China orders Meta to unwind $2 billion Manus acquisition after review

China’s order for Meta Platforms to unwind its $2 billion-plus purchase of Manus turned a routine-looking AI acquisition into a warning shot for every software company chasing talent, code, and partnerships across borders. The deal, announced in December 2025 and valued in the $2 billion to $3 billion range, is now being rolled back after regulators treated it as a sensitive transfer of AI assets and intellectual property.

That sensitivity only grew because Manus had already shut its China offices and moved operations to Singapore. Chinese authorities also barred co-founders Xiao Hong and Ji Yichao from leaving China during the review, showing how aggressively regulators can use travel restrictions and licensing pressure to shape a transaction. For deal teams, the message is clear: relocating a startup does not necessarily remove its Chinese exposure if the founders, IP, or talent still tie back to the country.

The review began in January 2026, when China started examining whether the Meta-Manus transaction complied with export control laws. Meta said in March that the deal “complied fully with applicable law” and that it expected an appropriate resolution. That did not stop China’s move to unwind the acquisition on April 27, a rare example of a completed tech deal being blocked after the fact. The White House said it would continue defending America’s technology sector against foreign interference, underscoring how quickly AI dealmaking has become a policy fight as much as a business one.

For monday.com, the practical lesson sits closer to the product than to geopolitics. AI strategy is now inseparable from where data is stored, who can access it, and which jurisdictions can touch the work. That matters to product, security, legal, and enterprise sales teams because customers in Europe, the Middle East, Asia, and regulated industries are increasingly asking whether AI tools respect regional controls before they buy. The Meta-Manus case also raises the stakes for AI hiring: companies are no longer just recruiting engineers and researchers, they are evaluating where those people can work, what data they can touch, and whether an acquisition will survive scrutiny from multiple governments.

monday.com has already staked out a governance-first position. Its AI Trust Center says AI follows the same data residency policies as each account, processes and stores data within the designated region, and stays within existing account permissions. The company also says it does not use customer data or content to train its models. Its trust center lists enterprise frameworks including SOC 2 Type 2, ISO 27001:2022, GDPR, HIPAA, and the EU-US Data Privacy Framework, and its 2025 AI roadmap introduced monday agents with an emphasis on enterprise-grade governance.

That is the direction the market is moving in. Cross-border AI deals are getting harder, AI talent is becoming more strategic, and software companies that want global scale will need cleaner answers on privacy, compliance, and acquisition risk before they move on the next startup.