One of China’s leading AI labs, Moonshot AI, is setting its sights on a monumental $2 billion in annual revenue by the end of the year, according to Bloomberg. This ambitious target reflects the success of its K3 model, which, despite a slight dip in usage, still generates up to 300 billion tokens daily, as per OpenRouter data.

Moonshot’s projections contrast sharply with those of its more established rivals, OpenAI and Anthropic, which are valued at $40 billion and $65 billion, respectively. However, the company’s revenue figures are significantly impacted by the fact that its models are open-source, offering much lower margins than their closed-weight counterparts.

The company’s model development practices have raised eyebrows, particularly due to accusations from Anthropic of a long-running model distillation campaign. This alleged campaign involved Moonshot routing nearly 300,000 requests from Kimi directly to Claude Opus, effectively using Anthropic’s models to train Moonshot’s own AI. Over 23 million responses were reportedly collected for training purposes.

Despite the controversy, Moonshot’s aggressive growth targets suggest there may still be significant potential in open-source AI, even if it’s a less lucrative path than the closed-weight models. It’s a complex landscape, one where the AI frontier is as much about ethics as it is about innovation.