Just one week into Donald Trump’s presidency, a little-known Chinese artificial intelligence company sent shockwaves through the global tech industry.
DeepSeek, a Beijing-based firm, launched its chatbot DeepSeek-R1 in January, quickly climbing to the top of Apple’s free app chart in the United States. The company claimed its product rivalled ChatGPT — and had been built for just $5.6 million, a fraction of the billions spent by American firms.
The reaction was immediate. Nvidia, the US chipmaking giant at the heart of the AI boom, saw $600 billion wiped from its market value in a single day — the largest one-day loss in US stock market history. Other AI-linked stocks tumbled in the fallout.
For years, Washington had considered China to be trailing in large language models. DeepSeek’s arrival challenged that view, with some analysts dubbing it an “AI Sputnik moment” — a technological wake-up call reminiscent of the Cold War space race.
Yet, six months later, DeepSeek is no longer dominating headlines. While it retains a loyal user base, concerns over its Chinese origins have led many US companies to block employee access, citing fears of data being routed to servers in the People’s Republic of China. Some users have turned to running the app locally to avoid sending information abroad.
The episode has nonetheless shifted industry thinking. US AI development had been focused on bigger, more resource-intensive models. DeepSeek showed that smaller, more efficient systems could match or even surpass performance benchmarks, forcing American firms to rethink their approach.
OpenAI recently released its first freely downloadable AI models in five years — a move some experts see as a direct response to DeepSeek’s success. However, the major US players have not abandoned their scale-first strategy. Days later, OpenAI unveiled GPT-5 alongside expanded data centre investments, echoing similar moves by Meta and other tech giants. Nvidia’s stock, which initially plunged, has since rebounded to record highs.
The Biden administration has scrutinised DeepSeek’s ties to Beijing, with officials warning the company may support Chinese military and intelligence operations. DeepSeek has denied wrongdoing but confirmed in its privacy policy that user data may be stored on servers in China.
Meanwhile, DeepSeek’s momentum faces new headwinds. Its planned follow-up, DeepSeek-R2, has reportedly been delayed due to a shortage of advanced semiconductors — the same hardware at the heart of the US-China technology rivalry.
“DeepSeek revealed the competitiveness of China’s AI industry,” said Marina Zhang of the University of Technology Sydney. “But maintaining that edge will be far harder in the face of intense competition and supply constraints.”
For now, the company remains a symbol of how quickly global tech power balances can shift — and how quickly they can shift back.
