The Beijing-based creator of the Kimi chatbot is targeting $2 billion in annualized revenue by the end of 2026, Bloomberg reported. Its revenue run rate reportedly climbed from $300 million in June to more than $1 billion in August following the July release of its open-weight Kimi K3 model. For channel partners, that growth could expand opportunities around enterprise AI hosting, integration, and managed services.
Founded in 2023 by researcher Yang Zhilin, the startup is seeking a $50 billion valuation as it prepares for a potential initial public offering in Hong Kong, Bloomberg reported. It is also evaluating a listing on Shanghai’s STAR Market.
Moonshot’s K3 model operates at an unusual scale, with 2.8 trillion parameters. Unlike closed models from OpenAI and Anthropic, Moonshot makes K3’s weights available for developers to download and modify. That approach can support wider adoption, although open-weight models generally produce lower margins than proprietary systems. OpenRouter metrics show users generating as many as 300 billion tokens daily through K3 models, TechCrunch reported.
To monetize that traffic, Moonshot is pursuing enterprise deployments with customers including AsiaInfo Technologies and Kingsoft Cloud Holdings. Under K3’s license, model-as-a-service businesses earning more than $20 million in aggregate revenue over any consecutive 12-month period must reach a separate commercial agreement with Moonshot. The company is also seeking revenue-sharing terms of up to 30% with Microsoft, Amazon, and Alphabet’s Google, Channel Insider previously reported.
Behind the price advantage
Moonshot’s trajectory signals a fundamental shift in how businesses choose artificial intelligence providers. By pricing high-end AI below many U.S. alternatives, Moonshot gives enterprise buyers another way to reduce inference costs. However, channel partners must weigh those savings against licensing conditions, data-governance requirements, infrastructure demands, and the possibility of future U.S. restrictions. If a low-cost, open-weight framework can handle production workloads, closed frontier labs will face mounting pressure to defend their premium rates.
Yet Moonshot’s rapid expansion carries serious operational exposure. Deploying a 2.8-trillion-parameter system at enterprise scale can require substantial computing capacity. Its open weights also allow other developers to inspect, modify, and build on the model.
Scrutiny and public market ambitions
A cloud of technical controversy also hangs over the company. Anthropic recently alleged that Moonshot routed nearly 300,000 Kimi customer requests to Claude during a 10-day period, with most directed to Claude Opus. Separately, Anthropic attributed more than 23 million alleged distillation exchanges to Moonshot between May and July 2026. Moonshot had not publicly responded to the latest allegations at the time of reporting.
Domestic rivals are accelerating their own commercial pushes as well. MiniMax reached an annualized revenue run rate of $800 million in August, while Zhipu AI’s Z.ai reached $1.6 billion, according to Bloomberg. Both are exploring STAR Market listings alongside Moonshot, Bloomberg reported. Meanwhile, Moonshot is working with Goldman Sachs, China International Capital Corporation, and Deutsche Bank on a Hong Kong share sale aimed at raising $3 billion.
Moonshot’s public-market plans matter most if they help K3 develop into a durable enterprise platform. Partners evaluating the model should watch its cloud agreements, licensing requirements, data-handling practices, and exposure to changing U.S. trade policy.
Read more: Cloudera and Mistral are bringing sovereign AI to private enterprise data, highlighting another model for partners delivering governed AI across cloud and on-premises environments.