Chris Wood of Jefferies has a warning for investors chasing AI stocks

The Wall Street's record run, built on the promise of artificial intelligence, risks ending in “a massive overinvestment bust” if investors ignore the lessons of recent developments in the sector, as capital spending by US technology giants accelerates at an unsustainable pace, Christopher Wood, global head of equity strategy at Jefferies, has cautioned.

Wood, speaking to ET Now, said the U.S. equity market’s momentum has little to do with trade tensions or tariffs, and is instead being driven by what he described as the “AI capex arms race”. Since early 2023, he noted, four US hyperscalers, along with Nvidia, have accounted for “nearly 50% of the gains in the S&P 500.”

The AI euphoria

“All the evidence right now today is that the AI capex is accelerating,” Wood said, pointing to forecasts that the four largest hyperscalers will spend about $350 billion this year on AI-related infrastructure. Investors, Wood argued, have brushed aside warnings from the launch of DeepSeek earlier this year, which showed that large language models could become “commodities.”

“The market has completely forgotten the lessons of DeepSeek, and is celebrating the fact that the four hyperscalers are guiding to spend 350 billion US dollars this year on AI-related capex, and everybody is celebrating the data centre construction. But in my view, this will end up in a massive overinvestment bust,” Wood said.

Retail-driven rally

The global head of equity strategy at Jefferies described the current momentum as a retail-driven phenomenon: “The market thinks AI is the next big thematic and probably many retail investors are buying stocks because the AI models are telling them to buy the stocks, which leads to a reflexive feedback loop.”

Valuations, he stressed, are stretched. “On a valuation basis, the U.S. market is trading at an all-time high price to sales,” Wood said, adding that reported earnings are often “artificially exaggerated” due to non-GAAP adjustments.

Beyond tariffs and trade

While much attention has focused on U.S. President Donald Trump’s tariff regime, Wood argued that the equity surge has little to do with trade policy. “The U.S. markets are not running on the tariff story; the U.S. market is running on the AI theme,” he said.

He contrasted the hyperscalers’ current business model with their earlier “asset light” approach. “These hyperscalers are moving from asset-light business models to asset-heavy ones,” he said, pointing out that even though they generate about $100 billion annually in profits, “$360 billion is a huge amount of money to be investing in capex.”

Also read | Ola Electric vs Ather Energy shares: Which EV bet looks stronger for your portfolio right now?

Risks of reversal

Wood maintained his long-held view that the U.S. market had already peaked as a share of global market capitalisation late last year. “When that theme unwinds, there will be a big correction,” he said, while emphasising that sentiment shifts would be the trigger to watch.

Also read | With gold prices at record highs, are gold loan lenders a better bet for your portfolio than jewellery makers?

For now, the rally remains supported by expectations of Federal Reserve rate cuts and by technical provisions in U.S. tax law that have delayed the accounting impact of the capital spending boom. But Wood’s warning was clear: “If the message of DeepSeek is correct, as these large language models are commodities and we have to remember DeepSeek is open-sourced, then it is clearly a risk that a lot of this spending will be wasted.”