one year on
Nvidia loses close to $600 billion in market cap as Chinese AI lab DeepSeek sparks sell-off
A single-day stock drop larger than any in U.S. history wipes out nearly $600 billion from the chipmaker as an open-source model built with cheaper hardware raises doubts about the spending assumptions driving the AI boom.
Nvidia lost nearly $600 billion in market capitalization Monday, the largest single-day loss for any company in U.S. history, as the emergence of Chinese AI lab DeepSeek rattled the technology sector. Shares of the chipmaker plunged 17% to close at $118.58, dragging the Nasdaq down 3.1%.
The sell-off began after DeepSeek’s free, open-source large language model — built in two months for under $6 million using less-capable Nvidia H800 chips — hit the top of Apple’s App Store, dethroning ChatGPT. The feat raised questions about whether the massive capital expenditures by U.S. tech giants on Nvidia’s most advanced GPUs are truly necessary. Broadcom fell 17%, while data-center hardware makers Dell, Hewlett Packard Enterprise and Super Micro Computer dropped at least 5.8%, and Oracle fell 14%.
Analysts at Cantor wrote that DeepSeek’s latest technology has caused ‘great angst’ about compute demand, but they argued the sell-off is overdone because AI advances will likely increase demand for compute, not reduce it. Venture capitalist David Sacks wrote on X that the AI race will be very competitive and that the U.S. cannot be ‘complacent.’ The model arrives despite a slew of recent U.S. curbs on chip exports to China.
The record
Wrote that DeepSeek has caused 'great angst' about compute demand, but argued the sell-off is overdone because AI advances will increase demand for compute, not decrease.
view the original post →Posted on X that DeepSeek's model shows the AI race will be very competitive and that the U.S. cannot be 'complacent'.
view the original post →One year later — open only if you can handle spoilers
The DeepSeek scare proved temporary for Nvidia's stock, which recovered most of its losses within three months as enterprise spending on GPU clusters continued to grow. However, the episode cemented DeepSeek as a legitimate competitor and forced U.S. labs to publicly justify their cost structures, while export control debates in Washington grew more heated.