Meta's acquisition of Manus halted; who will take over?
Core judgment
Preference for a model involving independent operation plus minority equity, contingent on resolving asset ownership and team contractual issues
Top recommendation
Immediately push for the inclusion of technical asset inventories and data compliance plans in regulatory pre-review, while simultaneously codifying business isolation clauses in shareholder agreements
Where they stand
Support 54.9% · Oppose 23.6% · Neutral 15.3% · Wait and see 6.2%
Based on 144 simulated statements by 19 virtual roles — not a real poll, and not the actual positions of these organizations or people
Camps and reasons
- Support: Leadership by local capital can fill gaps in the AI Agent ecosystem and ensure independent team operations for survival
- Oppose: Manus technology is deeply integrated into Meta's product lines, constituting an irreversible coupling where assets cannot be separated
- Neutral: Transactions must undergo legal review, with clear technical asset ownership and adequate data compliance, before proceeding
Biggest risks
- Hollowing out of technical assets: Core engineering capabilities are lost due to deep integration; what is acquired is merely a soulless shell
- Secondary cross-border compliance review: Squeeze play from security reviews by both China and the US could lead to indefinite Shelve of the transaction at the approval stage
- Collapse of team and control rights contracts: Verbal commitments cannot counteract ecosystem gravity; strategic divergence may lead to the departure of core team members and value destruction
Key uncertainties
- Whether the US will initiate a secondary review targeting technology repatriation (watch: Monitor whether CFIUS or BIS releases special review guidelines and changes in upstream supplier service terms)
- Compatibility between core team stability and control rights arrangements (watch: Monitor Singapore team turnover rates and the signing of draft shareholder agreements containing governance details)
- Compliance anchor points and payment paths for the $2 billion valuation (watch: Monitor issuance of asset valuation reports by licensed institutions and abnormal movements in cross-border fund approvals)
In their own words
The quotes below are what the virtual roles said in the simulation, reproduced verbatim; they are simulated dialogue, not real statements by these organizations or people.
“Manus must belong to Meta.” — Mark Zuckerberg (virtual role)
“Maintain control before discussing next steps. Who supports this, and who opposes?” — Zhang Tao (virtual role)
“Hurry up and take over; we really can't afford to wait any longer🙏” — Manus employee (virtual role)
“For those hoping to snap up bargains while Meta exits, save your breath.” — Xiao Hong (virtual role)
“Preserving operations and retaining staff is our bottom line.” — butterfly effect (virtual role)
“The priority liquidation right is non-negotiable. Anyone can take over, but the terms must pass our approval first.” — Benchmark (virtual role)
In December 2025, Meta announced the acquisition of AI agent company Manus for approximately $2 billion, but the deal was officially halted by China's National Development and Reform Commission in April 2026 on grounds of national security and technology flow risks, becoming the first foreign M&A case in the AI sector publicly prohibited since the implementation of the "Measures for the Security Review of Foreign Investment." As of July 30, 2026, market focus has shifted to rumors that a Chinese consortium led by Tencent intends to take over the stake at the same valuation. This simulation covers 19 key stakeholders, including regulatory bodies in China and the US, parties to the transaction, potential acquirers, and relevant industry chain participants, generating a cumulative total of 144 posts. Overall public opinion presents a clear trend of pragmatic survival coexisting with cautious compliance: Manus's parent company, employees, and most local capital explicitly support the notion that "another party should take over," viewing it as the only path to existence; Meta and some founders oppose this due to the depth of technology integration and demands for control; both Chinese and US regulatory authorities maintain procedural neutrality, emphasizing legal review rather than preset outcomes. The core contradiction in the current situation has transformed from a purely commercial M&A game into a complex negotiation seeking a balance point between regulatory red lines, technological sovereignty, and asset preservation. Label explanation: In the report, "material" refers to content from user-provided materials and the knowledge graph; "verified online" refers to external data found via online search during this session (with source URLs attached); "simulation" refers to statements and figures from virtual characters during the simulation, which are fictional Deduction s intended only for inspirational reference.
