Alibaba and Tencent just put real money behind a rival's video model instead of pouring more into their own. That says as much about Beijing's grip on foreign capital as it does about who's winning China's AI video race.
Alibaba and Tencent have joined a $2 billion financing round for Kuaishou's Kling AI, according to a Bloomberg report published July 2. The deal values the video generation unit at roughly $15 billion before the new money comes in, well below the $20 billion figure Kuaishou floated back in April when it first signaled plans to spin Kling out as an independent company. Baidu is also in the round, Bloomberg reported, putting three of China's largest tech companies behind a single AI video platform rather than each pushing their own competing model harder.
That's the part worth sitting with. Alibaba already has its own video generator, Wan, also known as WanX. Tencent has Hunyuan. Both are live, both are shipping updates, and both are still burning cash to keep pace with Kling. Instead of only funding their in-house teams, Alibaba and Tencent are now also funding the company beating them. That's not indecision. It's a hedge, and a fairly blunt admission of where the market has already settled.
Kling has earned that position with numbers, not hype. The unit posted an annualized revenue run rate near $500 million as of this spring, roughly double what it was before Chinese New Year, according to prior reporting cited by Startup Fortune. First-quarter 2026 revenue came in above 650 million yuan, about $96 million, with 75% of that coming from customers outside China. Kling crossed $100 million in annualized revenue about ten months after its June 2024 launch, a pace that took most enterprise software companies five to eight years to match.
The valuation drop from $20 billion to roughly $15 billion isn't a sign Kling is struggling. It reflects who was supposed to write the check. Bloomberg reported in June that General Atlantic, the US private equity firm behind early bets on Meta and Uber, was in talks to lead a $2 billion round for Kling at an $18 billion valuation. That deal became politically loaded fast. Beijing had already ordered Meta in April to unwind its own $2 billion acquisition of the Chinese AI startup Manus, citing concerns about ceding key technology to a foreign rival. With that precedent sitting fresh, a US firm leading a round for one of China's most valuable AI assets was never going to be simple, and Chinese regulators have separately told domestic AI companies to seek clearance before taking American capital at all. Alibaba, Tencent and Baidu stepping in reads less like enthusiasm and more like insurance: keep the money, and the control, at home.
Kling isn't winning by default either. On the Artificial Analysis video benchmark, Chinese models now hold seven of the top eight spots, with ByteDance's Seedance, Shengshu AI's Vidu Q3 Pro, Kling 3.0 and Alibaba's own Wan 2.6 all ranked ahead of most Western alternatives. OpenAI's Sora, which charged roughly $20 a month for basic access, has been undercut by Chinese competitors offering comparable or better output at API pricing 10 to 30 times cheaper. Sora has since been discontinued as a standalone consumer product. Runway's Gen-4.5, built in the US, still holds the top overall spot, but it's an outlier in a field China now dominates on both quality and price.
That pricing gap is the real pressure point for Google and OpenAI. Kling isn't just winning users in China. It's built an international base across the US, Europe and Japan, and three quarters of its first-quarter revenue already came from outside its home market. If Alibaba, Tencent and Baidu are willing to fund a single consolidated leader instead of three fragmented competitors, Kling gets the capital to keep undercutting Western pricing while also improving quality, which is a much harder combination for Google's Veo or any OpenAI successor to answer with margin intact.
Kuaishou still plans to take Kling public in Hong Kong, with a filing possibly coming as early as 2027. Whether that IPO lands at $15 billion, $18 billion or something else will depend on how the rest of this round fills out, and on whether General Atlantic or any other foreign investor ends up with a seat at the table at all. For now, the clearest signal isn't the price. It's who signed the check.
Also read: Jersey Mike's Mentioned AI 22 Times In Its Billion Dollar IPO Filing • Blackstone's QTS Walks Away From Its Massive Virginia Data Center Project • Crusoe is raising billions to prove investors still believe in AI's buildout
