China Poised to Lift Travel Ban on Manus Founders

Manus returned to Google News after China reportedly prepared to lift travel restrictions on two founders, five months after they were told to remain inside China.

The Financial Times report concerns Manus CEO Xiao Hong and chief scientist Ji Yichao. Chinese officials reportedly restricted their departure during a review of Meta’s acquisition of the AI agent company. The reported change has not received a detailed public explanation from Chinese regulators.

The timing matters more than the travel decision alone. China ordered Meta and Manus to reverse their completed transaction in April. Meta then separated Manus from its internal systems, while former investors explored financing an independent company.

A lifted restriction would therefore mark the closing stage of an extraordinary corporate reversal. A Singapore relocation and foreign acquisition did not remove the startup from Beijing’s reach. Instead, Chinese authorities treated its technology, founders, and earlier development as strategically connected to China.

The immediate contest is not Manus against another AI agent. It is offshore corporate ownership against a government’s claim over technology developed within its innovation system. Meta’s deal tested that boundary, and Beijing asserted that corporate domicile did not settle it.

What Changed in the Manus Travel Ban Report

The reported travel decision suggests Beijing has secured its preferred corporate outcome and no longer needs the founders’ movement as leverage.

According to the Financial Times headline carried by Google News, China is poised to lift the restrictions placed on the Manus founders. “Poised” matters because it describes an expected action, not a completed or formally announced decision.

Neither a public order nor a detailed regulatory statement accompanied the report when this article was prepared. Readers should therefore treat the expected lifting as reported information that still requires official confirmation.

The restrictions began during a much larger dispute. Chinese officials summoned Xiao Hong and Ji Yichao to Beijing in March, according to Reuters’ account of the regulatory review. They were questioned about the transaction that transferred Manus to Meta.

The founders were reportedly free to move within China but could not leave the country. Some reports described the measure as official guidance rather than a formally declared exit ban. That distinction affects legal terminology, but it did not change the practical result.

Hong and Ji were based in Singapore, where Manus had moved its headquarters and much of its workforce. Preventing them from returning created pressure on both the individuals and the acquisition they had completed.

The restrictions also arrived after China began examining whether the transaction complied with rules governing outward investment, technology exports, and cross-border data transfers. Those are separate regulatory pathways, although each can affect a technology acquisition.

China’s Ministry of Commerce had confirmed an investigation in January. A professor quoted in coverage of the initial inquiry identified technology export licensing as a central unanswered question.

The government escalated its response on April 27. In a published security-review decision, the National Development and Reform Commission prohibited the foreign acquisition of the Manus project and required the parties to withdraw the transaction.

That order did not provide a detailed public explanation. It cited applicable laws and regulations through the government body responsible for foreign investment security review.

The reported lifting now follows visible progress toward unwinding the acquisition. Meta separated Manus employees from internal systems and stopped its staff from using Manus tools for new work. The companies also halted data sharing, according to subsequent reports.

This sequence makes the change easier to interpret. The founders were restricted while Beijing examined the transaction and demanded its reversal. Their expected release comes after operational separation and negotiations over new ownership advanced.

However, correlation does not establish the precise bargain. No public source has confirmed that travel freedom was formally conditioned on a buyback, data separation, or another concession.

That verification gap should remain central to any assessment. The report signals movement, but it does not disclose every condition attached to the decision. It also does not establish when the founders can depart or where they will travel first.

For Manus users, the event is not merely diplomatic theater. Leadership mobility affects fundraising, customer negotiations, hiring, and the company’s ability to coordinate across Singapore and China.

For other founders, the message is broader. A company can change headquarters, establish an offshore parent, and move employees. Those steps do not guarantee that regulators will view its technology as detached from its country of origin.

Why the Google News Headline Matters Beyond Manus

The Manus case turns a travel restriction into a test of who controls mobile AI talent and intellectual property.

The primary Google News headline sounds personal because it focuses on two founders. The underlying dispute concerns ownership of an AI business that both China and Meta considered strategically valuable.

Manus develops a general-purpose AI agent. An agent is software that can plan and perform multi-step tasks, instead of only returning a conversational answer.

The product can work across activities such as research, coding, data analysis, and document preparation. Manus presents those actions through a consumer service rather than selling only a foundation model.

