NVIDIA's CEO Jensen Huang warns China is set to become superior to US in AI

NVIDIA CEO Jensen Huang warned of the U.S.’s AI dominance as Chinese AI chatbot DeepSeek is looking forward to optimising its models to run on Huawei chips instead of American hardware.

In an appearance on a podcast, Huang said that DeepSeek is planning to strategically shift to Huawei’s CANN platform instead of Nvidia’s CUDA.

“If future AI models are optimised in a very different way than the American tech stack, China will become superior to the U.S.,” he added.

DeepSeek is set to roll out its V4 multimodal foundation model later this month on Huawei’s Ascend 950PR processor.

The company has spent months improving its previous core code to integrate Huawei’s framework. The move is aimed at breaking the dependence on Nvidia that has locked most global AI development into the U.S. ecosystem.

Although Huawei’s processors are only around 60% as powerful as Nvidia’s H100, Huang pointed out that China has an “abundant supply of energy” and “a large number of AI experts” to offset any hardware deficiencies.

This creates an irony within the export policies of the United States. Policies aimed at curbing China’s AI capabilities have ended up boosting its domestic development capabilities instead, while U.S. representatives seek to add DeepSeek to the entity list.