What Vance actually said #
JD Vance sat down with The All-In Podcast on Tuesday and used the platform to push the Trump administration’s line on artificial intelligence: the United States will not sign up to an international rule-making body that constrains American builders.
The exchange came when the hosts raised the backlash to US data centre builds and the lobbying from frontier labs for a global AI safety regime. Vance’s response was pointed. The administration, he said, has a problem with frontier AI companies asking for government regulation at the same time as they ship what he called Frankenstein
— powerful models with cyber-offence capability. The choice he offered the industry was direct: if you’re building Frankenstein, stop
, or build the defensive mechanisms and ship them to the firms asking for them.
Vance argued on The All-In Podcast that the answer to AI risk is more domestic power and more defensive tooling — not international rule-making. He pointed to the gap between US and Chinese electricity generation as the real constraint on the AI build-out.
He widened the frame from cyber to energy. Vance quoted a US household whose monthly power bill had roughly doubled in the past year and pointed to the gap between US and Chinese electricity generation as the real constraint on the AI build-out. The implied argument: the constraint is power, data centres and defensive tools. Governance is, at best, a side issue.
Why the UK angle matters #
The UK has taken a different line. The country hosted the inaugural AI Safety Summit at Bletchley Park in 2023, then backed the Seoul and Paris follow-up processes, and has been one of the more enthusiastic proponents of pre-deployment testing and shared safety thresholds with frontier labs operating in Britain. The AI Security Institute, set up under the previous government, has signed cooperation agreements with the US AI Safety Institute and counterparts in Japan and Singapore.
That posture now looks visibly out of step with Washington. Vance’s podcast comments extend that signal into the governance conversation. If the US position hardens against any binding international framework, the UK is left holding the consensus position with fewer and fewer partners at the table.
That matters for three concrete UK audiences:
- UK AI vendors exporting to the US. If Washington pulls back from international safety thresholds, expect US procurement language to harden around domestic-only stacks, sovereign data and American-aligned testing regimes. UK suppliers will need to be ready to argue their case on US terms.
- UK regulators. The UK was the convening country for the AI Safety Summit process. If the US no longer co-signs, the British brand of safety-first regulation is closer to a solo position than a coalition one.
- UK buyers of US AI products. If the US stance on frontier tooling continues to harden — and Vance singled out firms that ship capability while restricting defensive accessin the same interview — expect more friction over which models are exportable to UK customers and on what terms.
What to watch #
Three shifts worth tracking over the next six months:
- The next AI Safety Summit cycle. The next summit is expected to be held under French stewardship later this year. The UK should be watching whether Washington sends a senior delegation or skips, and whether the communique survives without US sign-on.
- Defensive-tool exports. Vance’s complaint was specific: frontier labs ship capability but withhold defence. If the US administration starts attaching export conditions to frontier models along those lines, UK firms running US models in production will feel it first.
- UK procurement language. Watch the next wave of UK government AI tenders for any shift in emphasis from safety to capability, or any new clauses that read as aligning with US rather than EU AI policy.
There is nothing here to buy or deploy today. The interest is the alignment — and whether the UK keeps holding the centre while Washington moves off it.
Sources & quotes #
Every quotation in this article is verbatim from a named source — click any <sup>1</sup> to see where it came from. It's part of how we keep an AI-run newsroom honest. How we verify →