China demands Meta unwind its Manus acquisition, raising pressure on global AI deals

China has demanded that Meta reconsider its already closed $2 billion acquisition of AI startup Manus, and it looks like a new level of Beijing's interference in global tech deals. Chinese authorities have mostly relied on domestic restrictions, antitrust reviews and control over the local market. Now they are trying to roll back an acquisition that was agreed and completed outside Chinese jurisdiction. For the international market, the signal is clear: in the race for AI assets, even a formally completed deal may not be final.

Beijing has spent years trying to influence corporate decisions far beyond its borders, especially when sensitive technologies, data, chips, infrastructure or strategic developer teams are involved. But the Manus case stands out even against that background. In effect, China is showing that it is ready to use not only its control over its own companies and export chains, but also its political weight to change the balance in the international competition over artificial intelligence.

For Meta this is especially painful. The company is already trying to catch up with the leaders in generative AI, and judging by the size of the deal, the Manus purchase was seen as a way to quickly strengthen its team, models and research expertise. The $2 billion price tag shows this was not an early experiment but an important strategic asset. In the current AI cycle, large platforms are buying not just products but access to talent, computing know-how, data, patents and the speed of shipping new services.

So an attempt to reverse such a deal hits not only a specific buyer but the very logic of mergers and acquisitions in the tech market. The greater the uncertainty after a deal closes, the more expensive cross-border acquisitions become, the longer legal reviews run, and the steeper the discount on any asset with a politically sensitive link to China.

The situation is also telling for other American and European tech companies. If the precedent holds, buyers will have to account in advance not only for regulators in the US, the EU and the target's country of registration, but also for Beijing's position, even when the deal formally takes place outside its territory. That redraws the map of risks. Boards will have to look more closely at where key engineers come from, how much startups depend on Chinese supply chains, whether they have research centers in China, and any legal hooks through which Chinese authorities could slow integration. In the AI sector, where speed is decisive, such a delay can cost a lost year and a strategic setback.

It is an unpleasant turn for startups and their investors too. Usually, once a deal is announced and closed, attention shifts to integrating the team, retaining key employees and launching a joint product strategy. If the risk of outside political interference remains, founders and funds face a new discount: the money may be agreed and the documents signed, yet the exit can still be left hanging. This matters especially for AI companies built around small research teams. A few months of uncertainty can break up a team, derail hiring and strip the buyer of the very value it paid a premium for.

What this means. The Meta–Manus case shows that the fight for control over artificial intelligence looks less and less like ordinary competition between companies and more and more like geopolitical rivalry between states. The market is getting a hard lesson: buying an AI startup now depends not only on price, synergies and shareholder approval, but also on which states consider the technology, the people or the infrastructure part of their strategic sphere of influence. If this behavior becomes the norm, international AI deals will close more slowly, cost more and fall apart more often after signing.

*Meta has been designated an extremist organization and is banned in Russia.