
Chinaās Deepseek casts doubt on AI-driven semicon growth
The launch of an open-source Chinese AI chatbot is sending shockwaves throughout the tech industry. Said to rival OpenAIās o1 model, Deepseek was developed with a fraction of the resources that leading Western outfits employ, casting doubt on the premise that progress in AI requires massive amounts of cutting-edge chips. As a result, semiconductor stocks, including Dutch equipment makers ASML and ASM International, took a tumble on Monday.
Deepseek, a venture set up only last year by hedge fund billionaire Liang Wenfeng, claims it used a little over 2,000 Nvidia H800 GPUs to train its language model. The H800 is based on Nvidiaās H100, modified to comply with US export restrictions. Originally launched in 2022, itās no slouch, but being two generations behind Nvidiaās soon-to-be-launched Blackwell generation, neither can it be considered top-of-the-line. Other than on hardware, Deepseek is said to have spent less than 6 million dollars on training, compared to routinely 100+ million dollars by Western firms.
Such numbers call into question the US approach to AI, which has taken on the mechanics of an arms race: not to be outdone by the competition, Big Tech firms invest ever-larger sums of money, mostly on chips and data centers. Just last week, OpenAI, Softbank and Oracle announced a 500-billion-dollar investment in Stargate, the ālargest AI infrastructure project in history,ā according to the projectās top patron, President Trump. In another example, Microsoft said three weeks ago that it would spend over 80 billion dollars on AI this year, roughly double the 2024 amount.
With Deepseekās more cost-efficient approach, the lavish capex of what was until recently thought of as the worldās AI leaders may be up for re-evaluation. Deferred and/or reduced investment would be painful for the semiconductor industry, since AI is the only end-market keeping the semicon market going at the moment.
Some industry observers, however, point to a silver lining: competition will boost the development of AI and lowered cost will speed up its adoption, generating more reliable growth drivers for all AI-linked industries.