On November 5, 2026, ABB confirmed in an emailed statement what the Financial Times had reported hours earlier: China's National Development and Reform Commission would not approve SoftBank's $5.4 billion acquisition of ABB's robotics division. The deal, announced in October 2025, had cleared China's competition review in February. It was stopped at the next stage — the foreign investment security review.
The NDRC did not publish a detailed rationale. But the FT's sources identified the concern: robotics data being transferred overseas, where it could be used to train AI models, including those developed by OpenAI, in which SoftBank is a major investor. According to the FT's reporting, the review centered on where the data goes and what it trains.
SoftBank had expected the deal to close before year-end. It had been told the week before the FT report that the NDRC had prohibited the transaction. The Japanese investor is now examining whether it can seek approval again. ABB is assessing alternatives, including carving out the Chinese operations, selling to another buyer at a valuation of $2 billion to $3 billion, or reviving an earlier spin-off plan.
The block is the most consequential use of China's foreign investment security review against a technology deal since the mechanism was formalized. It also mirrors the reviews Washington runs through CFIUS — and demonstrates how the same legal instrument produces different outcomes when the reviewing state is the market the target cannot afford to lose.
Why China Had a Vote
China's approval was required because ABB's robotics business is not a foreign asset with a marginal Chinese footprint. China is its largest individual market, accounting for roughly 30 percent of the division's robotics business. ABB operates around 30 local companies in China with about 15,000 employees across nearly 130 cities. ABB Engineering (Shanghai) is home to the division's global robotics headquarters, as well as a major research and development center and production base.
The division generated $2.3 billion in revenue in 2024, about 7 percent of ABB Group sales, and employed around 7,000 people globally. Its manufacturing network spans China, Sweden, and the United States. China is not a satellite operation. It is the center of gravity.
That footprint gave the NDRC leverage that no other regulator could claim. A Swiss company selling a robotics division to a Japanese buyer is not, on its face, a Chinese transaction. But the division's Shanghai facility is the global headquarters. The data generated by the Chinese operations is part of the asset being sold. And the buyer is SoftBank, which holds a stake in OpenAI — the American AI lab at the center of the training-data competition.
The NDRC's review, which examines foreign investment for national security implications, was the mechanism that turned that footprint into a veto.
The Data Logic
The stated concern — robotics data being transferred overseas for AI training — reflects a specific theory of where value sits in the robotics stack.
Industrial robots generate continuous streams of operational data: motion trajectories, force feedback, failure logs, calibration records, and the tacit process knowledge of the factory floor. That data is increasingly valuable as training material for embodied AI — the class of models that learn physical tasks from demonstration and interaction.
If a buyer acquires the robotics division and consolidates its data outside China, the training material leaves the jurisdiction. If the buyer also holds a stake in an AI lab that trains frontier models, the data can be used to build models that the Chinese operations will later depend on — or compete against.
The FT's reporting identified SoftBank's OpenAI stake as part of the concern. The logic is not that SoftBank would deliberately transfer data to OpenAI. It is that the acquisition would create a corporate structure in which Chinese operational data and American AI training sit under common ownership. The NDRC's review treats that structure as the risk.
This differs from the framework the U.S. uses. CFIUS reviews foreign acquisitions of American companies for national security risk. The NDRC is reviewing a foreign-to-foreign acquisition because the target has substantial Chinese operations. The legal basis is the Chinese subsidiary rather than the acquirer's nationality. The effect is a veto exercised from the market side of the transaction.
The Pattern
The ABB block follows a sequence that has become more visible over the past year.
In April 2026, the NDRC ordered Meta to unwind its $2 billion-plus acquisition of Manus, a Chinese AI startup. That case involved an American acquirer and a Chinese target, the more conventional direction of a security review. The ABB case inverts the structure: a Japanese acquirer, a Swiss target, and a Chinese veto grounded in the target's Chinese operations.
The common thread is the technology. Manus was an AI startup. ABB Robotics is an industrial robotics business with a data pipeline that feeds embodied AI. Both sit at the intersection of physical systems and AI training — the layer where China has made deployment a strategic priority and where the U.S. has focused its own export controls.
The NDRC's review authority is not new. What is new is the willingness to use it against transactions that have already cleared competition review and are widely expected to close. SoftBank had expected approval until the week before the block. The timing signals that the security review is a separate gate that can close late, not a formality that follows the competition review.
The Carve-Out Option
The reported fallback — separating ABB's Chinese robotics operations from the asset sold to SoftBank — is an attempt to address the NDRC's concern without abandoning the deal.
Under a carve-out, a separately owned Chinese entity would hold the Shanghai headquarters, the local manufacturing, and the data generated by Chinese operations. SoftBank would acquire the non-China business: the Swedish and American manufacturing, the global sales network, and the international customer relationships. The Chinese entity would remain under different ownership, presumably Chinese or jointly held with a Chinese partner.
That structure could address the data-transfer concern. If the Chinese entity is not owned by SoftBank, its operational data does not flow into a corporate structure that includes OpenAI. The NDRC's stated concern would be mitigated.
But a carve-out would also change the economics of the deal. China is the division's largest market at roughly 30 percent of robotics revenue. Removing it from the acquired asset reduces the revenue base, the growth outlook, and the strategic value of the acquisition. SoftBank is driving the carve-out effort, according to Bloomberg, which suggests the firm still wants the non-China business enough to accept a smaller asset.
