This morning, Meta completed a significant acquisition, bringing the AI agent startup Manus under its umbrella. Details of the deal, such as the exact acquisition amount, remain undisclosed.

This AI agent star company, which initially had its roots in China and later moved to Singapore, precisely mirrors, through its growth path and final destination, the opportunities and challenges that Chinese entrepreneurs encounter.

For Chinese large - model startups that are currently at a crucial stage of technological breakthroughs and globalization exploration, this acquisition serves as both a reference point and a cautionary tale.

From a Chinese Startup to Acquisition by a Tech Giant: Manus's Transformation and Counter - attack

On December 30, 2025, Meta officially announced the completion of its acquisition of the AI agent company Manus. This acquisition ranks as the third - largest in Meta's history.

Manus was originally Beijing Butterfly Effect Technology Co., Ltd., founded in China with Xiao Hong as its core leader.

In August 2023, Manus registered its main entity in Singapore, laying the foundation for an early global expansion.

Its product officially launched in March 2025. In just eight months, it achieved an annualized revenue of $125 million, shocking the industry.

By the time of the acquisition, its platform had processed over 147 trillion tokens and created over 80 million virtual computers. It topped the RLI benchmark test with its "capability overflow" mining technology.

Backed by a strong capital lineup, including ZhenFund, Sequoia China, and Tencent, its valuation soared from $14 million to nearly $500 million.

All of this was preceded by a critical transformation in July 2025. During this time, Manus laid off most of its domestic staff and relocated its headquarters to Singapore.

This transformation was a result of multiple pressures. These included the need for U.S. government review after receiving dollar - fund investments, restricted access to advanced domestic computing power, and an immature local paid market.

Ultimately, this company with Chinese origins won the favor of a tech giant with a global presence. Xiao Hong was appointed as Meta's vice president, and the company maintained independent operations.

Survival and Breakthrough Strategies for Chinese Startups

Manus's success and transformation offer three key survival strategies for Chinese large - model startups.

Commercialization capabilities are the key to navigating through different economic cycles.

Meta's high - price acquisition of Manus was mainly due to its successful subscription - based business model, rather than just technological leadership.

Although financing remains active in China's AI sector, capital tends to favor companies with self - sustaining capabilities. Firms with inflated valuations and lacking practical applications are being gradually phased out.

Chinese startups must abandon the "technology - first, business - later" mindset. Instead, they should focus on rigid demand scenarios such as office automation and vertical industry solutions to quickly generate stable revenue.

Differentiated technological barriers are more important than "full - stack dominance."

With Meta and Google dominating foundational large models and domestic giants covering comprehensive layouts, startups do not need to reinvent the wheel.

Manus secured its position with unique technology for mining the latent capabilities of large models. This proves that vertical technological breakthroughs offer a shortcut to success.

Chinese startups can delve into niche areas, such as industrial - grade multimodal interaction and medical data processing, to forge irreplaceable technological identities.

Global layouts require early planning in terms of compliance and structure.

Manus's choice of a Singapore headquarters was crucial for avoiding geopolitical risks and securing dollar financing.

Chinese large - model companies face multiple obstacles when going global, including differences in cross - border data governance, complex international legal environments, and restricted access to computing power.

Establishing entities in neutral regions and building compliance frameworks in line with international standards have become essential for accessing global capital and markets.

Manus's case also exposes the unique challenges faced by Chinese large - model startups, necessitating targeted solutions.

Firstly, addressing geopolitical and compliance risks.

The U.S. continues to tighten investment and technological restrictions on China's AI sector. Data exports require strict filing. Standard contracts and filings are needed for exporting personal information of 10,000 to 1 million individuals annually.

Chinese companies can take advantage of scenario - based management policies in free trade zones. They can adopt techniques like data desensitization and federated learning to reduce compliance costs while avoiding reliance on a single market.

Simultaneously, it is essential to solve constraints on computing power and capital.

Restricted access to advanced domestic chips poses a significant bottleneck for technological iteration.

Startups can explore alternative domestic computing power solutions or deepen collaborations with cloud providers to optimize resource utilization.

On the capital side, flexibility is key. Aligning with RMB funds in the early stage lays a solid foundation. Later, introducing compliant international capital based on development needs can avoid over - reliance on a single funding source.

Balancing domestic markets and global layouts is also crucial.

Manus's decision to abandon the Chinese market is not ideal. China has 515 million generative AI users, making it an indispensable core market.

Chinese startups should establish a foothold domestically, leveraging policy dividends from the "AI Plus" initiative to deepen scenario - based implementations in industries like manufacturing, agriculture, and healthcare.

Going global should be a gradual process. Start with neutral markets like Singapore and Southeast Asia to accumulate experience, adapt to regional regulations and user habits, and avoid reckless expansion.

Additionally, drawing on Manus's negotiation wisdom is vital for securing core interests in potential mergers and acquisitions.

Acquisition by a tech giant is not the end goal. Retaining independent operational rights, core teams, and technological autonomy can enable scaled development through the giant's resources.

For Chinese companies, the choice lies in either becoming high - quality acquisition targets for giants with "technology + business" dual advantages or focusing on vertical areas to become indispensable complements in the ecosystem.