SINGAPORE — Chinese authorities say they have banned Meta’s acquisition of Manus AI, an artificial intelligence company founded in China — taking Beijing’s most aggressive step yet to stanch the loss of AI talent and resources to the United States and setting off what is likely to be a complicated legal and political fight.
Manus AI, which was created by Chinese engineers from Wuhan, relocated to Singapore before Meta acquired it in December for $2 billion. In January, Beijing began investigating Manus for compliance with export controls, and authorities barred two of the company’s leaders from leaving China several weeks ago as part of that probe.
On Monday, China’s National Development and Reform Commission (NRDC), which is in charge of foreign investment security reviews, issued a statement saying that it had decided to prohibit “the foreign acquisition of the Manus project” and that it would require “the parties concerned to unwind the acquisition transaction.”
The decision represents a major advance in Chinese efforts to claw back and contain key technologies it deems key to national security, and has sent shock waves through the AI industry in Asia.
The Chinese authorities’ attempt to wrest back the firm from a U.S. tech giant comes at a fraught moment in U.S.-China relations on AI, as the White House this week accused Chinese entities of running “deliberate, industrial-scale campaigns” to “extract capabilities” from U.S. AI models.
President Donald Trump and Chinese leader Xi Jinping are expected to meet in Beijing next month, where the two leaders will discuss AI among other agenda items.
It was not immediately clear whether Chinese authorities have the jurisdiction to reverse Meta’s acquisition, completed months ago, or how they could enforce their edict. Manus had incorporated as a Singaporean entity and shuttered its Chinese operations last year. Most Manus employees are now in Singapore and integrated into Meta’s offices here.
Manus did not respond immediately to requests for comment. In a statement, a spokesman for Meta, Andy Stone, said “the transaction complied fully with applicable law.”
Singapore’s Ministry of Digital Development and Information, which works closely with AI firms in the country, did not respond to requests for comment.
Faced with intensifying competition from the United States over AI, Chinese leaders are drawing what they say is a new red line against “China-shedding” — a practice in which homegrown companies sever ties with the country to compete for capital and other resources in the United States. Apart from Manus, Chinese authorities have warned at least one other prominent AI firm with Chinese roots — MiroMind — against sending valuable AI talent and resources out of the country.
The firms have become a growing headache for Beijing, as efforts to rein in the exodus with regulation lag behind the breakneck expansion of China’s AI sector and its thirst for capital.
In 2020, Beijing introduced a national security review mechanism to scrutinize foreign investments in Chinese firms deemed critical to national security. The rules give authorities the power to forcibly unwind any deal they do not approve — but it only applies to Chinese-founded firms within China.
Analysts say the Manus AI case is the first major test of whether China’s review mechanism can reach firms that shift operations overseas, in part to sidestep regulator scrutiny.
“The state has been sitting there looking at all of the regulations available and trying to figure out which one allows them extraterritorial jurisdiction over this, or can they make something that claims jurisdiction over Manus before it leaving. That part is missing from this so far,” said Kendra Schaefer, director of tech policy research at Beijing-based consultancy Trivium China.
While the current regulatory basis for unraveling the deal is flimsy, that could mean little against the Chinese government’s ability to pressure those Manus employees still in China, including Xiao Hong and Ji Yichao — the CEO and the chief scientist, respectively, who were summoned to Beijing last month and have been prevented from leaving China.
“If an authoritarian state is threatening your family, it doesn’t really matter what the law says,” said Chris McGuire, a senior fellow at the Council on Foreign Relations who oversaw technology and national security issues at the National Security Council during the Biden administration. He said the Manus AI case probably will have a chilling effect on Chinese founders.
“If you’re going to tie yourself to the Chinese AI ecosystem, you have to accept that there is no predictability. The Chinese government will always get its way,” he said.
Meta, which touted the deal to acquire Manus to investors in late December, may have its hands tied, and analysts say a resolution would have to be elevated to a political level.
“Mark Zuckerberg would have to go to Trump and say, ‘This has to be a top priority during your summit with Xi’; then it just gets thrown on the pile, and I very much doubt that this is going to be a top priority for the president when he talks to Xi,” said Sarah Bauerle Danzman, a senior fellow with the Atlantic Council and former staffer at the Committee on Foreign Investment in the United States.
As recently as last week, many working in AI in China and Singapore expressed doubt that the Chinese authorities would go so far as to attempt to unwind the Meta-Manus deal.
The NRDC decision is “a lot more ruthless than expected,” said a former Manus engineer in Wuhan, speaking on the condition of anonymity to avoid government reprisal. “I fear this could be the beginning of a mass crackdown, a campaign-style policy that tries to set things right by over-rectification，” the person said.
Matthias Hendrichs, a Singapore-based adviser to AI firms, said he has been perplexed by the level of scrutiny paid to Manus because it is not widely seen as providing critical technology. The company builds on top of large language models, or LLMs, to create AI programs that operate independently.
“I’m really surprised, as many are, about this decision,” said Hendrichs, who previously worked in China for Apple. While acquisitions are sometimes blocked by Chinese regulators for national security reasons, it’s “extremely rare” for a deal to be unwound after it’s done, he said.
“From a practical perspective, this is going to be very cumbersome,” Hendrichs said.
Lizzi Lee, a fellow at the Asia Society Policy Institute’s Center for China Analysis, said the Manus decision suggests that national security is becoming an “all-encompassing lens” for Chinese regulators when it comes to AI, and it is likely to send a chilling effect through the Chinese AI ecosystem.
“Beijing’s regulator is still playing this cat-and-mouse game between regulation and innovation,” Lee said. “I think it’s quite clear now that AI is treated by Beijing regulators much more like critical infrastructure rather than commercial technology.”
Rudy Lu in Taipei contributed to this report.
The post China says it will reverse major AI acquisition by Meta appeared first on Washington Post.
