductors.
As Goldman’s analysts noted, the prices of AI infrastructure stocks “have far exceeded the trajectory of near-term earnings,” which they consider a reflection of investor optimism.
Still, a slowdown in AI capex growth is on the way if Goldman’s team is right. Analysts “assume a sharp deceleration” beginning in the fourth quarter and lasting into 2026. And a big drop could pose what they describe as a “substantial downside risk to both the AI trade and the broad S&P 500.”
Capex spending by hyperscalers totaled $158 billion in 2022; for 2026, the consensus estimate is $433 billion. If capital expenditures are to revert to the 2022 level, it would pummel the broader S&P 500—the consensus estimate of $1 trillion in 2026 sales growth for companies in the index would sink by 30%.
“This extreme reduction in capex would likely be accompanied by a deterioration in the outlook for long-term AI-driven earnings growth, weighing on valuations as well,” Goldman wrote.
Though Wall Street expects capex to decline, the hyperscalers keep raising their spending expectations. Capex is on pace to grow about 50% this year, or roughly 2024’s rate. And their estimates have moved higher by $70 billion for 2026 since the start of the most recent second-quarter earnings season.
Third- and fourth-quarter earnings reports will be “key tests for the durability of hyperscaler capex,” as well as the valuations of AI infrastructure stocks, Goldman analysts wrote.
The impact of a capex slowdown on the hyperscalers is unclear. If the deceleration is viewed as a signal of slowing AI demand, long-term growth forecasts could take a hit.
At the same time, a reduction in capex “alleviate concerns about whether these companies can generate sufficient returns on their investments,” Goldman wrote.
Write to Mackenzie Tatananni at mackenzie.tatananni@barrons.com
