Meta’s $2 Billion Manus Deal Collapses Under Beijing’s Pressure: What It Means for AI, Investors and Global Tech

Meta has begun dismantling its $2 billion acquisition of Manus, a Chinese-founded agentic AI startup, after Beijing issued a divestiture order on national security grounds. The move marks the most concrete step yet toward compliance with the directive, with Meta cutting Manus off from internal systems and halting data sharing between the two companies. Employees are no longer permitted to use Manus tools for internal projects as the firms move toward a full operational separation. The decision underscores Beijing’s determination to retain control over strategically sensitive technology, regardless of where a company is incorporated offshore. For global technology firms eyeing acquisitions in China’s AI ecosystem, the episode signals a sharp escalation in regulatory risk and a narrowing window for cross-border deals.

The unwinding process comes as Manus continues to operate independently, rolling out new integrations with platforms such as Similarweb and Shopify. This suggests the company retains technical momentum even as its ownership structure is dismantled. Meanwhile, the co-founders of Manus have held preliminary discussions about raising roughly $1 billion from outside investors to reclaim the startup from Meta. Such a move could pave the way for a Chinese joint venture structure and an eventual listing in Hong Kong, where Chinese AI startups including MiniMax and Zhipu have seen a surge in listings this year. The potential recapitalization highlights the broader tension between global capital flows and national security imperatives in AI.

This episode is part of a broader tightening by Beijing, which has expanded travel restrictions on researchers and executives at private firms. Officials now require government approval before such individuals can travel abroad. Additionally, China is tightening its grip on foreign capital, with reports indicating that top AI firms such as Moonshot AI, StepFun, and ByteDance will need government sign-off before accepting U.S. investment. These measures add another layer to Beijing’s sweeping effort to control its AI sector and manage the flow of sensitive technology across borders.

The Manus Deal: From Viral Demo to Forced Divestiture

Manus first drew attention with a viral agent demo in late 2024, showcasing an AI system capable of autonomous web browsing, task execution, and third-party integrations. The company relocated its staff to Singapore in mid-2025, a move interpreted as an attempt to mitigate regulatory exposure while maintaining access to global markets. By December 2025, Meta announced it would acquire Manus for $2 billion, positioning the deal as a strategic bet on agentic AI and cross-border innovation.

However, Chinese regulators moved quickly to scrutinize the transaction. Early this year, authorities cited potential violations of technology export controls and foreign investment rules, setting the stage for intervention. The divestiture order, issued roughly two months before Meta began dismantling the deal, reflects Beijing’s broader campaign to assert control over AI technologies deemed strategically sensitive. The fact that the order targets a company incorporated outside China—yet founded by Chinese nationals and staffed with Chinese expertise—illustrates how nationality, talent, and technology access are increasingly intertwined in global AI governance.

Meta’s response has been operational: severing data flows, revoking system access, and initiating a legal and financial separation process. While the company has not publicly detailed the mechanics of the unwinding, the separation of systems and tools suggests a phased de-integration rather than an abrupt shutdown. This approach may help preserve Manus’s operational continuity, at least in the short term, even as its strategic alignment with Meta dissolves.

Beijing’s AI Policy: From Export Controls to Capital Controls

The Manus case is not an isolated incident but part of a wider policy shift in Beijing. Since late 2025, Chinese authorities have expanded the scope of technology export controls, particularly for AI models capable of autonomous reasoning, web interaction, or multi-step task completion. These capabilities are now treated as dual-use technologies, subject to stricter scrutiny regardless of where they are developed or hosted.

In parallel, China has tightened capital controls around AI firms. Reports indicate that leading AI companies—including Moonshot AI, StepFun, and ByteDance—must now obtain government approval before accepting foreign investment, especially from U.S. sources. This policy aims to prevent sensitive AI capabilities from flowing to overseas entities without state oversight. It also complicates fundraising for Chinese AI startups that rely on international venture capital, potentially redirecting investment toward domestic or neutral jurisdictions.

Travel restrictions have been another lever of control. Researchers and executives at private AI firms now require government approval before traveling abroad, ostensibly to prevent the unauthorized transfer of technical knowledge or personnel. While framed as a national security measure, the policy also limits the mobility of top talent and reduces opportunities for international collaboration—factors that have historically driven innovation in AI.

Together, these measures form a coordinated strategy to centralize control over China’s AI ecosystem, balancing openness to global markets with tight constraints on technology and capital flows. For foreign investors and acquirers, the message is clear: access to China’s AI sector is increasingly conditional on compliance with Beijing’s evolving rules.

What Happens Next to Manus: From Meta Exit to Hong Kong Listing?

Despite the divestiture order, Manus continues to operate and ship new features, indicating that its technical platform remains intact. The company’s recent integrations with Similarweb and Shopify suggest it is still pursuing commercial viability, even as its ownership structure is unwound. This operational continuity raises a key question: can Manus survive as an independent entity, or will it need to restructure under new ownership?

The co-founders of Manus have held preliminary discussions about raising up to $1 billion from external investors to buy back the company from Meta. Such a recapitalization would likely involve a Chinese joint venture structure, potentially with domestic partners, to satisfy regulatory concerns. An eventual listing in Hong Kong has also been discussed, aligning with a trend in which Chinese AI startups have increasingly turned to the city’s capital markets for funding and visibility.

