Opinion | DeepSeek success shouldn’t be seen as a win for China or loss for US

On January 1, people who don’t closely follow innovations in artificial intelligence (AI) would not have heard of DeepSeek. However, by late January, it became a widely celebrated sensation. US tech corporations such as Meta and OpenAI were shocked by the apparent breakthroughs accomplished by this hitherto obscure Chinese start-up.

In a comprehensive and publicly available post, the Chinese “David” revealed that it had managed to develop a R1 model with performance comparable to OpenAI’s o1, at a fraction of the cost that American “Goliaths” have poured into AI over the years. Yet DeepSeek’s triumphs should not be construed as a victory for China or a loss for the US. Instead, the breakthrough vindicates the importance of Sino-American cooperation over AI development and governance.

The China-located venture succeeded by fusing the best of Silicon Valley-style entrepreneurship, the grit and perseverance of home-grown Chinese talent and a stroke of genius rooted in its intellectual openness. As with many other top Chinese AI and advanced technological firms, the start-up has a uniquely horizontal and non-hierarchical corporate culture.

Its staff are driven by a common devotion to purpose but are granted considerable flexibility in the workplace. Indeed, founder Liang Wenfeng – an engineering-trained venture capitalist-turned-AI whizz – once declared, “experience is not that important”. DeepSeek had acquired 10,000 Nvidia A100 graphics processing units (GPUs) before export controls imposed by Washington.

In embracing the open-source ethos, however, the start-up eschewed an iron grip over its innovation and data, as embraced by many of its American peers.