The deal was reportedly worth approximately $2 billion. Meta wanted Manus, a general-purpose AI agent startup capable of autonomous coding and research, and by all visible indicators, Manus operated as a Singapore company. That framing did not hold up to China’s security review process.
According to reporting by AP and Reuters, China’s NDRC, acting through its Office of the Working Mechanism for Security Review of Foreign Investment, issued a prohibition order and directed all parties to withdraw. The NDRC did not publish the specific criteria it applied. It rarely does. What legal analysts familiar with China’s NSR framework have noted is that the framework does not limit its reach to Chinese-incorporated entities. The operative question is whether an acquirer is gaining access to assets, technology, or capabilities that touch Chinese national security interests, and whether the target’s origins can be traced to Chinese-registered entities.
According to reporting by Greenwich Time, Manus traces its origins to Beijing-registered entities. That single fact appears to have been enough.
The Domicile Question
Corporate relocation is not new in the AI startup ecosystem. Chinese founding teams have been establishing operations in Singapore, London, Abu Dhabi, and San Francisco for years, sometimes to access capital markets, sometimes to simplify international hiring, sometimes explicitly to reduce Chinese regulatory exposure. The working assumption in many of these moves has been that once a company is domiciled outside China, the Chinese regulatory grip diminishes.
The Manus ruling complicates that assumption. China’s NSR framework for foreign investment in security-sensitive sectors allows review authority when a foreign transaction involves assets or entities with Chinese national security relevance. The specific mechanism, the Office of the Working Mechanism for Security Review of Foreign Investment under the NDRC, has the authority to examine founding entity provenance, not just current incorporation status. A company can move its headquarters and still carry the founding registry links that trigger Chinese jurisdiction.
What “Chinese roots” means operationally in NSR practice is still not fully defined in public guidance. The framework’s opacity is deliberate, it preserves discretion. What is known is that the NDRC has the authority to look through the current corporate structure to the founding entities beneath it. For AI M&A practitioners, that means the diligence question is not only “where is this company incorporated today?” but “where were its founding entities registered, and do any of those registrations carry Chinese national security relevance?”
Stakeholder Map
The NDRC acted as the enforcement authority here, but the ruling has consequences distributed across multiple stakeholders in the AI M&A ecosystem.
Meta Platforms loses a significant AI agent acquisition at a moment when agentic AI capability is central to its product strategy. The company has not issued a public statement, which is consistent with active legal review rather than acceptance. Whether Meta pursues restructuring options, different deal structure, different acquiring entity, different asset scope, is a question without a public answer as of this writing.
Manus faces a materially changed position. Its most visible pending acquisition is blocked. Its Beijing founding entity connections are now publicly documented through the regulatory process. Other potential acquirers, particularly US and European companies subject to their own export control frameworks, will now have that documentation in front of them during any future diligence process.
Other frontier labs with cross-border M&A activity, including those that have already completed acquisitions of AI startups with Chinese-origin founding teams – now have a live precedent to evaluate against their existing portfolio and future deal pipeline. The question is whether the NDRC’s action here reflects a new enforcement posture or an application of existing authority that has simply not been tested at this asset class before.
AI startups with Chinese founding teams now operating outside China face a structural valuation question. If Chinese regulatory jurisdiction can follow founding provenance rather than current domicile, the pool of potential acquirers for those startups is effectively narrowed. Buyers subject to US or European export controls will conduct additional diligence. Buyers who might otherwise have proceeded without extensive NSR analysis now have a reason to pause.
US export control policy intersects here in a way that has not been fully mapped in public commentary. The ongoing evolution of US chip and technology export controls has focused primarily on restricting Chinese access to advanced AI capabilities. China’s NSR action on the Manus deal is the mirror image: China restricting foreign acquisition of AI assets that it considers within its own national security perimeter. Both moves operate on the same underlying logic, advanced AI capability is a national security asset, and both governments are asserting authority over its movement across borders.
What This Means for AI M&A Compliance
The practical implications are narrow but consequential. They do not affect the majority of AI M&A transactions, which involve targets with clean jurisdictional profiles. They are directly material for any transaction where the target company has a founding history that includes Chinese-registered entities, regardless of where the company operates today.
Due diligence teams conducting AI M&A work will need to add a founding provenance layer to their standard corporate structure review. The specific questions are:
Were any entities in the target’s founding corporate structure registered in China? Has the target maintained any ongoing operational, financial, or IP relationships with Chinese-registered entities since its founding? Does the target’s technology or data involve Chinese national security-relevant applications as defined under China’s NSR framework? Has any prior change-of-control transaction involving Chinese entities occurred in the target’s history that might have been subject to NSR review?
None of these questions are new to export control lawyers. What is new is that this ruling makes them directly relevant to transactions where the target has already been repositioned as a non-Chinese company. The repositioning does not dissolve the founding history.
The cross-jurisdictional compliance complexity of 2026 AI M&A is accumulating faster than due diligence frameworks are updating. Chinese NSR authority over founding provenance, US export controls on AI capabilities, and EU AI Act obligations on deployers are three separate regulatory vectors that can all apply to a single transaction. The Manus ruling adds a fourth dimension that was not in most practitioners’ standard checklist.
What We Don’t Know
The NDRC’s specific criteria remain undisclosed. Whether the blocking authority here rested primarily on the Beijing founding entity connection, on the nature of Manus’s agentic AI capabilities, on the identity of the acquirer (Meta specifically, given its scale and strategic position), or on some combination of all three is not publicly known. The NDRC does not publish its deliberations.
Meta has not responded. Whether the company is evaluating appeal options, a restructured transaction, or a strategic withdrawal is unknown. Deals of this size and complexity rarely end at the first prohibition without further legal activity.
Whether this action represents a new enforcement posture from the NDRC toward AI acquisitions specifically, or an application of long-standing NSR authority that happened to surface in a high-profile case, is a question the NDRC has not answered. The framework’s authority to reach this type of transaction has existed for years. The question of whether it will be applied consistently to similar deals, or whether the Manus case reflects specific facts unique to this transaction, will only be answered by the next comparable deal that enters China’s review process.
TJS Synthesis
The Manus ruling is not, by itself, a policy shift. China’s NSR framework has always had the reach to do what the NDRC did here. What the ruling does is make visible a jurisdictional theory, founding provenance over current domicile, that the AI M&A market had not been forced to test. It is now tested.
The strategic implication for compliance teams is this: the corporate relocation playbook that worked to simplify capital access and reduce operational friction has a regulatory seam in it. That seam runs along the line between where a company operates today and where it was founded. For AI acquisitions, China’s NSR framework can reach through that seam. Whether it consistently will is unknown. That it can has been demonstrated.