China has ordered Meta to unwind its $2 billion acquisition of Manus, the Chinese AI agent startup that drew widespread attention earlier this year for demos showing an AI completing complex, multi-step tasks with minimal human direction, according to TechCrunch's report.
Chinese regulators spent months reviewing the deal before blocking it. The decision is a direct setback for Meta's AI agents push - the company had positioned Manus as a key acquisition that would accelerate its own capabilities in this area.
AI agents are software that can take sequences of actions autonomously: browsing the web, writing and running code, completing multi-step tasks without a human approving every step. Manus's demos showed that working at a level that outpaced most comparable Western products when the acquisition was announced.
China's decision follows a broader pattern. Beijing has grown more protective of AI capabilities it considers strategically important, and Manus - with its agent architecture and training approach - apparently qualified. The review wasn't quick; months of scrutiny preceded the final block, suggesting this wasn't a routine rejection.
Meta doesn't lose the concept of AI agents. It loses the team, the proprietary methods, and months of integration work. The company has its own AI research division and the resources to pursue agents independently - it just has to do it without this acquisition, on a longer timeline.
For any US company eyeing Chinese AI acquisitions, the regulatory window for deals like this appears to be narrowing. The combination of export controls moving in one direction and China's outbound acquisition reviews tightening in the other is making cross-border AI M&A increasingly difficult to close.