China’s National Development and Reform Commission issued a rare order Monday forcing Meta to unwind its $2 billion acquisition of Manus, the Singapore-registered AI agent startup that caused a global sensation when it launched just over a year ago. The brief regulatory statement gave no detailed reasons, but the move effectively ends a deal that had drawn scrutiny from both Beijing and Washington since it was first announced last December.
For anyone tracking enterprise AI, this is a significant moment. Not because of Meta or Manus specifically, but because of what it signals about where the AI industry is heading.
What Manus Actually Built
Manus, developed by Butterfly Effect Pte Ltd, gained international attention in March 2025 when it launched what it called the world’s first general-purpose AI agent. Unlike a chatbot that answers questions, Manus could complete multi-step, real-world tasks end to end: screen job applicants, research investment theses, write and execute code, plan logistics. It outperformed OpenAI’s models on the GAIA benchmark at launch.
The startup is registered in Singapore but its founders are Chinese. CEO Xiao Hong and chief scientist Ji Yichao were reported to have been summoned to Beijing in March for a regulatory review, and told they could not leave the country while the probe was underway. China’s commerce regulators had been assessing the deal since January.
The underlying concern, at least from Beijing’s side, was technology leakage. Manus built its core capabilities on top of Anthropic’s Claude and fine-tuned versions of Alibaba’s Qwen, which gave it genuinely world-class agentic performance. Handing that to Meta meant handing it to the US tech complex.
The Bigger Pattern
This isn’t an isolated event. It’s part of a broader dynamic that is reshaping how companies think about AI infrastructure and vendor risk.
The US and China are each tightening control over their most capable AI systems. The US has export controls on chips and certain model weights. China is now showing it will use its own regulatory apparatus to block foreign acquisition of domestically-connected AI talent and technology, even when the company is technically registered offshore.
What’s different here is the target. Manus isn’t a semiconductor company or a state-adjacent research lab. It’s a startup that builds general-purpose AI agents for productivity and business automation. The kind of technology that Enterprise DNA clients are actively deploying right now.
If a general AI agent company operating out of Singapore with Chinese founders is now subject to Chinese national security review, then the question every business building AI infrastructure needs to ask is: where does my AI supply chain actually run through?
What This Means for Business
AI vendor geography is now a real consideration. Most enterprises haven’t thought carefully about where their AI providers are incorporated, where their training data comes from, or which governments have regulatory authority over the underlying models. That’s going to change.
The agentic AI market is strategic territory. Governments are treating general-purpose AI agents the same way they treat semiconductors and critical infrastructure. The fact that Beijing intervened on an agent company specifically tells you how seriously these capabilities are being taken.
Multi-vendor AI strategy becomes more important. If you’re building your operations on a single AI agent platform with opaque geopolitical exposure, you’re carrying risk that most boards haven’t priced in yet. A diversified approach, using platforms with clear sovereignty and data residency, is more defensible.
Tech transfer politics will touch M&A for years. The Manus deal took less than five months from announcement to forced cancellation. Any company considering acquisition of AI talent or technology that has connections, even indirect ones, to China or other sensitive jurisdictions should expect the same scrutiny.
The AI industry has spent the last three years focused almost entirely on capability benchmarks and model releases. The Manus decision is a reminder that geopolitics is now as important as compute.
For a deeper walkthrough of tools like this and how they fit together, the free Working With Claude field guide covers the ecosystem end to end. Get the guide.
Source
Bloomberg
Free Resource
Going deeper with Claude?
Get the free 32-page implementation guide for ANZ teams.
Your guide is ready
Check your downloads folder. If it did not open automatically, use the button below.
Download the Guide