Britain's AI Problem Isn't What It Builds. It's What It Exports.

Archie Hall's bear case for Britain and AI is worth reading for one argument most people making this case miss entirely.

It's not about whether Britain has AI companies. It does. It's about where the productivity gains from AI disruption actually land — and with a services export base dominated by legal work, financial back-office functions, and professional consulting, the answer is: largely not in Britain.

When a New York firm automates work that used to be done by London lawyers, the efficiency gain stays in New York. That's a different kind of exposure than losing a manufacturing contract. It's quieter, harder to see in the data, and less likely to generate a political response until it's already happened.

That asymmetry — between where AI disruption hits and where AI investment flows — is the argument worth keeping, and most commentary on AI and national competitiveness doesn't touch it.

Hall is weaker when he argues that British households are badly positioned for an AI boom because they hold too little equity. Hall treats retail savings allocation as a structural vulnerability, but British pension exposure is already globally diversified — most people with a workplace pension hold meaningful US equity whether they know it or not. The argument conflates where the average Brit parks their ISA with whether British households participate in AI-driven wealth creation. Those aren't the same thing, and the section doesn't do the work to show they are.

The footnote he almost buried is more valuable than any of the six named scenarios: he was more optimistic about Britain eight months ago because he over-counted the importance of having a domestic AI industry and under-counted whether the rest of the economy could adapt. That's the distinction that matters, and it's not in the headline.