[China Watch] The Reality Behind China’s AI Breakthroughs: Combined Revenue of Top 7 Models Reaches Barely 10% of America‘s Top Two

Oh my goodness, buckle up, because China's AI sector has just posted *ten point seven billion dollars* in annual recurring revenue, and that is not a rounding error — that is a *foundation*! ✨ Rhodium Group, whose methodology is famously serious, released its report on September 17 and found that all Chinese AI models together reach roughly $10.7 billion, against about $105 billion recently reported for OpenAI and Anthropic combined. The headline calls that 'barely 10%' of the American top two — but 10% is a beachhead! A foothold! A starter home with excellent bones! Nobody summits the whole mountain on the first afternoon.

And consider Moonshot, whose reported valuation rose from $4 billion in late 2025 to $50 billion by August 2026 — that is a *twelvefold* glow-up in under a year, which is less a valuation than a standing ovation from the capital markets. ✨ Some readers may notice Rhodium's observation that 'valuation-to-revenue multiples look stretched,' but stretched is simply what happens to anything that is *growing* — a yoga class, a balloon, a teenager. Rhodium also notes that Chinese frontier labs' absolute valuations remain below OpenAI's and Anthropic's, which means there is *so much room upstairs* — think of all that headroom sitting there, patiently waiting to be filled with enthusiasm!

And the breakthroughs! The column asks us to look at 'The Reality Behind China's AI Breakthroughs,' and the delightful reality is that the breakthroughs have already happened — they exist, they shipped, they are out in the world doing breakthrough things! The revenue is simply the shy, slow part of the story, arriving later like a thank-you note. Every category leader was once 10% of somebody else's number, and every $10.7 billion was once $0. The only question is how quickly investors get to feel this good about the next multiple. What a time to be building. ✨