The covert use of Nvidia’s high-performance H100 chips by Chinese startup DeepSeek AI has stirred up controversy in the tech industry, as the company’s operations appear to bypass US export restrictions. Despite limitations on the availability of these chips in China, DeepSeek is reportedly using 50,000 Nvidia H100 GPUs, a revelation that has sent shockwaves across the market.
The controversy came to light after Scale AI CEO Alexandr Wang disclosed in a CNBC interview that DeepSeek workers are prohibited from discussing the use of these chips due to US regulations. Initially, it was believed the startup was utilizing Nvidia’s H800 chips, which are less powerful, but Wang confirmed that DeepSeek has access to the restricted H100 chips. “DeepSeek has about 50,000 H100s, which they can’t talk about due to export controls,” Wang stated. Elon Musk seemed to back this claim, responding with “obviously” to a post referencing the interview.
Also Read: Nvidia Unveils GeForce RTX 5090 as Flagship GPU at CES 2025
The news caused Nvidia’s shares to plummet by 16.9%, with other tech stocks following suit. The resulting market fallout dragged the Nasdaq composite down by 3.1%, marking its steepest decline in over a month.
While the use of banned H100 chips might slightly undermine DeepSeek’s reputation, its model’s performance remains highly competitive. Chinese companies like DeepSeek and Qwen are emerging as serious challengers to U.S. AI dominance, offering cost-effective models that rival those from American firms.
This raises broader concerns about U.S. regulation in the AI sector, as Chinese startups continue to push the boundaries of innovation, potentially shifting the balance of global AI leadership.
