China Blocks Meta’s $2B Acquisition of Manus AI in Landmark Security Move

As the Sino-American technological rivalry intensifies within the sphere of artificial intelligence, the Chinese government’s oversight of core technical exfiltration has reached an unprecedented climax. According to the latest intelligence, the National Development and Reform Commission (NDRC) of the People’s Republic of China formally invoked the “Measures for the Security Review of Foreign Investment” on April 27, summarily halting Meta’s projected $2 billion acquisition of the prominent AI agent startup Manus AI and mandating a complete dissolution of the agreement.

This landmark intervention represents the first public prohibition of a foreign AI acquisition since the review mechanism’s inception in 2021. It serves as a stern communiqué to global capital markets: even if Chinese startups attempt to circumvent scrutiny by relocating their corporate headquarters to Singapore, they cannot bypass Beijing’s regulatory “red lines” so long as their technical core originates within China.

Founded in Beijing in 2022 by Xiao Hong and Ji Yichao under the parent entity “Butterfly Effect,” the firm introduced Manus AI in March 2025, heralded as the world’s inaugural general-purpose AI agent service. Its capacity to autonomously execute intricate tasks, such as programming and analytical research, swiftly garnered the favor of Silicon Valley investors, culminating in a $75 million Series B funding round led by Benchmark in April 2025.

Buoyed by this capital infusion, Manus AI embarked upon a radical “de-sinicization” strategy. In June 2025, the company officially transferred its headquarters to Singapore, drastically reducing its Beijing workforce by nearly two-thirds—retaining only 40 core personnel for the relocation—while simultaneously purging its Chinese social media presence and obstructing IP connections from mainland China. This maneuver, derisively labeled “Singaporean origin-washing” within the industry, initially appeared successful, leading to Meta’s multi-billion dollar acquisition offer by late 2025.

However, altering a corporate registry cannot erase the historical trajectory of technical research and development. Beginning in early 2026, the iron fist of Chinese regulation descended. The Ministry of Commerce first intervened to evaluate export controls, followed by the NDRC directly summoning executives for inquiries in March and imposing exit bans on the two founders, who had already taken up residence in Singapore. Ultimately, on April 27, the NDRC issued a laconic statement devoid of specific rationale, definitively declaring the transaction “prohibited.”

Analysts emphasize that Beijing’s scrutiny transcends geographical registration; the authorities are primarily concerned with Manus AI’s colossal data assets—encompassing 147 trillion tokens—and the core engineering prowess established prior to its relocation. These assets are deemed state strategic treasures, the clandestine flow of which to American tech titans via overseas acquisitions is strictly forbidden.

The dissolution of this merger is a profound setback for Meta. Since late 2025, the corporation has aggressively pursued large-scale acquisitions to fortify its position in the “AI Agent” domain against rivals like OpenAI and Google. Manus AI was envisioned as the cornerstone of Meta’s premium AI service offerings, with founder Xiao Hong slated to assume a Vice Presidency.

Now, with the transaction terminated, Meta faces a complex legal quagmire. Although the company’s official statement claims “full compliance with applicable regulations” and an expectation of “proper resolution,” the question of how to legally and technically bifurcate the “Manus from Meta” entity—already listed as a Singaporean affiliate on the App Store—remains a challenge without precedent.