--- title: "Can Kuaishou's Kling carry the company beyond short video?" date: 2026-10-24 category: Companies & Stack site: NeuroAI canonical: https://neuroai.site/a/na-corp-kuaishou-kling-ai-video language: en --- # Can Kuaishou's Kling carry the company beyond short video? > Kuaishou says its Kling video-generation model reached a US$240 million annual revenue run rate in December 2025, 19 months after its 2024 debut. A filmmaker in Lagos types a single sentence and watches a ten-second clip assemble itself, cut by cut. The result looks less like a slideshow and more like a rough first draft of an advertisement. The engine behind that clip is Kling, built by a Beijing company most global audiences know for endless short-video feeds rather than frontier research. Kuaishou has spent two years pushing the model from a lab curiosity into something creators actually pay for. For the parent company, the real test is no longer whether generative video can impress a demo crowd, but whether it can earn its place on the income statement while the compute (算力) bills keep climbing. ## Key takeaways - Kling reached a US$240 million annual revenue run rate in December 2025, Kuaishou said in its investor release. - The model crossed US$100 million ARR in March 2025, about ten months after its public launch. - Kling served more than 60 million creators and over 30,000 enterprise users worldwide as of December 2025. - Kuaishou confirmed it is weighing a restructuring of Kling's assets and outside funding. - Video generation is shifting from showcase to a booked revenue line at a major platform. ## From a side project to a revenue line Kling is a video-generation large model (大模型) that Kuaishou first opened to the public in 2024. What made it notable was timing: it arrived as a rare Chinese model that could stand beside better-known Western entrants on quality, and it did so while selling access rather than only chasing benchmarks. Kuaishou frames the commercial story in run-rate terms. In its January 2026 investor update, the company said Kling's monthly revenue exceeded US$20 million in December 2025, which it annualizes to roughly US$240 million (company-stated). That figure, if sustained, places Kling among the few generative-video products with a disclosed, recurring paid base at this scale. The ramp was steep by Kuaishou's own account. The company said Kling reached a US$100 million annual revenue run rate in March 2025, only ten months after launch, and then more than doubled that within nine months (company-stated). Kuaishou attributes the growth to a "high-velocity iteration cycle" rather than a single breakthrough, arguing that frequent model drops keep paying users from drifting to rivals. ## What Kling actually ships During an "Omni Launch Week" in early December 2025, Kuaishou unveiled a bundle of models it says lifts Kling from a text-to-video tool into a small production suite (company-stated): - **Kling Video O1** — described by the company as a unified multimodal video model with editing and video-referencing features. - **Kling Image O1** — a next-generation image model for still assets and concept frames. - **Kling Video 2.6** — adds simultaneous audio-visual generation so sound and picture are produced together. - **Kling Digital Human 2.0** — targets virtual presenters and spokes-characters for marketing and e-commerce. Kuaishou says these models serve professional creative work across marketing, e-commerce, film, short drama, animation and gaming (company-stated). The claim that matters for investors is not any one feature but the breadth: a paid toolkit that touches several stages of production instead of a single export button. ## The spin-off question Kuaishou has confirmed it is exploring a restructuring of Kling's related assets and the introduction of external financing, a plan the company described as preliminary. Chinese financial media have reported the move could precede a standalone listing, though those valuation and fundraising figures come from a single chain of reporting and should be treated as indicative rather than confirmed. The interest is understandable. Kling's disclosed run rate is small next to Kuaishou's overall business, yet it grows far faster than the core feed. That gap is exactly why a carve-out is tempting: it would let the model raise capital and be valued on its own trajectory rather than against a maturing short-video parent. ## Why the economics are still unproven The contrarian read is that a US$240 million run rate is impressive but early. Video generation is compute-hungry, and Kuaishou has said group capital spending will rise sharply as it builds out Kling's infrastructure. Until marginal revenue outruns marginal compute cost, "profitable product" and "profitable unit" are different claims. There is also a demand question. Kling's paying base skews toward "pro-sumers" — skilled creators who behave like businesses but buy like consumers. That segment is loyal and fast to adopt, yet it can shrink just as quickly if a rival model catches up on quality and price. Kuaishou's own figures show the per-quarter growth rate slowed in the back half of 2025, a hint that the easy adoption phase may be giving way to a harder retention phase. None of this dims Kling's standing as a commercial outlier. Most video models are still hunting for a business model; Kuaishou is the one publishing run-rate numbers and enterprise counts. The open question is whether a model that delights creators can also outrun the cost of serving them at scale. ## Honest limitations This article relies on Kuaishou's own investor disclosures for every revenue, user and model figure; the company did not publish an independently audited breakout for Kling. The reported spin-off valuation and fundraising size come from Chinese media and were not confirmed by a second source, so they are omitted or softened here. Run-rate math annualizes one month and can overstate a stable yearly result. I have not seen Kling's gross margin, churn, or per-video compute cost, which are the numbers that would actually settle the profitability debate. ## What readers can do now - **Operators:** test Kling against at least one Western video model on the same brief before committing a budget; judge on output consistency, not leaderboard scores. - **Investors:** track Kuaishou's disclosed capital spend versus Kling's run rate each quarter as the cleanest proxy for unit economics. - **Researchers:** treat Kling's Omni Launch Week suite as a case study in bundling generation, editing and audio into one paid workflow. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-corp-kuaishou-kling-ai-video Free to quote with attribution and a link to the original.