China's Moonshot AI eyes $2 billion annual revenue by year's end
Chinese AI lab Moonshot AI, maker of the popular Kimi model, is aiming to double its annualized revenue to $2 billion by the end of the year. The push comes as Anthropic accuses the company of improperly using its Claude Opus model to train Moonshot's own systems.
Moonshot AI, one of China's most prominent artificial intelligence labs, is setting an ambitious target of reaching $2 billion in annualized revenue by the end of the year, according to a Bloomberg report published Friday. The goal would represent double the revenue run rate the company reported for August.
The projection reflects strong momentum behind Moonshot's K3 model, released earlier this summer. Although usage of K3 has slipped somewhat in recent months, data from OpenRouter shows the model family is still generating as many as 300 billion tokens per day on the platform.
Still far behind Western rivals
Despite the rising numbers, Moonshot's revenue remains far smaller than that of leading Western competitors. Recent reports put OpenAI's annualized revenue at around $40 billion and Anthropic's at roughly $65 billion. Because Moonshot releases its model weights openly, the company operates on thinner margins than closed-weight rivals. Even so, the growth suggests open-weight AI models can still generate meaningful revenue, even if they trail the profitability of closed frontier models.
Accusations of improper model training
Moonshot's methods have drawn scrutiny. Earlier this week, Anthropic accused the company of running a long-term distillation scheme in which nearly 300,000 requests sent to Kimi were secretly routed to Anthropic's Claude Opus model instead, effectively having Opus answer on Kimi's behalf. Anthropic further alleged that more than 23 million responses were collected from its models and used to train Moonshot's own systems.
