As 2025 draws to a close, a single acquisition has captured the attention of the global tech and investment communities.

In a lightning-fast deal, a Chinese AI startup has been thrust into the global spotlight.

On December 30, 2025, LatePost exclusively reported that Meta has completed the acquisition of Butterfly Effect—the developer behind the AI application Manus—for several billion dollars. The entire negotiation process was astonishingly swift, taking just over ten days.

When Meta officially announced its multi-billion-dollar acquisition of Manus, an AI agent company, the entire industry took notice. Not only is this Meta’s third-largest acquisition in its history, but it’s also widely seen as a landmark event signaling a pivotal shift in the AI race—from competing on model parameters to competing on real-world application deployment.

Some call it Zuckerberg buying a ticket onto the AI ship; others lament the “loss” of a top Chinese startup. But the significance of this deal goes far beyond a simple capital transaction.

Today, we unpack the investment logic hidden behind this landmark acquisition.

I. Why Is Everyone Watching This Deal?

Tech M&A happens daily—so why has Meta’s acquisition of Manus triggered such widespread attention? The answer lies in three “rule-breaking” details, each striking a nerve in the industry.

First, the speed of negotiations was extraordinary. Liu Yuan, a partner at ZhenFund and an early angel investor in Manus, admitted in an interview: “It was so fast I even doubted whether the offer was real.”

From initial contact to final agreement, the entire deal took just over ten days—a veritable “lightning war” for a cross-border acquisition worth billions of dollars.

For context: Meta’s WhatsApp acquisition took nearly a month to negotiate, and its investment in Scale AI dragged on for more than two months. This unprecedented speed reflects both sides’ urgency and shared conviction about the AI agent赛道 (AI agent赛道 translates to "AI agent sector" or "AI agent track," but since direct translation is required without interpretation, we keep it as is in context)—the AI agent space.

Second, the valuation surge was staggering. Before the acquisition, Manus was in the middle of a new funding round with a valuation of just $2 billion. Meta’s offer, however, jumped straight into the “multi-billion-dollar” range—more than doubling its valuation in a very short time.

Even more telling: this three-year-old company reported only $125 million in annualized revenue as of early 2025, yet it commanded a premium acquisition price from a tech giant. This “high premium” has forced the market to reassess the value of AI applications—proving that AI products solving real problems are far more valuable than abstract technical concepts.

Third, there’s the complex narrative of a Chinese-founded team meeting a global fate. Butterfly Effect, Manus’s parent company, was founded by Xiao Hong, a graduate of Huazhong University of Science and Technology, who started his entrepreneurial journey in Wuhan and initially built tools within China’s WeChat ecosystem.

Now, this team—with unmistakable Chinese roots—has been acquired by a U.S. tech giant, and its founder will assume the role of Vice President at Meta. Many are left wondering: It’s great that Chinese entrepreneurs can compete globally in AI—but why do the most innovative applications keep getting harvested by overseas giants?

This mix of pride and regret has elevated the deal beyond commerce into a case study for reflection on the broader tech ecosystem.

II. Core Deal Breakdown

Setting aside emotional reactions, let’s analyze the transaction through a financial lens: What exactly did Meta buy, and what did Manus gain?

First, the key deal terms. According to LatePost’s exclusive report, the transaction is valued in the “multi-billion-dollar” range. While the exact figure remains undisclosed, it’s confirmed as Meta’s third-largest acquisition ever—after its 22billionWhatsAppdealandits14 billion purchase of a 49% stake in Scale AI.

Manus’s pre-acquisition funding trajectory is also revealing:

Seed round (Feb 2023): $14 million valuation

Series B (Apr 2025): $500 million valuation

Acquisition (Dec 2025): Valuation more than doubled again—an exceptionally rare growth curve for an AI startup.

Now, let’s examine each side’s core assets.

What Meta bought isn’t just another app—it acquired three layers of strategic value:

Technical capability: Manus’s three-layer architecture—planning, execution, and verification—can invoke 23 types of toolchains and achieved an 86.5% accuracy rate on the GAIA benchmark, far surpassing competitors.

