Chinese start-up DeepSeek has shaken the belief that only a few firms with huge budgets can compete in the artificial intelligence field — potentially challenging an investment case that rewarded the biggest players with rich valuations.
Investors are digesting the implications of DeepSeek’s breakthrough products offering comparable performance at seemingly a fraction of the cost — and what this means for US tech behemoths that have funnelled billions of dollars into cutting-edge AI processors.
“What makes Monday’s tech selloff so jarring is that the valuations of many of these AI and tech companies offer no margin of error,” said David Bahnsen, chief investment officer, The Bahnsen Group. “Excessive valuation always becomes a problem, eventually, but fundamental news becomes a heightened problem when it is combined with excessive valuation.”
Over the past 24 months, most of the Nasdaq 100 Index’s gains have come from a handful of AI names, inflating valuations in the process. Prior to Monday’s open, forward earnings multiples for Nvidia, Microsoft and Amazon.com swelled past 30x, while the tech-heavy benchmark itself traded at 27x — higher than its 10-year average.
Valuations have been staying high as analyst estimates for projected profits have risen slower than stock prices. Any hit to the bottom-lines of these companies could quickly render these valuations untenable.
“With many investors heavily exposed to AI’s biggest players, disruption in the sector could ripple through portfolios,” said Morningstar’s Kenneth Lamont. “While DeepSeek itself isn’t publicly investable, one way to guard against such risks may be diversifying away from the so-called Magnificent Seven tech giants — perhaps by reallocating to equal-weighted strategies.”
Stargate
The turnaround is particularly severe given Wall Street was welcoming AI investments as recently as Friday, when Meta Platforms’ stock hit a new high after the firm unveiled a capital expenditure target that blew past estimates.
The market moves also follow the Donald Trump administration’s high-profile announcement last week that saw OpenAI, SoftBank Group and Oracle pledge a US$100-billion joint venture called Stargate to build out data centres and AI infrastructure projects around the US.
Read: Mystery surrounds Nvidia’s plan for new CPU built with MediaTek
Worries that big tech firms may not need to spend as much to train their AI models have sunk the shares of Nvidia, which has been Wall Street’s biggest AI beneficiary over the past two years. The stock plummeted 11% as of 9.36am in New York — poised to shed more than $390-billion in market value. Meanwhile, ASML Holding — supplier of some of the world’s most advanced chip-making machines — dropped as much as 12% in Amsterdam.
DeepSeek “provides a wake-up call and somewhat of a question mark on how much needs to be spent in order to build a model, and whether quite the level of CapEx we have been seeing is really required,” said Ben Barringer, global technology analyst at Quilter Cheviot.
Some urged caution. “While DeepSeek’s achievement could be groundbreaking, we question the notion that its feats were done without the use of advanced GPUs,” Citigroup analyst Atif Malik wrote in note, referring to graphics processing units.
DeepSeek’s breakthrough shows that training costs can be “way lower” for smaller AI models that can handle some specific tasks, said Nori Chiou, an investment director at White Oak Capital Partners. Still, these models won’t be replacing bigger fundamental models anytime soon, he said. “Investors are primarily looking for a reason to lock in profits, especially after last week’s positive AI news.” — Subrat Patnaik and Henry Ren, (c) 2025 Bloomberg LP
Get breaking news from TechCentral on WhatsApp. Sign up here.