Kuaishou’s generative video unit has closed a new funding round at a valuation of about USD 18 billion. That is higher than the market capitalisation of Kuaishou itself, the Hong Kong-listed company that built it.
The immediate reaction is to call it a bubble. The more useful reaction is to treat it as the clearest available signal of how Chinese capital is repricing AI assets away from their parents.
The money going in
On 3 November, three separate AI announcements landed in the same news cycle. Kuaishou’s Kling closed its round. Tencent released Hyra 2.0 under its Hunyuan line. Alibaba’s Hong Kong shares rose on the back of a partnership with Apple.
The aggregate numbers behind those headlines are large enough to change the framing. Between July 2025 and June 2026, Chinese AI and application companies completed 3,297 financing rounds, according to data compiled by 21 Chuangtou. Of that total, 1,595 rounds closed in the first half of 2026, with disclosed proceeds estimated at about RMB 227 billion.
Internet platforms are the most aggressive participants. Kling’s revenue reached more than RMB 650 million in the first quarter of 2026 and more than RMB 850 million in the second, representing year on year growth above 300 per cent and 200 per cent respectively. Annualised recurring revenue climbed from USD 100 million in the first quarter of 2025.
The argument capital is ignoring
The sceptical case was made in October, and not by a marginal voice. Li Lu, the investor often described as China’s answer to Warren Buffett, used a Columbia lecture to argue that AI has moved beyond being a productivity tool and is travelling along a spectrum from tool to creator, acquiring reasoning, coding and collaboration capability, and edging past the threshold of the Turing test.
His conclusion was not that this will happen unopposed. It was that a technology capable of reaching that point will not be allowed to run all the way, either because it falls short technically or because society organises to stop it. Buffett and Duan Yongping have made versions of the same argument.
Capital is behaving as if that constraint does not bind. The reason is straightforward: for the platforms, AI spending has stopped being a discretionary bet and become a fixed cost of staying in the market. When capex moves from expectation to obligation, the valuation question changes shape. Nobody prices a survival tax on a discounted cash flow basis and expects a small number.
The question the Kling number actually asks
The interesting figure is not USD 18 billion against Kuaishou’s market cap. It is whether Kling’s revenue growth is a product of genuine demand or of platform subsidy, because that determines whether the valuation holds through a funding winter.
RMB 850 million in a quarter at those growth rates is real revenue, not a pilot programme. The open question is retention and pricing power once the initial novelty of generative video fades and the same capability is available from a dozen Chinese vendors.
The answer will not come from Kling’s valuation. It will come from the next two quarters of gross margin, and nobody has published that number yet.
Editor’s note: This is an adapted translation of the original 36Kr report. It has been trimmed and restructured for readability for an international business audience. The full original (in Chinese) is at https://36kr.com/p/4058156403234691.