Commercial validation: Its subscription model is already proven, with over $100 million in annualized revenue—enabling Meta to see near-term returns on its AI investments.

Team excellence: Founder Xiao Hong’s product intuition and the team’s execution prowess fill a critical gap in Meta’s engineer-dominated culture.

In return, Manus gains access to Meta’s ecosystem. Per the agreement, Butterfly Effect will remain independently operated—shielding the team from corporate assimilation—while gaining integration with Meta’s Llama large models, Quest VR hardware, and access to billions of global social users to scale its agent capabilities across new scenarios.

Xiao Hong stated plainly upon announcement: “Partnering with Meta allows us to build on a much stronger foundation without changing how we operate.”

III. Strategic Context

In a public letter this past July, Zuckerberg wrote: “Superintelligence will usher in a new era of personal empowerment—giving people greater agency to shape the world according to their own vision.”

For Meta, acquiring Manus is a critical step toward realizing this “superintelligence” vision, as it intensifies its AI race against OpenAI, Google, and Microsoft.

There are no random acquisitions. This “lightning marriage” stems from Meta’s urgent needs and Manus’s growth constraints—making their union almost inevitable.

Zuckerberg has declared AI Meta’s top priority and pledged $600 billion over the next three years to U.S. infrastructure projects, most tied to AI. This summer, Meta offered nine-figure annual salaries to top AI researchers and restructured its AI teams.

Let’s break down the motivations:

Meta’s AI anxiety: AI has surpassed even the metaverse as Meta’s “number one strategic priority.” To catch up with OpenAI and Google, Meta has open-sourced Llama, launched a superintelligence lab, and spent $14 billion on Scale AI. Yet a glaring problem remains: strong foundational technology but no killer application.

Most of Meta’s AI capabilities remain confined to labs, failing to translate into user-facing products or clear monetization paths. Investors are growing impatient—endless spending without returns isn’t sustainable. Enter Manus: a commercially validated solution that plugs Meta’s biggest gap.

Manus’s growth bottlenecks: Despite rapid growth, the startup faced three major challenges:

Soaring compute costs—as AI agents require massive computational resources that scale with user growth;

Lack of ecosystem defenses—difficult to compete alone against tech giants’ walled gardens;

Regulatory complexity—global operations demand robust compliance capabilities across jurisdictions.

All these are areas where a giant like Meta excels.

More broadly, the AI competition has entered a new phase. The early years focused on model size and compute power. By 2025, the consensus has shifted: the winner will be whoever makes AI genuinely useful in real tasks.

Gao Yuncheng, CEO of Hillhouse Capital, predicted early: “2026 will be the year of mass adoption for AI Agents.” Meta’s move is a preemptive strike to secure leadership in this golden赛道 (again, keeping original term)—ensuring it doesn’t fall behind in the next generation of AI.

IV. The Capital Journey

Butterfly Effect’s fundraising path reads like a textbook case of modern tech globalization:

Chinese team builds startup

Chinese capital provides early fuel

U.S. giant acquires at peak valuation

Timeline:

Feb 2023: ZhenFund leads seed round at $14M valuation

Aug 2023: ZhenFund again, now at $50M

Nov 2024: Sequoia China, Tencent, and other “Chinese tech + top VCs” join Series A at $85M

Apr 2025: Benchmark Capital (U.S.) leads Series B, rocketing valuation to ~$500M

Dec 2025: Pre-acquisition fundraising at $2B—then Meta swoops in with a “multi-billion-dollar” offer

V. Industry Implications

Meta’s acquisition of Manus isn’t an isolated event—it’s a turning point that will reshape competitive rules, supply chain structures, and investment logic across the AI industry.

First shift: A wave of M&A will sweep the AI Agent sector.

Meta’s move sets the tone: AI agents are non-negotiable battlegrounds. Expect Google, OpenAI, ByteDance, Tencent, and others to follow. Mid-sized AI startups will face a stark choice: get acquired or go deep in niche verticals.

The era of organic, independent growth is ending. Giants are now “harvesting” talent and tech directly. Competition is shifting from pure R&D to ecosystem integration.

