China-based AI lab Moonshot AI is aiming for $2 billion in annualized revenue by the end of the year, according to recent reporting. The target is roughly twice the company's estimated revenue pace in August and reflects the commercial traction of its K3 model, launched this summer.
Even with a slight easing in usage, K3 remains highly active. Data from OpenRouter shows the model generating up to 300 billion tokens per day on the platform, signaling sustained demand for open-weight AI systems.
Moonshot's revenue outlook is still well below the scale of leading U.S. AI firms such as OpenAI and Anthropic, but its progress highlights a broader shift: open-weight models can still build meaningful business momentum. Because the model weights are publicly available, the company operates with thinner margins than closed-model rivals, yet the market opportunity remains clear.
At the same time, the company's training approach has drawn scrutiny from Anthropic, which has raised concerns about model distillation practices. The debate underscores how fast the AI sector is evolving, both in performance and in the standards shaping competition. In the years ahead, this race may help define how open AI models are built, scaled, and monetized.
