Chinese AI Companies Struggling to Finance Competition with US
Chinese AI companies are finding it increasingly difficult to attract funding for model development and accelerator procurement. In 2025, private investments in the sector in China amounted to $12.4 billion, compared to $280 billion in the US.

According to Rhodium Group estimates, Chinese tech giants will double their capital expenditures this year, reaching $139 billion. Revenue from AI startups associated with them covers less than 8% of this amount, and by 2027, expenses could exceed $193 billion.
The gap is also evident in revenues: the six largest Chinese AI companies collectively earn about one-tenth of the revenue of OpenAI and Anthropic, despite American developers spending roughly 5–6 times more. Open models and price competition accelerate the spread of Chinese technologies but reduce companies' ability to profit from them.
Restrictions on Nvidia supplies are increasing costs. In China, a Nvidia DGX B300 server was priced at nearly $3 million—about eight times higher than in the US. Nvidia announced the cessation of H200 accelerator shipments to China. According to DeepSeek CEO Liang Wenfeng, achieving performance comparable to Nvidia’s GB300 would require four times as many Huawei 950 accelerators. Analysts believe that without large-scale and sustained state funding, the industry will struggle to balance expenses and revenues.



