China’s AI Ban Didn’t Kill This Startup. It Handed It to Tencent at a Discount.

You’ve probably been watching the AI wars thinking it’s a race to build the smartest brain. You’re looking at the wrong battlefield.

The story of Manus is the most cinematic saga in tech right now. Three founders build an AI agent in Singapore. Mark Zuckerberg loves it so much he buys it for $2 billion in just ten days. Beijing steps in, blocks the deal on national security grounds, and leaves the startup for dead. Then, Tencent swoops in to buy it back. It’s a modern epic of geopolitics, Silicon Valley hubris, and second chances.

But if you look closely at the $2 billion price tag, you’ll realize the most shocking part of this story: Manus isn’t actually a technology company.

In the age of AI, the algorithm is cheap. The user’s daily habit is the actual crown jewel.

Manus doesn’t have a secret, proprietary AI model. It runs on third-party APIs. Its core technology—optimizing context windows, using file systems for memory, learning from failed attempts—is impressive engineering, but it’s entirely replicable. Open-source competitors are already copying it. If you stripped away the hype, a pure tech analyst would say Manus has zero moat.

So why did Zuckerberg pay a 20x revenue premium? Why is Tencent fighting to lead a buyout right now?

Because Manus achieved something infinitely harder than building a model: it built a product moat. It reached $100 million in annual recurring revenue in just eight months. By the time regulators killed the Meta deal, its ARR had ballooned to $500 million. Millions of users weren’t just chatting with Manus; they were letting it autonomously execute complex, multi-step workflows. They were locked in.

When Beijing blocked the Meta acquisition in April 2026, the tech world gasped. They thought regulators had saved a Chinese startup from being swallowed by an American giant. But the reality is far more cynical.

The regulators didn’t protect Chinese technology; they subsidized Tencent’s shopping list.

Tencent has been pouring billions into AI infrastructure, poaching top talent from OpenAI, and building massive computing clusters. Yet, despite having the models and the money, they had a glaring hole in their ecosystem. They had AI tools tied to WeChat and enterprise software, but no standalone, globally capable, cross-platform AI agent. Manus filled that exact gap.

And here is the brilliant, accidental arbitrage of the whole situation: Manus is now being bought back at the original $2 billion valuation. But its revenue has quintupled since Meta first made the offer. By standard SaaS metrics, Manus should be valued much higher today. Tencent is getting a validated, high-growth asset at a massive, regulator-mandated discount.

This isn’t just a quirky startup drama. It’s a blueprint for the next decade of technology. The underlying AI models are becoming a commodity. The real war is over distribution, workflow lock-in, and regulatory positioning. Meta proved Manus had global product-market fit. Beijing forced the discount. Tencent is reaping the reward.

The next great tech war won’t be fought over who builds the smartest AI, but over who owns the daily habits of the humans using it.

FAQ

Q: If Manus doesn't have proprietary AI models, isn't it just a shell company?

A: No, it's an interface company. Building a foundational AI model is a math and compute problem; building a product that millions of people trust to autonomously execute their work is a human behavior problem. The tech is easily cloned, but the user habit is not.

Q: Why is Tencent buying it back at the exact same $2 billion valuation?

A: Because it's a steal. When Meta offered $2 billion, Manus had $100 million in ARR. By the time the deal was blocked and restructured, its ARR hit $500 million. Tencent is buying a hyper-growth asset at a 75% discount thanks to a regulator-mandated reset.

Q: Did Chinese regulators intentionally orchestrate this to help Tencent?

A: Probably not directly, but the result is the same. The ban was framed as protecting national security, but practically, it prevented a foreign giant from absorbing a domestic success story and handed a discounted, globally validated asset to a local titan.

📎 Source: View Source