Regulatory  ·  2026-06-02

China State Council Publishes Sweeping AI Outbound-Investment Rules — Technology-Tracing Framework Targets Cross-Border AI Deals Regardless of Corporate Domicile

RegulatoryHigh impactChina
China's State Council published new regulations on June 1, 2026 that take effect July 1, 2026, codifying and expanding the legal framework the NDRC used to force Meta to unwind its $2 billion acquisition of AI-agent startup Manus. The rules require prior authorisation for cross-border transfers of restricted technologies, goods, services, and related data — and, critically, assert Chinese jurisdiction based on where technology was developed and where key personnel built their expertise, not merely where the acquiring or target company is currently incorporated.
The 'Singapore-washing' or Cayman-Islands restructuring playbook — found a company in China, relocate offshore, sell to a US buyer — no longer reliably shields cross-border AI acquisitions from Chinese regulatory review. The rules also ban unauthorised cross-border talent transfers in sensitive sectors, include retaliation authority (blocking foreign entities from trading with China if their home country restricts Chinese investment), and establish a legal basis for unwinding already-completed transactions. AI is explicitly identified as a sensitive, national-security-critical sector.
Clients active in APAC M&A or with Chinese-origin AI IP in their supply chain should immediately engage legal counsel to map any pending or recently completed transactions against the new framework. Clients acquiring AI startups that relocated from China to Singapore or another offshore domicile within the last five years should specifically re-evaluate deal structure before July 1.
Sources
The Next Web — China formalises tougher outbound-investment rules after the Meta-Manus blockadeReuters — China toughens rules on outbound investment after Meta-Manus contentionNikkei Asia — China tightens grip on outbound investment after Meta-Manus deal fallout
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →