China Blocks Meta's $2 Billion Manus Acquisition, Escalating AI Tech War

Beijing escalated its use of national security and export control law as an AI trade weapon on Monday, ordering Meta to unwind its $2 billion acquisition of Manus — the Singapore-registered AI agent startup whose Chinese roots made it one of the most contentious cross-border AI deals in history.

China’s National Development and Reform Commission, the country’s powerful top planning agency, announced it had determined that foreign investment in Manus was prohibited under laws and regulations governing export controls, technology imports and exports, and overseas investment. The NDRC issued no detailed explanation, offering only a brief statement that cited the legal basis for its decision.

The move concludes a months-long probe that had already taken dramatic turns: both co-founders of Manus were reported to have been banned from leaving China during the investigation, a measure typically reserved for individuals suspected of criminal conduct or material to a state security inquiry.

How Manus Became a Flashpoint

Manus burst onto the global AI scene in March 2025 as one of the first AI agents — systems capable of autonomously performing complex, multi-step tasks on a user’s behalf — to achieve mainstream traction. The startup was originally rooted in Beijing, with its founding entities registered in China, though it later established a Singapore holding structure as it sought international investment and expansion.

The product spread virally. By December 2025, just eight months after launch, Manus claimed to have crossed $100 million in annual recurring revenue — a pace it said made it the fastest-ever startup globally to achieve that milestone from zero. The announcement turned Manus into one of the hottest properties in an AI market desperately seeking agentic applications that worked outside the lab.

Meta moved swiftly. The company announced a roughly $2 billion acquisition plan in December 2025, presenting the deal as a way to accelerate its AI agent capabilities and bring a proven agentic system into its sprawling product ecosystem. For Manus’s founders and early investors, it was a spectacular outcome.

But the deal immediately attracted scrutiny from Chinese authorities. In January 2026, the Ministry of Commerce announced it would conduct an assessment of how the acquisition complied with Chinese law. Weeks later, sources reported that co-founders Xiao Hong and Ji Yichao had been barred from traveling abroad — a chilling signal about how seriously Beijing was treating the review.

China’s Stated Rationale

The NDRC’s Monday statement framed the block as a straightforward application of Chinese law. The agency’s mandate covers export controls and outbound investment rules that are designed to prevent the transfer of technology deemed strategically important from leaving Chinese control — even when the technology is nominally housed in a foreign-registered entity.

Manus’s situation is precisely the kind of case those laws were written to address: a company founded in China, built primarily by Chinese engineers, developing AI agent technology — which Beijing considers a strategic priority — that was now on the verge of being absorbed by one of America’s most powerful tech companies.

The logic is not unlike the Committee on Foreign Investment in the United States (CFIUS), which regularly blocks Chinese and other foreign acquisitions of American tech companies on national security grounds. Beijing has increasingly mirrored CFIUS-style mechanisms in its own outbound investment and acquisition review processes, applying them with growing frequency to prevent technology transfer in the opposite direction.

Notably, the block covers Manus specifically. But the precedent it sets is broader: any Chinese-founded AI startup, regardless of where it is incorporated, may now face similar scrutiny if a US buyer comes calling.

The Messy Reality of Unwinding

The NDRC’s order creates a practical nightmare for Meta. Unlike a deal that was merely announced and then blocked, Meta had moved quickly after the December announcement to integrate Manus into its operations. According to reports, Manus executives had already formally joined Meta, and the technical teams had begun combining systems and infrastructure.

Unwinding that level of integration — untangling code repositories, data access agreements, employment contracts, and organizational structures — is a significant operational undertaking. The fate of Manus’s roughly $100 million-ARR business, its employees, and its technology pipeline is now uncertain.

Meta issued a carefully worded statement saying the transaction “complied fully with applicable law” and that it anticipates “an appropriate resolution to the inquiry.” That language suggests Meta may explore legal challenges or seek a negotiated outcome, but the NDRC’s order leaves little obvious room for maneuver.

Manus’s co-founders, meanwhile, remain under an apparent travel ban — a detail that raises questions about whether Chinese authorities view the executives themselves as potential targets beyond the scope of the acquisition review.

A Warning Shot to the AI Acquisition Market

The Manus decision has immediate implications for the M&A landscape in AI.

A significant subset of the world’s most promising AI startups have Chinese founders, Chinese engineering talent, or Chinese technical heritage — even when they are incorporated in Singapore, the Cayman Islands, or Delaware. Until now, acquirers largely assumed that structuring a deal through a non-Chinese holding company was sufficient legal insulation from Beijing’s regulatory reach. Monday’s decision suggests that assumption was wrong.

For strategic buyers at American tech giants — Google, Microsoft, Amazon, OpenAI, Salesforce — the calculus around acquiring Chinese-founded AI startups has fundamentally shifted. Deals that previously seemed viable may now require extended legal analysis of Beijing’s technology export frameworks. Some may not be pursued at all.

For the startups themselves, the chilling effect is significant. Chinese AI founders who have relocated to Singapore or other jurisdictions partly to make themselves more acquirable by US buyers now face uncertainty about whether that structural choice provides the protection they assumed.

US-China AI Competition Intensifies

The Manus block is the latest salvo in what has become an increasingly explicit battle for AI supremacy between the United States and China. Both governments have framed frontier AI as a national security priority, and both have been tightening the mechanisms by which they control the flow of AI technology across their borders.

The US has restricted AI chip exports to China — targeting Nvidia hardware and advanced packaging — and has used CFIUS to block or condition Chinese investment in American AI companies. China has responded with its own export control framework targeting critical minerals, battery technology, and now, increasingly, software and AI systems.

What makes the Manus case distinctive is that it targets an outbound flow of AI capability: not physical hardware, not raw materials, but the agentic AI intellectual property and talent that Beijing has decided should not become part of Meta’s empire.

The broader question now is whether Monday’s decision represents a one-off use of an existing legal tool, or the beginning of a more systematic Chinese policy of blocking US tech giants from acquiring Chinese-origin AI companies. Given the pace at which Chinese AI startups are achieving product-market fit — DeepSeek, Manus, MiniMax, and others — that question will define one of the most consequential battlegrounds in the global tech industry for years to come.