The combined annual recurring revenue (ARR) of the seven largest Chinese AI developers reached $10.7 billion. However, this is only about 10% of the figures reported by OpenAI and Anthropic, according to a report by Rhodium Group.
Some Chinese companies’ valuations are already comparable to or exceed those of American firms in terms of revenue, noted the specialists who studied China’s AI industry financing. They analyzed data from Alibaba, ByteDance, Kuaishou, DeepSeek, Moonshot AI, Z.ai, and MiniMax, comparing their performance with U.S. competitors.
Revenue Lags Behind Valuations
According to Rhodium Group, the combined ARR of Chinese AI models was about $10.7 billion from March to August 2026. In comparison, OpenAI and Anthropic reported a combined revenue exceeding $100 billion.
ByteDance led among Chinese companies, with its AI division reaching approximately $4 billion ARR in July, followed by Alibaba with $2.4 billion in August. The figures for other developers were significantly lower.
Meanwhile, the revenue of Chinese companies is growing rapidly. For instance, Z.ai’s ARR increased from $74 million in January to $1.6 billion in August—about a 20-fold rise. MiniMax’s figure grew approximately fourfold over the same period.
However, revenue growth does not yet equate to comparable profitability.
China Invests Hundreds of Billions in AI
Rhodium Group estimates that Chinese companies’ investments in AI infrastructure in 2026 amounted to 932 billion yuan ($139 billion)—more than double the previous year’s level.
In 2027, this figure could exceed 1.2 trillion yuan ($193 billion). Meanwhile, U.S. investments in data centers in 2026 are estimated at around $800 billion.
Thus, China is expanding its infrastructure much faster than the revenue growth of AI models. The combined free cash flow of the three largest Chinese hyperscalers—Alibaba, Tencent, and Baidu—turned negative in the first half of 2026, reaching minus 16 billion yuan. A year earlier, it was 170 billion yuan.
Chinese AI Companies Depend on Investors
The financial model of the Chinese AI sector differs from the American one. In the U.S., major tech companies increasingly rely on debt financing: the net inflow from debt issuance for the five largest American hyperscalers rose from $90 billion in 2025 to $163 billion in just the first half of 2026.
Chinese companies rely more on equity financing and bank loans.
In 2026, investments in the capital of China’s AI industry reached 282 billion yuan by August 21. For leading AI labs—Z.ai, MiniMax, DeepSeek, and Moonshot—the growth of private and venture financing is particularly notable. In 2025, their combined investments from PE/VC were about 9 billion yuan, while in the first eight months of 2026, funding from IPO and private placements totaled 179 billion yuan.
The state continues to play a significant role. About 25% of direct investments in China’s AI sector in 2026 came from state funds and related entities. In the chips and servers segment, the share of state and state-affiliated investors reaches 47%, with another 14% coming from banks, predominantly state-owned. Thus, more than 60% of investments in this segment have state or quasi-state origins.
High Valuations with Low Revenue
Another finding of the study concerns the valuation of Chinese AI companies. Rhodium Group compared their market capitalization and ARR, discovering that some Chinese developers trade at higher multiples than OpenAI and Anthropic.
For OpenAI and Anthropic, the authors estimate the company value to ARR ratio at approximately 34x and 21x, respectively. Z.ai’s ratio is about 46x, Moonshot’s is around 50x, and DeepSeek’s reaches 163x.
Moonshot is particularly notable: the company’s valuation rose from about $4 billion at the end of 2025 to $50 billion in August 2026.
However, the profitability of Chinese developers remains lower than that of their American counterparts. For example, DeepSeek’s gross margin at the company level was estimated at 45% in July, compared to about 65% for Anthropic. Z.ai and MiniMax had margins of about 26% and 18%, respectively, in the first half of 2026.
One reason cited by Rhodium Group is the lower prices of Chinese AI services. Even after recent price hikes, most Chinese frontier models cost about $0.04–0.50 per task, while top-tier Claude LLMs cost around $2–4, and the most expensive GPT models are $1–2.
AI Race Hinges on Funding
The study’s authors conclude that the Chinese AI sector faces the same fundamental issues as the American one: companies are simultaneously increasing capital expenditures, expanding computing power, and not yet generating sufficient cash flow.
However, the Chinese financing model has its own characteristics. To continue investing, companies will need to maintain access to equity markets and bank lending. If the valuation of Chinese AI companies continues to grow faster than their revenue, any deterioration in the stock market could limit their ability to attract new capital.
Rhodium Group expects further growth in AI investments in China in 2026–2027 but sees financing as one of the key constraints for scaling the industry, along with the availability of modern chips.
According to Morgan Stanley, by 2030, monetizable revenue from consumer AI in China could reach 294 billion yuan (~$44 billion).
