- DeepSeek’s new AI model has triggered a sell-off in top AI stocks.
- Former Intel CEO Pat Gelsinger thinks the markets have got it wrong.
- Gelsinger said DeepSeek’s claims of creating AI cheaply will drive demand for chips, not reduce it.
Former Intel CEO Pat Gelsigner has said that “the markets are getting it wrong” after investors drove a panic sell-off in top AI stocks in response to concerns that DeepSeek would weaken demand for advanced chips.
Top AI companies, including Nvidia, saw hundreds of billions of dollars wiped off their valuation on Monday after a new AI model from Chinese startup DeepSeek claimed to have emulated the success of a leading model released by OpenAI just months ago — and at a fraction of the cost.
DeepSeek’s new model R1, released on Donald Trump’s inauguration day, appeared to have accomplished the feat with fewer and less powerful chips than used by top AI labs in America. The development has raised concerns that chip stocks like Nvidia may see demand recede.
Nvidia, which has added trillions of dollars to its market capitalization since the start of the ChatGPT boom, suffered a wipeout of as much as $500 billion on Monday, triggering the biggest stock market rout in US history.
Gelsinger took to X on Monday to suggest that the market’s assumptions were wrong. He said that instead of reducing demand, making computing “dramatically cheaper” and more efficient to use — as DeepSeek appears to have done — “will expand the market for it.”
The former Intel boss, who retired from the company in December after struggling to capitalize on the AI boom, also suggested that the Chinese engineers at DeepSeek “had limited resources, and they had to find creative solutions” to squeeze performance out of their models.
The AI industry has insisted that models become smarter when powered with more computing power in the form of chips loaded in data centers, where they are trained and hosted.
Due to tough export controls, Chinese companies have struggled to access the best chips from America. That said, DeepSeek has not explicitly said how much computing power is behind the R1 model it released last week.
Gelsinger’s views echo those of others in the AI industry, who feel the market reaction to DeepSeek’s claims of greater computing power efficiency has been overblown.
Wharton professor Ethan Mollick wrote on X that he’s “not sure why people assume this will make compute less valuable,” adding that “more efficient models mean that those with compute will still be able to use it to serve more customers and products at lower prices & power impact.”