That distinction helped make the company attractive to Meta. AI agents can connect a model’s reasoning to commerce, advertising, customer support, and business workflows. Those functions align with Meta’s effort to put AI assistants across its applications.

Meta announced the acquisition in December 2025. The companies did not publicly disclose a price, although multiple reports placed the value around $2 billion or higher.

Before the deal, Manus said it had crossed $100 million in annual recurring revenue eight months after launch. Annual recurring revenue is a subscription business’s current revenue run rate projected across one year.

That number was a company-reported metric, not an audited annual result. Even with that limitation, it helped explain why Meta wanted the team and product.

The acquisition also came amid Meta’s wider spending on AI talent. It had made a large investment in Scale AI and recruited CEO Alexandr Wang to help lead its superintelligence work.

Manus offered something different. Scale centered on data infrastructure and research leadership. Manus brought an agent product with paying users and an established workflow interface.

Beijing saw another dimension. Manus had moved to Singapore, but its founders established predecessor entities in Beijing. Chinese investors had also supported its earlier development.

Meta said the transaction would leave no continuing Chinese ownership in Manus. It also said the company would stop services and operations inside China.

Those terms strengthened Meta’s separation argument. They simultaneously sharpened Beijing’s concern that Chinese-developed capabilities, talent, and economic value were transferring to a United States company.

The NDRC’s order demonstrated that incorporation alone would not decide jurisdiction. China examined the technology’s origin, its earlier corporate structure, its investors, and the movement of its team.

That approach creates pressure for founders long before an acquisition. Startups must now think about export reviews, data transfer controls, investor rights, employee locations, and regulatory approval while designing their corporate structure.

Investors face similar questions. A Singapore entity can provide access to global capital and customers. It cannot eliminate every claim created by earlier research, local subsidiaries, Chinese investors, or resident founders.

The case also places Singapore in a difficult position. The country remains a major base for companies operating across Asian and Western markets. Manus shows that Singapore incorporation cannot neutralize disputes between larger governments.

For American buyers, the lesson is equally direct. Acquiring a foreign company’s offshore parent may not provide clean control of technology that another government considers strategically domestic.

This complicates due diligence. Lawyers can verify ownership documents and corporate registrations. They must also examine where intellectual property was created, who developed it, and which approvals regulators may assert afterward.

The result is a new kind of acquisition risk. A completed deal can still unravel after employees move, systems connect, and investors receive proceeds.

Offshore Ownership Met Beijing’s Strategic Technology Claim

The central conflict is whether a startup’s legal home outweighs the national origin of its technology and talent.

Manus had a plausible case for being treated as a Singapore company. Its parent was based there, most employees reportedly worked there, and the service targeted customers outside China.

Meta also said the acquisition complied fully with applicable law. The company expected an appropriate resolution after China announced its investigation.

Beijing reached a different conclusion. Its April order prohibited the foreign investment and required the acquisition to be withdrawn. The decision established a practical claim over the transaction, regardless of Meta’s legal position.

The government did not publish a detailed technical analysis. That absence leaves uncertainty about which factor carried the greatest weight.

One possibility involves restricted technology exports. China maintains controls covering technologies that cannot leave the country without approval. Officials can examine whether software, training methods, or other know-how falls within those rules.

Another possibility concerns outward investment reporting. Regulators reportedly questioned whether the founders complied with requirements connected to transferring Chinese-linked assets into an offshore structure.

Cross-border data presents a third path. AI products can involve user information, training data, operational records, and model interaction logs. Moving those materials can trigger separate security and privacy reviews.

The case may also reflect a broader policy judgment. China increasingly treats advanced AI talent and intellectual property as national strategic assets, especially during competition with the United States.

The public record does not establish which argument was decisive. It does show that Beijing was prepared to enforce its position through both corporate and personal pressure.

This is the reversal at the center of the story. Manus moved outward to secure global customers and reduce geopolitical friction. The move instead became evidence in a dispute about whether strategic technology was being moved beyond China’s control.

Meta’s purchase then amplified the issue. A transfer to a neutral financial buyer might have drawn less attention. A sale to one of America’s largest technology platforms created a direct strategic contrast.

