THE DEAL
● More than $500 million, announced 8 October 2026.
● Led by Boyu Capital and IDG Capital. Existing investors Tencent, HSG and ZhenFund also took part.
● It only exists because Meta's acquisition fell through. Reporting puts the blocked Meta deal at around $2 billion.
● Valuation is not disclosed. A roughly $4 billion target has been reported but no party has confirmed it — treat the number with care.
What is confirmed, and what is not
Confirmed by TechNode citing Yicai: the amount is over $500 million, the leads are Boyu Capital and IDG Capital, existing backers Tencent, HSG and ZhenFund participated, and the announcement date is 8 October 2026. Butterfly Effect has not disclosed a valuation or what the money is for.
Reported but unconfirmed: a target valuation near $4 billion, and a share buyback by the founding team at roughly half that figure after the Meta deal unwound. Both circulate widely and neither has a named source attached. We are flagging them rather than repeating them as fact.
That distinction matters more than usual here, because the whole story is about a number that was prevented from happening.
The Meta deal that did not close
Meta moved to acquire Butterfly Effect earlier in 2026 in a deal reported at around $2 billion. Chinese regulators blocked it.
Read the raise against that backdrop and it changes shape. This is not a company that chose independence — it is one that had an exit removed and then had to go and finance itself. The investors who stepped in are domestic: Boyu, IDG, Tencent, HSG, ZhenFund. That is the predictable consequence of a blocked foreign acquisition, and it is the pattern to watch rather than the headline figure.
It also makes this a reasonable proxy for a larger question. If a Chinese AI application cannot be bought by a US company and can still raise half a billion dollars domestically at short notice, the capital constraint that was supposed to force consolidation is not binding.
What Manus actually is
Manus is an autonomous AI agent: you give it a task in plain language and it plans, browses, writes and executes without step-by-step prompting. It drew attention in 2026 both for genuinely capable demos and for a demonstration that raised questions about synthetic social-media content.
We have covered it in depth before, and the two pieces are the useful background here:
Why the raise matters beyond China
Agent platforms have a cost problem that chat products do not. An agent that works for twenty minutes burns tokens for twenty minutes, and the pricing that makes a chatbot viable makes an agent expensive. Half a billion dollars buys the runway to subsidise that gap while inference prices fall.
And they are falling fast. In the first week of October alone, Anthropic cut Haiku's input price by 90% to $0.10 per million tokens, and Google halved Nano Banana's per-image cost. An agent company raising a large round into a market where its main input cost is dropping every few weeks is in a considerably better position than the same round would have bought a year ago.
The question for Manus is not funding any more. It is whether an agent people trust to act on their behalf can be built by a company that most Western enterprises will not be permitted to buy from.
What to watch next
- A confirmed valuation. Until Butterfly Effect or a lead investor states one, $4 billion is a rumour with good distribution.
- Whether pricing changes. Half a billion dollars usually shows up as a cheaper free tier before it shows up as a better model.
- Enterprise availability outside China. This is the constraint the money cannot solve.
- Other blocked acquisitions. If this becomes a pattern, domestic mega-rounds follow each one.
Sources