Kling's $2.8 Billion Raise Is a Verdict on AI Video Hype
Kuaishou's Kling AI closed a funding round of roughly $2.8 billion at an $18 billion valuation ahead of a planned Hong Kong spin-off listing, backed by Tencent, Alibaba Cloud, Baidu, and Chinese state-linked funds. The raise lands five months after OpenAI shut down Sora, which generated just $2.1 million in lifetime revenue against roughly $1 million in daily operating costs. Kling's edge isn't a flashier model — it's pricing 65% below Sora and distribution embedded directly into Kuaishou's commerce ecosystem. For PMs, the lesson is that vendor selection in generative video should weight unit economics and distribution over demo polish, because the category is now punishing spectacle and rewarding plumbing.
This week, pull up every AI video vendor in your stack or shortlist and run the math your last vendor review probably skipped: cost per second at your actual production volume, not the teaser tier. Kling 3.0 undercuts Sora by roughly 65% and Runway by 44%, according to reporting on the round, and it's winning capital precisely because of that math, not because its outputs are the most cinematic. If you picked a vendor based on demo quality, you picked on the wrong axis. Then ask the uncomfortable question directly: if this vendor's own unit economics don't work, what's your fallback, and how much notice would you actually get? Disney had a billion-dollar Sora partnership and found out about the shutdown less than an hour before the public did. That's not a hypothetical anymore. It's the base rate in this category right now. To be fair, none of this means Kling is automatically the safer long-term bet. It's a pre-IPO Chinese entity mid-reorganization, chasing a Hong Kong listing on a 12-month clock, at a valuation that assumes a price-war business model scales cleanly into public-market scrutiny. Treat it as the current cost leader, not a permanent one. Build your integration layer so you can swap the underlying video model without rewriting your product, because in this market, the vendor you're using in December might not be the one you started the year with.
Kuaishou's Kling AI raised roughly $2.8 billion (potentially near $3 billion) at an $18 billion valuation, backed by Tencent, Alibaba Cloud, Baidu, and Chinese state-linked funds, per WSJ and SCMP reporting.
On March 24, 2026, OpenAI posted a two-line goodbye to Sora on X and walked away from a product that had cost roughly $1 million a day to run and had generated, over its entire life, $2.1 million in revenue, according to The Wall Street Journal. The web and app were dark by April 26; the API dies in September. Nine days into July, on the other side of the world, Kuaishou's video unit Kling AI closed a funding round of roughly $2.8 billion (some reporting puts the ceiling near $3 billion) at an $18 billion valuation, according to The Wall Street Journal and the South China Morning Post, with Tencent, Alibaba Cloud, Baidu, and a syndicate of Chinese funds writing the checks.
Same category. Same twelve months. Opposite outcomes.
The verdict: in AI video, being first mattered less than being cheap, embedded, and boring about unit economics.
The Money Chasing the Wrong Kind of First
Kling's round is not a modest top-up. Per WSJ and SCMP reporting, an initial consortium of roughly 23 investors committed about $2 billion, with 15 more adding another $766 million, pushing committed capital toward $2.8 billion against a $15 billion pre-money valuation, rising to $18 billion once the round closes. The investor list reads like a truce among rivals: CPE Yuanfeng, Guofang Venture Capital, Abu Dhabi's BlueFive Capital, the Zhongguancun Science City Fund, and CITIC Securities lead, with Tencent, Alibaba Cloud, and Baidu joining as strategics.
Kuaishou's own stake in the newly consolidated Beijing Kling entity dilutes from 100% to roughly 68.33% once the round fully closes. The plan: a nine-month reorganization folding all Kling assets into that single operating company, then a 12-month clock, starting from the round's final payment date, to list on the Hong Kong Stock Exchange.
That's real capital chasing a real spin-off, not a rumor. But it's landing in a market where, five months earlier, the single most recognizable name in generative video, the one that made "Sora" a household term, got switched off because nobody had solved the money question. Sora's downloads peaked above 3.3 million in November 2025 and had fallen to about 1.1 million by February; active users slid from near a million to under half a million.
Its billion-dollar partnership with Disney reportedly collapsed with the studio finding out about the shutdown less than an hour before the public did.
