China just reminded Silicon Valley that in the AI age, your past address may matter more than your current one.
Beijing’s decision to torpedo Meta’s multibillion‑dollar purchase of Manus AI doesn’t just kill a deal — it puts a bullet in the heart of an entire playbook for escaping Chinese scrutiny and dials up the U.S.–China AI rivalry another notch.
Manus’ rise was pure hype‑cycle rocket fuel. Founded in China by entrepreneurs Xiao Hong and Ji Yichao, the company burst onto the scene in March 2025 with a “general AI agent” capable of handling real‑world tasks like searching real‑estate listings, booking international flights and hotels, and even using tools and writing code on a user’s behalf. The system works as an “agentic wrapper” around models like Anthropic’s Claude 3.7 Sonnet, orchestrating multiple agents — a planner to assign tasks and an executor to browse the web, run software, and verify results.
The product landed at exactly the moment Silicon Valley decided “AI agents” were the next platform. By December 2025, Meta swooped in, announcing a roughly $2–2.5 billion acquisition, seeing Manus as a cornerstone of Mark Zuckerberg’s push to build “personal [AI] superintelligence for everyone.”
Meta moved fast. Manus’ team and tech were folded into key products, including Meta’s Ads Manager, the core tool that advertisers use to create and track campaigns on Facebook, Instagram, Messenger, and WhatsApp. At that point, the deal looked not just done — but strategically embedded. Undoing it would mean, as one analysis put it, trying to “unscramble the egg.”
There was a catch: Manus was born in China.
By the time Benchmark led a $75 million Series B in spring 2025, Manus was legally domiciled in the Cayman Islands but “was otherwise a Chinese company,” with around 100 employees in China. To calm both U.S. national security worries and Beijing’s tightening grip on advanced tech, the plan was classic 2020s corporate jujitsu: move the team to Singapore and operate as a global startup, not a Chinese one.
This so‑called “Singapore washing” became a popular move — shift headquarters to the city‑state to dodge political landmines from both Washington and Beijing while still tapping Western capital and markets. Manus’ staff “quickly” moved south to Singapore under this logic.
The strategy worked — until it didn’t. Benchmark’s investment even triggered a U.S. Treasury Department investigation into whether backing a China‑rooted AI firm violated outbound investment rules, underscoring that Washington was also watching these corporate gymnastics closely.
The real blow came from the other direction.
Although Manus had shifted its center of gravity abroad, China’s regulators concluded that origin still mattered. Because the company was founded in China, they argued, it remained subject to the country’s strict export control rules on selling domestic technology to foreign firms.
In January 2026, China’s Ministry of Commerce formally announced that it would investigate Meta’s acquisition, focusing on whether the transfer complied with technology import and export regulations. That move alone sent a chill through founders who had banked on distance and restructuring as protection.
Beijing’s scrutiny went beyond paperwork. In March, two of Manus’ co‑founders were reportedly barred from leaving China while the review was underway, a pointed reminder of how personal and coercive these fights can become.
On April 27, 2026, China’s National Development and Reform Commission (NDRC) dropped the hammer. It prohibited the investment outright and ordered Meta and Manus to unwind the deal, a late‑stage intervention that blindsided many in the tech and venture world. The NDRC said in a brief statement that its decision was made “in accordance with laws and regulations,” offering no granular explanation.
A parallel account from other coverage echoed the same outcome: China’s economic watchdog blocked Meta’s $2 billion purchase after months of scrutiny dating back to the December announcement, despite the fact that “it was largely complete and Manus is already integrated into some of Meta’s tools.”
Ars Technica captured the geopolitical stakes bluntly: China asked Meta to unwind the acquisition after deciding to “ban foreign investment in Manus based on national security concerns,” a move that “indicates how difficult it has become for US and Chinese tech companies to strike and sustain such deals” amid a deepening AI rivalry.
For Meta, the timing could hardly be worse. The company has tied a huge part of its post‑metaverse identity to AI agents and tools embedded across its platforms.
Meta has publicly maintained that it did nothing wrong. In a statement provided to Axios, the company insisted: “The transaction complied fully with applicable law. We anticipate an appropriate resolution to the inquiry.” A similar line went to the BBC, with a spokesperson repeating that the deal followed all relevant regulations and voicing confidence in a resolution.
But the confidence looks increasingly aspirational. Commentators now openly question whether Meta can truly unwind a company that, by its own account, is already “deeply integrated” into its operations. Manus isn’t a minor plugin; it’s fused into Meta’s ad infrastructure and AI roadmap.
Axios summed up the dilemma: “It’s unclear how or if Meta could unscramble the egg.”
Beijing, for its part, has barely bothered to dress the move in market‑friendly language.
Articles tracking the saga point out that the NDRC and related authorities justified the prohibition under export control rules and national security concerns, with little elaboration beyond formal references to acting “in accordance with laws and regulations.”
In substance, China is asserting three things:
If the intent was to send a deterrent message, it appears to be working. One analysis notes that the “late‑stage intervention involving the two non‑Chinese companies has drawn much alarm among tech founders and venture capitalists,” who are now re‑evaluating the supposed safety of Singapore‑washing.
Axios was blunt about the broader fallout: by ordering Meta to unwind its $2.5 billion acquisition, Chinese regulators have likely signaled “the end of ‘Singapore washing,’ a corporate sleight of hand” that let Chinese‑origin tech firms secure foreign investment while trying to sidestep both Beijing and Washington.
AI Magazine reached a similar conclusion, noting that Manus “is finding that a change of address is not enough to evade the regulatory eye in Beijing,” despite its relocation from China to Singapore.
The message to founders is stark:
That is not just a compliance headache; it is an existential business risk.
Zooming out, the Manus saga is less an anomaly than a preview.
China’s decision “shows how tech founders struggle to cut China ties” and “indicates how difficult it has become for US and Chinese tech companies to strike and sustain such deals” as the AI rivalry hardens. On the U.S. side, outbound investment screens and export controls are tightening. On the Chinese side, authorities are increasingly willing to reach across borders to stop deals between two non‑Chinese parties if they see Chinese‑origin AI as being quietly transferred.
Meanwhile, Manus itself is left in limbo. It had already claimed the title of the “fastest startup in the world” to surpass $100 million in annual recurring revenue just eight months after launch. Now its founders face exit bans at home, a blocked sale abroad, and a future where both superpowers treat their technology as too sensitive to freely trade.
Meta, for its part, has to decide whether to fight, re‑architect, or walk away — all while its competitors race ahead with their own AI agents.
If this is what happens over one AI agent startup, imagine what the next decade of AI platform battles will look like.
