As 2025 drew to a close, the global tech landscape was set ablaze by a monumental announcement.

Meta announced its acquisition of Manus, a general artificial intelligence agent company (parent company: Butterfly Effect), in a multi-billion-dollar deal.

This acquisition ranks as Meta's third-largest in history, trailing only WhatsApp and Scale AI.

A Striking Example of Entrepreneurial Breakthrough

Xiao Hong, the founder of Manus, will take on the role of Vice President at Meta. Post-integration, the Manus team will continue to develop general AI agent products and services while retaining independent operations in Singapore.

Since its launch in March 2025, Manus has reshaped the landscape of traditional AI assistants with its "hands-on, minds-on" philosophy.

Utilizing a multi-agent collaborative architecture, it autonomously plans and executes complex tasks within isolated virtual machines, outperforming OpenAI's equivalents in GAIA benchmark tests.

From a commercial standpoint, Manus employs a subscription-based model, with annual recurring revenue (ARR) exceeding $100 million. This achievement places it among the few profitable entities in China's AI sector, accomplished merely nine months after its March 2025 launch.

This "lightning-fast acquisition," spanning just over ten days, has catapulted the young Chinese team into the global AI limelight.

The ascent of Butterfly Effect embodies the breakthroughs achieved by China's new-generation entrepreneurs.

Established in 2022, Butterfly Effect did not emerge from a high-profile tech hub. Founder Xiao Hong, a graduate of Huazhong University of Science and Technology, embarked on his entrepreneurial journey in Wuhan.

His early ventures, including the WeChat Official Account formatting tool Yiban and the enterprise WeChat customer management tool Weiban, had already demonstrated his prowess in monetization.

In 2023, amidst the industry's fixation on foundational model R&D, Xiao Hong charted a different course by launching the browser AI plugin Monica. Despite facing "wrapper" criticisms, it emerged as one of China's few profitable AI products, thanks to its high-frequency, essential features, even attracting a $30 million acquisition offer from ByteDance.

In 2024, with the support of ZhenFund, core talents like post-90s serial entrepreneur Ji Yichao joined the team. They zeroed in on the AI agent sector and developed the groundbreaking product Manus.

This general-purpose AI agent, boasting a planning-execution-verification three-tier architecture, can leverage 23 toolchains and outperformed OpenAI's counterparts with an 86.5% accuracy rate in GAIA benchmark tests.

From a $14 million seed round at a $14 million valuation in February 2023 to a nearly $500 million Series B funding in April 2025, continuous backing from top-tier investors such as Sequoia China, Tencent, and Benchmark Capital has witnessed a staggering 35-fold valuation surge in just over two years. This underscores the triumph of a new entrepreneurial spirit that places less emphasis on "connections and credentials."

A Coveted Asset in the AI Race

Behind this acquisition lie profound shifts in AI industry trends and global resource allocation.

2025 has been hailed as the "commercialization starting point" for AI agents, with Gartner predicting them as a top technology trend.

As foundational models become increasingly standardized, application-layer companies with superior implementation capabilities and user experiences have become coveted assets for tech giants.

Meta's swift deal-making stems from Manus's precise alignment with Zuckerberg's vision of "superintelligence empowering individuals." Its cloud-based virtual machine architecture grants AI an independent file system and runtime environment, truly achieving "knowledge-action integration"—a critical piece in Meta's AI application layer puzzle.

For the Butterfly Effect team, joining Meta means tapping into global infrastructure, reaching billions of users, and accessing top-tier computational resources, potentially scaling from millions to billions of users.

More significantly, Manus's success validates a new paradigm for Chinese AI entrepreneurship: amidst the dominance of foundational models by overseas giants, Chinese startups can still secure pivotal roles in the global value chain through "scenario deep-dive + engineering innovation," leveraging global vision and product excellence.

This "born global" strategy positioned Manus within the global market system from day one. Its achievements—147 trillion tokens processed and 80 million virtual computers created—serve as a universal metric transcending cultural barriers.

Challenges Ahead, Yet Opportunities Abound

Despite the opportunities, Manus confronts multiple challenges that could offer vital lessons for China's AI globalization.

Under the agreement, Butterfly Effect will maintain independent operations, with Xiao Hong serving as Meta's Vice President. This "independent + integrated" model preserves the startup's agility while enabling strategic synergy. However, cultural integration remains the top challenge: how to sustain innovation within a corporate behemoth and avoid bureaucratic constraints is a conundrum Xiao Hong's team must navigate.

Technologically, large model hallucinations persist as a bottleneck, with reliability in high-precision sectors like finance and healthcare urgently needing enhancement.

Commercially, the subscription model must strike a balance between user growth and model invocation costs. In just the first two weeks post-launch, Manus's expenses on the Claude model surpassed $1 million.

Competitors are also evolving rapidly.

The market is fiercely competitive, with giants like OpenAI and Google encircling the AI agent space, while rivals like Cursor and Devin are rapidly iterating.

Moreover, global operations bring compliance pressures. Differences between the EU's AI Act, U.S. regulatory frameworks, and Chinese regulations impose stricter demands on data security and algorithmic transparency.

Nevertheless, Manus has charted a clear path for differentiated competition by retaining independent decision-making while gaining resource support. Its strategy of extending from general tools to vertical scenarios and building a developer ecosystem is evident.

Dai Yusen, Managing Partner at ZhenFund, captured the essence of this acquisition: "They achieved what previous generations of entrepreneurs couldn't or even dared not imagine."

After all, Meta's acquisition of Manus transcends mere commercial success—it's a testament to the Chinese AI entrepreneurial spirit. It proves that through pure product innovation, Chinese teams can ascend to the global pinnacle.

For China's AI entrepreneurial ecosystem, this case signals "the arrival of an era belonging to this generation of young Chinese entrepreneurs," illuminating the path for successors. On the global tech stage, technological prowess and product innovation remain the ultimate core competitiveness.

As AI agent technologies mature, Manus's prospects within Meta's ecosystem are promising. Its exploration of globalization paths, technological innovation models, and risk mitigation strategies will ultimately provide replicable blueprints for more Chinese AI firms venturing abroad.