Rhodium: Chinese AI Models Lag US Rivals in Revenue

BEIJING — China’s burgeoning artificial intelligence sector is experiencing a surge in model adoption, but this enthusiasm is not yet translating into significant revenue streams. This disconnect is casting a shadow over company valuations and prompting a closer examination of the financial viability of these rapidly developing AI entities.

According to a report released Thursday by the U.S.-based research firm Rhodium Group, the combined revenue generated by all of China’s AI models currently amounts to a mere 10% of what leading U.S. companies like OpenAI and Anthropic are reporting. This estimate is derived from industry-standard annual recurring revenue (ARR) figures, a metric designed to capture the rapid growth trajectory by projecting a recent monthly revenue figure over a 12-month period.

The Rhodium report highlighted that DeepSeek’s ARR stood at $500 million, placing it at the lower end among major Chinese AI companies. MiniMax followed with an ARR of $800 million, while Moonshot reported $1 billion.

In a recent investor update on Wednesday, Z.ai disclosed its latest ARR at $1.8 billion, as indicated by a transcript reviewed by CNBC.

Even when factoring in contributions from giants like ByteDance, with an estimated $4 billion, and Alibaba, at $2.4 billion, these figures pale in comparison to the $40 billion in ARR generated by OpenAI alone, let alone Anthropic’s $65 billion, according to Rhodium’s findings.

Crucially, this stark revenue gap is failing to keep pace with the stratospheric valuations that investors are currently assigning to these Chinese AI startups.

“Valuations relative to revenue appear exorbitant for Moonshot and DeepSeek at present,” the Rhodium report stated, with analysts noting estimated valuation multiples of 50x and 163x respectively for these two emerging players. These figures significantly outstrip the multiples of 34x for OpenAI and 21x for Anthropic.

It is important to acknowledge that the Rhodium analysis may be based on the most recent data available from earlier in the summer, and the adoption of Chinese AI models has seen exponential growth from nascent levels earlier this year.

Z.ai on Wednesday revised its end-of-year ARR forecast upwards to $3 billion, an increase from its previous projection of $2.4 billion. This upward revision signals growing confidence in the monetization potential of their AI offerings.

Rhodium also pointed to the strategic initiatives by Chinese AI labs to capture a larger share of revenue from third-party providers who integrate their models. The open-source nature of many Chinese AI models, while fostering rapid adoption, also allows for independent deployment by entities with sufficient computational resources, presenting a complex revenue-sharing landscape.

In contrast, leading U.S. models, such as those from OpenAI and Anthropic, are largely proprietary. This closed ecosystem often translates to a higher cost per task for end-users, a factor that may contribute to the revenue disparity. According to AI-comparison firm Artificial Analysis, the cost per task for these U.S. models significantly exceeds that of their Chinese counterparts.

“The financing gap means it will be far more difficult for Chinese frontier AI labs to scale sustainably,” Logan Wright, partner at Rhodium Group and co-author of the report with research analyst Endeavour Tian, stated. He elaborated that these labs will be “heavily dependent upon a favorable climate in the equity market—historically that’s not an easy bet in China.” Wright added that while government funding has been instrumental in the hardware build-out for compute capacity, it is unlikely to extend to direct funding for frontier AI labs, particularly given the current economic climate.

Rhodium’s analysis indicates that over 60% of equity investment in Chinese AI chips and servers originates from state-affiliated entities, underscoring the government’s strategic role in nurturing the foundational infrastructure of the AI ecosystem.

The year has been marked by considerable volatility for publicly listed Chinese AI companies.

Z.ai shares saw an uptick of more than 5% in Thursday morning trading, rebounding from a decline earlier in the week following the announcement of its second significant fundraising round in two months. The stock, traded in Hong Kong, had experienced a sharp decline, reverting to levels seen earlier in the spring after a brief but dramatic surge that more than tripled its price over the summer.

Similarly, shares of rival Minimax have struggled in recent months to maintain their gains from their initial public offering day. The stock experienced a significant spike in the spring, but has since faced challenges in holding onto those early gains.