China’s AI Startups Narrow Model Gap as U.S. Funding Lead Widens

China’s AI startups are narrowing the performance gap with U.S. frontier models while facing a much larger shortfall in the capital and infrastructure needed to build and operate them at scale.

U.S.-based AI startups have raised about $380 billion in AI-related venture capital since 2023, compared with roughly $39 billion for Chinese startups. The infrastructure gap is also wide: U.S. technology companies spent more than $400 billion on capital expenditures in 2025, versus $63 billion in China, and U.S. technology spending is projected to exceed $800 billion in 2026.

U.S. AI-optimized computing capacity was estimated at eight times China’s based on H100-equivalent GPU capacity. Training and serving large models requires sustained investment in chips, data centers, power and engineering talent, making access to capital a central constraint in the competition.

Chinese companies have continued to produce increasingly capable models. Moonshot AI’s Kimi K3 uses a mixture-of-experts architecture with 2.8 trillion parameters and a one-million-token context window. Its performance remained below the strongest proprietary systems while exceeding other evaluated open and proprietary models in testing.

Hong Kong’s public markets are becoming an important financing route for Chinese AI companies. Zhipu AI raised $558 million in a Hong Kong listing on Jan. 8, while MiniMax Group raised $711 million on Jan. 9. AI companies listed in Hong Kong raised $4.9 billion during December 2025 and January 2026, and about 20 AI companies were in the listing pipeline as of Feb. 3, 2026.

The pipeline could make Hong Kong an earlier funding channel for companies that need computing capacity and research capital before reaching the scale or profitability expected of more mature public businesses. Public listings provide access to equity and international investors, but they also expose younger companies to market pressure while spending remains high.

“New AI companies are going public in Hong Kong, [and] investors can now – for the first time – look beyond AI proxy stocks to direct investment opportunities in frontier companies from across the China AI value chain,” said Johnson Chui, HKEX’s head of global issuer services.

Those public-market proceeds remain small beside the largest U.S. private financings. OpenAI announced $110 billion in new investment on Feb. 27 at a $730 billion pre-money valuation, including commitments of $30 billion from SoftBank, $30 billion from Nvidia and $50 billion from Amazon.

OpenAI co-founder and CEO Sam Altman said, “We’re pushing the frontier across infrastructure, research, and products to make AI more capable, reliable, and broadly useful.”

The result is an uneven contest. Chinese companies are advancing in model capability and access to public financing, while U.S. companies retain a substantial advantage in the funding and computing capacity required to build AI systems at the largest scale.