The talks are preliminary. ABB is simultaneously assessing a sale to another buyer at $2 billion to $3 billion — less than half the original price — or a spin-off that would avoid a security review altogether. The carve-out is one option among several, and none of them restores the original deal.
The Cost of the Block
The financial consequences are already visible.
ABB had expected to use the proceeds from the sale for acquisitions. Citigroup analyst Martin Wilkie estimated that the blockage would reduce ABB's available acquisition capacity by half, to approximately $5 billion. ABB shares fell about 1 to 2 percent on the day of the announcement, according to DealStreetAsia and Bloomberg HT.
SoftBank shares dropped 14.4 percent on Thursday and a further 10.8 percent in early Friday trading in Tokyo. The decline was driven by a combination of factors: the broader technology selloff following the U.S. midterm elections, and the China block itself, which the FT reported had also weighed on market sentiment.
The strategic cost is harder to quantify. SoftBank CEO Masayoshi Son had pursued the ABB Robotics acquisition as a way to establish SoftBank as a core player in AI development beyond its software investments. The division would have given SoftBank a physical deployment channel — factories, logistics, and industrial customers — to complement its stakes in OpenAI and other AI companies. The block removes that channel from the deal, at least in its original form.
What the Block Reveals
The ABB decision demonstrates how China's foreign investment security review functions as a strategic instrument, not just a regulatory formality.
The review is nominally about national security. In practice, it is about control over the data and technology that sit inside Chinese operations. When a foreign acquirer wants a business with substantial Chinese assets, the NDRC has a veto over the transaction — not because the acquirer is Chinese, but because the target is partly Chinese. The leverage comes from the target's dependence on the Chinese market, not from the acquirer's nationality.
That is a different model from CFIUS. CFIUS reviews inbound investment: a foreign buyer acquiring an American company. The NDRC is reviewing a transaction between a Japanese buyer and a Swiss seller because the target has a Chinese subsidiary. The legal basis is the Chinese operations. The effect is a veto that reaches across borders.
For global technology M&A, the implication is that China is now a mandatory jurisdiction for any deal involving a target with significant Chinese operations and strategic technology. The review can come late, after competition clearance, and it can stop a deal that both parties expected to close. The carve-out option creates a path, but it requires separating the Chinese business from the global asset — a structural change that reduces the value of what is being sold.
SoftBank and ABB are still exploring that path. The original deal is blocked. The question is whether a smaller, China-excluded version can be salvaged.
Sources: Reuters via Zawya (November 5, 2026); Bloomberg (November 5, 2026); Financial Times via The Edge Malaysia (November 5, 2026); The New York Times (November 5, 2026); Anadolu Agency (November 5, 2026); South China Morning Post (November 5, 2026); Bloomberg via Investing.com on the carve-out discussions (November 7, 2026); Economic Times (November 5–6, 2026); BW Businessworld (November 5, 2026); Evertiq (November 5, 2026); S&P Global on the original deal terms (July 8, 2026); DealStreetAsia on ABB share movement (November 5, 2026); Bloomberg HT on ABB share movement (November 5, 2026); Dahui Law Firm analysis of the Manus case (2026); Yahoo Finance on SoftBank share movement (November 6, 2026).
Disclaimer
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations
China's NDRC did not publish its reasons, so the rationale described here is drawn from reporting by the Financial Times and other outlets.
ABB's public statement confirms only that the transaction was not approved; the carve-out and buyer alternatives remain talks, not decisions.
Deal value and revenue figures are as announced at signing and were not restated at the time of publication.
Underlying financial details and market share figures come from company disclosures and industry estimates that may be revised.
Reporting standards differ between Chinese and Western outlets, and figures cited here reflect the sourcing available at the time of writing.
Market size and shipment estimates vary by methodology; where estimates conflict, this article cites the figure used by the original source.
The sectors discussed are changing quickly, and some figures may be revised after publication.
Sources
- Reuters via Zawya (November 5, 2026)
- Bloomberg (November 5, 2026)
- Financial Times via The Edge Malaysia (November 5, 2026)
- The New York Times (November 5, 2026)
- Anadolu Agency (November 5, 2026)
- South China Morning Post (November 5, 2026)
- Bloomberg via Investing.com on the carve-out discussions (November 7, 2026)
- Economic Times (November 5–6, 2026)
- BW Businessworld (November 5, 2026)
- Evertiq (November 5, 2026)
- S&P Global on the original deal terms (July 8, 2026)
- DealStreetAsia on ABB share movement (November 5, 2026)
- Bloomberg HT on ABB share movement (November 5, 2026)
- Dahui Law Firm analysis of the Manus case (2026)
- Yahoo Finance on SoftBank share movement (November 6, 2026).
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations: China's NDRC did not publish its reasons, so the rationale described here is drawn from reporting by the Financial Times and other outlets.; ABB's public statement confirms only that the transaction was not approved; the carve-out and buyer alternatives remain talks, not decisions.; Deal value and revenue figures are as announced at signing and were not restated at the time of publication.; Underlying financial details and market share figures come from company disclosures and industry estimates that may be revised.; Reporting standards differ between Chinese and Western outlets, and figures cited here reflect the sourcing available at the time of writing.; Market size and shipment estimates vary by methodology; where estimates conflict, this article cites the figure used by the original source.; The sectors discussed are changing quickly, and some figures may be revised after publication.