Moonshot is chasing a $50bn valuation, and it wants two funding rounds to get there

China’s Moonshot AI is raising money at a pace that would make most startups dizzy. The lab should close a round that values it at $31.5 billion in the coming days, Bloomberg reported. It then plans to open another almost immediately, targeting a valuation of as much as $50 billion.

Two rounds, one summer

The second raise would begin as early as August. It would be the final private infusion before an initial public offering in Hong Kong, which could come as soon as this year. SCMP reported the same August timing, describing it as Moonshot’s last private raise before listing.

The company is also unwinding its offshore “red-chip” structure by the end of this month. That clears the path for a domestic Chinese fundraising and the broader IPO push.

The Kimi K3 effect

The rush is built on one model. Moonshot released Kimi K3 last week, a 2.8-trillion-parameter open-weight system billed as the largest of its kind. On independent benchmarks reported by Bloomberg, it trailed only Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6.

The launch rattled US markets and revived talk of a “DeepSeek moment.” It also moved the numbers investors care about. Annual recurring revenue hit $300 million in June, up from $200 million in April. Daily sales have jumped at least six times since K3 arrived.

The reclusive founder

Behind it is Yang Zhilin, a 33-year-old who studied at Carnegie Mellon and worked at Meta and Google. He built the 300-person outfit with technologists from Tsinghua University. He named it after Pink Floyd’s “Dark Side of the Moon.” On Weibo, fans now call him a “millennial genius,” according to Business Insider. Existing backers include Tencent, Meituan, and state-linked funds.

The catch nobody is pricing in

Here is the awkward part. Open-weight AI has been a punishing business. Business Insider notes that rival Zhipu lost almost $500 million last year on about $107 million of revenue. Its shares have since fallen more than 40% in a month. Peer MiniMax lost $250 million on $79 million.

The problem is structural. Once a lab gives away its model weights, other firms run the model and collect the fees. Moonshot felt the flip side of its own success, too. Within days of launch it had to pause new sign-ups, short of the computing power to serve everyone.

Why raise so fast, then

Speed is the strategy. Chinese labs raise quickly when their sector runs hot, and Moonshot is the hottest name in it right now. The money buys compute and time. It also lands before any door shuts. Top US AI executives are warning loudly about Chinese models, and Washington is weighing curbs on their use.

Whether a $50 billion price tag survives contact with those economics is the real test. For now, investors are not waiting to find out.