However, the path forward is fraught with challenges. Any new investment would need to pass regulatory scrutiny, especially regarding technology ownership and data governance. A joint venture structure could help address Beijing’s concerns about foreign control, but it may also limit strategic flexibility and global reach. Investors would need to weigh the risks of regulatory reversal, capital controls, and geopolitical uncertainty against the potential upside of owning a leading agentic AI platform.

For Meta, the unwinding process is likely to result in financial losses, reputational costs, and strategic setbacks. The company had positioned the Manus acquisition as a cornerstone of its agentic AI strategy, aiming to integrate autonomous capabilities into its ecosystem. With the deal reversed, Meta must now reassess its approach to agentic AI, potentially accelerating in-house development or pursuing alternative partnerships—especially in regions with more predictable regulatory environments.

Broader Implications for Global AI M&A and Investment

The collapse of the Manus deal signals a new phase in global AI dealmaking, where geopolitical risk has become a primary factor in transaction viability. For technology companies considering cross-border acquisitions—especially those involving AI models with autonomous capabilities—the Manus case serves as a cautionary tale. Even well-structured deals with clear commercial logic can be derailed by sudden regulatory intervention, particularly when national security concerns are invoked.

Investors, too, face heightened uncertainty. Venture capital firms that backed Manus, including Benchmark, have already received proceeds from the acquisition, but Asian backers such as Tencent, Hillhouse Capital Group, and ZhenFund have indicated they will cooperate with the unwinding process. This suggests a recognition that regulatory compliance now outweighs financial returns in certain markets. For limited partners and LPs, the episode underscores the need to factor geopolitical risk into due diligence, particularly in AI and other dual-use technologies.

The tightening of capital controls in China also complicates exit strategies for investors. A traditional IPO in the U.S. or Europe may now be less viable for Chinese AI startups, pushing them toward domestic listings or alternative markets like Hong Kong or Singapore. This shift could reshape global capital flows, with investors increasingly favoring jurisdictions with clearer regulatory frameworks and more stable geopolitical conditions.

The Agentic AI Race: Who Benefits from the Manus Unwinding?

Agentic AI—systems capable of autonomous task execution, web browsing, and multi-step reasoning—has become a critical frontier in the AI arms race. With Manus sidelined, at least temporarily, the competitive landscape shifts. Companies with strong agentic capabilities, such as domestic Chinese players and international firms operating outside China’s regulatory reach, may gain ground.

For Chinese AI firms that remain compliant with Beijing’s rules, the Manus episode could create an opening. By aligning with national security priorities—such as data localization, joint venture structures, or restricted model capabilities—they may secure regulatory approval for foreign investment and partnerships. This could accelerate consolidation in China’s AI sector, with larger, state-aligned firms absorbing smaller, high-risk startups.

Internationally, the vacuum left by Manus may benefit companies in the U.S., Europe, and Southeast Asia that are developing agentic systems without direct exposure to Chinese regulatory risk. Firms like Adept, Inflection AI, and Mistral AI have already made strides in autonomous AI, and could see increased interest from enterprises seeking alternatives to China-linked solutions.

However, the agentic AI race is far from decided. Regulatory uncertainty in China, combined with ongoing U.S.-China tech tensions, means that no single player can claim dominance. The Manus unwinding is likely just one chapter in a longer story about how AI governance, national security, and global competition intersect.

Practical Takeaways for Businesses, Investors and Policymakers

For businesses considering AI acquisitions or partnerships involving Chinese entities, the Manus case highlights the need for rigorous geopolitical due diligence. Legal teams should assess not only commercial and technical risks but also regulatory exposure in both the target company’s home jurisdiction and the acquirer’s. Contracts should include clear termination clauses and compliance mechanisms in case of sudden regulatory changes.

Investors should diversify geographic exposure and avoid over-concentration in high-risk markets. While China remains a critical hub for AI talent and innovation, the Manus episode demonstrates that capital and technology flows are increasingly subject to state control. Limited partners should pressure general partners to model geopolitical scenarios and stress-test portfolios against regulatory shocks.

Policymakers in open economies should consider how to support AI innovation without compromising national security. Measures such as export controls and investment screening are necessary, but they must be balanced with incentives for domestic AI development and international collaboration. A fragmented regulatory landscape risks stifling innovation and pushing critical AI research to less transparent jurisdictions.

Conclusion

Meta’s decision to unwind its $2 billion acquisition of Manus marks a turning point in the global AI landscape. It reflects Beijing’s determination to assert control over strategically sensitive technology, even when it involves offshore entities or foreign acquirers. The episode is part of a broader tightening of AI governance in China, encompassing export controls, capital restrictions, and travel limitations.

While Manus continues to operate independently, its future ownership and structure remain uncertain. A recapitalization led by Chinese investors or a Hong Kong listing could offer a path forward, but only if aligned with Beijing’s evolving priorities. For global tech firms and investors, the Manus case is a reminder that in AI, technology and geopolitics are inseparable. The road ahead will require careful navigation—balancing ambition with compliance, innovation with oversight, and opportunity with risk.