A chatbot made by Chinese artificial intelligence startup Deepseek has shot to the top of the Apple App Store charts in the United States this week, dethroning OpenAi's chatgpt as the most downloaded free app. The eponymous ai assistant is powered by Deepseek's open source models, which the the company says It can be trained at a fraction of the cost using far fewer chips than the world's leading models. The claim has riled financial markets, with Nvidia's share price falling more than 12 percent in pre-market trading.
Deepseek also claims to have needed only about 2,000 specialized Nvidia chips to train V3, compared to the 16,000 or more required to train core models, according to <a target="_blank" href="https://www.nytimes.com/2025/01/23/technology/deepseek-china-ai-chips.html”>he New York Times. These unverified claims are leading developers and investors to question the compute-intensive approach favored by the world's global ai companies. And if true, it means Deepseek engineers had to get creative in the face of trade restrictions meant to ensure ai dominance.
Nvidia, Microsoft, OpenAi, and Meta are investing billions in ai data centers: $500 billion for the Stargate project alone, of which $100 billion is believed to be earmarked for NVIDIA. Investors and analysts are now wondering if that's money well spent, with Nvidia, Microsoft and other companies with substantial bets on maintaining the ai status quo all trending lower in pre-market trading.
