Beijing Unwound a Closed $2 Billion Deal. Closing Certainty Is Now a Political Variable.

要約： Beijing’s order to unwind Meta’s closed Manus deal shows closing certainty now depends on the founders’ home state, so cross-border sellers must price political reversal into signing.

The Deal Closed, Then a One-Line Order Undid It

Meta announced its purchase of Manus, a Singapore-based developer of general-purpose artificial intelligence (AI) agents, on Monday 29 December 2025. Meta kept the terms private, and The Wall Street Journal reported that Meta closed the deal at more than $2 billion¹. A Meta spokesperson added that there would be “no continuing Chinese ownership interests in Manus AI” and that the platform would end its services and operations in China.

Ten days later the Chinese state answered. At its 8 January 2026 press conference, the Ministry of Commerce (MOFCOM) said it would, with other departments, assess and investigate the acquisition’s consistency with export-control, technology import and export, and outbound-investment rules². On 27 April the office that runs China’s foreign-investment security review, housed at the National Development and Reform Commission (NDRC), published its decision. It runs to a single sentence: the office prohibited the investment and required the parties to revoke the acquisition³. It gives no reasons. Meta’s position, as relayed to CNBC⁴, was that the transaction “complied fully with applicable law”.

Compliance followed within weeks. In June, Meta completed an operational split from Manus⁵, telling its staff to stop using Manus tools and cutting Manus employees off from its internal data systems. A Chinese law firm quoted in the same report described the April order as an unprecedented move under the security-review mechanism, one that set in motion the unravelling of a completed deal. It is reported as the first closed transaction China has reversed under these rules.

A share purchase agreement spends most of its pages on the gap between signing and closing: conditions, long-stop dates, break fees, material-adverse-change clauses. Every one of those tools assumes that once the price is paid, the regulatory question is settled. Manus passed signing and closing, and the risk arrived afterwards, from a government standing outside both documents.

Moving to Singapore Left the Leverage in Beijing

Manus had done what a cautious founder team is advised to do. It was founded in China, and according to e27 it moved its team to Singapore and shut its China operations in July 2025⁶. The practice is common enough to carry a nickname, “Singapore-washing”⁴: relocating from China to the city-state to step outside the scrutiny of both Beijing and Washington. Singapore kept its side of the bargain. The same e27 report quotes Singapore minister Chee Hong Tat, who said that Beijing’s intervention had left Singaporean law intact; in his words, it “had not violated Singaporean law”.

The intervention worked anyway, because the levers sat away from the registered office. A home state reaches a company through three attachments that a change of address leaves in place: the founders, who remain its citizens; the technology, whose development history sits inside its export-control perimeter; and the capital, held by investors domiciled within its reach. Manus’s early backers reportedly included Tencent, ZhenFund and HSG¹. MOFCOM’s January statement covered the technology, data and outbound-investment attachments when it reminded companies that outbound investment, technology export, data export and cross-border mergers must follow Chinese law and complete the statutory procedures².

Beijing has since written the improvisation into law. According to CNBC, rules taking effect on 1 July 2026 give China, for the first time, a formal legal basis to force the unwinding of completed overseas transactions⁵, and require approval for cross-border talent transfers in sensitive sectors.

e27 reports that the round of more than $500 million announced on 8 October was led by Boyu Capital and IDG Capital, with Tencent, HSG and ZhenFund joining⁶. Nine months after a buyer promised “no continuing Chinese ownership interests”, e27 describes the company as largely back in the hands of Chinese capital. For Neumarz, the lesson for sellers is direct: after a redomicile, the founders’ home state stays a party to the closing in everything but the signature block.

Four Regulators Now Decide Whether a Deal Closes

A two-column checklist, covering the buyer’s investment screening and antitrust, leaves the Manus risk off the page. The case shows the list needs four columns, and that one of them sits with a third state: the one the founders, the code and the early capital came from. Neumarz uses a closing-certainty map to make the four visible at the term-sheet stage. Each regime is scored on the strongest power it holds over the transaction.

