Chinese startup DeepSeek's launch of its latest AI models, which it says are on a par or better than industry-leading models in the United States at a fraction of the cost, is threatening to upset the technology world order.

The company has attracted attention in global AI circles after writing in a paper last month that the training of DeepSeek-V3 required less than $6 million worth of computing power from Nvidia H800 chips.

U.S. President Donald Trump has described it as a "wake up call" for U.S. industries to be "laser focused on competing to win".

DeepSeek's success has already been noticed in China's top political circles. On January 20, the day DeepSeek-R1 was released to the public, founder Liang attended a closed-door symposium for businessman and experts hosted by Chinese premier Li Qiang, according to state news agency Xinhua.

DeepSeek: Chinese AI firm making waves in Silicon Valley

So, will DeepSeek's latest AI model spark a race with China and the U.S.? And will they prompt a move away from huge investment in high-powered semiconducter chips?

Kayla Blomquist, an AI technology and geopolitics expert on US-China relations at the Oxford Internet Insitute at Oxford University, said it's a good motivator for the rapidly developing industry, particularly in the U.S., but issued a warning about the "hype cycle".

"Some people say it's a ‘Sputnik moment’.. but we want to be careful of not taking instances as this, the hype cycle, and using that to further fuel extremely rapid development without necessary safeguards in place," Blomquist said.

DeepSeek's AI Assistant, powered by DeepSeek-V3, has overtaken rival ChatGPT to become the top-rated free application available on Apple's App Store in the United States.

Silicon Valley rattled by low-cost Chinese AI

This has raised doubts about the reasoning behind some U.S. tech companies' decision to pledge billions of dollars in AI investment and shares of several big tech players, including Nvidia, have been hit.

The DeepSeek-R1, released last week, is 20 to 50 times cheaper to use than OpenAI o1 model, depending on the task, according to a post on DeepSeek's official WeChat account.

But some have publicly expressed skepticism about DeepSeek's success story.

"If its true, it may indicate a potential paradigm shift to a less compute-intensive era of AI development," Blomquist said. "I think there’s a lot of temptation to react quite strongly to that, and at the end of the day, compute is just fundamentally going to matter even if companies find more efficient ways to use it."

(Production: Lucy Marks, Matt Stock)