China’s National Development and Reform Commission issued a one-line statement on Monday, April 27, 2026, saying it would prohibit foreign investment in the acquisition of the Manus project and require all parties to withdraw from the deal.
The statement did not name Meta. It did not explain the reasons. It cited laws and regulations and offered nothing further.
That single sentence ended a $2 billion acquisition that Meta had announced in December 2025 and had been defending to regulators on both sides of the Pacific ever since. China blocked it. The deal is dead.
Here is the full story of what Manus is, why Meta wanted it, how China responded, and what this means.
What Is Manus?
Manus launched publicly on March 6, 2025. Its creator described it in a launch video as more than “just another chatbot or workflow. It’s a completely autonomous agent.”
Chinese state television cheered its debut as potentially “China’s next DeepSeek moment.”
The company’s name comes from the Latin word for “hand,” the idea being that this was AI that could actually do things rather than just discuss them.
The distinction matters. Most AI tools are reactive, you ask, they answer. Manus was designed to be proactive and autonomous, executing complex multi-step tasks without requiring the user to intervene at each stage.
At launch it demonstrated an AI agent that could screen job candidates, plan vacations with full itineraries and budget breakdowns, analyze stock portfolios and build financial models, conduct market research, write and debug code, and plan travel arrangements end to end.
It claimed at launch to outperform OpenAI’s Deep Research. Microsoft began testing it in Windows 11 PCs by October 2025.
The company behind it was Butterfly Effect Pte Ltd, founded in Beijing in 2022 and also known as Monica.Im.
As US-China technology tensions escalated through 2024 and 2025, the founders recognized that staying headquartered in Beijing created significant risks for their global ambitions.
In mid-2025, Manus relocated its headquarters to Singapore, laid off most of its Beijing staff in July, and repositioned itself as a Singapore-based company.
The move was designed to make it more accessible to American investors and potential acquirers.
It worked… almost. Benchmark, the prominent Silicon Valley venture firm, led a $75 million Series B in April 2025 at a $500 million valuation.
By December 2025, Manus had millions of subscribers, more than $100 million in annual recurring revenue, and a revenue run rate exceeding $125 million, eight months after launch.
It had processed more than 147 trillion tokens and supported over 80 million virtual computers.
Why Did Meta Seek To Acquire Manus?
Meta announced the acquisition on December 29, 2025. The company did not disclose financial terms, but the Wall Street Journal reported the deal was valued at over $2 billion.
The Wall Street Journal also noted that Manus had been seeking a fresh fundraising round at a $2 billion valuation when Meta approached it.
The strategic logic was clear. Meta has been investing aggressively in artificial intelligence, it invested $14.3 billion in Scale AI in June 2025, bringing founder Alexandr Wang into Meta’s AI leadership team, and acquired AI-wearables startup Limitless in December 2025.
Manus represented something different and specific: a proven, revenue-generating AI agent system with millions of paying customers and demonstrated capability at autonomous task execution.
While Meta’s own Meta AI chatbot serves users across Facebook, Instagram and WhatsApp, the agentic layer, the ability to actually do things rather than answer questions, was where Manus was already winning.
Meta said it would continue operating and selling the Manus service after the acquisition, integrate its technology into Meta AI and its family of apps, and keep Manus based in Singapore.
Meta also told Nikkei Asia that after the acquisition, Manus would have no ties to Chinese investors, a statement clearly aimed at pre-empting the US national security objections that the deal’s Chinese origins invited.
Why Did China Block The Manus Acquisition?
Shortly after the acquisition was announced in December, China’s commerce ministry said it would investigate whether the deal complied with local laws and regulations.
In January 2026, the investigation expanded, China raised questions about compliance with technology export controls, whether user data could be shared with Meta, and issues involving cross-border currency flows and tax accounting.
The central concern, stated plainly in Chinese media and regulatory framing, was technology leakage.
The argument was that Manus’s AI technology had been developed in China, by Chinese founders, with Chinese investment, and that relocating to Singapore and selling to an American company represented a transfer of strategic technology out of China’s orbit regardless of where the company’s legal address was registered.
Beijing’s position, reinforced by restricting the founders’ travel, was that relocation does not exempt companies or individuals from domestic oversight.
In March 2026, the Financial Times reported that CEO Xiao Hong and chief scientist Ji Yichao, both normally based in Singapore, had been summoned to a meeting in Beijing and told they were not allowed to leave China while the review was ongoing.
The two founders of a Singapore-registered company were effectively held in China during a regulatory review of a transaction they had already announced.
Meta told CNBC in March that the acquisition “complied fully with applicable law” and expected a timely resolution.
The resolution arrived Monday. It was a prohibition.
The Geopolitical Frame
The Manus deal attracted scrutiny from both directions, which is itself a sign of how charged the US-China technology competition has become.
On the American side, Senator John Cornyn of Texas, a Republican and senior member of the Senate Intelligence Committee and one of Congress’s most vocal China hawks, had already criticized Benchmark’s investment in Manus back in May 2025, when the deal was still just venture capital flowing toward a startup rather than a $2 billion acquisition.
US regulators ultimately approved the deal, viewing the Singapore relocation as adequate distance from Chinese control.
The FTC, which has been scrutinizing Meta over its market power in social media, did not block it.
On the Chinese side, the calculus was the opposite. What Washington saw as an American company buying a Singapore-based AI startup, Beijing saw as an American company extracting Chinese AI technology through a jurisdictional shell move.
The National Development and Reform Commission, China’s top economic planning agency, is one of the most powerful bodies in the Chinese government.
Its one-line statement on Monday carries the full force of that authority.
The statement asked the parties to withdraw. The deal is unwound. Manus, whose CEO and chief scientist were restricted from leaving China for weeks during the review, is no longer a Meta acquisition.
What happens to the company now, whether it seeks a different acquirer, raises another round of funding, or operates independently, has not been announced.
What This Means For Meta
Meta shares fell 0.2 percent in premarket trading on Monday following the news.
That is a relatively muted reaction for a $2 billion deal falling apart, which reflects either confidence that Meta can find another path to agentic AI capability or recognition that the Manus deal, given its regulatory difficulties, had already been discounted by the market.
The loss of Manus does not cripple Meta’s AI strategy. The company has Scale AI’s technology, Alexandr Wang’s involvement in its leadership, an ongoing Llama development program, and billions in capital to continue acquiring capabilities.
Manus was a specific thing, a proven agentic AI product with paying customers and demonstrated real-world utility at scale, something Meta’s own AI division has not yet produced at equivalent revenue.
Finding that elsewhere will take time.
The broader implication is the one that applies beyond Meta specifically. China has now demonstrated that it will use its regulatory authority to prevent Chinese-founded AI companies from being acquired by American firms, even when those companies have relocated to third countries and restructured their ownership.
The Singapore maneuver, which Manus executed deliberately and which worked well enough for US regulators, did not work for Chinese regulators.
That will inform how the next wave of Chinese-origin AI startups approach their international expansion, and how American companies approach acquiring them.