An illustrative case shows how the map reads in practice. Take a hypothetical Zurich software company, 120 staff, acquired by a listed European group. Its two founders hold a passport from a state that runs export and outbound-investment controls, its core code was first written in that state, and a fund domiciled there holds 15 per cent. In this illustration, buyer screening in the acquirer’s home state scores 1: a filing, with no approval required for a private buyer. Target screening scores 0, because the Swiss Act reaches only foreign state investors and the buyer is privately held. Antitrust scores 2 if the parties’ turnover crosses the merger-control thresholds. The founders’ home state scores 3: the founders, the code and the minority shareholder each give it a route to the deal after closing.

A two-column checklist would record this deal as a 2 and schedule it around the antitrust filing. The map records a 3 and moves the work forward: the home-state question is answered in diligence, written into the conditions and priced in the fee before the buyer’s board approves the price. The same company sold to a buyer from the founders’ own state would score lower in that column and higher in the buyer column, which is why the map belongs in the choice of bidders as well as in the drafting.

Two rules make the map useful. First, the deal’s closing certainty equals its highest score: three regimes at 1 and one at 3 make a reversible closing, however reassuring the average looks. Second, a 2 is manageable and a 3 is the one to price, because a 2 resolves before the money moves while a 3 can arrive after the integration has started. The founders’ home state is the column a two-column checklist leaves blank, and after the Manus order and China’s July rules it is the column most likely to hold a 3.

The map also changes who carries the risk. A buyer’s screening and antitrust sit on the buyer’s side of the table, where conditions precedent and reverse break fees already allocate them. A home-state reversal attaches to the seller’s people and history, which gives the seller both the better information and the larger exposure: the seller gives back the price, and the company returns with its integration half done and its buyer pool reduced.

Switzerland’s Own Screening Law Arrives in 2027

Swiss sellers will soon score the second column at home. Parliament adopted the Investment Screening Act on 19 December 2025, with entry into force expected in 2027⁷, according to the State Secretariat for Economic Affairs (SECO), which will run it and currently accepts no applications. The Act applies to takeovers of Swiss companies by foreign state investors, meaning acquirers controlled by a foreign state.

The adopted text in the Federal Gazette sets the scope in Article 3 in two tiers. The first covers companies with at least 50 full-time posts or CHF 10 million of turnover⁸ in defence-critical goods, electricity transmission and large power plants, high-pressure gas pipelines, water supply for more than 100,000 people and security-relevant IT for federal authorities. The second covers companies with at least CHF 100 million of turnover in hospitals, pharmaceuticals, medical devices and vaccines, transport hubs, railways, food distribution centres, telecoms networks, systemic financial-market infrastructure and systemically important banks. Article 3(3) lets the Federal Council add categories for up to twelve months, renewable once, so the list at signing can grow before closing.

The Swiss design differs from the Chinese one in the place it puts the risk. Under Article 8(4), a takeover that needs approval takes civil-law effect only once it is approved⁸, so the political decision lands before closing, where a timetable can absorb it. Article 8(2)(b) sends any decision “of considerable political significance” to the Federal Council itself, which writes the political variable into the statute openly. The reversal power exists for deals that skip the queue: Article 19 lets the Federal Council order the measures needed to restore the lawful position, including divestment, when an approvable takeover closes without approval, when an approval rested on false information, or when a condition attached to it is breached, and Article 20 sets an administrative sanction of up to 10 per cent of the Swiss company’s average worldwide annual turnover⁸ over the two financial years before the takeover.

For a Swiss seller, the reading is two-sided. The Swiss column will usually score 0 or 2: visible, front-loaded and priceable, and a 3 only for parties who close without the required approval, obtain it on false information or breach its conditions after closing. Those three triggers belong in the warranties and covenants of any Swiss deal with a state-controlled buyer. The column that can still score 3 after a clean Swiss closing is the founders’ home state, which is why a Swiss company with founders, engineers or early capital from a state that runs export and outbound controls needs the map more urgently than one facing the new Swiss Act.

What to Write Into the Signing Documents

The response belongs in the documents, before the first draft circulates. Each of the five measures below converts a political reversal from an open-ended exposure into a priced, allocated term, and each names the party that acts on it first.

