In late April 2026, China’s National Development and Reform Commission (NDRC) formally ordered Meta Platforms to unwind its roughly $2 billion acquisition of Manus, a Singapore-incorporated agentic AI startup with deep roots in China. The decision represents one of Beijing’s most significant interventions in a cross-border technology deal, and has sent shockwaves through the global AI industry. At its core, the case raises a question that no government has yet answered definitively: when a startup’s founders, engineers, and foundational code originate in one country but its corporate entity is registered in another, who has the authority to decide whether that company can be sold abroad?
In this article, we’ll discuss the timeline of events leading to China’s unprecedented decision, what Manus AI actually does, why Meta wanted it so badly, how the concept of “Singapore washing” has been exposed as a fragile strategy, and what all of this means for the future of AI investment across geopolitical fault lines.
TL;DR Snapshot
China’s NDRC has ordered Meta to fully reverse its acquisition of Manus, an agentic AI startup that can autonomously execute complex, multi-step tasks like coding, data analysis, and web automation. The ruling follows a months-long investigation, exit bans imposed on Manus’ two co-founders, and growing criticism within China that the deal amounted to handing over strategic technology to a geopolitical rival. The move has disrupted the so-called “Singapore washing” playbook that many Chinese-founded startups had been using to attract Western capital and acquirers.
Key takeaways include…
- China’s NDRC ordered Meta to fully unwind its $2 billion acquisition of Manus, marking one of Beijing’s most aggressive interventions in a U.S. tech deal and establishing new precedent for how China treats AI technology outflows.
- The Manus co-founders, CEO Xiao Hong and Chief Scientist Ji Yichao, were barred from leaving China in March 2026 and remain unable to join Meta’s global engineering teams, effectively stalling integration efforts.
- The case has shattered confidence in the “Singapore washing” strategy, where Chinese-founded startups relocate their legal entities to Singapore to sidestep scrutiny from both Beijing and Washington, sending a clear warning to founders and investors across the AI landscape.
Who should read this: AI founders, venture capitalists, M&A professionals, tech policy analysts, and anyone following the U.S.-China technology rivalry.
What Is Manus, and Why Did Meta Want It?
Manus, whose name comes from the Latin word for “hand,” is an agentic AI platform built by Butterfly Effect Pte Ltd. It officially launched in March 2025 and quickly became one of the most talked-about AI products in the world. Unlike traditional chatbots that respond to prompts with text, Manus operates as an autonomous agent. Give it a high-level goal, and it can break that goal into sub-tasks, navigate the web, write and execute code, analyze data, and deliver a finished result without requiring human oversight at every step. According to Manus’ own blog, the platform processed more than 147 trillion tokens and powered over 80 million virtual computers in its first months of operation.
The technology was built on a multi-agent architecture, leveraging models from Anthropic and Alibaba’s open-source Qwen family. An MIT Technology Review assessment noted that Manus could break tasks down into steps and autonomously navigate the web to gather information, positioning it alongside tools like OpenAI’s Deep Research and Anthropic’s Computer Use, but with a heavier emphasis on full autonomy.
For Meta, the acquisition was a strategic play. The company was looking to move beyond passive generative AI and into the agentic paradigm, where AI systems don’t just assist users but actively perform work on their behalf. Analysts at CX Today described the deal as Meta’s attempt to transition from “assistance to labor substitution,” with plans to integrate Manus’ agent technology into products like WhatsApp Business and the Meta AI assistant. In a competitive landscape where Google and OpenAI were advancing rapidly, Manus represented a shortcut to capabilities that could take years to build in-house.
A Deal Unravels: The Timeline From Acquisition to Cancellation
Meta announced its acquisition of Manus in December 2025 for a reported $2 billion to $3 billion, and in the aftermath things moved fast. According to TechCrunch, about 100 Manus employees relocated to Meta’s Singapore offices by early March 2026, with CEO Xiao Hong reporting directly to Meta COO Javier Olivan. Meta publicly stated there would be no continuing Chinese ownership interests and that Manus would discontinue its services and operations within China. To outside observers, it looked like a clean break.
Beijing, however, saw it differently. Within weeks of the announcement, China’s Ministry of Commerce launched a formal review under export control and technology transfer regulations. According to CNBC, the review focused on whether AI technologies developed by Manus while it was based in China fell under national security or technology export rules.
The situation escalated dramatically in March 2026. As Euronews reported, Manus co-founders Xiao Hong and Ji Yichao were summoned to Beijing for questioning by the NDRC and subsequently barred from leaving mainland China. It was the first known instance of Beijing using exit bans on executives to directly impede a multibillion-dollar deal with a U.S. tech company.
Then, on April 27, 2026, the NDRC issued a one-line order: the acquisition was prohibited, and both parties were required to withdraw the transaction. No further explanation was provided. A CNN report noted that the decision arrived just weeks before a high-profile summit between U.S. President Donald Trump and Chinese leader Xi Jinping, adding a layer of geopolitical timing to the regulatory action.
Meta responded by stating that the transaction “complied fully with applicable law” and that it anticipated “an appropriate resolution to the inquiry.” However, unwinding a deal this far along will be extraordinarily complex. As The Next Web observed, the Manus case has become the foundational event prompting China to formalize an entirely new regulatory framework for technology outflows at the intersection of AI, foreign investment, and national security.
