When a deal is rejected by regulators, the asset's price typically declines. The pool of potential buyers shrinks, exit avenues narrow, and potential bidders withdraw—a phenomenon well-documented in finance as the "liquidity discount."

Manus, however, followed an opposite trajectory.

The company launched in March 2025 as the "world's first general-purpose AI agent," attracting over 10 million visits to its official website within four hours. Invitation codes were resold for as much as RMB 100,000 on secondary platforms. On December 29, 2025, Meta announced its acquisition of Manus in a deal valued at over USD 2 billion, with the founding team joining Meta and founder Xiao Hong becoming a Meta vice president. On April 27, 2026, China's Foreign Investment Security Review Office (housed under the National Development and Reform Commission) prohibited the investment, demanding the parties revoke the transaction and restore the pre-deal status. This marked the first publicly halted foreign acquisition in the AI sector since the implementation of the *Foreign Investment Security Review Measures* in 2021.

Over five months later, on October 8, Manus' parent company, Butterfly Effect, announced the completion of a new financing round exceeding USD 500 million, led by Boyu Capital and IDG Capital, with participation from existing investors Tencent, Sequoia China, and ZhenFund. According to multiple media reports, including Bloomberg, the post-money valuation reached approximately USD 4 billion.

The same asset doubled in price after being barred by regulators from selling to the highest bidder.

This case demands scrutiny from industry insiders: Why did the price rise? What did regulators gain? And two unanswered questions remain.

01 Two Prices: The Shift in Pricing Power

The valuation ceiling for Chinese AI companies was historically set by U.S. dollar funds and buyers. Manus itself exemplifies this: In April 2025, it completed a USD 75 million financing round led by Benchmark, with a post-money valuation of approximately USD 500 million—priced by Silicon Valley venture capital.

This time was different. The lead investors, Boyu and IDG, are market-driven PE/VC firms, with Tencent, Sequoia China, and ZhenFund following. The approximately USD 4 billion valuation was determined by Chinese-backed capital. The valuation anchor shifted from "what U.S. giants are willing to pay" to "what Chinese capital deems it worth."

Second, the discount transformed into a premium.

Under conventional financial logic, an asset barred from sale to the wealthiest buyer should trade at a discount. Instead, Manus secured a premium. The rationale is straightforward: An asset explicitly prohibited from "going offshore" becomes scarce domestically. Potential buyers knew competitors like Meta or other U.S. giants would no longer participate—essentially a scarcity premium superposition ( superposition : superimposed) policy premium.

Third, the bidders were financial investors, not state-backed entities.

The lead investors were market-driven institutions like Boyu and IDG, not state-owned platforms. This investor structure addresses speculation about "policy-driven arrangements": Pricing resulted from market-based investment decisions, not administrative bailouts.

However, a nominal price doubling does not equate to a per-unit price doubling.

In December 2025, Manus announced annual recurring revenue (ARR) exceeding USD 100 million, with 105 employees and over 147 trillion tokens processed on its platform. By June 2026, its ARR reached USD 400–500 million, according to Caixin and *The Information*. Roughly calculating from Meta's disclosed acquisition price range of USD 2–3 billion, the acquisition multiple was approximately 20–30x ARR. At the current USD 4 billion valuation, the multiple contracts to 8–10x ARR.

Revenue growth outpaced valuation gains. While domestic capital offered a higher total valuation, the per-unit price became more conservative, aligning valuations with revenue metrics.

An unavoidable temporal factor: Most of the ARR growth from USD 100 million to USD 400–500 million occurred while Manus was nominally under Meta's ownership. Some analysts attribute this growth to Meta's advertising systems and key client resources. Post-independence, customer acquisition capabilities will be validated by next quarter's data.

02 From Rejection to Redirection

Halting an acquisition does not, in itself, set a precedent. "Prohibiting the sale" is the easiest step. The real challenge lies in what follows: Will the team disband? Will the company collapse? Will future buyers shy away from Chinese AI?

