China is rapidly expanding its artificial intelligence infrastructure as competition with the United States intensifies, with Beijing investing heavily in data centres, domestic AI chips, research talent and open-source models.
The expansion is particularly visible in regions such as Inner Mongolia, Ningxia, Gansu and Guizhou, where large data-centre facilities are being constructed or expanded to support the country’s growing AI industry.
According to Arise News, hundreds of data centres are being developed or expanded in these regions, involving technology companies including Huawei, ByteDance and DeepSeek. The areas offer relatively cheap land, cooler conditions and access to growing renewable-energy resources, factors that can help reduce the cost of operating energy-intensive data centres.
China has also been increasing efforts to develop its own AI computing infrastructure as US restrictions limit Chinese access to some advanced American chips and semiconductor manufacturing technology.
Reuters reported on September 17 that Huawei currently cannot produce enough AI computing equipment to meet demand in China. The company said it is therefore limiting its international expansion while prioritising the domestic market.
Huawei is accelerating development of its Ascend AI processor family, with the company saying its Ascend 960DT chip is expected in the first quarter of 2027. Huawei also unveiled a computing architecture designed eventually to connect as many as one million processors in a single system.
The company said more than 1,000 systems based on earlier Ascend processors have already been deployed, while more than 5,200 developers are active monthly on software for its Ascend platform. Reuters noted, however, that Nvidia continues to hold a major advantage through its widely used CUDA software ecosystem.
China’s strategy extends beyond hardware. Beijing is also promoting open-weight and more affordable AI models, which Chinese officials and technology companies see as a way of accelerating adoption and reducing dependence on foreign technology.
Recent analysis by the Center for Strategic and International Studies said China’s rapid progress has been driven partly by the indigenisation of technology and the country’s growing emphasis on open-weight AI models.
However, the developments do not mean China has overtaken the United States across the entire AI sector.
A September analysis by the Rhodium Group estimated that China’s AI capital expenditure remains at roughly 15–20% of US levels. The group expects China’s AI capital expenditure to rise significantly, but the gap in investment remains substantial.
Similarly, Boston Consulting Group’s 2026 assessment found that the United States continues to hold advantages in areas including AI talent, capital deployment and frontier AI models, while China has been rapidly closing the gap in computing capacity and expanding AI adoption across its economy.
The growing competition is also extending beyond the two countries. Southeast Asian and other emerging markets are increasingly becoming arenas where American and Chinese companies compete for AI partnerships, infrastructure projects and technological influence.
The developments show that the US-China AI competition is increasingly about more than individual AI models. Data centres, semiconductor supply chains, electricity, computing capacity, software ecosystems and research talent are becoming critical components of the broader technology race.
As both countries continue to invest in AI infrastructure and seek greater technological independence, the competition is likely to shape the development and availability of artificial intelligence technology globally.
