After writing off 2 billion USD, is Manus planning to list in Hong Kong to make a comeback?

null

Text | Shadow Memo

April 27, 2026, a notice from the Office of the Foreign Investment Security Review Mechanism of the National Development and Reform Commission ignited a frenzy in the tech circle.

Meta’s proposed acquisition of AI star project Manus for over $2 billion was officially banned after months of review. The technical team, ad management tools, and AI strategic layout that Meta spent months integrating all had to be rolled back to the starting point.

Once the news broke, the market was in an uproar. But the real dramatic turn came less than a month later—Manus’s three founders Xiao Hong, Ji Yichao, and Zhang Tao were reported to be discussing raising about $1 billion from external investors to buy back the company from Meta, with a target valuation of at least $2 billion, and plans for a subsequent IPO in Hong Kong.

From being acquired to being forced to divest, then to self-funded buyback and seeking independent listing, Manus’s script was completely rewritten in just five months.

For Manus, going public is not an option but a must-answer question.

A $2 billion “shotgun marriage” halted by a red card

To understand why Manus must go public, we need to go back to the acquisition that was once dubbed “the fastest record of domestic AI team being acquired by a giant.”

On December 30, 2025, Meta officially announced the completion of its acquisition of Manus, with a deal valuation exceeding $2 billion. It was disclosed that Manus’s recurring revenue had already surpassed $100 million annually before the acquisition, only eight months after its product launch.

An AI product built by a Chinese post-90s team, from zero to $100 million in annual revenue in just eight months—this number is unparalleled in the global SaaS industry.

How exaggerated was the acquisition price? Sources revealed that negotiations from initial contact to signing took only about ten days. Liu Yuan, partner at ZhenFund, even said, “I was so surprised I thought it was a fake offer.” Meta’s eagerness stemmed from its inability to afford losing.

At that time, in the AI battlefield, OpenAI and Google were competing over parameters and foundational models, but Zuckerberg’s biggest trump card was still the open-source Llama, while Manus was a brand that could fight—it had already validated a complete commercialization path for AI Agents among demanding global paying customers.

Manus’s product differs from mainstream chatbots. It adopts a “large model + cloud virtual machine” architecture, capable of autonomously understanding tasks, calling tools, and completing complex deliveries, closing the loop from instructions to results.

In the field of general AI Agents, Manus, Genspark, and Flowith are regarded as the main axes by Tsinghua research institutions. This kind of “truly capable AI,” is exactly the capability puzzle Meta urgently needs to fill.

However, this seemingly perfect “shotgun marriage” soon hit an insurmountable regulatory wall.

On January 8, 2026, China’s Ministry of Commerce publicly stated that it would conduct an evaluation and investigation with relevant departments on the consistency of this transaction with laws and regulations such as export control and technology import-export.

On April 27, the review conclusion was finalized: the Office of the Foreign Investment Security Review Mechanism of the National Development and Reform Commission officially made a decision to prohibit the investment, requiring the parties to revoke the acquisition.

Notably, this is the first publicly disclosed security review decision involving artificial intelligence since the implementation of the “Measures for the Security Review of Foreign Investment” in 2021, and it was made with a “ban” rather than “conditional approval.”

Previously, Manus had chosen to block some Chinese regional users to reduce regulatory attention, but ultimately still failed to evade the security review. A founder of a Beijing AI startup told the media directly: “The rejection of Manus’s acquisition has great policy educational significance for the entire industry.” He added, “The industry has a real need for global capital and market channels, but the bottom line of security in key technology fields must not be touched.”

Why was Manus stopped? Because it touched a seemingly abstract but actually clear red line. Although Manus had moved its headquarters to Singapore and made organizational internationalization adjustments, its core technology R&D, data accumulation, and business origins are closely linked to China.

From the regulators’ perspective, changing the company’s registration location does not change the true ownership of technology and data, nor does it constitute grounds for circumvention of review.

This “substance over form” judgment essentially blocks a common cross-border M&A route: establishing offshore structures or “relocating headquarters” to transfer technology assets cross-border and exit capital.