Five Terms Turn Political Reversal Into a Priced Risk

1. Map the origins (seller and adviser). Record, for the target, where each founder holds citizenship and residence, where the core technology was developed, and where each shareholder is domiciled. Every state on that list gets a column in the closing-certainty map.
2. Make home-state approval a condition of closing (buyer). Require approvals from the founders’ home state, or a written confirmation that none is needed, before completion, and size the long-stop date to the slowest of the four regimes.
3. Name political reversal in the reverse break fee (seller and buyer). Draft the trigger to cover a prohibition issued by any of the four regimes, including one issued after signing, and state in writing which side bears a reversal ordered by the seller’s home state.
4. Add an unwind covenant backed by escrow (buyer). Set out in advance how shares, price, data, intellectual property and staff return if a reversal order arrives after closing, using the separation Meta carried out in June as the work list, and hold part of the price in escrow for the period in which an order remains possible.
5. Keep a fallback buyer warm (seller and adviser). Run a second bidder through to closing, ideally one whose own home state the founders’ state accepts. A deal of the Manus size sits in the single-digit-billion band that megafunds now leave to a narrower set of owners, so the second bidder is worth securing early, on the two-track logic of Neumarz’s buyer map for a repriced credit market.

Political reversal will remain a feature of cross-border technology deals for as long as states treat founders, code and capital as strategic assets. The deals that close and stay closed will be the ones that scored all four columns at the term sheet and wrote the answer into the signing documents.

Neumarz, a Kainjoo SA venture, advises founders, acquirers, investors and family offices on cross-border transactions, including the regulatory map and the signing terms that hold a closing in place. Parties preparing a cross-border sale or acquisition can request a confidential review of their closing-certainty map.

This article has been reviewed by Haider Alleg, Managing Partner at Kainjoo SA (Terraxis).

This article is general commentary prepared for professional and qualified contacts. It is general information within the meaning of the Swiss Financial Services Act (FinSA), separate from any offer, solicitation, investment advice or personal recommendation, and advice on individual circumstances remains a matter for a dedicated mandate. The closing-certainty map is an illustrative framework. Mandate-specific information is shared under a non-disclosure agreement (NDA) only.

参考文献

1. Associated Press, via ABC7. “Meta buys startup Manus in latest move to advance its AI efforts,” 30 December 2025 (citing The Wall Street Journal). https://abc7news.com/post/meta-buys-startup-manus-latest-move-advance-artificial-intelligence-efforts/18333425/
2. Ministry of Commerce of the People’s Republic of China. “商务部召开例行新闻发布会（2026年1月8日）” (regular press conference, 8 January 2026). https://www.mofcom.gov.cn/xwfbzt/2026/swbzklxxwfbh2026n1y8r/index.html
3. National Development and Reform Commission, Office of the Foreign Investment Security Review Working Mechanism. “对外资收购Manus项目作出安全审查决定” (security review decision on the foreign acquisition of Manus), 27 April 2026. https://zfxxgk.ndrc.gov.cn/web/iteminfo.jsp?id=20623
4. CNBC. “China blocks Meta’s $2 billion takeover of AI startup Manus,” 27 April 2026. https://www.cnbc.com/2026/04/27/meta-manus-china-blocks-acquisition-ai-startup.html
5. CNBC. “Meta reportedly begins dismantling $2 billion Manus deal on Beijing’s orders,” 12 June 2026. https://www.cnbc.com/2026/06/12/meta-reportedly-begins-dismantling-2-billion-manus-deal-on-beijings-orders.html
6. e27. “Manus came home, and Singapore couldn’t stop it,” October 2026. https://e27.co/manus-came-home-and-singapore-couldnt-stop-it-20261013
7. State Secretariat for Economic Affairs (SECO). “Investment screening.” https://www.seco.admin.ch/en/investment-screening
8. Swiss Confederation. Bundesgesetz über die Prüfung ausländischer Investitionen (Investitionsprüfgesetz, IPG) of 19 December 2025, Federal Gazette BBl 2026 31, Articles 3, 8, 19 and 20. https://www.fedlex.admin.ch/eli/fga/2026/31/de