Beijing — China’s Ministry of Commerce has initiated a regulatory review of Meta Platforms Inc.’s acquisition of Manus, a high-profile AI agent developed by a startup with Chinese roots, citing potential violations of technology export controls.
The deal, announced on December 30, 2025, and reportedly valued at more than $2 billion, marked a significant vote of confidence in Manus, widely regarded as one of the world’s first fully autonomous AI agents capable of understanding, planning, and executing complex tasks. However, the transaction has drawn scrutiny amid heightened geopolitical tensions over advanced technologies.
In a January 8, 2026, press briefing, the Ministry of Commerce stated it would coordinate with relevant authorities to evaluate whether the acquisition complies with China’s laws on export controls, technology trade, and outbound investment.
Manus was originally developed by Wuhan-based Monica, founded in 2022 by entrepreneur Xiao Hong and colleagues. The company initially built AI-integrated browser tools before pivoting overseas. In March 2025, its Singapore-registered entity, Butterfly Effect, launched Manus, which quickly gained traction in international markets.
Backed by a $75 million funding round led by Benchmark Capital in April 2025—at a valuation nearing $500 million—Manus achieved rapid commercialization success, surpassing $100 million in annualized recurring revenue within months. To support global expansion, the company relocated its primary operations to Singapore in mid-2025 and discontinued services in China.
Industry observers note that Manus is tailored for Western users, leveraging overseas large language models (such as those from Meta and OpenAI) and integrating tools popular in global markets. A standout feature is its ability to generate standalone web pages for task outputs, complete with traffic analytics—a design choice that aligns closely with international software preferences.
The regulatory probe centers on China’s updated Catalog of Technologies Prohibited or Restricted from Export, revised in late 2023 and effective from 2024. Key provisions target “artificial intelligence interaction interface technologies,” encompassing natural language understanding, task-driven dialogue management, and related architectures.
Even though Manus relies primarily on foreign foundational models, regulators are likely examining whether core elements—such as prompt engineering frameworks, task orchestration logic, or data-driven personalization features—were developed in China and qualify as restricted technologies. Additional clauses covering large-scale data processing and personalized information services could also apply.
Under Chinese law, the transfer of controlled technologies from domestic entities to overseas affiliates, including through acquisitions, is treated as an export requiring approval. Analysts suggest the review will focus on the origins of Manus’s underlying code and training processes, much of which may trace back to its early development in China.
The case highlights growing challenges for Chinese AI entrepreneurs pursuing global ambitions. While offshore incorporation has become common to access international capital and markets, it does not automatically exempt domestically originated intellectual property from export rules.
Market reaction has been muted so far, with Meta declining to comment on the ongoing review. The Manus website now displays a “From Meta” badge, though it remains unclear whether the overseas portion of the transaction has closed.
The outcome could set precedents for future cross-border AI deals, underscoring the need for early compliance planning in areas like technology provenance and jurisdictional structuring. For now, the Manus saga exposes the delicate balance between commercial innovation and national security priorities in China’s thriving AI sector.
