Manus acquisition halted: A comparison of China and US frontier technology acquisition reviews

Author: Zhang Feng

1. The Origins and Development of the Manus Acquisition Case

In March 2025, a general-purpose AI agent product called Manus suddenly emerged. This product, developed by Chinese founder Xiao Hong and his team within China, exploded onto the global market with the ability to “think independently, plan, and end-to-end execute complex tasks.” It quickly gained widespread attention upon launch, with annualized revenue surpassing $125 million. In just nine months, Manus became a phenomenon-level star product in China’s AI field and was regarded as one of the representative works of domestic AI “overtaking on the curve.”

However, after its impressive debut, Manus quickly initiated a rapid and covert strategic repositioning. In June 2025, the company moved its headquarters from China to Singapore, changing its operating entity to Singaporean company Butterfly Effect Pte, wholly owned by the Cayman Islands parent company. Subsequently, the Chinese team was significantly downsized, with only about 40 core technical staff remaining to relocate to Singapore from the original 120 employees. Domestic social media accounts were cleared, and the official website blocked access from Chinese IPs. On December 30 of the same year, US tech giant Meta announced the acquisition of Manus’s parent company, Butterfly Effect, for about $2 billion, with founder Xiao Hong scheduled to become Meta Vice President. This deal became Meta’s third-largest acquisition ever, after the 2014 $19 billion purchase of WhatsApp.

Following Meta’s announcement, Chinese regulators quickly intervened. On January 8, 2026, the spokesperson for the Ministry of Commerce, He Yadong, publicly stated during a routine press conference that relevant departments would evaluate the consistency of Meta’s acquisition of Manus with laws related to export controls, technology import/export, and foreign investment. Subsequently, the National Development and Reform Commission (NDRC) summoned senior executives from both sides, highlighting risks related to technology transfer and data security. Ultimately, on April 27, 2026, the Office of the Foreign Investment Security Review Mechanism (NDRC) made a legally compliant decision to prohibit Meta’s investment in Manus, requiring the parties to revoke the transaction. This was the first publicly halted foreign investment in the AI sector since the implementation of the “Measures for the Security Review of Foreign Investment” in 2020, and it represented the strictest review outcome under that framework.

The core controversy of the Manus case lies in: core technology developed by a Chinese team domestically, but through a chain of “domestic R&D → offshore shell change → foreign acquisition,” attempting to transfer control of the technology abroad without legally reporting for foreign investment security review. Regulators focused on the question of “when, how, and what is transferred out”—changing registration location does not exempt the activity from Chinese legal jurisdiction.

2. Legal Basis: Differences Between Chinese and U.S. Legal Systems

China’s legal system. China’s foreign investment security review system is based on the top-level design of the “Foreign Investment Law,” establishing a management model of “pre-establishment national treatment plus negative list.” The core implementing document is the “Measures for the Security Review of Foreign Investment,” issued in 2020 and effective in 2021. The review scope covers military industry, key technologies, infrastructure, important information technology and internet products/services, and critical financial services.

It is important to emphasize that China’s review mechanism adopts a “piercing” review principle. As exemplified by the Manus case, regulators do not focus solely on whether the “investment entity remains a Chinese company in form,” but rather investigate the “origin and growth path” of the technology—so long as the core R&D results are completed within China, even if the registration location is changed offshore, the activity may still fall within Chinese regulatory scope.

U.S. legal system. The U.S. foreign investment security review system is based on the “Defense Production Act of 1950,” specifically Section 721, and has been significantly expanded by the “Foreign Investment Risk Review Modernization Act” (FIRRMA). FIRRMA is the most substantial legislative expansion of U.S. foreign investment review in the past 40 years, broadening the scope of review—before FIRRMA, the Committee on Foreign Investment in the United States (CFIUS) mainly reviewed “control transfer” transactions; after FIRRMA, even non-controlling investments that do not constitute control transfer are subject to review.

CFIUS is a multi-agency committee led by the Department of the Treasury, with members including the Departments of Defense, State, Commerce, Homeland Security, Justice, and other key agencies. A key difference is that the U.S. review system institutionalizes a differentiated approach based on the “source country”—in 2025, the Trump administration launched the “America First Investment Policy,” which restricts foreign adversaries’ investments in critical technologies, infrastructure, and personal data, while creating “fast-track” pilot programs for allied capital.

