Meta has started dismantling its reported $2 billion acquisition of Manus, marking the clearest sign yet that the deal is being unwound under pressure from Chinese regulators. According to reporting from Bloomberg and the Wall Street Journal, the two companies are now in the process of separating operations, cutting off shared systems and preparing for what could become a complete reversal of one of the most closely watched AI transactions involving China-linked technology.
The move follows a divestiture order from Beijing issued about two months ago on national security grounds, according to the source material. It also highlights a much wider shift in China’s approach to artificial intelligence: authorities are no longer just policing chip exports and model training, but are now moving deeper into ownership structures, capital flows and the cross-border movement of talent.
For Meta, the reported deal unwind is a striking turn after what was framed as a high-profile expansion into agentic AI. For Manus, the startup that drew global attention with a viral demo, the situation threatens to upend a fast-moving rise that had already taken the company from relative obscurity to the center of a geopolitical fight over the future of AI.
What Meta has reportedly done so far
The most immediate step appears to be an operational split. Bloomberg reported that Meta has cut Manus off from internal systems, preventing employees from using Manus software or tools on internal projects as the separation proceeds. The companies are also said to have stopped sharing data, which is a crucial move in any breakup involving software infrastructure, product development and model integration.
That matters because AI acquisitions often begin with tight technical coupling. Data pipelines, product interfaces, internal testing environments and support systems can all be intertwined very quickly after a deal closes. Once those links are severed, reversing course becomes more than a legal exercise; it turns into a practical engineering project.
Meta has not publicly described the process in detail, and neither Meta nor Manus responded immediately to requests for comment outside normal business hours, according to the source material.
Why Beijing stepped in
The central reason for the unwind is Beijing’s intervention. Chinese regulators reportedly ordered the deal to be divested roughly two months ago, citing national security concerns. The case shows how China’s government is increasingly treating AI not only as a commercial opportunity but also as a strategic asset that should remain under state oversight, especially when foreign ownership or offshore structures are involved.
The pressure appears to have been triggered by Manus’ corporate background and the cross-border nature of the transaction. Manus was founded in China through a parent company called Butterfly Effect, but its structure, personnel and financing profile drew scrutiny in both China and the United States. That made it vulnerable to criticism from two directions: Chinese officials worried about technology leakage, while U.S. policymakers and observers raised questions about American money flowing into a Chinese-linked AI company.
Beijing’s response suggests that, in its view, the location of incorporation matters less than the strategic sensitivity of the underlying technology.
That distinction is increasingly important. As AI systems become capable of carrying out multi-step tasks, accessing enterprise data and serving as digital labor, governments are looking at them less like ordinary software and more like infrastructure with national-security implications.
How the Manus deal became a geopolitical flashpoint
Manus attracted extraordinary attention after a viral demo showcased its agentic AI capabilities. The startup became part of a broader wave of excitement around AI agents: systems designed not just to answer questions, but to complete tasks, browse tools, connect services and act on behalf of users. In a market crowded with chatbot announcements, that made Manus stand out.
The company then relocated staff to Singapore in mid-2025 before Meta announced the acquisition in December, according to the source material. That move likely reflected an effort to position the company for global expansion and cleaner international operations, but it did not prevent regulators from examining the deal more closely.
Chinese authorities later scrutinized the transaction, citing possible violations of export controls and foreign investment rules. In practical terms, the concern seems to have been that advanced AI capabilities, especially those with agentic or enterprise-facing functions, should not be allowed to move too freely across borders without direct oversight.
The result is a transaction that briefly symbolized the globalization of Chinese AI talent and startup value creation, but is now becoming a case study in how fragile such deals can be when they intersect with state policy.
What happens to Manus now
Even as Meta steps away, Manus is not standing still. The startup has reportedly continued to launch new features, including integrations with Similarweb and Shopify. That suggests the team is still trying to position the product as a useful, evolving platform rather than a company defined solely by a failed deal.
According to earlier reports cited in the source material, Manus co-founders have been in preliminary discussions about raising about $1 billion from outside investors in an effort to buy back the startup from Meta. If that effort advances, it could lead to a new ownership structure involving a Chinese joint venture and, eventually, a listing in Hong Kong.
That possibility is especially notable because Hong Kong has become a favored venue for Chinese AI companies seeking public-market capital. The city has seen a wave of listings this year from firms such as MiniMax and Zhipu, reflecting investor interest in AI despite broader geopolitical tensions.
Whether Manus can follow a similar path is uncertain. The company would need to satisfy regulators, reassure investors and rebuild a clean ownership story after a deal that has now become politically sensitive on both sides of the Pacific.
China is tightening its AI grip beyond one company
The Manus case does not exist in isolation. It arrives alongside a series of broader restrictions that show Beijing is tightening control over the AI sector at multiple levels.
According to the source material, Chinese authorities have expanded travel restrictions for researchers and executives at private firms, requiring government approval before they can travel abroad. That is a major signal that the state wants tighter visibility into who is moving, where they are going and what expertise may leave the country.
