Moonshot AI d new subscribers to its hit chatbot two days before Bloomberg reported the Beijing startup is negotiating a fundraising round that could value it at $50 billion.
The timing is almost comic. On July 19, Moonshot AI told users it was closing new sign-ups for its Kimi K3 model. Demand had surged sixfold in 48 hours. Its GPUs were maxed out. Two days later, Bloomberg reported the startup is preparing to start talks in August on what would be its final private fundraising round before a Hong Kong listing, this time targeting a valuation of up to $50 billion.
That is not a typo. Moonshot closed a round at a $31.5 billion valuation just weeks earlier, according to Bloomberg. Now investors want to pay considerably more for a company that just told its own customers to wait.
Kimi K3 is the reason. Released July 17, the 2.8 trillion parameter model is, by Moonshot's account, the largest open-weight frontier model built so far. According to benchmark results Moonshot has publicized, K3 outperformed Anthropic's Opus 4.8 and OpenAI's GPT-5.5 on a range of coding and agent tasks. The South China Morning Post reported that demand pushed Moonshot's compute close to full capacity within two days of launch. That forced the company to split its membership into a general Kimi Membership and a separate Kimi Code Membership for developers, while it scrambles to add servers. Moonshot has said it plans to release the model's full weights by July 27.
You don't usually see a startup ration access to its own product while its valuation triples in six months. But that's exactly the bet investors are making. Kimi K3's performance, at a fraction of the compute cost of its American rivals, is apparently worth paying up for even while the company can't fully serve the customers it already has.
Moonshot's valuation math has moved fast even by AI standards. TechCrunch reported the company raised $2 billion at a $20 billion valuation in May. By June, Bloomberg reported Moonshot was in talks for a valuation north of $30 billion. That round closed at $31.5 billion. Now the company wants to go back to the well at $50 billion, according to Bloomberg's July 21 report, in what would be the last private round before an IPO that could land before the end of the year.
Moonshot's annual recurring revenue has climbed past $200 million. The company has also been unwinding the offshore holding structure Chinese startups typically use to raise foreign capital, a step it flagged in May to satisfy Beijing's tightened rules on overseas listings. It has held talks with China International Capital Corp and Goldman Sachs about the offering.
Kimi K3's launch rattled markets outside China too. The Nasdaq slipped roughly 1% in the days after the model's release, and chipmakers including Nvidia took a hit, as investors weighed whether a Chinese lab could match frontier performance at a lower cost. It's the same performance-per-dollar argument that rattled Wall Street when DeepSeek's R1 model surfaced last year. This time it's landing on a startup asking to be valued at $50 billion instead of one nobody had heard of twelve months earlier.
Washington has noticed. The Trump administration is weighing an executive order that would limit how freely open-source AI tools reach Chinese developers, on top of the chip export controls already barring Chinese firms from Nvidia's newest accelerators. A bill called the AI Overwatch Act is also on the table. Not everyone in the administration agrees. White House AI adviser David Sacks has warned that tighter rules risk, in his words, entangling the country in unnecessary complications and ceding competitive ground. That tension is playing out in real time as Moonshot tries to close a round that assumes none of those restrictions bite.
Frankly, the GPU shortage is the more interesting signal here. A company that can't onboard new paying subscribers because it doesn't have enough chips is not the picture of a startup coasting on hype. It built something people actually want to use, badly enough that they'll wait for a server slot to open. Investors betting $50 billion on Moonshot are betting that gap between demand and capacity closes, not that it was a mirage.
Whether Moonshot actually lands that valuation in August, or whether a Hong Kong IPO arrives this year at all, depends on factors well outside the company's control, including how far Washington is willing to go to slow it down. The number on the table right now is $50 billion. For a company whose own product currently comes with a waiting list, that's a remarkable vote of confidence.
Also read: SentinelOne Founders Raise $100 Million to Secure AI Agents Inside the Enterprise • Microsoft Commits Billions More to Mistral's European GPU Buildout • BrainCo's Bionic Hands Are Selling to Robot Makers, Not Just Amputees