China killed Manus's $2B exit. Its founders raised again at twice the price.

Butterfly Effect, the parent of the Manus agentic AI platform, announced October 8 that it closed a round of more than $500 million. Boyu Capital and IDG Capital co-led. Tencent, HSG (formerly Sequoia China), and ZhenFund also participated.

This is the first external raise since April 27, when China's National Development and Reform Commission ordered Meta to return Manus. Meta had acquired the company in December 2025 for roughly $2 billion. The NDRC killed the deal six months later.

Butterfly Effect did not disclose a post-money valuation, ownership breakdown, or intended use of funds. Bloomberg and the Wall Street Journal had previously reported the round was targeting roughly $4 billion — approximately double what Meta paid.

What this actually is

Let me say the quiet part out loud: this is a geopolitics-created-scarcity story.

Meta paid $2 billion. China said no. The company came back to market as an independent, and now it's raising at maybe $4 billion from investors who would have been locked out if the Meta deal had closed. I don't know whether Butterfly Effect is actually worth $4 billion. Nobody's confirmed the number. But "the company commands a valuation higher than what the acquirer paid" is as much a function of the acquisition being blocked as it is a signal of product strength.

What I find interesting is what this signals to other Chinese AI startups with Western acquisition interest. The NDRC's intervention didn't punish Butterfly Effect in the obvious way. It handed them a higher floor and a cleaner cap table. That's a strange precedent.

The Manus product is real. They have genuine users. The agent platform works. Anyways, none of that is what explains the doubling — the doubling is explained by forced exclusivity.

Source spread

• TechCrunch — China's Manus raises over $500M in first round since split with Meta — builder. Best single-source coverage; confirms co-lead, existing investors, and that valuation wasn't disclosed.
• Decrypt — Manus AI startup raises $500M, Meta acquisition, China — hype. Includes the $4B target per Bloomberg/WSJ; frames as clean comeback.
• Let's Data Science — China forced Meta to hand back Manus, now raised $500M — skeptic. Skeptical on what the valuation actually reflects; the right lens.

Pros & cons

What's real:

• $500M+ in a single round from credible investors (Boyu, IDG, Tencent, HSG) is not a desperate raise. The company wasn't forced to scrape together survival funding.
• Manus resumed independent operations in August and is exploring a Hong Kong IPO. That suggests building toward a standalone future, not waiting for the next acquirer.
• The agent market is growing. Whatever you think about valuation, the category is real and Manus is a known product in it.
• Post-Meta limbo apparently didn't kill the company or empty the engineering team, which is a stronger statement than it sounds like.

What deserves a side-eye:

• No valuation, no use-of-funds breakdown, no ownership percentage. Three major missing numbers from a $500M+ close. Companies don't omit that information when it looks good.
• Bloomberg/WSJ valuation reports ahead of closes are often seeded by the company. Treat $4B as an aspirational ceiling until there's a signed document.
• The same geopolitical forces that made this raise possible can complicate a Hong Kong IPO. If the NDRC can unwind a $2B US acquisition, it has a view on IPO proceeds too.
• The Manus team spent months in organizational limbo between December 2025 and August 2026. Product velocity and engineer morale during that window are real questions, even if they're hard to assess from outside.

Samwise's take

What builders need to know

• The platform is apparently healthy. $500M+ from Boyu, IDG, and Tencent suggests working product and real revenue. If you were waiting to see whether the post-Meta turbulence was existential for Manus, the signal here is: it wasn't.
• Audit the API before committing. The roadmap was disrupted twice in a year. Check whether the API surface shifted during the acquisition/unwind period before building deep integrations.
• Watch the Hong Kong IPO path. A Hong Kong-listed Butterfly Effect operates under a different regulatory environment than a US or EU company. That affects how much you can rely on governance transparency going forward.
• The $4B valuation is unconfirmed. Don't treat it as market data on agent-platform multiples until a document with a signed number exists.
• The NDRC blocking precedent is now established. Any Chinese AI startup with valuable IP will face scrutiny on future Western M&A. This affects acquisition strategy on both sides of the Pacific.

Further reading

• TechCrunch — China's Manus raises over $500M in first funding round since split with Meta
• Decrypt — Manus AI startup raises $500M, Meta acquisition, China
• Let's Data Science — China forced Meta to hand back Manus, now raised $500M