Kimi-maker Moonshot AI targets $2B in annual revenue

Moonshot AI announced an aggressive revenue target of $2 billion for the coming twelve months, citing the commercial traction of its K3 language model released this summer. The company claims the model now processes roughly 300 billion tokens per day on OpenRouter, a usage level that underpins the projected sales despite a modest dip in recent weeks. Because Moonshot distributes K3’s weights freely, its profit margins are far slimmer than those of closed‑weight rivals such as OpenAI and Anthropic, whose annual revenues sit near $40 billion and $65 billion respectively. The open‑weight approach, however, still appears capable of generating sizable cash flow, provided the firm can monetize token consumption through services, licensing, or cloud partnerships.

The push for $2 billion places Moonshot squarely in the middle tier of the AI ecosystem, where open‑source‑style models compete with proprietary, high‑margin offerings. The Chinese AI sector has long favored more permissive licensing to accelerate adoption, but the revenue gap with Western incumbents remains stark. Moonshot’s strategy reflects a broader industry experiment: extracting value from freely available model weights by bundling them with infrastructure, support, or domain‑specific fine‑tuning. At the same time, the company’s alleged practice of routing hundreds of thousands of user queries to Anthropic’s Claude Opus—collecting over 23 million responses for training—highlights a contentious gray area where open‑weight developers may lean on competitor APIs to bootstrap performance, raising questions about intellectual‑property boundaries and fair competition.

If Moonshot can sustain token volume and convert it into paid subscriptions or enterprise contracts, the $2 billion goal is plausible; however, the Anthropic lawsuit introduces legal uncertainty that could stall growth or force costly settlements. Regulators in China and abroad may also scrutinize cross‑border data flows and model‑distillation tactics, especially as governments tighten AI governance. Observers should monitor the outcome of the Anthropic accusations, Moonshot’s pricing and partnership announcements, and any shifts in token‑usage trends that could signal whether open‑weight models can scale profitably alongside closed‑weight behemoths.