Manus provides a closed-loop solution:

Identify sensitive assets → Halt foreign M&A via security review → Enable domestic market-driven capital acquisition → Assign higher valuation → Continue commercialization along original path

Actual execution involved an additional step. After the halt, the founding team, Union ( Union : jointly) with existing investors like Tencent, Sequoia China, and ZhenFund, repurchased shares from Meta at a valuation of approximately USD 2 billion (Bloomberg, July). Tencent acquired Benchmark's stake, becoming the largest external institutional investor. Subsequently, Boyu and IDG led the new financing round. On September 1, Manus resumed independent operations under its original founding team of Xiao Hong, Zhang Tao, and Ji Yichao. Domestic capital's acquisition unfolded in two phases: repurchase, then capital increase.

This closed loop ( closed loop : closed loop) transformed regulatory action from a "destructive rejection" into "targeted redirection." The signal sent by regulators was not "AI assets cannot be traded," but rather "AI core assets cannot leave China, but can be properly priced domestically."

As a "precedent," this path can be directly applied to future cases involving foreign acquisitions of model companies, agent firms, or data assets: Halt the deal, guide domestic capital to acquire, and enable market-based pricing.

For companies, the calculus changes. If staying in China yields comparable valuations, the incentive to "pre-emptively relocate offshore" diminishes. Manus' prior path served as a cautionary tale: In June 2025, it relocated its headquarters to Singapore, renamed its operating entity to Butterfly Effect Pte, drastically cut its domestic team, and ceased all services and operations in China. State media later labeled such moves "wash-and-go offshore strategies" and assessed them with eight words: "This path will not work."

Regulators aim to force companies considering this route to recalculate before departing.

03 Demonstration Effect

This narrative conceals a stress test.

If Manus had failed to secure financing post-rejection, suffered a valuation collapse, or lost its team, the regulatory move would have backfired, creating a negative demonstration effect.

The current outcome—asset retention, valuation doubling, and a stronger investor lineup—ensures a positive effect.

Regulatory efficacy ultimately hinges on the capital market's capacity to acquire assets. The NDRC's one-page prohibition order was symbolic; Boyu and IDG's USD 500 million commitment validated the pathway. The "precedent" hinges on domestic markets' ability to price AI assets—the load-bearing structure of the entire closed loop ( closed loop : closed loop).

According to reports, Manus is exploring restructuring to prepare for a Hong Kong IPO. Acquirers face clear exit deadlines, making IPOs inevitable. If the Hong Kong market absorbs the company, the closed loop ( closed loop : closed loop) is complete; if not, "proper domestic pricing" remains confined to the primary market.

04 Unresolved Questions

The first question concerns jurisdictional reach.

Manus' operating entity had relocated to Singapore before the transaction, yet regulators still halted the deal, likely citing "technology, team, and data developed in China." Shen Ziying, a senior partner at Beijing Jincheng Tongda & Neal Law Firm, explained publicly that China's foreign investment security review is "penetrating" and adheres to a "substance over form" principle: "While Meta's acquisition of Manus appeared to involve two offshore entities, the underlying assets—technology, algorithms, data, and talent—all originated in China."

This jurisdictional logic has far-reaching implications: Any AI company developing core technologies in China could fall under review, regardless of its registered location. If confirmed, this would fundamentally alter structural choices for Chinese AI firms.

The second question pertains to valuation transparency.

Whether the approximately USD 4 billion valuation includes liquidation preferences or old share transfers remains undisclosed. Such details affect the credibility of the "100% premium" claim. The multiple contraction (from 20–30x ARR to 8–10x) suggests the nominal doubling masks tighter per-unit pricing. Additionally, the consideration paid by existing shareholders during the repurchase (approximately USD 2 billion) may not align with the new capital's valuation metrics.

Thus, "repricing" accurately refers to total valuation, not unit price. This transaction establishes that "domestic capital can assign higher total valuations to Chinese AI assets," but the portion attributable to scarcity and policy—versus what the market will validate post-IPO—remains unclear until the Hong Kong Stock Exchange discloses prospectus details.

On September 28, Xiao Hong wrote upon releasing Manus 2.0: "I once thought our journey would end as a footnote, but now we have a chance to write a chapter." Originally about the product, this sentiment also applies to the year's capital trajectory. Yet the questions of who prices this chapter and where it concludes are no longer solely up to its authors.