How Beijing’s Block of Meta’s Manus Deal Is Redrawing the Map of Global AI

China’s rejection of Meta’s $2 billion acquisition of agentic AI startup Manus is not merely a regulatory footnote. It is a strategic declaration—one that signals AI has crossed a threshold from commercial asset to national sovereign resource, and that the old playbook for internationalizing Chinese technology is finished.

By the middle of 2025, a new term had entered the quiet lexicon of venture capitalists navigating the US-China technology rift: “Singapore-washing.” The concept was simple enough. A promising Chinese AI startup, wary of rising geopolitical risk, would shift its legal registration to a neutral Southeast Asian jurisdiction, relocate a portion of its team, raise Western capital, and present itself to global markets as a de facto global firm—one that happened to be founded by Chinese engineers but was now safely incorporated outside the dragon’s reach.

Manus looked, for a time, like the paradigmatic success story of this model. Founded as the Monica.im project by Beijing Butterfly Effect Technology, it released its first general-purpose AI agent in March 2025 to breathless comparisons with DeepSeek. The startup had passed $100 million in annual recurring revenue by December 2025—eight months after launching a product—claiming at the time to be the fastest company in the world to reach that milestone from zero. CNBC It had moved staff to Singapore mid-year, raised $75 million in a round led by Silicon Valley’s Benchmark, and attracted the attention of Meta Platforms, which announced a $2 billion acquisition in December 2025.

Then, on April 27, 2026, Beijing slammed the door.

China’s National Development and Reform Commission ordered the deal’s cancellation in a brief statement, citing laws and regulations, without elaborating further. Bloomberg The message, however, was anything but brief in its implications.

The Anatomy of a Blocked Deal

To understand why Beijing intervened, one must look past the Singapore address on Manus’s incorporation documents. According to Chinese regulatory findings, Manus’s core technologies were developed in China and involve processing massive amounts of user data. Its China-based affiliated entities—Beijing Red Butterfly Technology and Beijing Butterfly Effect Technology—remained active, and the technical origin and domestic entities had not been legally separated. TechNode

In Beijing’s legal reading, this was not a Singaporean company selling itself to an American buyer. This was Chinese intellectual property—developed on Chinese soil, using Chinese data infrastructure—attempting to depart through a side door without paying the sovereign toll.

Under China’s Measures for the Security Review of Foreign Investment, the Catalogue of Technologies Prohibited and Restricted from Export, and the Measures for Security Assessment of Data Export, core AI algorithms fall under restricted export technologies, requiring compliance with technology export licensing procedures and data security assessment requirements. TechNode The parties, regulators alleged, had circumvented these procedures entirely.

The human consequences were stark. The Financial Times reported in March that Manus co-founders Xiao Hong and Ji Yichao had been subjected to exit bans—barred from leaving China as the investigation deepened. Meanwhile, Manus employees had already moved into Meta offices in Singapore, capital had been transferred, and exiting investors including Tencent, ZhenFund, and Hongshan had received their proceeds. Business Standard Unwinding the deal, as the NDRC now demands, will be legally and logistically Byzantine.

Singapore-Washing: A Model Under Siege

The Manus affair has exposed what was always the fundamental vulnerability in the Singapore-washing strategy: geography is not sovereignty. A legal address in one of the world’s most business-friendly jurisdictions cannot override the reality of where technology was conceived, where engineers were trained, and where the underlying data originates.

Beijing is sending a message, and it reads clearly: residency in Singapore will no longer insulate Chinese-founded companies from regulatory scrutiny, forcing founders to choose between Western capital and Chinese ties. Yahoo Finance

This represents a seismic shift in the operating assumptions of Chinese entrepreneurship over the past decade. The variable interest entity (VIE) structure, red-chip listings, offshore incorporation, talent relocation—these were all instruments of a system that allowed Chinese innovation to access global capital while remaining, in practice, deeply embedded in China’s domestic ecosystem. Beijing tolerated this ambiguity when the technology in question was consumer internet. It is no longer prepared to do so when the asset in question is frontier artificial intelligence.

The reasoning, while heavy-handed in execution, is not without internal logic. According to Stanford University’s 2026 AI Index report, the long-standing performance gap between US and Chinese AI systems has effectively disappeared, with models from both countries now competing at comparable levels—and Chinese systems like DeepSeek-R1 at times matching leading US models. International Business Times When you are operating in a near-peer technological competition with the world’s largest economy, allowing your most valuable intellectual assets to be acquired by a direct rival is not merely bad commercial strategy. From Beijing’s perspective, it is a national security failure.

The Broader Crackdown: Capital Controls Enter the AI Age

The Manus block did not arrive in isolation. It is the visible tip of a rapidly submerging policy architecture.

Chinese agencies including the NDRC have told several private firms in recent weeks that they should reject capital of US origin in funding rounds unless explicitly approved. Yahoo Finance The scope of these instructions is remarkable. Moonshot AI, which is considering an initial public offering, was among those that received guidance from the state planner. Fellow startup StepFun received similar instructions. Regulators have also decided on similar restrictions for ByteDance, the owner of TikTok, blocking secondary share sales to US investors without government approval. Yahoo Finance

Consider the scale of disruption this implies. Moonshot AI is seeking to raise as much as $1 billion in a funding round that would value the startup at approximately $18 billion. StepFun, which is considering a $500 million float in Hong Kong, is in the process of unwinding its overseas entities and onshoring capital to meet regulatory requirements. Yahoo Finance

The new restrictions risk further isolating China’s recovering tech sector from the venture backing that has underpinned it for two decades, much of which was sourced from American pensions and endowments. Yahoo Finance The parallel move to restrict red-chip firms from seeking Hong Kong IPOs simultaneously closes off two of the primary channels through which Chinese companies have historically accessed Western capital markets. The architecture of financial globalization that powered China’s technology rise is being deliberately disassembled, brick by brick.

