According to TheRegister.com, Meta is acquiring the Chinese AI startup Manus, which only launched in March 2025. The deal, announced by both companies, will see Manus’s staff join Meta, though the financial terms were not disclosed. Manus, created by a company called Butterfly Effect, claims an astonishing $100 million in annual recurring revenue just eight months after its debut and says it serves millions of users and businesses worldwide. The startup’s technology focuses on “general agents” that can perform complex, multi-step tasks like evaluating job applications from a .ZIP file. This marks Meta’s fifth AI-related acquisition in 2025, following buys like PlayAI and Rivos. The move comes as CEO Mark Zuckerberg pursues a “superintelligence” and Meta plans over $70 billion in capital expenditure for 2025, largely for AI datacenters.
Meta bets on agents, not just chatbots
Here’s the thing: this isn’t just another chatbot company. Manus is pitching something fundamentally different from your average ChatGPT conversation. They’re talking about AI that can actually do things—open files, analyze data across multiple documents, and produce actionable results. Basically, they’re selling AI as a worker, not just a talker. That’s a huge shift. And it fits perfectly with Zuck’s vague but expensive vision of a “superintelligence” that “knows us deeply” and helps achieve goals. You can see the pitch on their comparison page versus ChatGPT. It’s all about action over chat.
The $100 million question
Now, let’s talk about that $100 million ARR claim. It’s… aggressive, to say the least. Eight months in? That’s a rocket ship trajectory that makes most Silicon Valley startups look sleepy. They posted about this milestone on their blog, but it raises eyebrows. Is it real sustainable revenue, or is it a flashy number built on heavy discounts and pilot projects? Either way, it was clearly a compelling data point for Meta’s acquisition team. In the hyper-competitive AI land grab, a claimed user base of “millions” and a big revenue number—even if it’s fresh—is a trophy asset.
A Chinese AI firm under Meta’s wing
The geopolitical angle here is subtle but fascinating. Manus is “made-in-China,” founded by Butterfly Effect, which has moved its HQ to Singapore—a classic maneuver for Chinese tech firms navigating international tensions. So Meta isn’t buying a mainland Chinese entity directly; it’s buying a Singaporean one with deep Chinese roots. This is probably the only way a deal like this gets done. It lets Meta tap into top-tier Chinese AI engineering talent and IP without walking directly into the US-China tech cold war. Smart, if not a little tricky.
Where does this fit Meta’s plan?
So what’s Zuck going to do with it? Look, Meta is throwing mountains of cash—$70 billion+ this year alone—at AI infrastructure. But so far, they don’t have a major paid AI product. The rumored “Meta AI+” subscription needs a killer feature beyond a slightly better chatbot. Manus’s agent technology could be that feature. Imagine a business subscription where an AI agent manages your ads, analyzes your performance reports, and hires your next employee. That’s the dream they’re buying. As they stated in their announcement blog, the goal is to expand to “billions of people on Meta’s platforms.”
The bigger picture
This is another brick in the wall of Meta’s closed ecosystem. They’re not just building AI; they’re buying every piece of the stack—chips (via Rivos), wearables (Limitless), and now advanced agent software. They’re trying to build an entire AI universe that lives inside Facebook, Instagram, and WhatsApp. The risk, of course, is that they’re spending like there’s no tomorrow. $70 billion is a staggering bet. But if AI agents become the next platform shift, Meta doesn’t want to be left behind like they were with mobile. They’re buying their way to the front. Whether Manus is the key piece or just another expensive experiment, we’re about to find out.
