China blocked Meta on Monday from acquiring the agentic platform Manus, a move that underscores both Manus’s strategic value and the increasingly fraught geopolitics of AI.
The company behind Manus, Butterfly Effect, moved part of its operations, including its registered headquarters, from Beijing to Singapore after raising funding, including $75 million from the American VC firm Benchmark Capital in 2025.
China contends that the “core DNA” of Manus was developed domestically. That claim underpins the government’s effective blockade of Meta’s proposed $2 billion acquisition of Butterfly Effect and Manus, with regulators placing the platform’s underlying algorithms under strict technology export controls.
Analysts say Beijing is sending a clear message to Chinese AI startups: Relocating to a neutral country like Singapore—sometimes called “Singapore washing”—does not necessarily free them from Chinese oversight or exempt them from Chinese law.
Manus is an agentic orchestration layer, sometimes called a harness, that can sit atop multiple frontier AI models, including Anthropic’s Opus and Alibaba’s Qwen. It is available globally through a web app as well as iOS and Android apps.
Manus can plan and reason through complex tasks, then deploy agents and subagents to execute them step-by-step on a user’s behalf. Its interface offers users a transparent view of the agent’s desktop and decision-making process, enabling human oversight without disrupting autonomous workflows.
Users can also assign “dark” tasks, in which agents and subagents complete complex projects—such as financial modeling or competitive market research—in the background, delivering a finalized output all at once rather than a running chat history.
What differentiates Manus is the maturity of its platform and the reported accuracy of its agents, which have performed strongly in multiple benchmark tests. Enterprises worldwide are betting that AI agents will soon mature enough to take over major business functions—from operations and strategic planning to decision support and customer relations.
The blocked deal could represent a significant setback for Meta’s AI ambitions. The social media giant spent billions last year to reload and refocus its AI strategy in hopes of competing more directly with OpenAI, Anthropic, and Google.
Rather than building its own mature agentic platform from scratch, Meta had hoped to acquire one. Manus reportedly reached $125 million in annual recurring revenue, making it an especially attractive target.
Meta sees a major opportunity to reshape commerce across its platforms (Facebook, Instagram, WhatsApp, Messenger, and Threads) by deploying personal AI agents that guide users through shopping and purchasing decisions. That vision could fundamentally expand its advertising-based business model.
So far, Meta has remained relatively restrained in public statements about the setback, likely to avoid escalating tensions with Chinese regulators. The company has said the proposed deal complied with applicable laws and that it expects to continue working with regulators toward a potential solution.
China’s block on the Meta-Manus deal will likely be viewed as a new flashpoint in the escalating competition between the U.S. and China for AI dominance.
The global AI ecosystem is increasingly fragmenting. The U.S. has restricted exports of advanced chips critical to AI research, while China is now tightening controls over the movement of researchers, intellectual property, and capital tied to AI development.
By keeping Manus within its technological sphere of influence, China joins a growing list of nations seeking to build sovereign AI ecosystems that reduce dependence on Silicon Valley, still the dominant center of artificial intelligence research and commercialization.
Just weeks ahead of the U.S.-China summit in May, Beijing has effectively signaled that agentic AI is emerging as a new front line in the global technology and trade war.