One investor predicts: “In the next year, AI Agent startups will either secure backing from a giant or get acquired. Pure independence will become increasingly untenable.”

Second shift: Clear division of labor will emerge across the AI stack.

Previously, everyone tried to do everything—giants built models and apps; startups aimed for full-stack control.

Post-acquisition, a three-tier structure will crystallize:

Foundation layer: Large models (Meta/Llama, Google/Gemini)

Agent layer: Intelligent agent platforms (Manus-type companies)

Application layer: Vertical-specific solutions (healthcare, education, etc.)

Giants will dominate the bottom two layers and set the rules. Startups will survive only by building differentiated applications in specialized domains.

Third shift: A moment of reckoning for China’s tech ecosystem.

Manus’s journey is emblematic: Chinese team, Chinese early investors (ZhenFund, Sequoia China, Tencent), ultimately acquired by a U.S. giant at a massive premium.

This raises hard questions: Why can we incubate world-class AI application teams but not retain them? Early acquisition offers from Chinese tech firms were reportedly in the tens of millions—orders of magnitude below Meta’s bid. This valuation gap reveals a systemic lag in recognizing the true worth of applied AI innovation.

More profoundly: When China’s best AI outcomes consistently flow overseas, does our tech ecosystem suffer from structural failure?

Risks to watch:

Antitrust scrutiny: U.S. and EU regulators are highly sensitive to big tech’s consolidation in core AI sectors. This deal could trigger investigations, delaying integration.

Technology misalignment: The AI agent paradigm is still evolving. If Manus’s approach diverges from mainstream trends, Meta’s multi-billion-dollar bet could go sour.

VI. Industry Ripple Effects

Research estimates that while a DeepSeek conversation uses ~1k tokens, a single Manus agent task consumes up to 100k tokens—100x more compute.

Agent-driven applications will dramatically increase inference compute demand, creating long-term tailwinds for domestic (Chinese) AI chipmakers.

Moreover, Manus-style “AI democratization” could spark a new wave of entrepreneurship. By enabling model composition and multi-model orchestration, it allows developers to build transformative products without deep expertise in model internals or fine-tuning.

Industries like healthcare, education, and finance—rich in proprietary data—will reduce reliance on big model vendors. Using open-source models and simple API calls, they can build high-performance agents tailored to their needs.

Conclusion: In the AI Era, Tickets Aren’t Free

Behind Meta’s acquisition of Manus lies a giant’s bet on the future—and an inevitable stage in AI’s evolution. This deal sends a clear message: the AI race has moved from “technical arms race” to “application trench warfare.” Real-world problem-solving AI products are where real value lies.

For investors, two opportunities stand out:

Platform companies with defensible AI agent technology

Vertical leaders with proven application落地 (application implementation)

Meanwhile, undifferentiated small AI firms will likely be squeezed out.

Rather than mourn Manus’s “departure,” we should reflect: How do we build an ecosystem where world-class AI application teams choose to stay and grow locally? That’s far more important than merely producing a few headline-grabbing startups.

True technological strength isn’t measured by how many teams we create for others to acquire—but by how many we retain and empower to become global leaders in their own right.

This acquisition isn’t an end—it’s the beginning of the AI Agent era.

In the year ahead, expect more giant moves, more consolidation, and continuous reshaping of the tech landscape. For all of us, understanding this transformation is the only way to avoid being left behind in the AI age.

The AI era encompasses not just new forms of AI civilization and science, but also enduring disciplines—philosophy, politics, economics, and business—that demand deeper foundational thinking.

Against this backdrop, BijiXia has launched China’s first PPE (Politics, Philosophy, Economics) program designed specifically for entrepreneurs—aimed at returning to the source of decision-making, rebuilding foundational cognitive frameworks, and equipping leaders with strategic clarity for the next five years.

We invite you—if you are:

A founder or CEO leading your company at the industry frontier

A decision-maker in a market-leading enterprise or a key leader within a large organization

Someone eager to navigate the AI wave with precision and make smarter, more efficient strategic choices

Curious about business, passionate about deep thinking, and convinced of the power of rational decision-making

Willing to exchange ideas and co-explore the future with fellow leaders

As 2025 draws to a close, a single acquisition has captured the attention of the global tech and investment communities.

