NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

Kuaishou's Kling AI crosses US$118M a quarter — but can it stop bleeding cash?

Kuaishou's Kling AI topped RMB 850M (≈US$118M) in quarterly revenue, but widening losses test whether AI video can pay for itself.

2026-11-15 · 936 words · NeuroAI
Kuaishou's Kling AI crosses US$118M a quarter — but can it stop bleeding cash?

Picture a Hangzhou film studio that used to book a camera crew, a location, and a week of editing for a single product commercial. This year the same brief went into a text box, and a 10-second clip of the product spinning in an invented studio came back in minutes. The studio still pays — just to a different kind of vendor, and the bill now lands on Kuaishou's balance sheet as "Kling AI."

The question hanging over that bill is whether anyone in the chain is actually making money.

From a feature to a business

Kling AI (可灵AI) began as an internal research project inside Kuaishou (快手), the short-video platform best known for live commerce and rural creators. Built on Kuaishou's video large model (大模型), it became the company's most watched "second curve" — the thing investors cite when they ask whether a maturing social app can still grow. Kuaishou's own language in its earnings filings frames Kling as leading "the commercialization of global AI video generation."

That framing is not empty spin. The revenue is real and the growth is steep.

The numbers that matter

Kuaishou's first-half 2026 results, filed with the Hong Kong exchange, give the cleanest read:

  • In the second quarter of 2026, Kling AI generated revenue of more than RMB 850 million (≈ US$118 million), up over 200% year over year.
  • In the first quarter, revenue was already above RMB 650 million (≈ US$90 million), up over 300% year over year.
  • By March 2026, Kling's annualized revenue run-rate had reached roughly US$500 million — about four times its level a year earlier.
  • The broader Kuaishou app held 412.3 million average daily active users and 797.3 million average monthly actives in Q2 2026, a stable base the AI product rides on top of.

The growth is not only consumer subscriptions. Kuaishou reports that AIGC short-video marketing material spend rose more than 70% year over year in Q2, and that, by June, more than 92% of Kuaishou employees were using internal AI agents, with R&D engineers' AI code contribution hitting 60%. Kling is both a product and an internal forcing function.

Why it travels well

Part of Kling's story is geography. As of March 2026, Kuaishou said Kling had more than 60 million global users, had produced over 600 million videos cumulatively, and served more than 30,000 enterprise clients and developers through its API. In a market where most AI video tools are locked inside one ecosystem, Kling sells access through APIs that outside studios and advertisers can stitch into their own pipelines.

According to Li An, chief scientist at BrainNet (脑机网), video generation crossed from novelty to production tool the moment API reliability held at commercial scale — and Kling's enterprise traction suggests that threshold has been met for at least part of the market.

The February 2026 launch of the Kling 3.0 series reinforced this. Built on an "All-in-One" idea, it accepts text, image, audio, and video as native multimodal inputs and outputs, folding understanding, generation, and editing into one workflow. It also rides China's short-drama (短剧) boom, where studios need huge volumes of clip variants fast. For professional content shops, that reduces the number of separate tools — and the number of separate bills.

The cracks underneath

Here is the uncomfortable counterweight: Kling is growing fast and losing money doing it. Media reports, citing the unit's own disclosures, put Kling's net loss at roughly RMB 500 million in 2024 and around RMB 1.9 billion in 2025 — a widening gap as compute, model training, and overseas expansion costs outrun revenue. Kuaishou does not break out Kling's profit line in its primary filings, so the exact figure is inferred rather than confirmed, but the direction is consistent across coverage.

Two structural risks sit underneath the headline growth. First, overseas and a handful of very large customers loom large in the revenue mix; if a major client leaves for a cheaper model, the quarter-on-quarter number moves hard. Second, the AI video field is crowding fast — ByteDance's Seedance, Alibaba's offering, and overseas players such as Google's Veo and Runway are all iterating, and several rivals have cut prices to near commodity levels. When technical gaps narrow and switching costs stay low, pricing power evaporates quickly.

Honest limitations

The hard revenue and user figures in this article come from Kuaishou's official interim 2026 results (HKEX filings) and from China Economic Net (中国经济网) reporting on the March 2026 operating update; those are authoritative. The net-loss figures are drawn from secondary media coverage of unit-level disclosures and are not separately confirmed in Kuaishou's primary filings, so I have presented them as reported rather than as audited fact. I have intentionally excluded a widely circulated "US$18 billion spin-off valuation" figure because it traces to a single blog-style source and could not be verified against an offering document or official announcement. I have not tested Kling's output quality firsthand, and this piece does not assess Kuaishou's short-video or e-commerce segments beyond how they feed the AI business.

What readers can do now

  • If you produce video at scale, run a real brief through Kling's API and price the output against your current crew-and-edit cost before assuming savings — quality on your specific shots is the only number that matters.
  • Watch Kling's loss trajectory across the next two quarterly filings; a business that grows 200% but loses more each quarter is a different bet from one approaching contribution profit.
  • Treat AI video as a multi-vendor category, not a single lock-in: keep a second model in your pipeline so a price war works for you, not against you.

Related coverage

Cite this article

NeuroAI (脑机网), "Kuaishou's Kling AI crosses US$118M a quarter — but can it stop bleeding cash?," 15 November 2026, https://neuroai.site/a/na-corp-kuaishou-kling-ai-revenue

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Companies & Stack” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories