Kling AI Q2 revenue: RMB850M+, up 200%+ YoY and about 30% QoQ
RMB19B external equity round into Beijing Kling at US$18B post-money valuation
Kuaishou’s stake diluted from 100% to 68.33%, external investors hold 16.67%
Adjusted net profit: RMB3.9B, down 30.3% YoY as AI investment absorbed margin
Kling ARR trajectory: US$240M to US$500M in three months (Dec 2025 to Mar 2026)
Kuaishou is a Chinese short-video and e-commerce platform with 412M app DAU and 797M MAU, according to Kuaishou. Kling AI is Kuaishou’s video generation model, launched in preview in June 2024 and now in the 3.0 series.
Beijing Kling Intelligent Technology (Beijing Kling) is the subsidiary that will hold Kling AI’s assets, models, and R&D teams under a nine-month restructuring transfer. Kuaishou reports Kling revenue inside "Other services".
Two documents define the analytical picture. One is the Q2 release. The other is the July 2 HKEX filing on Beijing Kling’s external financing, which arrived days after quarter-close.
Kling monetized while everyone else was still testing
Kling AI generated more than RMB850M in Q2 revenue, up more than 200% YoY and about 30% QoQ, according to Kuaishou. Q1 revenue was RMB650M, up more than 300% YoY per TechNode. First-half revenue crossed RMB1.5B.
The annualized recurring revenue trajectory tracked even faster. Kling ARR roughly doubled in Q1 2026, moving from US$240M in December 2025 to US$500M by March, per TFN. The Kling 3.0 launch drove the acceleration.
Product depth expanded through Q2. Kling AI shipped native 4K video output, industry-first for a generation model. Kling 3.0 Turbo followed, cutting production costs while retaining dynamic quality and audio-visual synchronization. Kling MCP and Kling CLI launched, enabling AI agents to orchestrate Kling for batch content creation.
Creative recognition arrived alongside product. Kling-generated advertising videos took one Silver Lion and two Bronze Lions at the 2026 Cannes Lions International Festival of Creativity. Multiple Kling works were nominated at the 2026 Beijing International Film Festival.
Kling’s target segment is professional. The 4K output is aimed at film, television, and advertising studios that would otherwise commission post-production. MCP and CLI make Kling usable inside enterprise automation workflows, not just individual creator interfaces. Professional positioning explains the willingness-to-pay curve that shows up in the ARR trajectory.
The commercial context makes the numbers sharper. OpenAI shut down the consumer version of Sora in April 2026, per TFN. Sora had burned roughly US$1M per day in compute against approximately US$2M in total revenue. Runway pivoted away from commercial video in the same window. Kling was quadrupling revenue as its Western analogues retreated from the category.
The RMB19B round validated the monetization thesis
On July 2, 2026, Kuaishou disclosed to HKEX that Beijing Kling had signed a capital increase agreement with external investors, per Kuaishou’s filing. Initial investors committed RMB13.8B (~US$2.04B). Fifteen additional investors subscribed a further RMB5.22B. Total committed capital reached RMB19.04B (~US$2.8B) with a ceiling of RMB20.45B if additional investors accede within 60 days.
The valuation was US$15B pre-money, approximately US$18B post-money. On Kling’s US$500M March ARR, that implies roughly a 36x ARR multiple. The multiple is sharp. It rests on disclosed revenue rather than adoption metrics.
Kuaishou’s stake in Beijing Kling fell from 100% to 68.33%. External investors took 16.67%. Share participation schemes for Beijing Kling talent absorbed the remaining 15%. Kuaishou continues to consolidate Beijing Kling’s results into its own financial statements, per Kuaishou.
Investor stakes were disclosed. Alibaba took 0.87%. Shanghai Qishan, a Tencent subsidiary, took 0.44%. Baidu took 0.22%. Combined, the three top Chinese internet groups hold about 1.5% of Beijing Kling. State-backed vehicles also participated, including the Beijing Information Industry Development Investment Fund and the Beijing AI Industry Investment Fund, per 36Kr.
The restructuring framework agreement gives Beijing Kling nine months to consolidate the Kling AI stack, per Kuaishou’s July 2 filing. Assets in scope include models, algorithms, R&D teams, and overseas operations. The structure is pre-IPO by design.
Tencent absorbed its AI cost on its own P&L in Q2. Tencent disclosed a RMB10.5B new AI products loss rather than rent its compute at 30%+ margins available today, per prior CIW analysis. Kuaishou took a different route. Kuaishou secured RMB19B of external equity capital into Beijing Kling to fund Kling’s forward compute build. Consolidation stays. Future upside is diluted in exchange for present capital. Different capital structure, same underlying problem: how to fund AI ambition without shrinking the parent.
The core Kuaishou business showed the cost of the AI shift
Q2 group revenue was RMB35.5B, up 1.4% YoY, per Kuaishou. Adjusted net profit was RMB3.9B, down 30.3% YoY. Operating profit fell 29% YoY to RMB3.8B. Adjusted net margin held at 11%, keeping the parent profitable.
Gross margin fell 410bps to 51.6%, from 55.7% a year earlier. Revenue-sharing costs and related taxes rose to 34.7% of revenue from 30.1%, the dominant driver of the gross margin decline. AI capex shows up on the balance sheet, not in the gross margin lines this quarter.
R&D expenses rose 34.7% YoY to RMB4.6B, and Kuaishou attributed the increase explicitly to AI investment including training expenditures. Property and equipment on the balance sheet reached RMB34.1B at June 30, up 49% since December 2025, reflecting the capex ramp for AI infrastructure.
Live streaming revenue fell 13.5% YoY to RMB8.7B. The decline is a deliberate quality shift, not AI-driven. Kuaishou has been throttling live streaming to protect ecosystem health for multiple quarters.
Reported cumulative Kling losses reached roughly RMB2.4B over two years, per BigGo. Those losses remain on the parent’s P&L until the Beijing Kling asset transfer completes. The RMB19B external equity funds forward burn. Historical burn stays with Kuaishou.
Cash position held up despite the margin pressure. Total available funds reached RMB121.3B at June 30, up from RMB117.7B at March 31. Kuaishou repurchased 43.3M Class B shares in the first half for HK$1.97B and paid a final HK$3.0B dividend for 2025 in July. Shareholder returns kept pace even as the AI drag widened. The parent still generates enough cash to absorb the transition period.
Internal AI deployment as the cost-out story
Kuaishou disclosed internal AI adoption metrics that sit outside the Kling story. MyFlicker, Kuaishou’s proprietary general-purpose agent product, reached 92% employee adoption in June 2026, per Kuaishou. AI-generated code hit 60% of R&D team output, per Kuaishou.
Commercial-facing tools scaled similarly. Spending on AIGC short-video marketing materials grew more than 70% YoYin Q2. More than 850,000 merchants used Kuaishou’s free AI business tools in H1 2026. Use cases spanned product listing, marketing materials generation, business analysis, smart placement, and AI-powered customer service.
Kuaishou Vanchin, an enterprise-grade large model service platform, supports both internal AI applications and external enterprise clients. The Vanchin disclosure is thinner. Kuaishou did not release revenue, seat counts, or client names for the platform.
All four data points are Kuaishou-only disclosures. Independent verification is not available for the 92% MyFlicker adoption or the 60% AI-code figures. Read as adoption signals rather than benchmarks.
The strategic significance sits in the offset. Kuaishou is deploying AI internally to compress opex growth while the AI capex builds up on the balance sheet. R&D headcount productivity rising via AI-assisted code means the +34.7% R&D expense growth funds capability rather than headcount. AIGC marketing materials generation cuts creative production costs across the online marketing services business. Together, the internal AI story is a partial answer to the margin compression showing up on the P&L.
What the BAT investment reveals about the Chinese AI supply chain
The most analytically interesting feature of the RMB19B round is who invested. Alibaba, Tencent, and Baidu are the three largest Chinese internet groups. All three have their own generative AI programs. All three invested in Kling anyway. Each has a different reason.
Tencent contributed roughly US$200M to the Kling round, per CNBC citing the July 2 regulatory filing. Tencent operates Hunyuan Video, its own generative video model. Investing in a competitor buys commercial distribution access to Kling’s user base.
The stake also reads as a defensive position against ByteDance’s Jimeng AI (or Dreamina outside mainland China), Kling’s closest domestic rival.
Four days after the Kling round closed, Tencent sold 273M Kuaishou Class B shares for HK$11.8B (~US$1.51B), per Bamboo Works. Tencent’s Kuaishou parent stake fell from 15.68% to 9.37%. The US$1.51B raised on the parent sale was roughly 7x Tencent’s US$200M Kling contribution. On net, Tencent extracted capital from the Kuaishou complex in the same week it added Kling exposure.
Alibaba launched HappyHorse in April 2026, a competing video generation model. HappyHorse is led by Zhang Di, per Bloomberg and Sina Finance coverage. Zhang was Kuaishou’s VP and technical architect of Kling AI before joining Alibaba in late 2025.
The architect of Kling now competes with Kling. Alibaba’s Kling investment reads as a hedge. If HappyHorse falters, Alibaba retains upside through its Kling stake. If HappyHorse wins, the Kling exposure is small enough to absorb.
Baidu has Ernie Bot and no dominant video generation offering. The Kling stake is the purest commercial exposure of the three top Chinese internet groups. A category bet where the buyer is not competing directly.
State-backed fund participation signals policy support for a Chinese AI video champion positioned to compete with Sora, Runway, and other Western platforms. Two Beijing state vehicles joined the round: the Beijing Information Industry Development Investment Fund and the Beijing AI Industry Investment Fund.
The Yangtze River Delta Digital and Intelligent Culture Industry Private Equity Investment Fund also participated. State capital in a vertical AI round of this size marks Kling as strategically significant beyond commercial merit.
Abu Dhabi’s BlueFive Capital also joined, per Bloomberg, extending sovereign-capital reach into Chinese vertical AI. Entertainment industry investors including Huace Film & TV and Mango Investment took stakes reflecting Kling’s utility as a production tool. Twenty-plus co-investors round out the syndicate.
The ByteDance dynamic sits underneath the whole round. ByteDance’s Jimeng AI is Kling’s closest domestic competitor, embedded in the Douyin ecosystem. All three top Chinese internet groups plus state capital coalescing around Kling reads as a structural response to Jimeng AI. Chinese AI supply-chain coordination against a single-firm rival is not new. Coordination at this scale in vertical AI is.
A pattern is forming. The Beijing Kling deal opens a spin-out-plus-strategic-capital route that pure-play Western competitors like Runway lacked. Revenue traction attracts strategic investors. Strategic investors provide the compute funding without loading the drag onto the parent’s P&L. Consolidation preserves upside. IPO becomes the exit valve. Whether other Chinese vertical AI leaders follow is the next question.
Kuaishou has monetized AI where the hyperscalers are still spending. It secured RMB19B of external equity into Kling that dilutes Kuaishou’s ownership from 100% to 68.33% while preserving consolidation.
Q3 tests are cleaner than the Q2 disclosure. Whether the core business stabilizes margins. Whether Kling holds the 30% QoQ Q1-to-Q2 pace. Whether the RMB19B external capital deploys into Kling capex or sits as a cash buffer. The spin-out defined the capital structure. The next two prints define whether it works.