We use cookies to ensure you get the best experience on this and other minimal.tech websites. For more information on how we use cookies, please see our cookie policy.

4 November 2026 · Eddie Bullock

Minimal awarded compute grant from the UK's AI Research Resource

Minimal has been awarded a compute grant from the UK's AI Research Resource (AIRR). Today, our autonomous driving models start training on Isambard, the supercomputer at the University of Bristol.

Access to supercomputing of this scale is a major step for UK industry. It gives commercial teams the infrastructure needed to train frontier AI models at speed, ensuring the UK stays at the front of future mobility and autonomous technology.

Building our own driver

We are developing our own self-driving model. It reads the street, makes decisions and drives our autonomous vehicles. It practises in our World Model, a city in simulation, before it ever touches a real road.

With access to compute on Isambard, our self-driving model improves faster, with more accuracy on the roads. This allows Minimal to bridge the gap between silicon training and real-world driving, deploying autonomous vehicles around the world at scale.

Where it leads

All of this work goes into our first autonomous vehicle. Our bot will be small, self-driving and built to bring food and groceries to the door.

The vehicle will be unveiled for the first time in San Francisco next week, at Micromobility America.

"We're excited to be training on Isambard at the University of Bristol. It gives our models more practice than we could ever give them alone. Thanks to UKRI and the University of Bristol for opening it up to teams like ours." says Liam Sill, Minimal CTO.

Training starts today. There's more to come from our autonomy team soon.