In a lightning-fast deal, a Chinese AI startup has been thrust into the global spotlight.

On December 30, 2025, LatePost exclusively reported that Meta has completed the acquisition of Butterfly Effect—the developer behind the AI application Manus—for several billion dollars. The entire negotiation process was astonishingly swift, taking just over ten days.

When Meta officially announced its multi-billion-dollar acquisition of Manus, an AI agent company, the entire industry took notice. Not only is this Meta’s third-largest acquisition in its history, but it’s also widely seen as a landmark event signaling a pivotal shift in the AI race—from competing on model parameters to competing on real-world application deployment.

Some call it Zuckerberg buying a ticket onto the AI ship; others lament the “loss” of a top Chinese startup. But the significance of this deal goes far beyond a simple capital transaction.

Today, we unpack the investment logic hidden behind this landmark acquisition.

I. Why Is Everyone Watching This Deal?

Tech M&A happens daily—so why has Meta’s acquisition of Manus triggered such widespread attention? The answer lies in three “rule-breaking” details, each striking a nerve in the industry.

First, the speed of negotiations was extraordinary. Liu Yuan, a partner at ZhenFund and an early angel investor in Manus, admitted in an interview: “It was so fast I even doubted whether the offer was real.”

From initial contact to final agreement, the entire deal took just over ten days—a veritable “lightning war” for a cross-border acquisition worth billions of dollars.

For context: Meta’s WhatsApp acquisition took nearly a month to negotiate, and its investment in Scale AI dragged on for more than two months. This unprecedented speed reflects both sides’ urgency and shared conviction about the AI agent赛道 (AI agent赛道 translates to "AI agent sector" or "AI agent track," but since direct translation is required without interpretation, we keep it as is in context)—the AI agent space.

Second, the valuation surge was staggering. Before the acquisition, Manus was in the middle of a new funding round with a valuation of just $2 billion. Meta’s offer, however, jumped straight into the “multi-billion-dollar” range—more than doubling its valuation in a very short time.

Even more telling: this three-year-old company reported only $125 million in annualized revenue as of early 2025, yet it commanded a premium acquisition price from a tech giant. This “high premium” has forced the market to reassess the value of AI applications—proving that AI products solving real problems are far more valuable than abstract technical concepts.

Third, there’s the complex narrative of a Chinese-founded team meeting a global fate. Butterfly Effect, Manus’s parent company, was founded by Xiao Hong, a graduate of Huazhong University of Science and Technology, who started his entrepreneurial journey in Wuhan and initially built tools within China’s WeChat ecosystem.

Now, this team—with unmistakable Chinese roots—has been acquired by a U.S. tech giant, and its founder will assume the role of Vice President at Meta. Many are left wondering: It’s great that Chinese entrepreneurs can compete globally in AI—but why do the most innovative applications keep getting harvested by overseas giants?

This mix of pride and regret has elevated the deal beyond commerce into a case study for reflection on the broader tech ecosystem.

II. Core Deal Breakdown

Setting aside emotional reactions, let’s analyze the transaction through a financial lens: What exactly did Meta buy, and what did Manus gain?

First, the key deal terms. According to LatePost’s exclusive report, the transaction is valued in the “multi-billion-dollar” range. While the exact figure remains undisclosed, it’s confirmed as Meta’s third-largest acquisition ever—after its 22billionWhatsAppdealandits14 billion purchase of a 49% stake in Scale AI.

Manus’s pre-acquisition funding trajectory is also revealing:

Seed round (Feb 2023): $14 million valuation

Series B (Apr 2025): $500 million valuation

Acquisition (Dec 2025): Valuation more than doubled again—an exceptionally rare growth curve for an AI startup.

Now, let’s examine each side’s core assets.

What Meta bought isn’t just another app—it acquired three layers of strategic value:

Technical capability: Manus’s three-layer architecture—planning, execution, and verification—can invoke 23 types of toolchains and achieved an 86.5% accuracy rate on the GAIA benchmark, far surpassing competitors.

