Late in 2025 it was announced that Meta would be paying a princely sum – ‘billions,’ though never reported exactly – to acquire Chinese AI startup, Manus. That deal has just fallen through. China put a block the Meta AI deal in a move that looks a lot like the state tightening up control over strategic tech.
The deal would have given Meta access to Manus’s advanced AI agent technology – software capable of autonomously carrying out complex tasks such as research, planning, and analysis.
China’s top regulators barred the deal on national security grounds, arguing that foreign ownership of such technology risked transferring valuable intellectual property and talent abroad at a time of intensifying US–China tech competition.
And it’s not just a simple rejection. Authorities have instructed both parties to reverse the acquisition entirely, potentially requiring the removal of any integrated technology and the restoration of Manus’s assets to their original state.
Manus was founded in China but later moved to Singapore where U.S. restrictions on Chinese tech wouldn’t interfere with business. Singapore also put the company outside of Beijings reach… or so they hoped.
Chinese authorities rejected that framing, applying a ‘substance over form’ approach that treats Chinese-founded technology as subject to domestic oversight regardless of where it is registered.
The case also involved direct intervention in staff. Key executives linked to Manus were reportedly prevented from leaving China during the investigation – proof that Chinese regulators are not messing around when it comes to cross-border AI transactions.
The Dao view: China putting a block on the Meta AI deal is just one piece of an AI puzzle
This certainly complicates Meta’s push toward AI-driven products. Manus’ tech would have given Meta tech a very practical edge. But the real news here is what’s happening at the Chinese end.
In the same week that Chinese regulators blocked the Manus deal, Chinese courts ruled that companies cannot fire employees simply to replace them with artificial intelligence. The message is that while AI is advancing quickly, the boundaries around how it is deployed – and who gets to control it – are being drawn just as deliberately.
