Chinese AI Startups Generate 10% of OpenAI Revenue Despite Sky-High Valuations

China's largest artificial intelligence companies together generate only about one-tenth of the revenue that OpenAI reports alone, according to new research that highlights a widening commercial gap between Chinese and American AI labs.

Rhodium Group published estimates Thursday showing OpenAI's annual recurring revenue at $40 billion and Anthropic's at $65 billion. By contrast, ByteDance leads Chinese AI firms with $4 billion in ARR, followed by Alibaba at $2.4 billion. Z.AI reported $1.8 billion in ARR to investors Wednesday, according to a transcript reviewed by CNBC. Moonshot recorded $1 billion, MiniMax $800 million, and DeepSeek $500 million.

Why it matters

The revenue disparity becomes more striking when measured against investor valuations. Chinese AI startups are commanding valuation multiples that far exceed their American counterparts despite generating a fraction of the revenue — a pattern that raises sustainability questions as these companies prepare for public listings. The structural financing gap could limit China's ability to compete in the capital-intensive race to scale frontier AI models.

Valuation disconnect

Rhodium Group calculated that Moonshot carries a valuation-to-revenue ratio of 50x, while DeepSeek's reaches 163x. Those figures dwarf OpenAI's 34x multiple and Anthropic's 21x ratio. "Valuations relative to revenue appear exorbitant for Moonshot and DeepSeek at present," the report stated.

Logan Wright, a Rhodium Group partner who co-authored the analysis with research analyst Endeavour Tian, told CNBC the funding shortfall creates fundamental challenges. "The financing gap means it will be far more difficult for Chinese frontier AI labs to scale sustainably," Wright said. "They will be heavily dependent upon a favorable climate in the equity market — historically that's not an easy bet in China."

Monetization challenges

Chinese AI laboratories face additional revenue headwinds from their open-source model strategies. Rhodium observed that the open-source nature of these systems allows third parties with adequate hardware to operate the models without routing usage through the original developer — eliminating a key revenue stream.

American AI models remain predominantly closed-source, and top U.S. models command higher per-task pricing than Chinese alternatives, according to AI-comparison firm Artificial Analysis.

IPO pipeline builds

Several companies covered in the Rhodium analysis are advancing toward public markets. Moonshot has filed confidentially for a Hong Kong IPO targeting $3 billion, Reuters reported. DeepSeek has engaged CITIC Securities to prepare for a listing on Shanghai's STAR Market, also according to Reuters.

OpenAI has pushed its IPO timeline to next year, while Anthropic is reportedly preparing to go public in the U.S. next month, per CNBC.

Rhodium acknowledged one significant limitation: the figures reflect data from this summer, when Chinese AI adoption was still climbing from relatively low levels earlier in the year. Z.AI told investors Wednesday it has raised its year-end ARR target to $3 billion from a previous projection of $2.4 billion.

The analysis was first reported by Quartz and published by Rhodium Group on Thursday.