Manus is independent again, and this time it has half a billion dollars to prove it can stay that way.
On Thursday, October 8, Butterfly Effect — the parent company of the Chinese AI agent startup Manus — said it had completed a funding round of more than $500 million. The round was co-led by Boyu Capital and IDG Capital, with existing investors Tencent, HSG, and ZhenFund also participating, according to Reuters.
The announcement closes a strange chapter for one of the most closely watched startups in the AI agent race. Manus resumes independent operations just months after Beijing ordered Meta to unwind its acquisition of the company, a deal reportedly worth more than $2 billion. In August, Manus said it would begin operating independently again and delete some user data as part of its separation from Meta.
A round shaped by geopolitics
Few funding rounds carry this much political baggage. According to Reuters, Beijing ordered Meta to unwind its acquisition of Manus in April 2026, amid tightening scrutiny of U.S. investment in Chinese startups developing advanced AI. The reversal turned what looked like a clean Big Tech exit into a forced breakup — and the new round is, in effect, the market’s answer to the question of whether Manus can make it on its own.
The new investor list offers an answer. Boyu Capital and IDG Capital co-leading, alongside Tencent and ZhenFund, means the company now has deep-pocketed backers on both sides of the Pacific divide — but notably, the Chinese half is dominant. Tencent, which has been expanding its own AI footprint this year, was already a shareholder and has now doubled down. That matters, because the company’s future regulatory life will almost certainly be lived in China.
Manus builds general-purpose AI agents designed to autonomously carry out tasks such as research and automation with minimal human input. Its agent went viral in 2025 on the strength of demos showing it completing multi-step knowledge work — browsing the web, assembling reports, and chaining tools together — without constant prompting. It was one of the first products to make “agentic AI” feel real to a mainstream audience, and its invite-only launch turned scarcity into buzz.
The numbers behind the raise
Butterfly Effect did not disclose the exact size of the round or the valuation it raised at. But one figure floating around the deal suggests the business is no longer just potential: The Information reported in June that Manus’ annualized revenue run rate had surged to about $500 million, up from roughly $100 million at the time of the Meta acquisition.
If that figure is accurate, it explains why investors were willing to write a nine-figure check to a company that had just been through a forced divestiture. A 5x revenue jump over roughly the deal period is the kind of trajectory that keeps funds competing with each other — and it suggests the agent product is finding paying enterprise users, not just demo-day admirers.
There is also a longer play in motion. Reuters reported that Manus has been considering a joint-venture structure incorporated in China, which would pave the way for a Hong Kong listing. A person familiar with the matter said the company will not begin the IPO process until at least 2027. The timeline suggests the fresh capital is meant to buy runway: scale the business, settle the corporate structure, and time a public debut after the dust from the Meta episode fully settles.
Agents remain the industry’s hottest bet
The Manus round fits into a wider pattern of enormous capital flowing into autonomous AI systems. Just last month, Instinct raised $1 billion at a $10 billion valuation for a personal AI agent that promises to actually complete tasks rather than just answer questions. On the voice side, Inworld acquired Ultravox to fix AI’s interruption problem in conversations. The common thread: investors are betting that the next platform shift is not a smarter chatbot but software that does things.
Manus sits squarely in that thesis. Unlike model labs that compete on benchmarks, agent companies compete on workflow — how many real tasks they can complete without a human hovering over them. That makes the revenue run-rate figure especially significant: it implies customers are already handing the agent real work and paying for the outcome.
What the Meta reversal signals
The unwinding also carries a lesson for Big Tech’s M&A playbook in AI. Meta’s acquisition of Manus was meant to fast-track its agent ambitions with a team that had already demonstrated the product-market fit others were still chasing. Beijing’s intervention — ordering the deal unwound on national-security grounds — shows that cross-border AI acquisitions now face a regulatory environment that can reach backward in time, not just block new deals.
For Manus, the irony is rich: it was briefly owned by one of the world’s largest tech companies, was forced back into independence by its own government, and then raised one of the largest AI rounds of the year anyway. Tencent’s continued participation underscores the point — the company may be pivoting its center of gravity east, but its investor base still sees a global opportunity.
What to watch next
Three things will determine whether this round becomes the story of a company that turned a regulatory setback into a win. First, whether the reported $500 million revenue run rate holds up under scrutiny — revenue claims from private companies always deserve a healthy dose of skepticism, and Manus has not confirmed the figure publicly. Second, how the planned China-incorporated structure takes shape, and whether it complicates the company’s ability to serve users outside China. Third, how the agent market itself evolves: with giants like Tencent — already reportedly involved in a $7 billion AI cloud deal with Oracle — investing heavily on both sides of the agent-infrastructure divide, independent players will need more than a viral launch to keep their edge.
For now, Manus has something most startups never get: a second first chance, funded at nine figures. The market will decide whether it deserved one.
Sources: Reuters (via Economic Times, October 8, 2026); Nikkei Asia (October 8, 2026).