The End of “Singapore Washing”
Perhaps the most consequential ripple effect of the Manus case is what it means for the practice known as “Singapore washing.” This term refers to a strategy used by Chinese-founded startups that relocate their legal headquarters to Singapore in order to access deeper Western capital pools and avoid regulatory friction from both Beijing and Washington. Manus followed this playbook precisely. It was founded in China, moved its headquarters and core team to Singapore around mid-2025, raised $75 million from the prominent U.S. venture capital firm Benchmark, and was then acquired by Meta.
When the deal was first announced, it was widely celebrated as a validation of this model. CNBC reported that for Chinese founders striving to build products that could rival American peers, the deal felt like proof that an intricate offshore structure could successfully circumvent scrutiny from both sides. But China’s intervention quickly dismantled that perception. A Silicon Valley-based seed investor in AI told CNBC bluntly, “The path taken by Manus, people will not go down that route anymore.”
Legal experts reinforced the message. Beijing-based lawyer Yuan Cao of Yingke law firm told CNBC that the case should be seen as red flag (no pun intended) for companies that develop technology in China before transferring assets to an overseas entity through restructuring. Matthias Hendrichs, a Singapore-based adviser to global AI firms, was even more direct, stating that simply setting up a local legal entity and hiring a handful of local staff is nowhere near sufficient.
The implications extend well beyond Manus. Chinese authorities have signaled that they’ll look past corporate registration and examine where the code was written, where the data was generated, and where the talent was trained. For any Chinese-founded AI startup considering a similar international exit, the roadmap has fundamentally changed.
What Comes Next for Meta, Manus, and the AI Industry
For Meta, the practical fallout is significant. They paid $2 billion for a team and technology platform that it must now, at minimum, attempt to disentangle from. Over 100 Manus employees have been working from Meta’s Singapore offices, though co-founders remain stuck in China. The technology has reportedly already been partially integrated into Meta’s internal systems which will require a complicated and potentially costly rollback.
As CNN noted, this could represent a major missed opportunity for Meta as it races to compete against Google and OpenAI in the rapidly evolving AI agents space. For the broader AI industry, the Manus case has accelerated the bifurcation of the global technology ecosystem. Invezz reported that the abrupt halt is likely to have wide-ranging implications for future mergers and acquisitions in the AI space, highlighting the growing complexity of navigating regulatory environments amid geopolitical tensions.
Venture capitalists, particularly those investing in Chinese-founded AI startups, now face a new layer of due diligence risk. According to CNBC, Alex Ma, managing partner at Singapore-based family office Alpha Omega Holdings, warned that Chinese authorities would “look past the Singaporean facade and dig into the root of the company.” At the same time, Ma acknowledged that Beijing may not want to “over-punish success,” as doing so would discourage founders and distort incentives.
What remains clear is that the old playbook is gone. The Manus case didn’t just block one deal, it redrew the boundaries of what’s possible in cross-border AI transactions. And maybe more importantly, it forced governments, investors, and founders to confront the reality that the rules governing AI technology transfer are being written by geopolitics in real time.
Frequently Asked Questions
Manus is an agentic AI platform developed by Butterfly Effect Pte Ltd. The name comes from the Latin word for “hand.” Unlike traditional AI chatbots that simply generate text in response to prompts, Manus is designed to act autonomously. It can receive a high-level goal, break it into sub-tasks, browse the web, write and execute code, analyze data, and deliver completed work without step-by-step human supervision. It launched in March 2025 and quickly gained international attention for its capabilities.
Agentic AI refers to artificial intelligence systems that can operate autonomously to accomplish goals. Rather than waiting for a user to provide instructions at every step, agentic AI can plan, reason, use software tools, and execute multi-step workflows on its own. This makes it fundamentally different from a standard chatbot, which is primarily designed for conversational back-and-forth.
The NDRC, or National Development and Reform Commission, is China’s top economic planning agency. It plays a central role in shaping industrial policy, approving major investment projects, and overseeing foreign investment security reviews. In the Manus case, its Office of the Working Mechanism for Foreign Investment Security Review issued the formal order blocking Meta’s acquisition.
Singapore washing is a term used to describe a strategy where Chinese-founded technology companies relocate their legal headquarters and some operations to Singapore. The goal is typically to access international capital markets, attract Western investors, and distance the company from regulatory scrutiny in both China and the United States. The Manus case has demonstrated the limits of this approach, as Chinese authorities asserted jurisdiction based on where the technology was originally developed, not where the company was legally incorporated.
Other Enterprise AI Articles You May Be Interested In
DeepSeek V4 Changes the AI Pricing Game: Full Breakdown and Analysis
GPT-Image-2 Is Here: What Marketers, Designers, and Developers Need to Know
SpaceX Eyes $60 Billion Cursor Acquisition to Challenge Anthropic and OpenAI in AI Coding
Google’s Eighth-Generation TPU: A New Era of Specialized AI Chips
Can Flexible Data Centers Fix the AI Energy Crisis? A New Santa Clara Pilot Aims to Find Out
