- Published on
China Blocked SoftBank's $5.4 Billion Robot Deal. The Reason Is the Story
- Authors

- Name
- AbnAsia.org
- @steven_n_t
Listen to this article · China Blocked SoftBank's $5.4 Billion Robot Deal. The Reason Is the Story
0:00 / 0:00
Watch this article · China Blocked SoftBank's $5.4 Billion Robot Deal. The Reason Is the Story
ABB confirmed on 5 November that its plan to sell its robotics division to SoftBank has been prohibited in China — not by a competition authority, but by the country's foreign investment security review.
The deal was signed in October 2025, at an enterprise value of $5.375 billion, and was expected to close in mid-to-late 2026. China's competition regulators cleared it in February. What stopped it, roughly nine months later, was a different question entirely: who should own the factories that build robots.
What ABB actually said
ABB's own statement is short and precise. The company said it had been informed by SoftBank that "the Office of the Working Mechanism for Foreign Investment Security Review in China has prohibited the acquisition of the ABB Robotics business in China."
Two details in that sentence matter. First, the prohibition is about the business in China — the local entity and its assets, not a general ban on the corporate transaction elsewhere. Second, the process was SoftBank's to run: ABB noted that the regulatory approvals were managed by the buyer, and that the two companies remain "in close dialogue" and are "working constructively on the transaction."
Reporting by the Financial Times, summarised by Mobile World Live, attributes the decision to the National Development and Reform Commission — the body responsible for supply chain security — and says people close to SoftBank believed until the previous week that the deal would clear.
Antitrust was never the hard part
The sequence is the point. Competition review asks whether a merger would harm consumers in a market. Security review asks something else: whether the resulting owner should be allowed to hold the asset at all.
Those two questions can give opposite answers about the same transaction, and increasingly do. A merger can be perfectly competitive and still be refused, because the objection is not market power. It is control.
For anyone who has watched cross-border deals over the last few years, this is now a familiar shape. The binding constraint on acquisitions in semiconductors, cloud infrastructure, ports, telecoms and robotics is rarely the antitrust file. It is the national security file, and it is usually the one that is least visible until the day it decides.
Why robotics is now in that category
The numbers explain the sensitivity. Per FT-sourced reporting, around 90% of ABB's robots are manufactured in China, in what is the company's biggest market, with roughly 12,700 employees there. The stated concern was the transfer of technology developed in China to a Japanese owner.
That reframes the deal completely. This was not a Swiss company selling a business to a Japanese buyer. It was a transaction that would move a large block of China-based robotics engineering and manufacturing capacity into SoftBank's portfolio — and SoftBank's robotics holdings already include SoftBank Robotics, Berkshire Grey, AutoStore and Agile Robots.
Robotics sits where AI meets physical infrastructure. An industrial arm is not a chatbot; it is a component of a factory, a warehouse, a port or a hospital. Governments treat the ability to build and service those systems the way they treat any other strategic supply chain, because the dependency is structural and slow to replace.
The uncomfortable asymmetry
There is a lesson here that outlives this one deal.
Ownership is a permission, not a purchase. Buying an asset and being allowed to operate it are separate approvals. A deal can clear every commercial hurdle and still fail on a question the buyer never got to argue.
Where you build becomes who controls you. The 90% figure is not a footnote about cost efficiency. It is the reason the review exists, and it is the reason the review has teeth.
The best time to think about this is before you need to. Companies discover their geographic concentration during a crisis — a review, a sanction, an export control, a regulator's letter. By then the layout is already built.
What happens next
The reported path forward is a carve-out: sources suggest parts of the business may need to be spun off for the transaction to proceed. ABB and SoftBank describe continued dialogue rather than a termination, which leaves room for a restructured deal.
It is also worth noting what the deal was for. SoftBank has been assembling a physical-AI position across warehouse automation, storage systems and industrial robotics. The strategic logic of the acquisition has not changed. Only the number of governments that have to agree to it has.
For anyone building on top of robotics — integrators, logistics operators, manufacturers planning automation — the practical takeaway is unglamorous. Your robot fleet has a nationality: where the hardware is made, where the firmware is signed, where the service contracts sit. Knowing that, before a regulator asks, is cheaper than discovering it afterwards. Software you can rewrite; a supply chain takes years.
Sources: ABB Group press release, "Update on divestment of ABB Robotics", 5 November 2026; Financial Times reporting on the decision, summarised by Mobile World Live, 5 November 2026.