US panel credits China’s AI edge to open-source models, manufacturing dominance

China’s strategy of promoting open artificial intelligence (AI) models and leveraging manufacturing dominance is “mutually reinforcing”, forming a feedback loop that could challenge US dominance in AI, according to a new report by a US congressional advisory body.

“It is the intersection of these two loops – one digital, one physical – that gives China’s open strategy its compounding force and poses the most serious long-term challenge to US AI leadership,” the United States-China Economic and Security Review Commission wrote in its report.

The two countries have diverged in their approaches to AI to date: the US has focused on leading technological breakthroughs, while China has prioritised rapid, widespread adoption. The report highlighted that, despite limited access to advanced AI chips due to US export controls, China has embraced low-cost open-source models, optimising them for mass deployment and scaling.

“This open ecosystem enables China to innovate close to the frontier despite significant compute constraints. Chinese labs have narrowed performance gaps with top Western large language models,” the report said.

Meanwhile, US firms such as OpenAI, which developed ChatGPT, and tech giants like Google remain at the forefront of model development, giving them a lead in cutting-edge capabilities.

05:04

China creates analogue AI chip said to be 1,000 times faster than Nvidia GPU

How much Chinese firms are spending on AI in total, compared to the hundreds of billions of US dollars their American competitors are spending, remains difficult to determine, with investment often opaque and some companies likely benefiting from state subsidies.

South China Morning Post