Manus quagmire: The long arm of China’s cross-border strategic reach

China’s decision to block Meta’s acquisition of Manus should not be treated as a routine intervention in a corporate deal. Beijing is signalling that relocating an AI company overseas does not necessarily remove it from Beijing’s strategic concerns.

Manus is a Chinese-founded, Singapore-headquartered AI company focused on AI agents: systems that automate tasks and workflows. The transaction might have looked like an offshore acquisition of an offshore target, but Beijing saw something else.

Beijing’s longer reach

For Chinese regulators, the question is no longer simply where a company is incorporated or where its shareholders sit. Regulators will look past the company’s mailing address. They will ask where the technology was built, how the company developed its capabilities and whether a sale of the entity would strengthen a strategic competitor.

Chinese-founded AI companies should take the hint. Redomiciling overseas may reduce some risks, but it will not necessarily sever the company’s connection to China in Beijing’s eyes. Incorporation in Singapore, Delaware or London may be relevant under corporate law. It matters far less for national security, technology transfer, data governance and industrial policy.

That is why the Manus decision matters beyond Meta. It suggests that a company can legally leave China while remaining within Beijing’s strategic reach. If its founders, engineers, datasets, research history or early commercial life remain tied to China, Beijing may continue to regard the company as part of China’s technological ecosystem.

China is not alone in this shift. The US has long used the Committee on Foreign Investment in the United States to review foreign acquisitions of companies with sensitive technology, data or infrastructure. Washington has also created an outbound investment regime covering certain China-related investments in semiconductors, quantum information technologies and AI. Once technology has national security consequences, governments stop treating private deals as purely private.

Europe has taken a less blunt but recognisably similar path, using foreign investment screening and the AI Act to bring strategic technology under closer public oversight.

China’s intervention goes further because it reaches outward. US and EU screening regimes often focus on foreign control over domestic assets. In the Manus matter, China appears to have asserted an interest in a company that had already positioned itself outside the mainland. That distinction matters. Beijing may increasingly treat Chinese-origin AI capability as a strategic asset even after a company has moved offshore.

There is a cost for China, too. Blocking deals like this may help Beijing keep strategic AI capabilities from moving abroad, but too much control can make Chinese-founded companies harder to fund, harder to scale and less attractive as partners. 

Advanced AI depends on the movement of people, ideas, capital and customers across borders. If founders believe that a Chinese origin will follow them wherever they incorporate, some may choose to build outside China from the start or avoid China-linked investors and research altogether. That could make the system more cautious and more fragmented, with more state direction and less room for private companies to test global markets. 

AI agents as strategic assets

Beijing has likely had several overlapping reasons to intervene. AI agents are strategically sensitive. They are increasingly designed to perform tasks such as writing code, conducting research, automating business processes and operating across digital environments. A company with credible agent technology is not merely building another consumer application. It may be building systems that businesses, researchers and governments will come to rely on.

China also has an industrial policy reason to object. China has invested heavily in AI and does not want promising companies, engineers or technical know-how absorbed by American platforms. In an earlier era, the sale of a Chinese-founded company to a US technology giant might have been celebrated as entrepreneurial success. Today, Beijing may see the same transaction as the loss of a strategic capability.

The Manus episode narrows the perceived options for Chinese founders. Some entrepreneurs believed that relocating early to Singapore or elsewhere could reduce geopolitical and regulatory risk. It suggests that efforts to reduce China-related regulatory risk must begin earlier and run deeper to be persuasive.

When origin matters more than domicile

But even early separation may not be enough. AI capability is harder to relocate than a factory or a trademark. It is bound up with people, training, research culture, data access and technical judgment. If the founders were trained in China, the early research took place there, the engineering team remains largely Chinese, or the company drew on Chinese commercial relationships, regulators may still see continuity. In AI, corporate nationality is not only a matter of incorporation. It also reflects where the technology was developed.

That is what makes Manus different from TikTok. TikTok raised questions about whether a Chinese-owned platform abroad could be trusted with foreign users’ data and political discourse. Manus raises a different question: whether a Chinese-origin AI capability can be sold to a leading American company after it has moved offshore. The concern is no longer only Chinese control over foreign users. It is also a foreign acquisition of a Chinese-origin capability.

Today, technology decoupling is no longer driven only by Washington, which worries that Chinese firms may obtain foreign data and sensitive technology. Beijing worries that American firms may absorb Chinese talent and frontier capability. Both sides accuse the other of politicising commerce. In practice, however, both are increasingly doing so.

Investors should draw a practical lesson from this. Any acquisition of a Chinese-founded AI company will require more than ordinary legal and financial due diligence. Buyers will need to examine ownership, founder history, technical lineage, data sources, prior Chinese operations, employment footprint and regulatory exposure. In frontier AI, due diligence now has to include geopolitics.

Harder road for founders and investors

For founders, the choice is more difficult. They must decide earlier whether they are building primarily for China, for the US-led technology ecosystem, or for a genuinely global market. A company seeking Chinese talent, US capital, a Singaporean domicile and global customers may discover that each advantage brings a different regulator to the table.

The old model of globalisation is becoming harder to sustain. For years, companies could incorporate in one jurisdiction, hire in another, raise capital elsewhere and sell worldwide. Frontier AI puts that model under strain. States now care where capability begins, who controls it and where it may end up.

China’s blocking of the Meta-Manus deal is best read as a warning. Beijing will not object to every offshore transaction involving a Chinese-founded technology company. But where the company sits near the frontier of AI, has strategically valuable talent and is being acquired by a major US platform, regulatory intervention is now foreseeable.

Chinese-founded AI companies will still globalise. Many will continue to raise foreign capital and build for global users. But they will do so under closer scrutiny. They may be able to move their legal domicile. But for advanced AI companies, their origins may prove harder to leave behind.

Related: How Manus went from AI superstar to a geopolitical problem | Manus plight: Should AI companies start in China or overseas?