Executive brief
Core judgment: Under the base case, the takeover transaction is inclined to be implemented within a compliant framework through an independent operation plus minority equity model, but this is premised on resolving the contractual issues of technology asset ownership and team stability, rather than simply inheriting the original acquisition plan.
Stance landscape: Most public opinion supports a takeover led by local capital to ensure survival, and the founding team is also willing to cooperate. However, Meta takes a hardline stance, and there are still significant disagreements among supporters regarding control arrangements and valuation rationality.
Greatest risk: Hollowing out of technology assets, meaning core engineering capabilities have been substantially lost due to deep integration. This means that even if equity changes hands, what is acquired is merely an empty shell devoid of its soul.
Turning-point signals: If licensed institutions issue asset value reports, abnormal movements appear in cross-border fund approvals, or regulators release compliance guidelines for AI technology imports and exports, indicating that the situation is deviating from the base-case judgment.
Primary recommendation: Immediately push to include the technology asset inventory and data compliance plans in regulatory pre-review, while simultaneously solidifying legally binding business isolation clauses in the shareholder agreement.
This brief is a summary of the full text's conclusions; see the corresponding chapters in the main body for the basis and sources of each judgment.
Distribution of key stakeholder stances: Coexistence of consensus on survival and divergence on control rights
This chapter focuses on the distribution of stances regarding the proposition that 'another party should take over Manus, which Meta has halted.' The simulation statistics show that among 144 valid posts, supporters account for 54.9% (79 posts), forming the current mainstream public opinion; opponents account for 23.6% (34 posts), primarily coming from Meta and its associated stakeholders; neutral and wait-and-see parties combined account for 21.5%, mostly regulatory agencies and some financial investors. The overall pattern presents a landscape of 'strong consensus for survival, but sharp divergence in control ownership.' (Source: simulation)
Support Camp: Local Capital and Founding Team Form a 'Survival Alliance'. A simulation stakeholder modeled after Tencent explicitly expressed willingness to lead the takeover. Its core logic is not bottom-fishing, but strategic positioning to fill gaps in the general AI Agent ecosystem. It committed to keeping Manus operating independently post-transaction, with Tencent acting only as the largest minority shareholder. Stakeholders such as Manus parent company Butterfly Effect and founders Xiao Hong, Zhang Tao, and Ji Yichao highly supported a Chinese takeover in the simulation, but set rigid prerequisites: the founding team must retain voting and veto rights, product lines must not become appendages of giants, and technical routes must not be halved. They generally believe that given regulators have officially halted Meta's acquisition, Tencent is the only pragmatic choice combining financial strength, ecosystem synergy, and respect for team independence. Early-stage investors like Sequoia China and ZhenFund also tend to support this plan, viewing it as a window for compliant exit and track positioning, but emphasizing that valuation needs renegotiation and priority liquidation rights must be guaranteed. (Source: simulation)
Opposition Camp: Meta Insists on 'Indivisibility of Assets'. A simulation stakeholder modeled after Mark Zuckerberg repeatedly emphasized throughout 8 rounds that 'Manus must belong to Meta.' Its core basis is that Manus technology has been deeply integrated into Ads Manager within 7 weeks and spawned a Business Agent, constituting irreversible technical coupling. This stakeholder denied claims of 'exit' or 'weakening bargaining chips,' arguing that regulatory halting is merely a compliance procedure rather than transaction termination, and questioned whether Tencent's ecosystem could truly fulfill the 'independent operation' commitment. Notably, while some Manus employee stakeholders emotionally leaned toward Tencent taking over for stability, they showed strong anxiety regarding Meta's stance of 'delaying, not abandoning,' reflecting grassroots fears that a transaction deadlock might lead to team dissolution. (Source: simulation)
Wait-and-See and Prudent Camp: Regulators and State-Owned Capital Focus on Compliance Boundaries. Simulation stakeholders modeled after China's NDRC and Ministry of Commerce did not directly express support or opposition to specific takers, but repeatedly emphasized the principle of lawful review: if the transaction structure is clear, technical asset ownership is defined, data compliance is in place, and foreign access red lines are not touched, it belongs to normal commercial behavior; however, no formal application materials have been received yet, making prediction difficult. The Wuhan state-owned capital stakeholder acknowledged Manus's industrial value but held a prudent attitude toward the $2 billion valuation, pointing out that Meta has already integrated some technologies, making divestiture uncertain, and clarifying the boundaries of receivable assets is needed to avoid taking a 'hot potato.' External stakeholders like Anthropic and US Senator John Cornyn raised concerns from the perspective of AI safety and technology repatriation risks, implying that even with a Chinese takeover, secondary review pressures regarding cross-border technology transfer remain. (Source: simulation)
In summary, although 'taking over by another party' receives majority support in quantity, there are significant internal divergences among supporters regarding control arrangements, valuation reasonableness, and feasibility of technical asset segmentation; meanwhile, Meta's hardline stance and the incomplete status of regulatory procedures keep the transaction in a highly uncertain range. Verified online materials also corroborate that Tencent's proposed plan is indeed designed around 'largest minority shareholder + independent operation' to avoid compliance risks triggered by single-entity control (Source: wap.eastmoney.com), which aligns highly with the focus of Game (game) among various parties in the simulation.
