Meta's $2B Manus Unwind Sets New Precedent for Cross-Border AI M&A Risk

For legal professionals advising on cross-border technology transactions, the Meta-Manus unwind represents a paradigm shift in regulatory risk assessment. Beijing's first-ever use of its foreign investment security review mechanism to retroactively reverse a completed acquisition establishes that even closed deals involving Chinese-origin AI assets remain vulnerable to sovereign intervention. This precedent demands that M&A attorneys fundamentally reassess the adequacy of deal protection clauses, regulatory risk allocation provisions, and post-closing integration planning in any transaction touching strategic technology with Chinese connections.

Meta Platforms has begun systematically dismantling its $2 billion acquisition of Manus, marking the most concrete action yet in what has become a watershed moment in the escalating U.S.-China technology conflict. The operational separation — which includes cutting off Manus from Meta's internal data systems and prohibiting employees from using Manus tools — comes in direct response to Beijing's April 2026 order to reverse the deal under China's foreign investment security review mechanism. This unprecedented regulatory intervention has transformed what was celebrated as a landmark exit for Chinese AI into a cautionary tale that is already reshaping how cross-border technology transactions are structured, priced, and risk-assessed.

“

Meta Platforms has begun systematically dismantling its $2 billion acquisition of Manus, marking the most concrete action yet in what has become a watershed moment in the escalating U.S.-China technology conflict.

The forced divestiture represents the first time Chinese regulators have compelled the unwinding of a completed technology acquisition on national security grounds, invoking a legal framework that until now had been viewed primarily as a screening mechanism rather than an enforcement tool. The foreign investment security review mechanism, which China has steadily expanded in scope since its formalization in 2019, has typically been used to impose conditions on pending transactions or, in rare cases, to block them before closing. Retroactively reversing a consummated deal signals a dramatic escalation in Beijing's willingness to intervene in transactions involving what it deems strategic technology assets — and a willingness to accept the legal and commercial chaos that unwinding a completed merger entails.

The operational mechanics of the separation reveal the complexity inherent in disentangling two entities that have already integrated. Meta has been ordered to block Manus employees from accessing its internal data systems, and Meta staff have been instructed to cease using Manus tools for internal projects. This data separation is particularly fraught in the AI context, where intellectual property boundaries blur once engineers gain deep access to proprietary systems. As Matthias Hendrichs, a Singapore-based advisor to global AI firms, noted, 'once another company's engineers have been inside your stack, you can delete the repository, but you can't make them unsee what they've seen.' This observation crystallizes the core dilemma: even a technically complete operational separation cannot undo the knowledge transfer that has already occurred, raising questions about whether Beijing's stated national security concerns can actually be remedied through structural separation.

The dismantling of the Meta-Manus deal has triggered a cascade of secondary regulatory actions from Beijing that extend far beyond this single transaction. In the weeks following the April divestiture order, Chinese authorities tightened technology export controls specifically targeting cross-border AI transactions, imposed new travel restrictions on researchers and executives at private AI firms requiring government approval for international travel, and signaled that top AI companies including Moonshot AI, StepFun, and ByteDance will now require government sign-off before accepting U.S. investment. These measures collectively represent a comprehensive regulatory tightening of China's AI sector — one that fundamentally alters the risk calculus for any foreign entity seeking to acquire, invest in, or partner with Chinese-origin AI companies, regardless of where those companies are incorporated.

For Manus, the path forward remains uncertain. Reports indicate that the company's co-founders have held preliminary discussions about raising approximately $1 billion from outside investors to reclaim the startup from Meta, a move that could pave the way for a Chinese joint venture structure and an eventual listing in Hong Kong. The Hong Kong venue has seen a surge in AI listings this year, including from Chinese AI firms like MiniMax and Zhipu, making it a plausible exit route that satisfies both Beijing's desire to keep strategic assets within its regulatory orbit and investors' need for liquidity. However, raising $1 billion at a valuation that makes the buyback economically viable for all parties — particularly Meta, which paid $2 billion just months ago — presents formidable challenges in a geopolitical environment where the very assets that gave Manus its value are now subject to regulatory restrictions that limit their commercial exploitation.