Will Manus’ sale to a U.S. company trigger a butterfly effect? Experts say it depends on the company’s size and business.

On Wednesday, December 31, Meta, the parent company of Facebook, announced its acquisition of Manus, an artificial intelligence (AI) startup originally founded in China and later headquartered in Singapore. Pictured is the Meta logo at the Paris Expo Porte de Versailles in France. (Reuters)

Founded in China, relocated its headquarters to Singapore, and then acquired by U.S. tech giant Meta at a premium price—this is the remarkable path of globalization and turnaround taken by Chinese AI startup Manus.

AI is a key battleground between China and the United States. As geopolitical tensions between the two countries intensify, observers are watching closely whether Manus’ successful rebranding and exit strategy will become a model for other Chinese AI companies seeking to go global and monetize their technology.

Experts interviewed say whether the “Manus model” can be replicated depends on a company’s size and business scope. Smaller firms that have not yet drawn the attention of the Chinese government, are not backed by major domestic platforms, and heavily rely on accessing overseas large language models are more likely to adopt the Manus approach. “For companies that have already reached a certain scale—like DeepSeek—if they try to relocate overseas, it’s basically impossible,” one expert noted.

Western media outlets including The Wall Street Journal and Bloomberg reported on Wednesday, December 31, that Meta had announced the acquisition of Manus. The purchase price was not disclosed, but reports suggest it could exceed $2 billion (SGD 2.572 billion), with the deal reportedly finalized in just 10 days.

A Meta spokesperson said that upon completion of the transaction, “Manus will no longer hold any Chinese ownership interest and will terminate its services and operations in China.”

Manus founder and CEO Xiao Hong, born in 1992, graduated from Huazhong University of Science and Technology and started his entrepreneurial journey in Wuhan. His two co-founders are Ji Yichao, a “post-90s” entrepreneur who graduated from Beijing Information Science and Technology University, and Zhang Tao, a “post-85s” graduate of Chongqing University of Posts and Telecommunications.

The general-purpose AI agent Manus was launched in March 2025, under its parent company Butterfly Effect, founded in 2022. Manus is said to autonomously understand goals, break down tasks, invoke tools, and complete complex workflows with minimal human intervention. Unlike large models such as ChatGPT or DeepSeek, AI agents like Manus can think and act independently, rather than relying solely on predefined rules and algorithms.

In April 2025, Manus raised $75 million in a funding round led by U.S. venture capital firm Benchmark. According to U.S. media reports, the U.S. Treasury Department launched a review of this investment due to the “reverse CFIUS” rule that took effect in January 2025, which restricts American capital from investing in Chinese AI, semiconductor, and quantum information technology sectors.

It is widely believed that a key factor behind Meta’s interest in Manus was the company’s June 2025 announcement to move its headquarters to Singapore and lay off its China-based staff—a move that significantly diluted its Chinese identity and paved the way for global expansion.

Shen Meng, Executive Director at Chanson Capital, told Lianhe Zaobao that Manus’ indirect route through a third country before being sold to a U.S. company reflects both geopolitical considerations and commercial interests.

He said that for other Chinese AI firms to follow the Manus model, they would need two characteristics: first, they must be relatively small, not yet on the Chinese government’s radar, and not affiliated with major Chinese tech platforms; second, they should operate in the AI agent space—autonomous software systems built on large models. Given that China’s large models still lag behind international counterparts in overall capability, relocating overseas becomes a strategic choice.

Shen added that once a company reaches a certain scale—or, like DeepSeek, achieves global recognition immediately upon launch—“the government would never allow it to relocate.” He explained that Manus chose to move abroad primarily because it needed to access multiple large-model APIs, and staying in China would limit its ability to freely connect with international models.

Associate Professor Gu Qingyang of the Lee Kuan Yew School of Public Policy at the National University of Singapore described the acquisition as a pragmatic response by a Chinese high-tech firm amid escalating technological decoupling and intensified geopolitical containment. “It reflects both the无奈 (helplessness) of navigating external barriers and the urgent need for Chinese technological achievements to gain global market recognition,” he said.

Gu predicted this could signal a new paradigm for Chinese tech companies going global: under a “China + N” operational logic, they establish neutral-country hubs to downplay geographic labels, thereby avoiding geopolitical scrutiny while leveraging technical strength to tap into international capital and markets—a new pathway to break through current constraints.

“This choice may represent an exploration of how Chinese technological value can secure broader development space in an era of deglobalization,” he said.

Associate Professor Fu Fangjian of the Lee Kong Chian School of Business at Singapore Management University assessed that the Chinese government has no intention of restricting AI companies from going overseas. “If it wanted to control them, it could have blocked the move when these companies tried to relocate to Singapore,” he said, noting that models like DeepSeek and Qwen are open-source, reflecting China’s relatively open stance toward AI development.