3. Review Procedures: Voluntary Declaration vs. Passive Tracking

Chinese review process. China’s foreign investment security review process consists of three stages: the first is preliminary review, where within 15 working days of receiving compliant application materials, a decision is made whether to initiate security review; the second is general review, completed within 30 days of review initiation; if there are concerns about national security impact, a third special review is initiated, to be completed within 60 days of start, with possible extensions. The review conclusions are categorized as: approval, conditional approval, or prohibition.

The Manus case’s review process is unique—it is a post-facto review of “unreported” activity. Since the transaction parties did not proactively report, regulators discovered clues through public information and lawfully initiated investigation, following the full procedural steps, ultimately resulting in a prohibition decision.

U.S. review process. CFIUS’s review also has two stages: initial review and investigation. The initial review lasts 45 days; if national security risks are identified, it proceeds to a second 45-day investigation. If issues remain unresolved, the President can be advised to intervene, with a 15-day window for a final decision. For certain transactions, if the U.S. target company’s key technology requires export licenses to the buyer’s home country, the parties must submit a mandatory declaration to CFIUS.

A key difference from China’s mechanism is the retrospective capacity—CFIUS can review completed transactions. Courts have even approved CFIUS’s post-divestiture orders, such as in the case of HieFo-Emcore, where the acquisition was completed over two years prior. Additionally, CFIUS has a dedicated “unreported transaction team” that proactively identifies transactions within its jurisdiction that were not voluntarily reported.

4. Review Focus: Data, Technology, and National Security

Core concerns in China’s review. Chinese review focuses on three main points: 1. Loss of key core technologies. In the Manus case, whether Manus’s core AI technology falls under the “Information Processing Technology” control points listed in the “Catalogue of Technologies Restricted for Export” is a key judgment. 2. Data export security. During training and operation within China, Manus collected large amounts of Chinese user data, raising issues of personal information export compliance. 3. Substantive control over technology transfer, regardless of form. Regulators penetrate Manus’s offshore structure to directly examine “when, how, and from whom” the technology is transferred out.

Core concerns in the U.S. review. CFIUS’s review is risk-based, assessing threats posed by foreign investors, vulnerabilities of the U.S. target company, and potential consequences of the transaction. Specifically, CFIUS focuses on three dimensions: 1. Technical sensitivity, whether the target involves critical technologies like semiconductors, AI, quantum computing covered by export controls; 2. Data risks, whether sensitive personal data flows to foreign adversaries; 3. Source country differentiation, applying stricter review standards to investments from China and similar countries. Moreover, CFIUS’s review increasingly intertwines with U.S. industrial policies—such as the recent steel acquisition case, where concerns extend beyond national security to labor rights and industrial competitiveness.

5. Penalties: Prohibition and Divestment

Chinese penalty framework. Under the “Measures for the Security Review of Foreign Investment,” prohibition is the most severe review outcome—explicitly stating “no investment shall be made in prohibited activities.” If the involved party refuses to comply, the mechanism office can order rectification, impose sanctions on equity or assets, and record bad credit information in national credit systems for joint sanctions.

The Manus case’s enforcement was highly specific: In terms of equity, if Meta had completed the share transfer, it must transfer all Manus shares back to the original shareholders or domestic entities, completing registration changes. In terms of funds, Meta was required to fully refund the approximately $2 billion paid, with foreign exchange authorities monitoring the fund flow to prevent capital flight under the guise of terminating the deal. In data and technology, Meta had to delete all collected domestic data, Manus had to restore data localization, terminate all technology licensing and code transfers to Meta, and Meta’s dispatched technical personnel had to withdraw. The review office coordinated inspections with multiple departments including development, commerce, cybersecurity, and foreign exchange.

