Beijing Breaks the Playbook

Executive summary

Beijing's order to reverse Meta's $2 billion acquisition of AI startup Manus has landed in uncharted legal territory — and the American public doesn't quite know what to make of it. This is the first time China's Foreign Investment Security Review mechanism has ever forced the unwinding of a deal that had already closed, and 126 respondents surveyed in June 2026 split almost perfectly three ways: a third called it normal business, a third worried about government overreach, and nearly a quarter backed the idea of protecting national tech interests.

The most consequential signal isn't any single opinion — it's the absence of a dominant one. With no majority sentiment, public reaction is highly susceptible to framing. Companies, policymakers, and communicators fighting for the narrative on cross-border AI deals are working with an unusually malleable audience.

Meanwhile, a striking paradox cuts through the data: the respondents most supportive of protecting national tech interests were also the most likely to say deal reversals hurt innovation. Low trust in big tech compounds the picture — but it doesn't push people toward welcoming government intervention. The public is anxious about the whole ecosystem, not just one player.

Context

On April 27, 2026, China's National Development and Reform Commission issued an order that had never been issued before: it told Meta to give Manus back. The $2 billion acquisition, which had closed in December 2025, was retroactively prohibited under China's Foreign Investment Security Review measures — a regime that had existed since 2020 but had never been used to unwind a deal after the fact. Within days, Meta cut Manus employees off from internal systems and data pipelines. The teams that had already relocated to Meta's Singapore office were effectively stranded mid-integration.

The Manus case exposes the limits of a structuring strategy that had become standard practice in cross-border AI deals. Manus was a Chinese-founded AI agent startup — its research staff, roughly two-thirds of whom still work out of Beijing and Hangzhou, built some of the most capable autonomous agent technology in the world. To make the Meta deal viable, founders relocated the company's legal home to Singapore, turning down requests from Chinese authorities for meetings and investment along the way. Beijing's NDRC called that move a 'washing-style' overseas expansion and declared it non-compliant. The Singapore-redomiciliation playbook, widely used by Chinese tech companies seeking U.S. capital, is now effectively closed.

At the same time, the United States has been tightening its own cross-border review apparatus. CFIUS — the Committee on Foreign Investment in the United States — has dramatically expanded its scope since the Foreign Investment Risk Review Modernization Act of 2018, and it has shown a growing willingness to block or condition deals involving AI, semiconductors, and sensitive data. Legal analysts at Morgan Lewis describe the simultaneous tightening of FISR and CFIUS as 'bilateral convergence' — a narrowing of the space where cross-border AI transactions can actually get done.

This survey captured public sentiment at the precise moment that convergence became impossible to ignore. The 126 respondents reflect a general American audience processing a regulatory event with no precedent, no established cultural script, and enormous downstream implications for how AI companies raise capital, structure deals, and operate across borders.

Findings

Conclusion

The Meta-Manus unwind is not a one-off dispute between two companies and one government. It is the opening move in a new era of bilateral regulatory competition over AI assets — one where both Beijing and Washington are simultaneously asserting jurisdiction over deals that were once considered safely cross-border. China's FISR mechanism has now demonstrated that it can reach backward through time to reverse a closed transaction. CFIUS has spent six years expanding toward the same capability in the opposite direction.

For companies structuring cross-border AI deals, the Singapore-redomiciliation strategy is gone. For investors, the risk calculus on any deal with Chinese-origin technology or talent has permanently shifted. For communicators and policymakers, the public is unusually persuadable right now — a 60% 'it depends' majority is an audience without a settled frame, ready to be shaped by whoever tells the clearest story first.

Watch for three developments in the months ahead: whether CFIUS moves to formalize analogous post-closing review authority; whether other Chinese AI startups with overseas structures face similar NDRC scrutiny; and whether the innovation-anxiety signal detected among national-tech-interest supporters grows into a coherent political constituency that pushes back on deal-blocking as national security strategy. The regulatory rulebook for cross-border AI is being rewritten in real time — and public opinion is still catching up.