The acquisition promised Meta control of Manus’s product, team, and commercial momentum. China’s order turned that promise into an expensive separation exercise.

By June, Meta had reportedly cut Manus employees off from internal systems. Meta staff could no longer use Manus for internal work, and active projects were moved toward Meta’s own infrastructure.

That operational separation addressed one part of Beijing’s demand. It reduced ongoing data exchange and made future integration less likely.

It could not undo everything that happened after the acquisition. Employees had already entered Meta offices, teams had interacted, and technical knowledge may have crossed organizational boundaries.

No public account explains how regulators will verify that separation. Code repositories, model access, internal documentation, and employee knowledge all require different controls.

Software is particularly difficult to return. Physical assets can change ownership through signed documents. Technical ideas remain with the people who reviewed, discussed, or implemented them.

That reality makes the Manus outcome more complex than a canceled merger. The transaction reportedly closed before the Chinese order, so the parties must reverse ownership after integration began.

Former investors also complicate the process. Some had already received acquisition proceeds. A buyback requires new financing, renewed governance agreements, and a workable ownership structure.

Reports said the founders explored raising roughly $1 billion from outside investors. They could contribute additional personal funds to reach a valuation matching Meta’s purchase price.

Former backers, including Tencent, were later linked to discussions about buying Manus back. Those negotiations were reported and remained subject to change.

A new Chinese joint venture has also been discussed as a possible structure. Such an arrangement might satisfy Beijing’s demand for domestic influence while allowing Manus to maintain operations abroad.

Yet that solution would revive the problem Manus tried to avoid. Greater Chinese ownership may increase scrutiny from American regulators, business customers, and infrastructure partners.

The company therefore faces two incompatible compliance environments. China wants assurance that strategic technology has not escaped its control. Western stakeholders may demand proof that the same company remains independent of Chinese state influence.

A lifted travel restriction would resolve the founders’ immediate mobility problem. It would not resolve this structural conflict.

What the Report Does Not Confirm

The expected release should not be mistaken for proof that the acquisition, buyback, or product transition is complete.

The Financial Times headline offers one reported development. It does not identify a public decision date, describe binding conditions, or confirm the founders have already departed China.

That leaves several verification questions.

First, it remains unclear whether both founders will receive unrestricted travel rights at the same time. Authorities could also attach reporting requirements or other conditions that are not publicly disclosed.

Second, the ownership transition remains opaque. Reports have described financing talks, investor consortiums, and an eventual independent Manus. A final shareholding structure has not been comprehensively published.

Third, Meta’s technical separation needs independent verification. The company reportedly removed Manus from internal systems, but no public audit has documented which data or code moved during integration.

Fourth, the customer impact is still unfolding. Manus users have reported receiving notices connected to account backup, temporary access changes, and service restoration.

Those reports are useful signals, but they are not substitutes for a complete company disclosure. Customers should rely on notices displayed inside their accounts and official support instructions.

The product’s current website has also continued to describe Manus as part of Meta during portions of the unwind. Public-facing language can lag behind legal and operational changes, so it cannot settle ownership alone.

This uncertainty matters for enterprise buyers. They need to know which entity processes their information, where data is stored, and which company can access agent activity.

An AI agent may handle source documents, account credentials, private research, and business outputs. Ownership changes can alter the legal parties responsible for those materials.

Customers should review data retention settings and export essential work when a provider announces a migration. A local knowledge base can reduce dependence on one cloud service’s account structure.

The Manus case also raises a larger product continuity risk. AI agents often depend on several external models, cloud providers, browser environments, and business integrations.

Changing ownership can disrupt any layer. A new investor may renegotiate infrastructure contracts. A forced separation may remove access to internal models or deployment systems.

Meta also faces unanswered questions. The company wanted Manus to support business agents across its platforms. It must now decide whether to rebuild those functions internally, acquire another team, or use partnerships.

OpenAI, Google, Anthropic, and several smaller companies continue developing tools that perform actions beyond chat. Manus’s separation does not slow that competition.

It may instead weaken Meta’s position. The company loses direct control of a mature agent product while rivals keep shipping their own computer-use and research capabilities.

However, Manus also loses Meta’s distribution and resources. Independence can preserve the product’s identity, but it creates new financing and infrastructure requirements.

