Peter Kyle told the BBC this week that he would have intervened to block the sale of ARM Holdings had he been in government at the time. ARM was bought by SoftBank in 2016 for £24 billion. It is now worth £285 billion on the New York Stock Exchange.
The statement costs nothing. ARM is famous, profitable, and already lost. Regretting its sale under a different government a decade ago is the political equivalent of knowing the answer after the exam.
The easy regret
The proposed fix is a cross-government concierge service, bigger public investments, and a higher risk threshold. Create the circumstances, the argument goes, where companies do not want to leave in the first place.
That is more scaffolding around a building that has already been sold.
The problem with ARM was never that nobody tried to block. The problem was that no British institution existed with the architecture to hold it. Patient capital, operational conviction, generational horizon. The Wallenbergs would not have sold ARM. Not because of sentiment. Because Investor AB is built to hold industrial IP across generations. Britain had nothing built to hold.
The actual test
The test of whether a government understands industrial capability is not whether it can name the last company it lost. The test is whether it can see the next one.
Nanoco Group is listed on the London Stock Exchange. It holds enabling intellectual property in quantum dot technology, with applications across semiconductor photonics, display, and defence sensing. It is precisely the kind of deep-tech linchpin that sits in the supply chain where value concentrates and sovereignty resides.
Walk into a ministerial office today and say “this company holds IP in cadmium-free quantum dot synthesis that positions it in the semiconductor value chain.” The response is recognition or it is referral. The assessment skill required to make it recognition is not in the room. It is not in the concierge service. It is not in the pipeline that produces the people who staff these functions.
The concierge problem
A government concierge service will be staffed by the same pipeline that missed ARM, missed DeepMind, and cannot currently identify Nanoco. Strategy consultants, policy graduates, civil servants. Capable people, selected for a capability that does not include reading a production system, a supply chain position, or a qualification pathway.
A concierge who cannot recognise the guest is a receptionist with a nicer title.
The manufacturing engineer, the industrialist, the operational allocator, these are the people who can see a Nanoco before the BBC writes it up. They read the technology, not the headline. They sit in gate reviews, not press conferences. They are absent from every room where the concierge service is being designed.
What the verb looks like
Industrialists read supply chains. They identify where value concentrates before it becomes famous. They assess production system maturity, not pitch deck market sizing. They know the difference between an IP position that holds and one that gets sold at a discount because nobody in the room could price it.
That is the skill the proposed architecture is missing. Not more money. Not more concierge. The people who can see the next ARM while it is still worth £24 billion, not £285 billion.
Kaipability works at this interface: the gap between what the technology is worth and what the allocator can see. The analysis is the starting point.