UK FCA AI review urges new powers over tech and cloud giants

The Mills Review is the first regulator-led deep dive into how advanced and agentic AI will reshape consumer finance, and it lands with sharper teeth than many expected. By explicitly floating the idea that the FCA should gain direct powers over critical third parties—AI labs and cloud providers that sit outside traditional financial regulation—it signals a regulatory perimeter that stretches all the way into the core of the AI stack. That matters because a handful of U.S. and UK firms currently dominate both frontier models and the hyperscale compute they run on.

For the AGI race, this is an early template for how non-tech regulators will try to rebalance power between finance users and AI infrastructure owners. If the FCA gets the authority it’s hinting at, firms like Anthropic and OpenAI could find themselves supervised not just by generic AI rules, but by sector-specific regulators that care about model behaviour in lending, advice and fraud scenarios. That could mean audits of training data relevant to consumer finance, mandated red-teaming against scams, and potentially even restrictions on which models can be embedded in retail products.

Over time, similar logic can spread from finance to health, energy and critical infrastructure. The message is clear: frontier AI isn’t just a tech story; any sector that leans on autonomous agents could see its regulators reach upstream into model vendors.