In March 2025, an AI agent product named Manus, also known in Chinese as Butterfly Effect, emerged as a major technological breakthrough. Developed by a China-based team, the system moved beyond traditional conversational AI by autonomously planning tasks, invoking tools, and delivering complete outputs. It was widely regarded as a milestone in the development of general-purpose AI agents. Within days of its release, the product attracted millions of users to its waiting list, invitation codes surged in secondary markets, and the company’s valuation climbed rapidly from under $100 million to around $500 million, making it one of the most prominent AI innovations in China at the time.
As the product gained traction, capital quickly followed. In April 2025, Manus secured investment from a U.S. venture capital firm. However, the deal soon drew scrutiny from U.S. regulators, as artificial intelligence had already been classified as a sensitive sector under tightening outbound investment restrictions. This external pressure, combined with the company’s ambition to expand globally, prompted a strategic shift.
Between June and July 2025, Manus undertook a major restructuring. The company relocated its headquarters to Singapore and established Butterfly Effect Pte. Ltd. as its new global operating entity. At the same time, its domestic team was significantly downsized, with core members relocating overseas, while its presence in China, both operationally and online, was gradually reduced. These moves effectively repositioned the company as an offshore entity in legal terms, aiming to separate its corporate identity from its origins.
Following the relocation, Manus continued to grow rapidly. By the end of 2025, it reported annualized revenue exceeding $100 million, alongside substantial increases in data processing volume and user activity. Against this backdrop, Meta initiated acquisition talks in December 2025 and moved quickly to finalize a deal valued at over $2 billion. The transaction was structured to avoid directly acquiring core assets or intellectual property, thereby attempting to minimize regulatory obstacles.
However, the regulatory environment shifted in early 2026. Chinese authorities began reviewing the deal, focusing on issues such as technology export, cross-border data flows, and changes in control. On April 27, 2026, China’s Office of the Working Mechanism for Foreign Investment Security Review formally prohibited the acquisition, requiring the parties to terminate the transaction and carry out follow-up compliance measures. This decision effectively ended the deal and marked a significant case in China’s scrutiny of foreign acquisitions in the AI sector.
From a regulatory perspective, the decision was not aimed at a single corporate action but reflected a broader set of considerations. First, the origin of the technology remained a decisive factor. Manus’s core algorithms and research capabilities were developed in China, and its technological foundation was built upon domestic engineering resources. Relocating the corporate entity abroad did not alter this underlying reality. Second, data governance emerged as a critical issue. As an AI agent platform, Manus processed vast amounts of user data, including a substantial portion originating from Chinese users, raising concerns over cross-border data transfers and compliance. Third, at a strategic level, general-purpose AI agents are increasingly viewed as potential infrastructure, making their ownership and control matters of national interest.
In this context, regulators adopted what can be described as a “capability-level” review approach. Rather than focusing solely on transaction structures or asset transfers, the assessment centered on whether critical technological capabilities, core teams, and strategic directions would effectively shift under foreign control. This helps explain why the case ultimately fell under the foreign investment security review framework, instead of being handled purely through technology export or data regulations.
The case also highlights the complex landscape facing technology companies pursuing global expansion. On one hand, governments around the world are strengthening controls over critical technologies, data resources, and high-skilled talent. On the other hand, companies still need access to international capital and markets to remain competitive. In such an environment, relying solely on corporate structuring to navigate regulatory systems is becoming increasingly ineffective.
For companies, the implications are clear. Globalization strategies must be grounded in comprehensive compliance frameworks. Key areas including corporate structuring, cross-border data management, technology transfer, and capital operations require thorough upfront assessment and ongoing oversight. Compliance is no longer a secondary consideration but a central capability that determines long-term viability.
More broadly, the case reflects the evolution of regulatory systems toward greater sophistication and earlier intervention. While innovation and international collaboration continue to be encouraged, higher expectations are being placed on safeguarding control over critical technologies and ensuring national security. Going forward, companies will need to navigate multiple regulatory regimes simultaneously and strike a balance between global ambition and legal constraints. Operating within established rules, rather than attempting to circumvent them, is likely to become the defining approach for technology firms expanding onto the global stage.
Source: sina, guancha, xinhua, 21jingji
