Why China Blocked Meta’s Manus Deal: When AI Agents Become a Security Issue

What Meta was really trying to buy

Manus has been described as a general AI agent capable of automating complex tasks, including S&P 500 analysis and drafting sales pitches. That matters because an AI agent is not just another chatbot feature. In commercial terms, it can become a tool for carrying out multi-step work across consumer and business products.

For Meta, the value of the deal was not only the startup itself. BBC reported that Meta had said Manus’s agents would be used to enhance AI across its platforms; CNBC said Meta planned to integrate advanced automation into consumer and enterprise products, including the Meta AI assistant; and TechCrunch reported that Meta planned to fold Manus’s agent technology directly into Meta AI.

The reported price tag has varied. BBC put the deal at about $2 billion, Reuters described it as more than $2 billion, TechCrunch gave a range of roughly $2 billion to $3 billion, and Foreign Policy referred to a $2.5 billion figure. So it is fair to call this a $2 billion-plus transaction, but not to treat any single valuation as settled.

How Chinese regulators stepped in

The intervention did not come out of nowhere. CNBC reported that China’s Ministry of Commerce said in January 2026 it would assess whether the acquisition complied with rules on export controls, technology import and export, and overseas investment; a Meta spokesperson told CNBC the transaction “complied fully with applicable law.”

Later, BBC and The Guardian reported that China’s National Development and Reform Commission, or NDRC — the country’s powerful state planning and economic policy agency — prohibited foreign investment in the deal and required the parties to withdraw the acquisition. Reuters, citing The Wall Street Journal, reported that Meta was preparing to unwind the acquisition after China blocked it on national security grounds. Bloomberg described the move as China pressing Meta to cancel a deal that had already been completed months earlier, underscoring Beijing’s effort to extend its influence over business transactions beyond its borders.

Why an AI agent became a national security concern

1. Meta could scale the technology quickly

The sensitivity of Manus lies in the kind of capability it reportedly offered: automated agent technology that could be integrated into Meta AI and into Meta’s consumer and business products. Once a capability like that is absorbed by a platform the size of Meta, it is no longer just the asset of a small startup. It can become part of a product layer reaching huge numbers of users and enterprise customers.

Public reports do not disclose the full internal reasoning of Chinese regulators. But the sequence is telling: first, scrutiny around export controls, technology transfer and overseas investment; then, a reported block on national security grounds. That suggests the technology itself was central to the review.

2. A Singapore headquarters did not remove the China link

Manus’s corporate location did not take the deal outside Beijing’s field of vision. TechCrunch reported that Manus was founded in 2022 and moved its headquarters from China to Singapore around mid-2025; Foreign Policy described it as an AI company founded by Chinese entrepreneurs and based in Singapore.

Yahoo Finance framed the restrictions as reflecting concern that Chinese-founded startups seeking global opportunities could lead to leakage of homegrown technology abroad, while also preventing U.S. investors from entering sensitive sectors treated as national security priorities. In other words, for sensitive AI businesses, regulators may look beyond the current headquarters. Founder background, technology origins and the path of cross-border capital can all matter.

3. The buyer was Meta

The identity of the acquirer was a decisive part of the story. Reuters described Meta as a U.S. tech major, while CNBC’s analysis said China’s determination to stop the technology from being absorbed became clear once Meta emerged as the buyer.

CNBC also reported that, according to multiple reports, the decision to block the acquisition was elevated beyond economic regulators to China’s National Security Commission, the Communist Party body chaired by Xi Jinping that oversees national security strategy. The Guardian reported that Beijing said domestic tech companies must seek explicit government approval before accepting U.S. investment. Taken together, those reports point to a broader rule of thumb: in the context of U.S.-China AI competition, a major American platform buying China-linked AI capabilities is far more likely to trigger security scrutiny than an ordinary financial investment.

What it means for Meta and AI startups

For Meta, the immediate question is practical: how to undo or restructure a deal that had already been announced. Reuters reported that Meta was preparing to unwind the acquisition, while Fortune noted that it remained unclear how Meta could actually do so. TechCrunch said losing access to Manus’s agent technology could be a serious setback for Meta’s ambitions in the fast-moving AI agents market.

For AI founders and investors, the signal is wider. A transaction involving China-linked founders, strategic AI capabilities, cross-border financing and a U.S. buyer may be reviewed through a national security and technology-leakage lens. That does not mean every AI company with Chinese roots is unable to raise overseas capital or sell to a foreign buyer. It does mean that in sensitive technology sectors, the headquarters address is only one variable. Regulators may also scrutinize intellectual property routes, team origins, financing structures and the ultimate destination of the technology.

What remains unclear

Several important details are still unresolved. First, the valuation is not uniform across public reports: estimates range from about $2 billion, to more than $2 billion, to roughly $2 billion to $3 billion, with Foreign Policy citing $2.5 billion. Second, if the acquisition is unwound, the mechanics are still uncertain; Fortune specifically noted that it was not clear how Meta could unwind the deal.

The safest conclusion is that China did not block Meta’s Manus acquisition simply because it was a costly foreign takeover. The deal brought together four elements that now sit at the heart of tech geopolitics: AI agent capabilities, China-linked founders and technology, cross-border capital flows, and a U.S. platform buyer with global scale. For future AI acquisitions, incorporation in another country may not be enough. The deeper questions will be where the technology came from, who built it, who funded it and who ultimately controls it.