Moonshot AI has closed a $2 billion funding round at a $20 billion valuation, a fivefold jump in just six months. The raise, led by Meituan’s venture arm with participation from Tsinghua Capital, China Mobile, and CPE Yuanfeng, signals a decisive shift in investor conviction: open-weight Chinese AI models are not only technically credible but now commercially scaled.
The surge in valuation is tied to the economics of open-weight systems. While these models often trail frontier closed models on raw benchmarks, they are dramatically cheaper to run. That cost advantage is proving commercially meaningful. Moonshot’s Kimi line—most recently the K2.5 and K2.6 models—has already reached an estimated $200 million in annual recurring revenue as of April and has become the second-most used model on OpenRouter. Revenue traction of that magnitude this early in the cycle is unusual in the foundation model market and central to why capital is concentrating quickly.
The momentum extends well beyond a single company. DeepSeek is reportedly raising at a valuation near $45 billion. Zhipu AI has reached a public market capitalization of $55.9 billion. MiniMax is valued around $33 billion. Across these labs, valuation step-ups are being driven by rapid model iteration, open-weight distribution strategies, and rising enterprise demand for lower-cost inference options. Investors are essentially betting that economic efficiency, not peak benchmark performance, will determine large portions of model adoption.
For global investors, the competitive implications are notable. Chinese labs are proving that open-weight architectures can scale into real businesses, challenging assumptions that the future belongs exclusively to closed, compute-intensive Western models. Moonshot, founded in 2023 by former Meta and Google researcher Yang Zhilin and backed by Alibaba, Tencent, and HongShan, is now positioned as one of the clearest signals that the market is rebalancing. Capital is flowing toward models that are fast to deploy, cheap to run, and flexible enough for enterprises to integrate deeply.
The result is a shifting landscape: open-weight players are no longer alternative bets but contenders shaping the long-term structure of the LLM market.