Understanding this, the question of “Singapore company acquired by a U.S. company, how will Chinese regulators handle it” no longer holds—regulatory review of foreign investment is misunderstood as a territorial registration review, ignoring the substantive review focus: whether control rights or key assets originating from China are substantially transferred to foreign investors.

In other words, Manus’s “exit” is not just a business act. Its team’s exit and technology export constitute a complete chain that triggers national security review in the eyes of regulators.

The suspension of the acquisition deal is a blow to Meta’s strategic chess piece; for Manus’s founding team and investors, it’s more like a nightmare falling from the clouds.

According to media reports, if the acquisition is revoked, the efforts of investors like ZhenFund and Sequoia Capital over more than two years will be wasted, the unrealized gains of top institutions will vanish, and Manus itself will fall from its peak valuation of $2 billion back to the level before its 2025 Series B financing.

But for the Manus team, the hardest part isn’t the money lost, but the feeling of “neither being able to leave nor come back”—it has already been absorbed into Meta’s ecosystem. To survive independently, all assets must be dismantled and rebuilt. This is the huge challenge facing today’s buyback plan.

A $1 billion buyback, a “self-redemption” risking their own assets

The halt is not the end but the beginning of another story.

In May this year, the buyback plan of Manus’s three founders began to surface. According to Bloomberg, Xiao Hong, Ji Yichao, and Zhang Tao are discussing raising about $1 billion from external investors to buy back the company from Meta, with a valuation at least matching the original $2 billion.

If there is still a funding gap, the founders may also chip in personally to make up the difference. Sources revealed that after the buyback, Manus plans to establish a Chinese joint venture with the investors, then list in Hong Kong.

It’s important to understand that by the time the news of the acquisition being halted broke, this deal had already been deeply completed in practice. Meta and Manus announced the acquisition as early as December 2025. By February this year, Manus team members had already moved into Meta’s Singapore office, with access to Meta’s corporate accounts and internal systems.

Over 100 Manus employees were integrated into Meta’s Superintelligence Labs department in early March.

This means Manus’s tech stack, data, and team are already deeply embedded in Meta’s system. To disentangle, everything must be dismantled—terminate employment, revoke system access, reassign personnel—all requiring the consent of involved individuals.

Some analysts bluntly said that revoking this already completed deal would be “time-consuming, complex, and difficult” in practice.

More tricky is the question of how high Manus’s self-developed barriers really are, which remains controversial. Manus itself does not own a self-developed foundational model; its Agent capabilities are built on third-party models like Claude, raising questions about technical barriers. Academic analysis points out that Manus’s $2 billion valuation is more driven by FOMO and optimistic growth narratives rather than sustainable financial fundamentals.

From the competitive environment, in the year-plus since Manus’s overseas expansion, acquisition, and buyback, the landscape of domestic AI Agent track has been thoroughly reshaped.

Zhipu launched AutoGLM 2.0 this year, equipped with cloud phones and cloud computers for cross-end autonomous execution; ByteDance’s Koushi and Alibaba’s Wukong have entered different application scenarios; Tencent’s Workbuddy and other products are rapidly iterating.

All these competitors share a common point: they possess self-developed models, have already gone public in Hong Kong, and have strong financial reserves.

If this is the case, even if Manus successfully buys back from Meta, it’s no longer the market it once dominated in 2025. Its first-mover advantage has been eroded by relentless industry evolution.

Yet, even so, the founding team still chose the buyback route. Why fight so hard? Because not buying back means outright exit. If the forced split is not taken over by anyone, Manus’s value will evaporate quickly.

The founders are willing to bear huge financial pressure and operational difficulties to buy back, which is less a business choice and more a survival instinct. Only by regaining control of the company can they have a chance to prove themselves again in the market.

According to reports, Manus’s projected revenue in 2026 will reach about $1 billion, providing strong confidence for investors.

HKEX, the “golden landing spot” for AI unicorns

Where to go after the buyback? The answer is already on the table: Hong Kong.

