Will Beijing meddle in the Meta-Manus deal?

This note was originally published for subscribers of Trivium Tech Daily on January 9. It has been republished here with minor changes and edits.

On December 29, Meta announced it would acquire Chinese AI startup Manus AI – reportedly for USD 2.5 billion.

Inside China, the deal was widely applauded as a canonical home run: a blockbuster exit for startup founders who built a breakout agentic AI product, and a vindication for early venture investors who placed the right bets.

But here’s the bad news: Beijing will get involved. 

On January 8, Ministry of Commerce (MofCom) spokesperson He Yadong put it plainly:

• “The Ministry of Commerce will, together with relevant departments, conduct an assessment and investigation into whether [the Meta-Manus deal] is consistent with relevant laws and regulations, including those on export controls, technology import and export, and outbound investment.”

At this point, the deal is no longer just a tech acquisition. It has morphed into a US-China geopolitical zero-sum game – similar to the forced sale of TikTok.

• To understand why Beijing is stepping in, and what this could mean for the deal’s outcome, we need to take a closer look at Manus AI’s origin story.

Manus AI’s back story

Manus AI’s parent entity, Butterfly Effect, was founded in 2022 by serial entrepreneur Xiao Hong in Beijing, with operations in Wuhan. It was originally a fully Chinese company. It was founded in China, operated domestically, and was run by Chinese founders.

• The company’s first product was Monica, an AI-powered browser extension, after which it pivoted to Manus, an agentic productivity tool for high-value knowledge work.

Launched in March 2025, Manus became an overnight success. It quickly made headlines both inside and outside China and was even dubbed the “ChatGPT of agentic AI products.” Within nine months, Manus had reached a USD 100 million annual revenue run rate, making it one of the fastest-growing AI products ever built by a Chinese team.

What set Manus AI apart from most Chinese startups was its overseas-first strategy. From day one, the company targeted international users rather than the domestic market. Xiao explained the rationale in a podcast:

• “Overseas users’ willingness to pay for software may be five times that of Chinese users, and payments can be denominated in US dollars. With an exchange rate of seven [RMB to USD], that is 5 × 7 = 35. So the market is at least 35 times bigger.”

To that end, Manus AI took a series of decisive steps that ultimately paved the way for Meta’s acquisition. These included:

• Rejecting capital from state-owned investors while raising a USD 75 million round led by famous US venture capital firm Benchmark in April
• Gutting its China operations and relocating to Singapore in June

Had Manus AI not made these early moves to sever its China ties, selling the company to Meta would likely have been far more difficult. Even then, Manus did not fully escape regulatory risk.

• Benchmark’s investment reportedly drew scrutiny under the Biden administration’s outbound investment rules – often referred to as “reverse CFIUS” – which restrict US capital flows into sensitive sectors such as AI, semiconductors, and quantum computing.
• That scrutiny, however, appears to have faded after Manus completed its relocation to Singapore. 

What pathways can Beijing take to intervene?

Considering that Manus has already moved operations outside of China, does Beijing even have any leverage left to intervene?

• Yes, yes it does.

Beijing is likely to invoke the very same legal mechanism – the Regulations on Technology Import and Export Administration – that it used to block the forced sale of TikTok, and apply it to the Meta–Manus deal.

The key thing to understand here is that Beijing is not reviewing the acquisition itself. Instead, it is scrutinizing the transfer of technology from Manus AI’s China entities to its Singapore entity as an indirect way to exert influence over the transaction.

• Under Chinese regulations, companies must obtain regulatory approval to transfer certain software technologies abroad.
• As the MofCom spokesperson indicated, Beijing could also rely on separate export control and outbound investment rules, giving it several avenues to get involved.

Beijing’s logic was best articulated by Cui Fan, a professor at the University of International Business and Economics and a close advisor to MofCom, in an article he published on January 3, one week before the ministry’s public statement. Cui laid out how such an investigation could be conducted:

• “What Chinese regulators primarily need to examine is not whether technology owned by Manus’s entities in Singapore or the Cayman Islands is being transferred to US entities, but rather when, in what manner, and which technologies Manus’s entities within China – including both natural persons and legal entities – have transferred to parties outside China.”

He also spelled out the potential consequences:

• “If it is confirmed that restricted technologies were exported without authorization, the relevant entities may be required to bear legal liability… [and] criminal liability could be pursued in accordance with the law.”

Why would Beijing bother to intervene?

So… why is Beijing worried about this particular acquisition? Why bother getting involved? After all, a USD 2.5 billion deal is not a blockbuster, and there’s no strategic technology issue here.