At the same time, reports indicate that top AI companies including Moonshot AI, StepFun and ByteDance may need government approval before taking U.S. investment. That would add another layer of supervision over foreign capital, making it harder for American money to flow into strategically important AI businesses without state review.
Taken together, these measures suggest Beijing is pursuing a comprehensive policy: keep key AI talent close, keep foreign capital filtered and keep sensitive technologies under domestic control. The Manus unwind fits that pattern neatly.
What Beijing appears to be prioritizing
- Control over strategic AI technology
- Limits on foreign ownership and foreign funding
- Closer supervision of executives and researchers who travel internationally
- Protection of domestic AI champions from politically sensitive cross-border deals
That policy direction could reshape how Chinese startups raise money, expand abroad and structure acquisitions with foreign buyers.
Investors are already cashing out
Despite the continuing uncertainty, some financial pieces of the deal appear to have already been settled. The source material says Manus investors, including California-based Benchmark, have received their proceeds from the acquisition. Asian backers such as Tencent, HSG and ZhenFund have reportedly indicated they will cooperate with the unwind.
That means the cleanup is not just theoretical. Capital has already moved, and the parties involved are now dealing with the consequences of a transaction that may not survive its own political scrutiny.
For investors, the situation underscores a growing reality in cross-border AI deals: even when the money clears and the paperwork is signed, the final outcome may still depend on regulators whose priorities are shaped by strategic rivalry rather than by market logic alone.
Why this matters for the global AI market
The potential collapse of the Meta-Manus deal is about more than one startup. It reflects a broader collision between the speed of AI innovation and the slower, more forceful logic of national policy. AI companies can scale globally in months, but governments can reassert control just as quickly once they view a transaction as too sensitive.
For U.S. and Chinese companies alike, the episode is likely to reinforce several lessons. First, ownership structures that look workable in ordinary venture-backed technology may become liabilities when the technology involved is considered strategically important. Second, cross-border deals can trigger separate, and sometimes conflicting, concerns from multiple governments. Third, the market for AI talent is becoming inseparable from debates over sovereignty, capital access and export control.
The unwind could also chill future acquisition discussions involving Chinese-founded AI startups with overseas ambitions. If firms believe a deal may later be blocked or reversed, they may lean more heavily toward minority investment, local joint ventures or regionally separated business lines instead of full takeovers.
Potential market consequences
- More caution around Chinese AI acquisitions by U.S. companies
- Greater use of joint ventures instead of direct control
- Increased scrutiny of capital sources for AI startups
- Higher demand for “clean” corporate structures before fundraising or M&A
A timeline of the Manus-Meta saga
| Date | Event | Significance |
|---|---|---|
| Mid-2025 | Manus relocates staff to Singapore | Signals a push toward international expansion and a more global operational base |
| December 2025 | Meta announces a $2 billion acquisition of Manus | Positions Manus as one of the most valuable emerging AI startup deals |
| Early 2026 | Chinese regulators begin scrutinizing the transaction | Raises concerns about export controls and foreign investment rules |
| Around April 2026 | Beijing reportedly orders divestiture | Marks the start of the forced unwind on national security grounds |
| June 2026 | Meta begins operational separation and data shutdown | Most concrete evidence yet that the deal is being dismantled |
The U.S.-China AI rivalry is moving into dealmaking
The Manus episode also says something important about the next phase of the U.S.-China technology rivalry. The competition is no longer limited to chips, model performance or app launches. It is now reaching into venture deals, acquisition structures, executive travel and overseas listings.
That means AI companies are not only competing to build better products; they are also navigating a legal and political landscape in which corporate ownership itself can become a strategic issue. In that environment, the most promising startups may find that growth is constrained not by model quality, but by where their founders were born, where their investors are based and which government has the final say.
Manus is a vivid example. What looked like a landmark exit for Chinese AI now appears to be collapsing into a forced separation, with regulators in Beijing asserting control and Meta retreating from the transaction.
The episode captures a new reality for AI dealmaking: in the era of strategic technology, regulators can be as influential as venture capitalists.
What to watch next
The next developments will likely determine whether Manus can recover as an independent company or through a new investor consortium. The most important questions now are whether the co-founders can assemble enough capital to buy back the startup, whether Beijing will permit a revised ownership arrangement and whether Hong Kong becomes the eventual listing venue that backers are reportedly considering.
Another open question is how Meta will reposition itself after the unwind. The company has spent heavily to expand its AI capabilities, and a failed deal of this scale may push it toward more cautious partnerships, internal builds or minority stakes rather than outright acquisitions in geopolitically sensitive markets.
For now, the story stands as one of the clearest examples yet of how AI dealmaking can be reshaped by state power. What began as an ambitious purchase has become an ordered retreat, with China, not the market, setting the terms of the exit.
As Beijing tightens its oversight of capital, talent and technology, the Manus case may prove to be an early sign of a broader pattern: the era of borderless AI expansion is giving way to one in which every major transaction must pass a geopolitical test.
