Meta has paid more than $2 billion for Manus, a Singapore-based startup that built what many consider the most advanced autonomous AI agent on the market.
China’s Ministry of Commerce announced it will investigate whether Meta’s acquisition of AI startup Manus violates export control laws since Manus was originally founded in Beijing before relocating to Singapore. That review could take up to 90 days, and Meta expects to close in Q1 2026.
But I want to focus on what Manus does and why Meta wanted it.
A worthy reminder:, one of my very early episodes of AI SIX was on Manus AI. If you want to go back and listen to it, it’s episode 44. At the time, it was an impressive demo. Now Meta owns it.
What Manus actually does
Manus is a general-purpose AI agent. Not a chatbot. An execution engine.
You give it an objective e.g. “research the top 10 competitors in this market and produce a report” or “screen these 50 CVs against this job description” – and Manus plans the task, coordinates tools, files, and code, and delivers completed work. Market research reports. Data analysis. Travel itineraries. Code. Recruitment screening.
The difference from ChatGPT or Claude is that Manus doesn’t just answer your question. It does the work.
The commercial traction speaks for itself. Manus launched in March 2025, hit $100 million in annual recurring revenue within eight months, and signed up millions of paying subscribers. By December, the revenue run rate exceeded $125 million. The company has processed more than 147 trillion tokens and created over 80 million virtual computing environments, which is 80 million times the number of systems spun up a virtual computer to complete a task for a user.
Why Meta wanted it
Meta has been spending $60 to $65 billion in 2025 on AI infrastructure, with $600 billion committed through 2028. Investors have been asking: when does this generate returns?
Manus is already profitable. That’s part of the answer.
But more importantly, Meta now has an agent layer to sit on top of its Llama models. Meta has strong AI research and open-source models, but until now it lacked the orchestration and execution capabilities that turn model outputs into completed tasks. Llama can think. Manus can act. Now Meta has both.
What this means for Meta’s products
Here’s where it gets interesting for anyone using Meta’s platforms.
Meta plans to integrate Manus across Facebook, Instagram, WhatsApp, and Meta AI, while keeping Manus running as a standalone subscription service. The goal is AI that handles end-to-end workflows rather than just answering questions.
For small businesses on WhatsApp Business or Instagram, think about what becomes possible:
- An AI agent that creates and manages ad campaigns, not just suggests copy, but actually builds, launches, and optimises them
- Customer service that doesn’t just draft replies but handles entire conversations and resolves issues
- Research and reporting that runs in the background and delivers finished documents
Meta has talked about “every business having a business AI.” Manus is how they deliver that.
For consumers using Meta AI, this means moving beyond “ask me anything” to “do this for me.” Book travel. Compare products and make purchases. Manage schedules. Produce content. The assistant becomes an agent.
The execution layer matters
There’s a broader point here about where value sits in the AI stack.
We’ve spent two years focused on foundation models, which have the best LLM, the highest benchmark scores, and the largest context window. But Manus built a $100 million revenue business in eight months without training its own model. It used Claude, it used open-source tools, and it focused entirely on the execution layer, making AI actually do things reliably.
Meta clearly agrees that’s where the value is. They didn’t buy a model company. They bought an environmental company. An orchestration company. The layer that turns AI capability into AI action.
For anyone building AI products or thinking about AI strategy, that’s worth noting. The model matters, but the ability to reliably execute multi-step tasks in the real world might matter more.
So let’s wait and see if the regulatory barrier is overcome and how Meta evolves with Manus.
📚 Want to master AI for your organisation? Get my book, AI Marketing Playbook – practical strategies you can implement today.
📰 Stay informed weekly: Subscribe to Human in the Loop, my LinkedIn Newsletter for AI insights that matter.
🎓 Go deeper: The AI Playbook Diploma is CPD Accredited and gives you the structured learning to truly understand, use, and govern AI.
📞 Do you want to talk about custom in-house training or building your own AI solutions? Book a call at digitaltraining.ie or publicsectormarketingpros.com
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