That tradeoff makes celebratory interpretations premature. A founder leaving China does not mean the business has escaped regulatory pressure. It may mean the company has accepted a structure that Beijing considers acceptable.

There is also no basis for assuming the founders faced criminal charges. Reporting focused on regulatory questioning and restrictions connected to the transaction.

Language matters here. Describing the restriction as punishment states a motive that authorities have not publicly confirmed. Describing it as leverage is an inference from the sequence of events.

The safest conclusion is narrower. Beijing restricted the founders during its review, ordered the deal reversed, and is reportedly preparing to restore their ability to travel as the unwind advances.

That sequence is still consequential. It demonstrates how personal mobility can become entangled with corporate compliance in strategic technology disputes.

What to Watch After the Latest Google News Report

Three signals will show whether the reported travel change closes the Manus dispute or only moves it into another phase.

The first signal is direct evidence that Xiao Hong and Ji Yichao have left mainland China. A public appearance in Singapore would provide clearer confirmation than another anonymous report.

Such an appearance would strengthen the view that authorities consider the immediate compliance dispute resolved. Continued absence would weaken that interpretation, especially without an official release order.

The second signal is a final ownership disclosure. Investors, customers, and employees need to know who controls Manus after Meta withdraws.

Reports have linked former backers to a proposed buyback. Those discussions do not establish the final valuation, investor group, voting rights, or governing jurisdiction.

A completed financing round would strengthen the independent-company narrative. It would also show which investors are willing to carry geopolitical and operational risk.

A structure dominated by Chinese capital would satisfy one concern while raising others. United States regulators could revisit the company’s access to American infrastructure, models, or customers.

A broadly distributed consortium might provide more balance. It could also make governance slower when Manus needs to rebuild quickly.

The third signal is stable product operation after the separation. Users should watch account access, data restoration, model performance, integrations, and published privacy terms.

A smooth transition would show that Manus can operate without Meta’s internal infrastructure. Persistent outages or missing features would reveal deeper dependencies created during the acquisition.

Enterprise customers should also watch the identity of the contracting entity. Terms of service and privacy notices can reveal an ownership transition before marketing pages are updated.

These three signals matter more than speculation about political reconciliation. Travel, ownership, and product continuity are observable. They can confirm whether the corporate reversal works in practice.

The next one to three months will also clarify Meta’s response. The company may disclose how it plans to replace the agent capabilities it expected Manus to provide.

Meta’s earlier statement said its transaction complied with applicable law. The April order and subsequent operational separation show that legal compliance can remain contested across jurisdictions.

Other technology buyers will respond quietly. They may add closing conditions for Chinese regulatory approval, even when a target is incorporated elsewhere.

As Reuters reported in its analysis of the blocked transaction, the decision heightened the risks facing global investors considering advanced-technology companies with substantial ties to China.

Investors may also demand earlier reviews of technology provenance. Where code was written can become as important as where the holding company is registered.

Founders will face difficult timing choices. Moving a company early can reduce later entanglements, but relocation brings costs and does not erase earlier connections.

Waiting creates a different risk. A successful product can become strategically important before its corporate structure is prepared for cross-border investment.

The Manus case gives both governments reasons to harden their positions. China demonstrated that it can challenge the offshore sale of Chinese-developed AI. The United States can answer with stronger investment and technology restrictions.

That escalation would make neutral operating models harder to sustain. Startups may have to choose markets, investors, and infrastructure earlier than their business plans require.

For knowledge workers, the lesson is practical. AI services can change control with little warning, and regulatory disputes can affect access to stored work.

Users should maintain exports of important projects, document model dependencies, and avoid keeping the only copy of critical material inside one agent platform. A personal knowledge system can make provider changes less disruptive.

The Google News report is therefore not simply an update about two founders. It marks the possible end of the personal restrictions used during a cross-border technology confrontation.

The corporate conflict remains active until Manus publishes stable ownership, Meta completes its withdrawal, and customers can verify uninterrupted control over their data.

Watch those outcomes, not only the founders’ first flight. If all three arrive, Beijing’s pressure campaign achieved a complete reversal. If one fails, the Manus dispute has only changed form.