Commercial validation: Its subscription model is already proven, with over $100 million in annualized revenue—enabling Meta to see near-term returns on its AI investments.

Team excellence: Founder Xiao Hong’s product intuition and the team’s execution prowess fill a critical gap in Meta’s engineer-dominated culture.

In return, Manus gains access to Meta’s ecosystem. Per the agreement, Butterfly Effect will remain independently operated—shielding the team from corporate assimilation—while gaining integration with Meta’s Llama large models, Quest VR hardware, and access to billions of global social users to scale its agent capabilities across new scenarios.

Xiao Hong stated plainly upon announcement: “Partnering with Meta allows us to build on a much stronger foundation without changing how we operate.”

III. Strategic Context

In a public letter this past July, Zuckerberg wrote: “Superintelligence will usher in a new era of personal empowerment—giving people greater agency to shape the world according to their own vision.”

For Meta, acquiring Manus is a critical step toward realizing this “superintelligence” vision, as it intensifies its AI race against OpenAI, Google, and Microsoft.

There are no random acquisitions. This “lightning marriage” stems from Meta’s urgent needs and Manus’s growth constraints—making their union almost inevitable.

Zuckerberg has declared AI Meta’s top priority and pledged $600 billion over the next three years to U.S. infrastructure projects, most tied to AI. This summer, Meta offered nine-figure annual salaries to top AI researchers and restructured its AI teams.

Let’s break down the motivations:

Meta’s AI anxiety: AI has surpassed even the metaverse as Meta’s “number one strategic priority.” To catch up with OpenAI and Google, Meta has open-sourced Llama, launched a superintelligence lab, and spent $14 billion on Scale AI. Yet a glaring problem remains: strong foundational technology but no killer application.

Most of Meta’s AI capabilities remain confined to labs, failing to translate into user-facing products or clear monetization paths. Investors are growing impatient—endless spending without returns isn’t sustainable. Enter Manus: a commercially validated solution that plugs Meta’s biggest gap.

Manus’s growth bottlenecks: Despite rapid growth, the startup faced three major challenges:

Soaring compute costs—as AI agents require massive computational resources that scale with user growth;

Lack of ecosystem defenses—difficult to compete alone against tech giants’ walled gardens;

Regulatory complexity—global operations demand robust compliance capabilities across jurisdictions.

All these are areas where a giant like Meta excels.

More broadly, the AI competition has entered a new phase. The early years focused on model size and compute power. By 2025, the consensus has shifted: the winner will be whoever makes AI genuinely useful in real tasks.

Gao Yuncheng, CEO of Hillhouse Capital, predicted early: “2026 will be the year of mass adoption for AI Agents.” Meta’s move is a preemptive strike to secure leadership in this golden赛道 (again, keeping original term)—ensuring it doesn’t fall behind in the next generation of AI.

IV. The Capital Journey

Butterfly Effect’s fundraising path reads like a textbook case of modern tech globalization:

Chinese team builds startup

Chinese capital provides early fuel

U.S. giant acquires at peak valuation

Timeline:

Feb 2023: ZhenFund leads seed round at $14M valuation

Aug 2023: ZhenFund again, now at $50M

Nov 2024: Sequoia China, Tencent, and other “Chinese tech + top VCs” join Series A at $85M

Apr 2025: Benchmark Capital (U.S.) leads Series B, rocketing valuation to ~$500M

Dec 2025: Pre-acquisition fundraising at $2B—then Meta swoops in with a “multi-billion-dollar” offer

V. Industry Implications

Meta’s acquisition of Manus isn’t an isolated event—it’s a turning point that will reshape competitive rules, supply chain structures, and investment logic across the AI industry.

First shift: A wave of M&A will sweep the AI Agent sector.

Meta’s move sets the tone: AI agents are non-negotiable battlegrounds. Expect Google, OpenAI, ByteDance, Tencent, and others to follow. Mid-sized AI startups will face a stark choice: get acquired or go deep in niche verticals.

The era of organic, independent growth is ending. Giants are now “harvesting” talent and tech directly. Competition is shifting from pure R&D to ecosystem integration.

