U.S. technology firms plunged in premarket trading, as Chinese startup DeepSeek sparked concerns over competitiveness in AI and America’s lead in the sector, triggering a global sell-off.
Shares of chip designer Nvidia, a huge beneficiary of the AI hype, were down 11.8% at 07:05 a.m. ET ahead of the market open. Netherlands-based chip companies ASML and ASM International tumbled 8.9% and 13.6% respectively in European trade, while in Asia, Japanese chip-related stocks were broadly lower.
DeepSeek launched a free, open-source large-language model in late December, claiming it was developed in just two months at a cost of under $6 million — a much smaller expense than the one called for by Western counterparts. Last week, the company released a reasoning model that also reportedly outperformed OpenAI’s latest in many third-party tests.
The developments have stoked questions about the amount of money big tech companies have been investing in AI models and data centers.
“DeepSeek clearly doesn’t have access to as much compute as U.S. hyperscalers and somehow managed to develop a model that appears highly competitive,” Srini Pajjuri, semiconductor analyst at Raymond James, said in a note Monday.
On one positive implication, Pajjuri said DeepSeek could “drive even more urgency among U.S. hyperscalers” — which are large computing infrastructure players like Amazon and Microsoft — to leverage their advantage in having access to graphics processing units (GPUs) to set themselves apart from cheaper options. GPUs are a key part of the infrastructure required to train huge AI models. Nvidia is the market leader in GPUs.
Analysts at Citi said DeepSeek’s large-language model had “prompted investor inquiries
around cost of compute.”
They note that, “while the dominance of the US companies on the most advanced AI models could be potentially challenged,” the access these tech firms have to advanced chips poses an advantage. “Thus, we don’t expect leading AI companies would move away from more advanced GPUs.”
The recent announcement of President Donald Trump’s $500 billion Stargate AI project is a “nod to the need for advanced chips,” they added.
Analysts at Bernstein meanwhile expressed doubt over whether the DeepSeek tool was actually built for under $6 million, arguing the figure “does not include all the other costs associated with prior research and experiments on architectures, algorithms, or data.”
They stressed that DeepSeek’s “models look fantastic but we don’t think they are miracles,” and said that panic about the “death-knell of the AI infrastructure complex as we know it” was “overblown.”
— CNBC’s Lee Ying Shan and Michael Bloom contributed to this story.
This breaking news story is being updated.
Read the original article here