U.S. penalty framework. CFIUS’s enforcement has also intensified recently. In recent years, CFIUS has set new records for enforcement actions, with fines totaling nearly $88 million, and a maximum single fine of $60 million. Before 2024, only six fines were publicly disclosed, with the largest at $1 million. In February 2026, the U.S. Department of Justice initiated the first enforcement lawsuit against a CFIUS divestment order under the Defense Production Act. From a retroactive perspective, the U.S. is even more aggressive than China— in January 2026, President Trump issued an executive order requiring Chinese-controlled HieFo to divest its U.S. acquisition of digital chip business from Emcore within 180 days, even though the deal was completed over two years earlier.

6. Focused Sectors: Common and Divergent

Shared focus areas. The core intersection of Chinese and U.S. review mechanisms lies in the high-tech sector. The semiconductor supply chain is the most frequently scrutinized and vetoed area—before the Manus case, U.S. CFIUS had already effectively blocked many semiconductor transactions. AI and related intelligent technologies are also top priorities—Manus’s case marked China’s first prohibition decision in AI, contrasting with the long-standing strict U.S. review.

Differentiated concerns. China’s review emphasizes “technology flowing out of China” rather than “foreign capital entering China.” The main regulatory concern is preventing “domestic R&D, offshore cash-out” models—core technology assets losing control is the primary focus. In contrast, the U.S. review is more about restricting foreign capital’s access to U.S. technology, with geopolitical factors integrated earlier into the assessment—such as the “fast-track” process for certain foreign investors from allied countries.

7. Global Impact: Reflection of Tech Competition and Regulation

The Manus case and the release of DeepSeek V4 occurred on the same day, forming a meaningful contrast. The rise of DeepSeek V4 signifies that domestic AI can stand firm among the world’s top tiers; meanwhile, the halt of the Manus acquisition indicates that the country will not allow core AI assets to be lost through mergers and acquisitions. Together, they outline two sides of China’s AI strategy: one actively supports independent innovation and growth, the other safeguards core assets.

The global influence of the Manus case manifests at least in three levels. First, the China-U.S. tech rift deepens, prompting multinational capital to reassess cross-border AI M&A risks. Second, foreign investment regulation worldwide is expanding in a coordinated manner—across the EU, UK, Japan, and others, key sector review frameworks are tightening. Third, security review is replacing traditional anti-monopoly review as the primary regulatory barrier for frontier tech M&A transactions.

8. Future Trends and Corporate Responses

Regulatory trends. Looking ahead, China and the U.S. are expected to continue strengthening their foreign investment security review regimes. China’s Manus case, as the first AI sector foreign acquisition halted, will likely promote further refinement of reviews concerning key technologies and data security. The U.S. will continue to differentiate based on investment sources, tighten controls on sensitive sectors, and strengthen industrial chain-related reviews. The “Comprehensive Foreign Investment National Security Law” passed in December 2025 further expands the U.S. legal toolkit for technology protection.

Corporate compliance advice. For enterprises engaged in frontier technology cross-border transactions, the Manus case offers clear compliance warnings:

— Proactive declaration. Chinese regulators have zero tolerance for “unreported” activities; proactive reporting is the most effective way to reduce regulatory uncertainty. Parties should communicate with the security review office through transparent compliance pathways to manage risks.

— Careful transaction structuring. Make compliance review a red line in transaction design. Use phased transaction arrangements, with regulatory approval as a prerequisite for closing. Any offshore structuring intended to evade review may be pierced and invalidated.

— Technology export compliance. For cross-border transfer of key technologies, conduct pre-approval assessments against regulations like the “Catalogue of Technologies Restricted for Export” and apply for export licenses when necessary.

— Data cross-border compliance. For data involving Chinese users, ensure legal procedures such as security assessments are followed for data transfer and export.

— Special considerations for U.S. investments. Besides CFIUS review, carefully evaluate whether export control systems could cause dual compliance conflicts, and design detailed risk-sharing provisions in transaction documents.

A comparison of China and the U.S. makes clear that the halt of Manus’s acquisition is not an isolated event but a typical case reflecting the tightening trend of foreign investment review in frontier technology sectors worldwide. While their review systems differ in focus and methods, the underlying logic is highly consistent: core technologies are strategic national assets, and sovereignty over key data must be maintained. Understanding this fundamental logic may be more important than merely familiarizing oneself with the texts of the laws of both countries.