(Bloomberg) — Meta Platforms Inc. has agreed to buy Manus, a popular Singapore-based artificial intelligence agent with Chinese roots, in its effort to build a business around its massive AI investment.
The deal values Manus at more than $2 billion, according to people familiar with the matter. It marks a rare US acquisition of an Asian tech company and the latest multibillion-dollar AI bet from Meta Chief Executive Officer Mark Zuckerberg. The agreement was struck in about 10 days, the people said, asking not to be identified as the details are not public.
Most Read from Bloomberg
Meta intends to continue operating and selling the Manus service while also integrating the technology into its products, it said in a statement. Backed by some of China’s biggest names including Tencent Holdings Ltd., ZhenFund and HSG, Manus shot to prominence early this year not long after DeepSeek’s debut. All of its existing investors have been bought out in Meta’s takeover, one of the people said. It’s unclear whether severing Chinese ties will ease government concerns at a time when the Asian nation and the US are vying for AI dominance.
“There will be no continuing Chinese ownership interests in Manus AI following the transaction, and Manus AI will discontinue its services and operations in China,” a Meta spokesperson said. Manus’ parent company, Butterfly Effect Pte, was founded in China before moving to Singapore.
Zuckerberg has made AI his company’s top priority, and is spending billions to hire researchers, build data centers and develop new models. Manus had an annual revenue run rate of $125 million earlier this year from selling its AI agent to businesses via subscriptions, which could give Meta a more immediate return on some of its AI spending.
The Manus AI agent can complete a handful of general tasks, such as screening resumes, creating trip itineraries and analyzing stocks in response to basic instructions. Butterfly Effect raised money earlier this year at close to a $500 million valuation in an investment round led by US venture capital firm Benchmark.
What Bloomberg Intelligence Says
Meta’s purchase of Manus is likely aimed at expanding the former’s AI agent task capabilities, as it currently lacks applications built on its own foundation model compared to a broader application ecosystem for ChatGPT, Google Gemini and Anthropic Claude. While the deal marks an early step for Meta to build a subscription and API business around its AI investment, it could draw regulatory scrutiny given that Singapore-based Manus was founded in China.
— Mandeep Singh and Robert Biggar, BI analysts
AI agents are tools that don’t need human supervision to perform specific digital tasks. Enterprise software companies such as Salesforce Inc. and ServiceNow Inc. have heavily promoted their versions of agents as the most effective way for businesses to use the emerging technology, rather than generative AI features such as chatbots, which require user prompts and interaction.
Meta already has an AI chatbot, Meta AI, which is available through the company’s social media and messaging platforms — Facebook, Instagram and WhatsApp — in addition to its AI glasses. The US company is acquiring the technology and leadership group from Manus, though its statement didn’t detail where the new team will sit within the organization. The Wall Street Journal, which earlier reported the deal’s value, said Manus co-founder and CEO Xiao Hong will report to Javier Olivan, chief operating officer of Meta, citing people familiar with the acquisition.
Alexandr Wang, Meta’s Chief AI Officer who joined this summer as part of a high-profile investment into his startup, welcomed the Manus team of about 100 staff with a post on X. In his own message on the platform, Manus’ Xiao said that the deal would help his company expand the reach of its agents. “The era of AI that doesn’t just talk, but acts, creates, and delivers, is only beginning,” he wrote. “And now, we get to build it at a scale we never could have imagined.”
Meta’s aggressive spending to compete in the AI race is matched by rivals like OpenAI, Alphabet Inc.’s Google and Microsoft Corp. Zuckerberg has pledged to spend $600 billion on US infrastructure projects over the next three years, many of them expected to be AI-related. The company has hired an expensive team of researchers to develop a new state-of-the-art AI model it plans to debut next spring, and has faced some skepticism from investors who worry that the heavy spending won’t result in meaningful revenue anytime soon.
VC firm Benchmark was criticized earlier this year by lawmakers and other venture investors for backing an AI company with ties to China. “Who thinks it is a good idea for American investors to subsidize our biggest adversary in AI, only to have the CCP use that technology to challenge us economically and militarily? Not me,” US Senator John Cornyn, a Texas Republican, wrote in a post on X in May. Benchmark didn’t immediately respond to a request for comment on the Meta deal Monday.
When asked about the deal during a regular press briefing, a Chinese Foreign Ministry spokesperson on Tuesday referred reporters to relevant regulatory agencies without elaborating.
–With assistance from Vlad Savov and Philip Glamann.
(Updates with US-China context in third paragraph)
Most Read from Bloomberg Businessweek
©2025 Bloomberg L.P.