Key points of divergence and key uncertainties: Technical separation difficulty and regulatory approval pathways
Feasibility of Technical Severance: Cognitive Conflict Between Deep Coupling and Ordered Decoupling
The primary divergence hindering the completion of the takeover transaction lies in whether Manus's technical assets possess the physical and legal basis for complete separation from the Meta ecosystem. Simulation roles based on Meta insist that Manus technology has been integrated into Ads Manager within seven weeks and has spawned Business Agent; this deep embedding implies that forced severance would lead to "value destruction," with disputed intellectual property rights over derivative outputs (Source: simulation). However, simulation roles based on Tencent and the Manus founding team counter that Meta's integration constitutes secondary development based on acquisition expectations, the underlying general-purpose AI agent architecture had already taken shape prior to the acquisition, engineering-level integration can be separated through "ordered decoupling," and they explicitly acknowledge that derivative outputs like Ads Manager belong to Meta, claiming only the return of underlying technology and data sovereignty (Source: simulation). The essence of this divergence is: Is the boundary of technical assets defined by the "degree of physical code mixing" or by "legal ownership and R&D timeline"? Without reaching a technical consensus on severance standards, any valuation negotiation will lack a subject-matter basis.
Finality of Regulatory Efficacy: Misalignment Between Administrative Bans and Commercial Negotiations
The second core point of contention is whether the legal effect of the NDRC's ban on transactions on April 27, 2026, has been fully internalized by all parties. Although source materials indicate the NDRC has formally ordered the termination of the transaction, it should be noted that the specific date and document number have not yet been verified against original announcements in authoritative channels (the original wording in the material was inaccurate; verification failed to confirm).
In the simulation, roles based on Chinese regulatory authorities emphasize that this decision carries legal binding force and is not a negotiable "suspension procedure"; whereas roles based on Meta continue to use phrasing such as "transaction shelved" and "internal review," attempting to retain leverage for priority acquisition rights or control (Source: simulation). This cognitive misalignment leads to an unstable negotiation foundation: the Chinese consortium views the takeover as "disposing of assets according to law," while Meta still tends to view it as a "pause button in commercial Game." As long as Meta has not fully accepted the fact of the transaction's termination at the legal level, the due diligence and closing processes for the takeover party cannot substantially commence.
Key Uncertainty I: Security Review Loop for Cross-Border Technology Repatriation
Even if the Chinese consortium and Meta reach a commercial agreement, whether the U.S. side will initiate secondary reviews targeting "technology repatriation" remains the largest external variable. In the simulation, a role based on U.S. Senator John Cornyn explicitly states that there must be scrutiny of whether any U.S. intellectual property, training methods, or proprietary datasets flowed into the Manus architecture during Meta's integration period; these "U.S. technology traces" could trigger export control jurisdiction (Source: simulation). A role based on Anthropic also reiterates that service restrictions on Manus will not be lifted until the safety boundaries of AI capabilities and isolation mechanisms are confirmed to meet standards (Source: simulation).