In the past year, Hong Kong’s capital market has undergone a profound “tech transformation.” In 2025, IPO fundraising in Hong Kong surged by 200%, reaching HKD 285.8 billion, surpassing Nasdaq and reclaiming the top spot globally.

Among this wave of IPOs, AI companies contributed the most—ranging from GPU chips to foundational large models, companies along the AI supply chain rushed into Hong Kong’s Central with unprecedented density.

Specifically, in the first two weeks of 2026, Hong Kong Stock Exchange witnessed an unprecedented concentration of AI listings: on January 2, BeiRan Technology listed with 2,348 times oversubscription, freezing HKD 130 billion; on January 8, Zhipu and TianShu ZhiXin listed on the same day; on January 9, MiniMax’s first-day surge reached 109%, with a market cap exceeding HKD 100 billion.

Zhipu’s stock price exceeded HKD 400 billion in market value within a month of listing, with astonishing growth. By the end of March 2026, a total of 34 companies had listed in Hong Kong this year, raising over HKD 100 billion, a more than 460% increase compared to the same period last year.

Why are AI companies flocking to HKEX? The reasons are quite simple.

First, HKEX’s listing system is highly inclusive of Chinese-background AI companies. Deloitte China’s South China Capital Markets Partner Lu Zhihong pointed out that Chinese AI companies choose to list in Hong Kong because of its advantageous listing system and internationalized capital market, which can efficiently gather capital from mainland China and globally, providing an international shareholder base and valuation discovery platform.

In the context of geopolitical uncertainties, Hong Kong has become the most important offshore platform for Chinese tech firms seeking stable international financing channels.

Second, the Hong Kong market has a strong tolerance for “high growth and high R&D investment” AI companies. Data from listed AI firms show a common pattern of “high growth coupled with high losses”—Zhipu’s R&D expenditure in the first half of 2025 was over 8 times its revenue; MiniMax’s computing power costs accounted for 70-80% of R&D investment. Yet, investors in Hong Kong have shown considerable tolerance for these phenomena in their valuation.

Third, Hong Kong stocks have a mature “placing” mechanism. Deloitte’s Lu Zhihong explained that after listing, companies can efficiently and conveniently raise additional R&D funds through Hong Kong’s financing channels. This is crucial for AI firms that need continuous funding for R&D. HKEX not only provides a platform for financing but also a continuous blood supply.

For Manus, HKEX is not just a listing choice—it’s the only capital platform capable of supporting its subsequent financing needs.

Zhipu’s annual revenue is about HKD 7.24 billion (roughly $1 billion), with a market cap of HKD 400 billion; MiniMax’s annual revenue is less than $80 million, yet it has a valuation of hundreds of billions of HKD. In comparison, Manus had already achieved $100 million ARR by December 2025, with its monthly run rate still climbing.

If Manus can land on HKEX in this market sentiment, it has a real chance to obtain a valuation much higher than what Meta bid at that time.

Remember, the capital market is never priced based on your current revenue but on your story, your track, and your growth potential. Manus’s story in 2025 was “China AI going abroad,” but today, it is transforming into a new narrative: a leading AI Agent company that has experienced life-and-death trials globally, reclaimed control from giants, and is poised to rise independently.

Perhaps only by going public can Manus find another way out.

Many might ask: why doesn’t Manus accept other big tech’s acquisition offers? For example, Tencent is now one of its former shareholders. Or, why not continue with the old path of private fundraising?

The answer is very clear: for Manus, IPO is not just one of many options—it’s the only viable path.

First, the route of seeking to be acquired by giants is completely blocked. The signals from the suspension of Manus’s acquisition are very clear: in today’s global tech competition where AI is at the core, any Chinese-tech-origin AI company selling to U.S. giants will face extremely strict national security reviews.

These reviews are not “conditional” but may be outright “prohibited.” Even if Manus relocates its headquarters and internationalizes its team to “de-Chinese-ify,” it cannot evade the principle of “substance over form” in review.

