In Depth: Chinese Startup Manus Swept Up in Global M&A Frenzy for AI

In the final days of 2025, Meta Platforms Inc. rapidly acquired Chinese AI startup Manus for $2–3 billion, showcasing the fierce competition in Silicon Valley to secure dominance in the AI sector. This deal reflects a wider industry trend of tech giants absorbing innovative startups across the AI ecosystem—from semiconductor supply to language models and consumer applications. Notably, the acquisition of Manus followed Meta's earlier $14 billion purchase of the data-labeling firm Scale AI, signaling aggressive expansion strategies targeting both advanced technologies and key AI talent[para. 1][para. 2][para. 3].

The acquisition occurred over a swift 20-day period, with Meta first contacting Manus about a month before the deal, and negotiations happening in Singapore. The offer came just two weeks before the year’s end, as Meta pushed to close before 2026. Although valued at $2 billion prior to the deal, Manus had limited suitors: Google was developing a similar project internally, OpenAI was focused on restructuring and IPO plans, and other major players were unwilling to pay a premium. Chinese internet giants also showed little interest, as Manus's products targeted overseas markets. While some critics considered the sale rushed, others noted a shift—Chinese AI startups now aspire to be acquired by international giants at higher valuations, rather than domestic tech firms[para. 5][para. 6][para. 7][para. 8].

Manus, initially backed by Beijing Butterfly Effect Technology, saw rapid growth following its March 2025 product launch. By April, it secured $75 million in funding at a $500 million valuation, which attracted U.S. political attention. By July, the core team relocated to Singapore and ended operations on the Chinese mainland. By December 17, Manus had reached $100 million in annual recurring revenue (ARR) just eight months post-launch. Meta announced the acquisition on December 29, stating Manus would remain an independent entity in Singapore[para. 9][para. 10][para. 11].

Despite a lack of operations or intellectual property in China at the time of sale, Manus’s Chinese origins prompted scrutiny from Beijing’s Ministry of Commerce. The ministry initiated a review to determine if the deal complies with Chinese export control and technology transfer laws. While China generally supports lawful cross-border tech cooperation, such high-profile sales must follow Chinese regulations. Legal experts noted that regulating such transfers is complex, given challenges in investigating cross-border technology flows and identifying violators[para. 12][para. 13][para. 14][para. 15][para. 16].

Meanwhile, Nvidia made a headline-grabbing move by licensing intellectual property from Groq, an AI inference chip company, involving a transaction worth around $20 billion—Nvidia’s largest such deal. Though Nvidia clarified it had not acquired Groq in entirety, it licensed Groq's technology and hired key staff. Groq’s chip, launched in February 2024, outperformed major rivals, intensifying Nvidia's competitive posture amidst growing ASIC chip adoption by Google, Anthropic, and Meta[para. 17][para. 18][para. 19][para. 20][para. 21].

The ongoing acquisition wave is fueled largely by "acqui-hires," with companies seeking top AI talent rather than just products. Meta’s recent acquisitions, such as that of Scale AI and Manus, exemplify this, with strategic leadership shifts in its AI efforts following these deals. Nvidia, too, has been actively acquiring smaller firms to augment its AI capabilities and talent pool, and invested in 87 AI-related firms in the past year[para. 22][para. 23][para. 24][para. 25][para. 26][para. 27].

The underlying motivation for these frenzied deals is a collective fear among tech giants of being eclipsed by new competitors. While market saturation and slowed growth loom, companies with substantial cash reserves are strategically acquiring AI application firms, anticipating greater industry consolidation in the future[para. 28][para. 29].