- Watch for: Whether the Committee on Foreign Investment in the United States (CFIUS) or the Bureau of Industry and Security (BIS) under the Department of Commerce issues special review guidelines or inquiry letters regarding the Manus case; observe whether upstream suppliers like Anthropic adjust service terms for entities associated with Manus.
Key Uncertainty II: Compatibility of Core Team Stability and Control Arrangements
Whether the takeover plan can land ultimately depends on whether approximately 100 core technical personnel, already integrated into Meta's Singapore office, are willing to return or accept a new structure. The simulation shows that Manus employee roles expressed anxiety over long-term uncertainty, desiring to end the turbulent state as soon as possible; while the founding team, although expressing support for the Chinese takeover, insists on "independent operational rights" and "team control" as non-negotiable prerequisites (Source: simulation). Meanwhile, a role based on Tencent promised to hold no more than 50% equity and maintain Manus's independent operation, but this still creates subtle tension with the absolute control demanded by the founders. If personnel placement plans and control clauses cannot be finalized simultaneously, the "people" and "power" of the technical assets face dual risks of loss.
- Watch for: Personnel turnover rates and survey results on willingness to return among Manus's Singapore team;Watch for whether Tencent and the founding team sign draft shareholder agreements containing detailed corporate governance rules.
Key Uncertainty III: Compliance Anchors and Payment Paths for the $2 Billion Valuation
Although multiple parties mention that the takeover valuation of approximately $2 billion equals Meta's original acquisition price, this transaction information has not been officially disclosed or confirmed (this data point has not been traced to its original source). More critically, under the current foreign exchange management and cross-border investment regulatory framework, whether the path for such large sums of capital to exit or be swapped domestically is smooth. In the simulation, roles based on Wuhan State-owned Assets and Sequoia China both emphasized the need to prudently assess valuations based on due diligence results to avoid taking on a "hot potato" (Source: simulation). If asset pricing after regulatory suspension lacks support from independent third-party assessments, or if cross-border fund arrangements fail to pass compliance pre-review, the transaction may stall at the final stage due to "money not passing through" or "price non-compliance."
- Watch for: Whether licensed assessment institutions issue value reports for Manus assets; monitor abnormal movements related to this case in the approval dynamics of cross-border RMB settlement or QDLP/QDIE quotas.
Risk warning: Triple fracture risks that may be triggered by forced advancement
If the ongoing technical separation disputes and regulatory compliance requirements are ignored, and the acquisition is forced through or the negotiation deadlock is allowed to persist, the decision-maker must be vigilant against three types of fracture risks. These risks are not theoretical simulations but realistic threats derived from actual integration facts and legal procedures; the outbreak of any one of them could result in Manus assets being reduced to zero value or the transaction collapsing entirely.
Risk of "Hollowing Out" Technical Assets: Physical Separation Lagging Behind Legal Closing. Meta established a data firewall on June 11, 2026, and used the phrasing "sunsetting Manus," indicating it is executing an orderly exit procedure (Source: material). However, Manus technology had previously been integrated into Ads Manager and incubated Business Agent. This deep coupling means that even if equity changes hands, core engineering capabilities may have already substantively flowed away with personnel and code migration. In the simulation, the stakeholder prototype based on Manus employees explicitly expressed concern about being forcibly relocated to the Singapore office leading to team dispersion, while the stakeholder prototype based on Sequoia China also pointed out that due diligence must clarify the boundaries of assumable assets to avoid taking over a shell company that has lost its R&D core (Source: simulation). If the acquiring party only obtains legal ownership without simultaneously locking in the core technical team and underlying architecture, the $2 billion valuation will lack physical support, becoming merely a premium payment for "past reputation."