• Despite its phenomenal product success, Manus AI arguably does not possess core technology IP that China is worried it will lose control over.
• The company does not train its own foundational models, it builds applications on top of existing models.

The most obvious answer is that Beijing may be concerned that more Chinese startup founders will attempt to replicate the Manus playbook. As noted earlier, Manus AI has been widely held up as a canonical success story within China’s venture ecosystem. If relocating overseas becomes a trend among Chinese startups, Beijing is likely worried that a broader exodus of AI talent and ecosystem activity could follow.

There is another possible angle here: Beijing may see this acquisition as an outgrowth of US outbound investment rules, and may be reluctant to play ball.

This issue was raised by Professor Cui in his commentary on the deal. He said:

• “Manus’ step-by-step divestment of its Chinese connections was undoubtedly driven by US regulations restricting investment in China, [and the fact that] those restrictions could produce such an effect was not something I had anticipated.”

In other words, Beijing may now view Manus’s exit from China as a direct consequence of the reverse CFIUS rules and is realizing that it may have underestimated the impact of these rules.

• Reverse CFIUS does not just constrain China’s venture funding; it may also trigger a wave of AI startup outflow.
• As a result, the issue has become geopolitical. Beijing will want to prevent a clear US strategic win – and to signal to the startup community that global expansion cannot come at the expense of China’s strategic interests.

So, what happens next?

We doubt Beijing knows exactly how it will respond. Beijing will certainly want to send the message that Chinese-born tech companies must honor certain responsibilities to the state and the Chinese people. Incidentally, that’s the same message that Beijing sent to platform companies during the 2020-2022 tech crackdown.

That said, Beijing also faces a real risk of overplaying its hand. Punishing success would backfire badly by undermining confidence and distorting incentives. Manus AI represents a rare home run for both Chinese startup founders and venture capital investors. Penalizing the company would not only discourage entrepreneurs who see Manus as a role model, but also further damage China’s already struggling venture capital industry. That, in turn, would make investors even more reluctant to deploy capital – precisely the opposite of what Chinese policymakers have been trying to achieve for years.

It is also worth noting that China’s policy goal has never been to keep its companies confined to the domestic market. On the contrary, policymakers want Chinese firms to compete globally, build international brands, and capture overseas demand. Undermining successful global expansion after the fact would be counterproductive, weakening incentives for innovation rather than strengthening them.

• But there’s a difference between competing internationally and renouncing one’s corporate citizenship entirely.

Taking a step back from geopolitics, we must remember that the fundamental reason Manus AI wanted to move abroad was actually not US investment restrictions, but weak domestic demand – as Xiao himself explained in the podcast mentioned above. In a parallel universe where China’s AI market was as monetizable as the US market, companies like Manus might never have needed to go through the trouble of relocating to Singapore. This points to a deeper structural weakness in China’s economy – something that geopolitical maneuvering can’t fix. Only domestic reforms can.

Chinese policymakers appear to be aware of this. Cui himself acknowledged as much in his article:

• “Relevant authorities should also take this acquisition as an opportunity to recognize that certain fundamental problems in China’s economy urgently need to be addressed, [namely that] China has strong supply but weak demand.”
• “[We] should encourage Chinese enterprises to break through, go global, and serve international markets.”

But Cui added an important caveat:

• “However, we also should acknowledge that, in the face of a complex international environment and various constraints on China’s economic development – especially the intense and complicated competition in high-tech sectors – the Chinese government needs to exercise appropriate oversight over enterprises’ overseas expansion.”

The Manus deal, therefore, is likely still quite negotiable – much like the TikTok deal was.

• How that negotiation plays out is anyone’s guess at this time, but as policymakers assess the deal’s impact, there’s a window for stakeholders to influence the outcome. 

Future implications

It is also very possible that oversight of technology transfers abroad will tighten going forward, driven by Beijing’s concerns about potential AI talent outflows and the loss of core AI capabilities without government visibility.

If that happens, it could open a can of worms, raising questions like:

• Would hosting a Chinese-developed model in an overseas data center count as an export of technology? 
• Would fine-tuning or updating models abroad trigger approval requirements? 
• What about Chinese companies contributing to overseas open-source AI projects?

The danger is not just regulatory overreach, but uncertainty. If companies can’t clearly determine what constitutes a prohibited “technology transfer,” risk-taking will slow, compliance costs will rise, and founders may become overly cautious or even opt for preemptive relocation if the environment no longer feels conducive to building a company. In that scenario, tighter controls meant to prevent an exodus could end up accelerating it.

This is the tightrope Beijing now has to walk: asserting control without choking off the very innovation it wants to keep at home.