One investor predicts: “In the next year, AI Agent startups will either secure backing from a giant or get acquired. Pure independence will become increasingly untenable.”

Second shift: Clear division of labor will emerge across the AI stack.

Previously, everyone tried to do everything—giants built models and apps; startups aimed for full-stack control.

Post-acquisition, a three-tier structure will crystallize:

Foundation layer: Large models (Meta/Llama, Google/Gemini)

Agent layer: Intelligent agent platforms (Manus-type companies)

Application layer: Vertical-specific solutions (healthcare, education, etc.)

Giants will dominate the bottom two layers and set the rules. Startups will survive only by building differentiated applications in specialized domains.

Third shift: A moment of reckoning for China’s tech ecosystem.

Manus’s journey is emblematic: Chinese team, Chinese early investors (ZhenFund, Sequoia China, Tencent), ultimately acquired by a U.S. giant at a massive premium.

This raises hard questions: Why can we incubate world-class AI application teams but not retain them? Early acquisition offers from Chinese tech firms were reportedly in the tens of millions—orders of magnitude below Meta’s bid. This valuation gap reveals a systemic lag in recognizing the true worth of applied AI innovation.

More profoundly: When China’s best AI outcomes consistently flow overseas, does our tech ecosystem suffer from structural failure?

Risks to watch:

Antitrust scrutiny: U.S. and EU regulators are highly sensitive to big tech’s consolidation in core AI sectors. This deal could trigger investigations, delaying integration.

Technology misalignment: The AI agent paradigm is still evolving. If Manus’s approach diverges from mainstream trends, Meta’s multi-billion-dollar bet could go sour.

VI. Industry Ripple Effects

Research estimates that while a DeepSeek conversation uses ~1k tokens, a single Manus agent task consumes up to 100k tokens—100x more compute.

Agent-driven applications will dramatically increase inference compute demand, creating long-term tailwinds for domestic (Chinese) AI chipmakers.

Moreover, Manus-style “AI democratization” could spark a new wave of entrepreneurship. By enabling model composition and multi-model orchestration, it allows developers to build transformative products without deep expertise in model internals or fine-tuning.

Industries like healthcare, education, and finance—rich in proprietary data—will reduce reliance on big model vendors. Using open-source models and simple API calls, they can build high-performance agents tailored to their needs.

Conclusion: In the AI Era, Tickets Aren’t Free

Behind Meta’s acquisition of Manus lies a giant’s bet on the future—and an inevitable stage in AI’s evolution. This deal sends a clear message: the AI race has moved from “technical arms race” to “application trench warfare.” Real-world problem-solving AI products are where real value lies.

For investors, two opportunities stand out:

Platform companies with defensible AI agent technology

Vertical leaders with proven application落地 (application implementation)

Meanwhile, undifferentiated small AI firms will likely be squeezed out.

Rather than mourn Manus’s “departure,” we should reflect: How do we build an ecosystem where world-class AI application teams choose to stay and grow locally? That’s far more important than merely producing a few headline-grabbing startups.

True technological strength isn’t measured by how many teams we create for others to acquire—but by how many we retain and empower to become global leaders in their own right.

This acquisition isn’t an end—it’s the beginning of the AI Agent era.

In the year ahead, expect more giant moves, more consolidation, and continuous reshaping of the tech landscape. For all of us, understanding this transformation is the only way to avoid being left behind in the AI age.

The AI era encompasses not just new forms of AI civilization and science, but also enduring disciplines—philosophy, politics, economics, and business—that demand deeper foundational thinking.

Against this backdrop, BijiXia has launched China’s first PPE (Politics, Philosophy, Economics) program designed specifically for entrepreneurs—aimed at returning to the source of decision-making, rebuilding foundational cognitive frameworks, and equipping leaders with strategic clarity for the next five years.

We invite you—if you are:

A founder or CEO leading your company at the industry frontier

A decision-maker in a market-leading enterprise or a key leader within a large organization

Someone eager to navigate the AI wave with precision and make smarter, more efficient strategic choices

Curious about business, passionate about deep thinking, and convinced of the power of rational decision-making

Willing to exchange ideas and co-explore the future with fellow leaders