AI Video Split in Two: China Just Put $18 Billion Into a Champion; America's Leader Made $2 Million in Its Whole Life and Died.
The clearest picture of the US-China AI split isn't a benchmark — it's two numbers from one season. Kuaishou spun its Kling video model into a standalone company and raised ~$3B at ~$18B, with rivals Alibaba, Tencent, and Baidu plus Abu Dhabi money all backing it, on ~$500M ARR. The same stretch, OpenAI's Sora shut down: $1M/day to run, $2.1M lifetime revenue. China capitalized a champion; America buried its leader. The caveats: Kling's profitability is unproven (~36x ARR, proceeds to compute), and a China-plus-Gulf champion may never clear Western market access — so video may harden into two blocs.
The clearest picture of how AI is diverging between the US and China this year isn't a benchmark score. It's two numbers from the same season. Kuaishou just spun its Kling video model into a standalone company and raised roughly $3 billion at an $18 billion valuation — the largest dedicated capital event in AI video ever — with rivals Alibaba, Tencent, and Baidu all writing checks, plus Abu Dhabi money. Over the same stretch, OpenAI's Sora, the product that defined AI video in the West, shut down: about $1 million a day to run, $2.1 million in lifetime revenue, a $1 billion Disney deal collapsed, fewer than 500,000 users. China capitalized a champion. America buried its leader. Same technology, opposite outcomes.
This isn't a quality gap; it's a model gap, in both senses of the word. The two systems bet on different commercialization models and different capital models, and in video — the most compute-hungry consumer-AI category there is — China's worked and America's broke. That's the story, and it's the clearest case yet of Chinese consumer-AI commercialization beating the US outright rather than merely matching its leaderboard numbers.
Look at the East side first: consolidate and capitalize. Kling is running roughly $500 million in annual recurring revenue, up about fourfold in a year, and it just got funded by its own rivals and a Gulf sovereign fund, with Kuaishou keeping 68.33% control and a Hong Kong IPO teed up inside twelve months. And it isn't alone — ByteDance's Doubao and Seedance operate at genuine consumer scale (on the order of 180 trillion daily token calls) at roughly a tenth of US pricing. China has multiple funded video contenders with built-in distribution through Kuaishou and Douyin and e-commerce to monetize them, and it's converging that crowded field into fundable champions.
Now the West: exit and commoditize. Sora is dead on consumer unit economics, and the survivors are fighting on price rather than building war chests. Grok Imagine undercuts Sora by 86% at $4.20 a minute — pricing quietly subsidized by an Anthropic compute deal, which is its own circular-financing tell. Google's Gemini Omni has the one genuine Western structural advantage worth naming: YouTube Shorts distribution. But no Western pure-play video company has anything close to Kling's $18 billion war chest. The West is commoditizing the exact tier China is capitalizing.
The sharpest part is the investor list, because rivals funding a rival isn't ordinary venture capital. Alibaba, Tencent, and Baidu bankrolling a competitor's spinoff is three things at once: strategic optionality (buy a stake in the likely winner of a crowded field so you can't be locked out of it), a shared bet against US video models, and a forced consolidation of China's fragmented video sector into one fundable national champion. Add Abu Dhabi's BlueFive and you get a China-plus-Gulf capital axis — the same Big-Tech-plus-Gulf template that funded DeepSeek. This is what industrial coordination looks like when it's wearing a cap table instead of a five-year plan.
Two things keep this from being a clean "China won," though. First, Kling's profitability is unproven: $18 billion is about 36 times ARR, no margins have been disclosed, and the IPO proceeds are earmarked for compute and data-center buildout — which means the cost side that killed Sora, that $1 million a day, is precisely what Kling is about to spend into at scale. This is a growth-and-geopolitics valuation, not a proven-profit one. Second, distribution and access: Google's YouTube Shorts is a Western structural moat no standalone Kling has over here, and it may never get to try, because a Chinese video model with Gulf sovereign backing runs straight into the same wall DeepSeek hit with US bans. The divergence is in capital and commercialization model, not settled dominance.
Which makes the real question the bloc question. Watch three things: whether Kling's Hong Kong IPO prices near $18 billion or the growth-versus-profit gap finally bites; whether a China-plus-Gulf video champion can sell into the West at all, or whether AI video hardens into two non-overlapping internets — Chinese-Gulf on one side, US-distribution-embedded on the other; and whether any-to-any, audio-native generation (Gemini Omni, NVIDIA's Cosmos) resets the quality race regardless of who has the capital. My read: China has demonstrably solved the thing the US couldn't — a consumer-AI-video business that patient, coordinated capital will fund all the way to an IPO — while the US retains the thing China hasn't proven it has, which is a path to global distribution without a bloc problem. The likeliest outcome isn't one side winning outright; it's two AI-video markets that don't touch. What would change my mind is Kling's IPO pricing at $18 billion on real margins, or a China-plus-Gulf champion actually clearing Western market access. Until then, the tell of the season is simple: the same product category produced a record fundraise in Beijing and a shutdown notice in San Francisco.