A Mirror Image: The Decoupling Is Now Bilateral

To appreciate the full symmetry of what is happening, one must look westward as well. Washington did not wait for Beijing to move first. The US Treasury Department finalized rules in 2025 restricting American outbound investment into Chinese companies operating in AI, semiconductors, and quantum computing. Congressional proposals including the COINS Act have sought to formalize and expand these restrictions. The Committee on Foreign Investment in the United States (CFIUS) has increasingly scrutinized technology transactions involving Chinese counterparties, regardless of where the formal acquirer is registered.

In other words, Beijing’s move is not an unprovoked act of protectionism. It is, in considerable measure, a mirror response to the regime Washington has been constructing for years. China’s escalation mirrors steps taken by Washington months earlier, when US authorities restricted outbound investment into Chinese companies operating in AI, semiconductors, and quantum computing on national security grounds. FX Leaders Both governments now operate from the same strategic premise: advanced artificial intelligence is not a commercially neutral technology. It is a lever of national power, and allowing adversaries to access it—through acquisition, investment, or talent migration—is a strategic error.

What has changed, dramatically and perhaps irreversibly, is the speed at which this mutual calculus is hardening. The Manus affair compressed years of latent tension into months of regulatory escalation. What once required a geopolitical incident to trigger can now be set off by a single startup’s term sheet.

The Chilling Effects on Innovation and Entrepreneurship

None of this, however, occurs without cost—and Beijing’s calculus, however strategically coherent, carries real dangers for the ecosystem it claims to protect.

The founders of Manus did not set out to betray China. They built a remarkable technology, attracted global capital, and attempted to navigate a world in which their homeland’s commercial and regulatory environment had become increasingly inhospitable to internationally ambitious startups. The exit bans imposed on Xiao Hong and Ji Yichao—whatever the legal justification—send a chilling signal to every talented Chinese engineer who might contemplate building for a global market.

Chinese government scrutiny of AI companies could impact other Chinese startups’ strategies for expansion and funding in the United States. The Washington Post The subtler damage is harder to measure but likely deeper: the erosion of the entrepreneurial confidence that produced companies like Manus in the first place. If the reward for building a breakout AI startup in China is an exit ban and a forced transaction reversal, the rational response for ambitious engineers is either to stay entirely within the domestic system—or to leave China earlier and more completely than Manus ever did.

Neither outcome serves Beijing’s long-term interests. The first risks insulating Chinese AI development from the competitive pressure that drives frontier innovation. The second accelerates precisely the kind of talent drain the crackdown is ostensibly designed to prevent.

Implications for Global Investors and the Future of Agentic AI

For the global venture capital community, the Manus block is a hard lesson in jurisdictional risk. The Singapore-registration playbook worked, until it didn’t. Benchmark, which led Manus’s Series A at a $500 million valuation, now finds itself holding an asset caught in one of the most complex geopolitical unwindings in recent startup history. Future investments in companies with significant Chinese technical heritage will require a level of regulatory due diligence that venture firms have historically neither staffed nor priced.

The specific domain of agentic AI—autonomous systems capable of executing complex multi-step tasks across diverse environments—makes this regulatory conflict particularly consequential. Agentic AI is widely viewed as the next major frontier of commercial AI deployment, with applications spanning enterprise automation, scientific research, and consumer productivity. When Meta announced the Manus deal, it said it would look to accelerate AI innovation for businesses and integrate advanced automation into its consumer and enterprise products, including its Meta AI assistant. CNBC The block does not merely deprive Meta of a team and a technology. It delays and fragments the development of a genuinely transformative technology category at the precise moment when the competitive race is most intense.

Recommendations: Navigating the New Landscape

The structural forces driving this bifurcation are not going away. But they need not produce an outcome that is worse for everyone. Several principles should guide policymakers, investors, and technologists in the period ahead.

For governments: Both Washington and Beijing should establish clearer, more predictable frameworks for cross-border technology investment reviews. The opacity of China’s current approach—a one-line NDRC statement, exit bans without charges, informal guidance to portfolio companies—creates uncertainty that harms legitimate commercial activity far beyond the specific deals under scrutiny. Rules that are knowable in advance are less disruptive than rules that arrive by surprise.

For investors: Structural due diligence must now include technology provenance analysis—understanding not just where a company is registered, but where its core intellectual property was developed, where its data originates, and whether its founding team could face legal constraints in either jurisdiction. Geography of incorporation is no longer a sufficient proxy for legal exposure.

For technology companies: The era of the “borderless startup” in AI is functionally over. Companies with genuine global ambitions must make earlier, cleaner decisions about their primary regulatory home. Ambiguity that was once commercially convenient has become a liability that can be weaponized by regulators on either side of the Pacific.

The Longer View

History will likely record the Meta-Manus episode as one of those moments when the underlying logic of an era became suddenly, viscerally legible. For the better part of two decades, the world operated on the assumption that technology and capital were, at their core, cosmopolitan forces—that they would flow toward talent and opportunity regardless of national boundaries, and that this flow was ultimately good for everyone.

That assumption is not dead. But it is seriously wounded. As the NDRC tightens its grip, the digital arteries connecting US and Chinese tech sectors are being severed, one funding round at a time. TechStory

The Manus block is not the end of Chinese AI innovation—China’s engineers are too numerous, too talented, and too well-supported by state capital for that. Nor is it the end of Meta’s ambitions in agentic AI. But it is the end of the comfortable fiction that the US-China technology competition could be navigated by clever corporate structuring. The battle lines are now drawn at the level of technology itself—who built it, where, with whose data, and for whose benefit.

In that contest, there are no neutral flags of convenience, and Singapore is no longer far enough away.