Moonshot AI is negotiating a $50 billion valuation — 160 times its reported revenue

Moonshot AI is negotiating a final pre-IPO funding round that would value the Beijing lab at as much as $50 billion, according to Bloomberg reporting on talks the company plans to begin in August once its current financing closes. That target is more than triple the roughly $31.5 billion valuation Moonshot was closing at just weeks ago, and more than eleven times where it stood in December 2025. The company intends for this round to be the last outside money it takes before going public on the Hong Kong Stock Exchange, with a listing possible before the end of the year.

The catalyst is Kimi K3, the 2.8-trillion-parameter open-weight model Moonshot released on July 17 and billed as the largest open-weight system publicly available — the first to cross into what the company calls '3T-class' territory. Demand was immediate enough that Moonshot paused paid subscriptions for new consumer users on July 19, two days after launch, citing capacity constraints. Reported annual recurring revenue climbed from roughly $100 million in March to about $200 million in April and $300 million by June, a trajectory the company is now using to justify the size of the round.

The multiple underneath the headline number

A $50 billion valuation against $300 million in reported run-rate revenue works out to roughly 167 times revenue — a multiple that would be extreme even by the standards of a sector already trading well above historical software norms. It's worth being precise about what that revenue figure is and isn't: it's a run-rate estimate derived from recent monthly activity, not an audited annual figure, and Moonshot has not published GAAP-equivalent financials. Investors betting at this valuation are pricing in continued triple-digit growth off Kimi K3's momentum, not the business Moonshot reports having today.

The IPO structure adds a second layer worth watching. Moonshot is reportedly preparing to unwind its overseas 'red-chip' holding structure — the offshore-entity setup many Chinese companies use to list abroad — ahead of a Hong Kong listing underwritten by CICC and Goldman Sachs, with a filing possible within about six months. Removing that structure typically signals a company positioning for a listing that keeps more of its regulatory footprint onshore, a detail that matters for how the deal will be read by regulators on both sides of the transaction.

What has to be true for this to make sense

For a 167x-revenue valuation to hold up past the IPO, Kimi K3's demand spike has to be a durable step-change in Moonshot's business, not a launch-week bump that a capacity-driven subscription pause temporarily flattered. Open-weight releases draw an initial wave of downloads and trials that don't always convert into the paid, sustained usage a revenue multiple like this requires. The company's own numbers — $100 million to $300 million in three months — are real growth by any standard, but they're also the kind of curve that investors have been burned extrapolating in this sector before. There's also a structural tension worth naming directly: Kimi K3 is open-weight, meaning anyone can download and self-host it for free, which makes it a strange product to point to as the engine behind a subscription-revenue growth story. Moonshot's bet is that the model's popularity drives usage of its own hosted, paid API and consumer app even as the weights themselves are freely available — the same wager DeepSeek and Meta have made with their own open releases, with mixed results on how much of the resulting attention actually converts to revenue.

Moonshot's rush toward Hong Kong isn't an isolated move. Chinese AI and chip companies have been turning to Hong Kong listings in waves through 2026 — GPU maker MetaX confidentially filed for its own Hong Kong IPO this same week, following chipmakers Biren Technology, Iluvatar CoreX, and Moore Threads, which have all listed in Shanghai or Hong Kong since late 2025. The pattern traces back to Beijing's semiconductor and AI self-reliance push: access to capital markets has effectively become policy support for domestic firms racing to build alternatives to US-controlled technology while export controls hold the newest American hardware out of reach. Moonshot, as a model developer rather than a chipmaker, sits one layer up from that specific fight, but it's riding the same capital-markets wave and the same state-linked investor base.

Who's actually funding this

Moonshot's cap table blends state and private capital in a way that's become typical for China's frontier AI labs. IDG Capital-affiliated vehicles hold the largest institutional bloc, at roughly 16.5% combined, alongside stakes built up over multiple rounds by Alibaba, Tencent, Meituan, and Xiaohongshu, plus early venture backing from Sequoia China, ZhenFund, and 5Y Capital. State money runs through the structure too: China's National Social Security Fund holds a position via a Yangtze River Delta sci-tech investment vehicle, alongside smaller stakes from local industrial funds in Qianyan and Lingang. Founder Yang Zhilin retains majority control of the domestic entity at just under 52% ownership, with his co-founders holding another 13.8% between them — a concentration of founder control that's unusual at this valuation and gives Yang significant leverage over how, and whether, the $50 billion round actually closes.