Second, pure private funding cannot solve the fundamental problem. From a regulatory perspective, Manus’s structure needs recognition within China’s legal framework, and purely foreign capital backgrounds face many uncertainties in market access. Restructuring Manus into a joint venture and then listing in Hong Kong can both ensure greater compliance certainty and provide long-term capital support for its internationalization and R&D.

Insiders say that after the buyback plan is implemented, Manus will establish a domestic joint venture with new investors—essentially “reintegrating” the company into China’s regulatory system. And this path naturally points toward one key goal: going public.

Third, the essence of AI technology competition determines that only sustained, high-level R&D investment can maintain competitiveness. A report from Tsinghua University on general intelligent agents pointed out that although Manus is regarded as a representative of autonomous general AI Agents, with the explosion of open-source frameworks like OpenClaw, the industry’s underlying technology is rapidly iterating.

Data shows that by 2028, the share of enterprise software applications with embedded agent AI will jump from less than 1% in 2024 to 33%, and the proportion of daily decisions autonomously made by intelligent agents will rise from less than 1% to 15%. In this fast-evolving track, standing still means falling behind, and falling behind means extinction.

Currently, Manus’s competitive situation is not optimistic. There are reports that its unique visitors and session durations are declining, and user retention remains low. The technological paradigm impact brought by open-source frameworks like OpenClaw is forcing all general AI products to re-evaluate their value.

This indicates that Manus needs to significantly ramp up R&D and product iteration more than ever. And sufficient R&D requires sufficient funding. Going public is the most straightforward way to solve this once and for all.

Fourth, Hong Kong’s IPO market is in a “golden window” in recent years. Since 2026, the Hong Kong IPO market has continued to heat up. By the end of March, 34 new companies had listed, raising over HKD 100 billion.

Market liquidity has improved, global capital is re-evaluating Asian markets, and HKEX’s institutional tilt toward AI companies has created an excellent timing for Manus’s listing.

UBS Global Investment Bank’s Asia-Pacific Equity Capital Markets Co-Head Hu Linghan pointed out, “As the industry develops, these companies will inevitably have financing needs, and more AI supply chain-related companies are expected to list in Hong Kong in 2026.” Missing this wave could mean facing even more crowded competition.

In conclusion, the butterfly effect continues

Three years ago, Beijing Butterfly Effect Technology started from a residential house and created an AI product capable of truly “getting things done.” Three years later, this product experienced a rollercoaster: viral success, global expansion, acquisition by giants, regulatory halt, and buyback for self-rescue.

Manus co-founder Ji Yichao once said in an interview a thought-provoking sentence—“Healthy people are unkillable; those who can humbly stand up again and calmly view external changes, then feed back into decision-making.” Today’s Manus, this sentence fits perfectly.

From a $2 billion merger to being forcibly dismantled by regulators, then to financing buyback, returning to the Chinese market, and seeking independent listing, Manus has undergone a transformation from “being arranged” to “controlling its own destiny.”

This process is painful, complex, and full of uncertainties—but it has pushed Manus onto a more resilient path.

Because the true moat of AI technology has never been something that can be bought with a single acquisition. It requires continuous R&D investment, independent product iteration, and repeated validation and correction through market interaction.

What Meta paid $2 billion for was Manus’s achievement at that moment, not its future potential. When this opportunity is forcibly reset, the Manus founding team chooses to win back the future themselves.

Going public is not the goal but the only reliable channel to secure ongoing R&D investment. When Manus steps onto HKEX, it will gain not just hundreds of billions of HKD in valuation but a long-term ticket to the future.

This ticket doesn’t depend on the face of any giant, nor is it interrupted by acquisition negotiations. It only requires proving to the market that its products are strong enough.

And that’s exactly what Manus excels at.

The butterfly effect continues. Manus’s turning point is just a microcosm of the global AI race. But one thing is certain—it will not be forgotten, because it is one of the few survivors that can keep flapping its wings amid the storm.

Next, it all depends on when the gong of HKEX will ring for it.