Cheap, Embedded, and Proud of It
Here's the divergence that actually matters for builders: Kling isn't winning on being the most technically dazzling model in the room. Kling 3.0 is priced at roughly $0.07 per second of generated video: about 65% cheaper than Sora and 44% cheaper than Runway. And instead of shipping a standalone consumer app that has to win attention from scratch, Kling rides inside Kuaishou's existing short-video and livestream-commerce stack, where video generation isn't the product — it's a feature bolted onto ad creation, product demos, and livestream selling that millions of merchants already use daily.
It's the Quibi problem again, just mirrored. Quibi raised $1.75 billion, hired Hollywood's best, and folded in six months because it never answered who pays and why. Sora had the same problem wearing an OpenAI badge: extraordinary output, no durable paying use case wired into a workflow anyone already had open.
Kling skipped that entire standalone-app phase and wired itself straight into a distribution channel where the video doesn't need to go viral — it just needs to close a sale.
To be fair to OpenAI, Sora's underlying model pushed real technical ground: physics consistency, coherent motion over time. But its collapse was commercial, not technical. Building the best model and building a business that survives on it turned out to be two separate jobs, and OpenAI, mid-IPO-prep and under pressure to show profitability, chose to stop paying to close that gap.
What This Means for Your Roadmap
If you're picking a video-generation vendor for a product right now, the demo reel is the least useful signal you have. Enterprises adopting AI video are mostly doing it for corporate training, onboarding, product demos, and marketing: the boring stuff, where reported production-cost cuts of 70–90% matter more than one gorgeous 10-second clip. Price per second at your actual volume, contract stability, and whether the vendor's own unit economics can survive a price war should outrank "which model looked best in the side-by-side." Ask your vendor plainly: what happens to my roadmap if you're the one that gets shut down mid-quarter?
Sora's Disney partners found out the hard way that the answer can be "almost nothing, less than an hour's notice."
It's not that Sora lost and Kling won on capability. It's that the market just repriced what winning even means in this category — and the repricing punished spectacle and rewarded plumbing.
The harder question isn't whether Kling's number is real; WSJ and SCMP sourcing suggests it is. It's whether an $18 billion valuation built on being the cheapest option survives the scrutiny of an actual public listing, and whether Kuaishou, still holding roughly two-thirds of the entity, is spinning Kling out because it's a rocket ship, or because it needs outside capital to keep funding a price war Kuaishou can no longer carry alone on its own balance sheet. If you're building on any of these video models today, that's your real due-diligence question, not whether the output looks good in a demo.
Get this in your inbox. AI Rundown Daily delivers original briefings every morning — free. Subscribe →
Frequently Asked Questions
The scale is corroborated by multiple outlets: The Wall Street Journal and the South China Morning Post both report a raise landing around $2.8 billion, with some accounts putting committed capital closer to $3 billion, at an $18 billion post-money valuation. The investor syndicate — Tencent, Alibaba Cloud, Baidu, plus state-linked funds like the Zhongguancun Science City Fund and CITIC Securities — is unusually broad for a single round, which suggests genuine competitive urgency rather than a single insider propping up a number. That said, the valuation still rests on a spin-off and Hong Kong listing that haven't happened yet, so treat the $18 billion figure as a funding-round mark, not a proven public-market price.
Kling 3.0 is priced at roughly $0.07 per second of generated video, which reporting places at about 65% below Sora's pricing and 44% below Runway's. For any workflow with real volume — product demos, ad variants, training content — that gap compounds fast, and it's a bigger differentiator than most side-by-side quality comparisons. The tradeoff is that Kling's distribution and tooling are built around Kuaishou's ecosystem rather than a Western-first developer experience, so factor in integration effort, not just the per-second rate, before committing production volume to it.
The most concrete risk is timeline and structural: Kuaishou still needs to complete a nine-month asset consolidation into Beijing Kling before the clock even starts on a 12-month push to a Hong Kong listing, and either phase can slip. There's also a valuation risk — $18 billion was set in a private round built partly on being the low-cost provider, and that story has to hold up once public-market investors start asking about margins in an active price war. Finally, there's the OpenAI/Sora precedent sitting right next to it: a well-funded, high-profile AI video product can still be shut down abruptly if the economics don't work, and Disney's billion-dollar partnership evaporated with almost no warning. None of that makes Kling a bad bet, but it argues for building an abstraction layer so you're not locked to any single vendor's survival.