Risk of "Secondary Review" for Cross-Border Compliance: Transaction Suffocation Under US-China Regulatory Pressure. Although the National Development and Reform Commission banned Meta's acquisition on April 27 (the specific date and document number of this ban could not be traced to the original source), this does not mean that subsequent domestic acquisition by Chinese parties automatically receives exemption. Stakeholders prototyped after China's Ministry of Commerce and NDRC both emphasized that if the new transaction structure involves foreign capital participation or technology import/export, it still needs to be re-declared and evaluated according to law (Source: simulation). Meanwhile, US Senator John Cornyn explicitly raised in the simulation that derivative IP generated during Meta's integration may constitute a Reflow of US technology, requiring CFIUS review (Source: simulation). This means the acquisition transaction may trigger security review mechanisms from both China and the US. Verified online materials also indicate that cross-border M&A of AI Agent enterprises must face multiple compliance thresholds such as export controls and cross-border data transfer; the faster the commercialization speed, the higher the risk of exposing flaws in the underlying legal architecture (Source: verified online, secrss.com). If the transaction design does not reserve buffer periods for bilateral compliance at the outset, it is highly likely to be indefinitely stalled during the approval process.
Risk of "Contract Collapse" for Teams and Control Rights: Oral Promises Cannot Counteract Ecological Gravity. The founding team views "independent operation" and "retention of control rights" as the bottom line for the acquisition. Stakeholders prototyped after Xiao Hong and Zhang Tao repeatedly emphasized that they cannot become appendages of capital (Source: simulation). However, the stakeholder prototyped after Meta questioned whether portfolio companies within Tencent's ecosystem can truly maintain independence, pointing out that when product directions conflict with platform synergy requirements, contractual constraints often yield to commercial realities (Source: simulation). If this trust deficit is not solidified in the shareholders' agreement with legally binding veto rights, board seats, and business isolation clauses, relying solely on the "largest minority shareholder" structural design is difficult to eliminate the concerns of the core team. Once strategic disagreements arise after the transaction, the founding team may choose to resign or perform duties passively, causing the core value of Manus as a talent-intensive AI asset to instantly collapse.
Recommended actions: Phased response strategies based on three scenarios
Regarding the proposition that 'other parties should take over Manus, which Meta has called off,' the current situation is characterized by 'regulatory red lines set, commercial Game unresolved.' Based on the preceding analysis of stance distribution, divergence focal points, and the triple fracture risks, decision-makers should not seek a single optimal solution. Instead, they should dynamically adjust response strategies among the following three scenarios based on regulatory approval progress and the reality of technological decoupling. All recommendations are based on verified material facts and key variables exposed in the simulation, aiming to provide executable operational frameworks for different trajectories.
1. Base Case: Implementation of an 'Independent Operation + Minority Equity' Transaction under a Compliance Framework
This scenario assumes that the safety reviews by the National Development and Reform Commission (NDRC) and the Ministry of Commerce ultimately approve Tencent's lead plan, and that Meta agrees to cooperate with an orderly exit. This is the path most anticipated by current market opinion and supporters in the simulation (accounting for 54.9%), but it presupposes resolving contractual issues regarding technical asset ownership and team stability. (Source: simulation)
- For Regulatory Authorities: It is recommended to make the 'Technical Asset Inventory' and 'Data Compliance Audit Report' prerequisites for transaction approval, rather than post-completion supplements. Given the fact that Meta has integrated Manus technology into Ads Manager and incubated Business Agent (Source: material), it is necessary to explicitly require legal confirmation of derived outcomes and underlying architecture before closing, to avoid triggering secondary reviews due to subsequent intellectual property disputes.
- For the Taking-Over Consortium: It is suggested to solidify 'independent operation rights' through legally binding clauses in the shareholders' agreement, including but not limited to board seat allocation, veto power over core businesses, and restrictions on related-party transactions. A simulation character based on Meta as a prototype once questioned the independence of invested enterprises within Tencent's ecosystem (Source: simulation). Only through rigid contracts can concerns of the founding team be eliminated, preventing asset hollowing out caused by talent drain.
- Signals to Watch: Monitor whether the NDRC releases an announcement of conditional approval; monitor whether the Manus website removes the expression 'has become part of Meta' (Source: verified online, manus.im); track changes in Tencent's financial report disclosures regarding investments in AI application layers.
