Moonshot AI is betting big on its open-weight strategy, aiming for $2 billion in annualized revenue by year-end. This target doubles the company’s reported run rate from August, signaling aggressive confidence in its K3 model. While usage has dipped slightly, OpenRouter data reveals a staggering 300 billion tokens generated daily by K3 models. That volume proves demand remains robust, even if the per-token economics are thinner than those of closed-source rivals.
The financial ambition stands in stark contrast to industry giants. OpenAI and Anthropic are reportedly chasing $40 billion and $65 billion revenue marks, respectively. Moonshot’s path is harder because it gives away model weights for free. Lower margins mean they need massive scale to compete. Yet, this push demonstrates that open-weight models can still drive serious cash flow, challenging the notion that only walled gardens yield profit.
Legal clouds loom over this growth spurt. Anthropic recently accused Moonshot of running a large-scale distillation campaign. The allegation claims Moonshot routed nearly 300,000 requests from its Kimi chatbot directly to Claude Opus. This trick allegedly harvested over 23 million responses to train Moonshot’s own systems. If true, it represents a significant breach of trust and potentially illegal data scraping practices.
This situation highlights a growing tension in the AI sector. Companies want the benefits of open collaboration but fear having their core assets stripped for competitors’ gain. Moonshot’s success depends not just on technical performance but on navigating these ethical and legal minefields. Can they hit that $2 billion mark while fighting a major lawsuit? The answer will define whether open-weight models can survive as standalone businesses or if they remain vulnerable to litigation.
