Chinese artificial intelligence firms are drawing users through affordable pricing and open-access frameworks, yet their earnings trail significantly behind those of their U.S. counterparts, including OpenAI and Anthropic. Together, the two American companies generate approximately $105 billion in annual recurring revenue (ARR), while China’s leading AI developers collectively earn about $10 billion, according to the research firm Rhodium Group.
This disparity is most pronounced when comparing individual companies: OpenAI’s ARR stands at roughly $40 billion, and Anthropic’s reaches $65 billion. Among Chinese firms, ByteDance leads with around $4 billion, followed by Alibaba at $2.4 billion. Z.ai, previously known as Zhipu AI, reported $1.8 billion in ARR earlier this month, with Moonshot AI and MiniMax trailing at $1 billion and $800 million, respectively. DeepSeek, the fastest-growing Chinese AI lab, records $500 million in ARR.
These revenue estimates are based on mid-year data rather than audited financial statements. The gap extends to company valuations, where Moonshot AI is valued at about 50 times its ARR, while DeepSeek’s valuation hits 163 times its revenue—markedly higher multiples than OpenAI’s 34 times or Anthropic’s 21 times. This contrast shows how Chinese AI firms prioritize accessibility over profit, frequently releasing open-weight models that developers can implement without subscription costs.
In contrast, OpenAI and Anthropic operate closed systems where users pay for access through application programming interfaces (APIs) or subscriptions. This approach has proven far more profitable, even as Chinese firms like DeepSeek gain recognition for developing competitive models using fewer computational resources. DeepSeek, established in 2023, has already secured $74 billion in its latest funding round and plans to launch an initial public offering on Shanghai’s STAR Market. Moonshot, the creator of the Kimi chatbot, has also filed confidentially for a Hong Kong IPO.
The move toward public markets aligns with broader developments in China’s AI sector. State-backed investments have fueled growth in domestic computing infrastructure, with over 60% of equity in AI chips and servers controlled by state-affiliated entities, according to Rhodium Group. Meanwhile, U.S. export controls restrict China’s access to Nvidia’s most advanced AI processors, compelling local developers to adopt domestically produced alternatives.
Despite the revenue shortfall, Chinese AI models are gaining domestic appeal. Z.ai’s shares rose over 5% in Hong Kong trading following its latest funding announcement, though volatility persists. MiniMax, which listed in January, has also experienced share fluctuations since its debut. Both companies are part of a broader trend among Chinese AI firms seeking capital as demand for models and computing power grows.