2. Worst Case: Transaction Failure Due to Technology Supply Cut-off or Approval Deadlock
Trigger conditions for this scenario include: upstream service providers such as Anthropic cutting off API access due to security concerns, cross-border capital paths being blocked, or regulators determining that the Reflux of technology poses uncontrollable risks and indefinitely suspending approval. The simulation shows that a simulation character based on Benchmark as a prototype explicitly warned that valuations would be re-evaluated if services were restricted (Source: simulation), and material verification indicates that multiple key transaction details have not yet been officially confirmed, meaning liquidity risks are real.
- For Manus Management: Immediately activate the 'Technology Backup Plan' and 'Cost Survival Mode'. Although the authenticity of measures restricting the core team from leaving the country has not been traced to original sources, the risk objectively exists. Priority should be given to ensuring salary payments and continuous social security contributions for the domestic R&D team, avoiding zeroing out technical capabilities due to personnel instability. Simultaneously, accelerate docking with domestic large model APIs as alternatives to Anthropic services to reduce dependence on external single suppliers.
- For Potential State-Owned Capital Participants: A simulation character based on Wuhan State-owned Assets as a prototype emphasized 'not taking hot potatoes' (Source: simulation). In this scenario, state-owned capital intervention should be limited to bridge financing or talent placement support to maintain basic operations, rather than assuming high-valuation equity. Wait for a third-party assessment agency to issue an asset value report before Assessment whether to proceed with low-cost restructuring for industrial strengthening purposes.
- Signals to Watch: Monitor changes in Anthropic's sales policies for AI services in China; watch for window guidance signals from foreign exchange management departments regarding cross-border investment payments; pay attention to negative public opinion on Manus employees' social media regarding unpaid wages or forced relocation.
3. Best Case: Regulatory Green Lights Combined with Ecosystem Synergy Release Excess Value
The trigger condition for this scenario is: the transaction is not only approved, but Manus successfully integrates into the WeChat/Cloud ecosystem to accelerate commercialization, while China and the US reach a Phased understanding on cross-border flow of AI technology. Although the probability is low, if such signals appear, the focus of decision-making should shift from 'risk prevention' to 'promoting integration'.
- For Industry Competent Departments: The Manus case could be used as a pilot sample for compliant Reflux of cross-border technology in the AI field, exploring the establishment of a new regulatory mechanism of 'whitelist + post-filing'. Multiple parties in the simulation mentioned the strategic value of Manus in making up for shortcomings in the domestic AI Agent ecosystem (Source: simulation). If its commercial closed-loop under independent operation can be verified, it can provide institutional references for similar assets returning in the future.
- For Investors and Founding Teams: While ensuring the stability of control rights, accelerate the preparation pace for an IPO in Hong Kong. However, vigilance against valuation bubbles is needed. A simulation character based on Sequoia China as a prototype reminded that 'exit paths must be verifiable' (Source: simulation), suggesting linking the listing timeline with product revenue milestones to avoid relying solely on capital narratives.
- Signals to Watch: Observe whether function entrances related to Manus are launched within the WeChat ecosystem; track whether the Ministry of Commerce or NDRC releases compliance guideline documents for the import and export of AI technology; pay attention to the valuation feedback of AI Agent concept stocks in the Hong Kong stock market.
Regardless of the scenario, decision-makers must bear in mind: the core contradiction of the Manus case has evolved from a simple commercial M&A into a compound Game of technological sovereignty and industrial security. Any action recommendation must be built on the basis of 'compliance first, team-oriented, clear distinction between virtual and reality', avoiding misleading judgments by unverified valuation figures or unilateral statements.
Materials and Sources
Verified Online Sources
- Tencent leads $2 billion "redemption" of Manus: A compliance case study for cross-border AI M&A _ Eastmoney
- From viral hit to Penetrate review: Manus's legal lesson for AI Agent companies - Security Insider | Decision-makers' cybersecurity knowledge base
- Sales Process Builder: AI-driven transaction tracking and forecasting
Source Materials