Chinese AI Model Providers Capture Roughly 10% of Combined OpenAI and Anthropic ARR

Rhodium Group’s most recent market study estimates that the combined annual recurring revenue (ARR) of the principal Chinese artificial‑intelligence model providers totals roughly $10.7 billion. This sum corresponds to about one‑tenth of the ARR that Rhodium calculates for OpenAI, which it places at $40 billion, and Anthropic, which it values at $65 billion.

Breakdown of Chinese AI Revenue

The Chinese cohort comprises six leading firms. Rhodium assigns $500 million of ARR to DeepSeek, $800 million to MiniMax, $1 billion to Moonshot, $1.8 billion to Z.ai, $4 billion to ByteDance, and $2.4 billion to Alibaba. These figures stem from a recent monthly ARR snapshot that has been annualised, meaning they do not necessarily represent cash actually received over a twelve‑month span.

Comparative Valuation Multiples

When the ARR numbers are matched with publicly disclosed or estimated market capitalisations, the resulting valuation multiples diverge sharply. Moonshot trades at roughly fifty times its ARR, while DeepSeek’s multiple climbs to 163. By contrast, OpenAI is valued at about thirty‑four times its ARR and Anthropic at twenty‑one times, according to Rhodium’s internal calculations.

The gap in multiples signals differing investor expectations regarding growth trajectories, market access and the scalability of each company’s business model. Higher multiples for Chinese start‑ups reflect optimism about rapid user adoption even though price points are generally lower.

Pricing Strategies and Open‑Weight Models

Chinese AI providers frequently compete on price, offering reduced subscription fees and, in certain cases, open‑weight models that developers can download and fine‑tune. While this strategy broadens the user base, it also complicates direct monetisation, forcing firms to extract revenue through ancillary services, premium features or enterprise‑level contracts.

Z.ai recently revised its internal forecast, lifting its year‑end ARR target from $2.4 billion to $3 billion. Rhodium stresses that a forecast is not a realised outcome, yet the upward adjustment underscores the company’s confidence in scaling both pricing and service offerings.

Limitations of Revenue as a Quality Indicator

Rhodium warns that revenue gaps alone do not determine model quality. The ARR gap primarily measures monetisation efficiency, distribution channels and access to capital, not the intrinsic performance of the underlying models.

OpenAI and Anthropic benefit from extensive venture backing, large‑scale cloud partnerships and a brand advantage that enables premium pricing. Chinese firms, while expanding quickly, often depend on domestic cloud ecosystems and financing structures aligned with government policy.

• DeepSeek – $500 million ARR
• MiniMax – $800 million ARR
• Moonshot – $1 billion ARR
• Z.ai – $1.8 billion ARR
• ByteDance – $4 billion ARR
• Alibaba – $2.4 billion ARR

The ARR methodology employed by Rhodium annualises a recent monthly snapshot, which can either over‑state or under‑state actual cash flow depending on seasonal contract signing patterns and the timing of revenue recognition policies across jurisdictions.

For organisations assessing AI partnerships, the data suggest that Chinese providers can deliver cost‑effective alternatives, particularly for workloads that are less sensitive to the latest model benchmarks. However, the lower ARR indicates that many of these firms are still constructing the enterprise‑grade support and compliance frameworks that larger U.S. players already possess.

In practical terms, an English‑speaking company contemplating a switch or diversification of AI services should weigh immediate cost savings against the maturity of service‑level agreements, data residency requirements and the long‑term sustainability of the provider’s revenue model.

The 10 % ARR share signals a growing yet still niche market segment that may be well suited for pilot projects or cost‑conscious deployments, while mission‑critical applications are likely to continue relying on OpenAI or Anthropic’s more established revenue streams and higher valuation multiples.

Implications for Future Market Dynamics

If Chinese firms maintain aggressive pricing and continue to refine open‑weight offerings, they could gradually erode the cost advantage of U.S. incumbents, especially in price‑sensitive regions. Nonetheless, scaling enterprise‑level support and meeting international compliance standards will be decisive factors in gaining broader market share.

Analysts note that the current ARR disparity also reflects differing regulatory environments. Chinese providers operate under a framework that encourages rapid deployment but may limit cross‑border data flows, whereas U.S. companies navigate stricter privacy regimes that can justify higher pricing.

Ultimately, the Rhodium Group’s figures highlight a market in transition: Chinese AI model providers are carving out a measurable revenue slice, yet the path to parity with OpenAI and Anthropic will require sustained investment in infrastructure, compliance and brand trust.