Sustainable AI Vote

Sovereign AI Awards 2026 · People's Vote

Sustainable AI Award

Reducing the environmental footprint of AI, from energy-efficient models to sustainable infrastructure. Our judges have shortlisted three finalists. Read about each one, then cast your vote.

01Meet the finalists 02Cast your vote 03Watch the live results
Vote now

The finalists

Finalists are listed in alphabetical order.

Isambard-AI

Bristol Centre for Supercomputing, University of Bristol

Nominated by Emily Coles, Bristol Centre for Supercomputing

Isambard-AI, the UK's most powerful AI supercomputer, is nominated as a model of sustainable supercomputer design. It sits in a compact, modular data centre on existing science park land near Bristol and uses direct liquid cooling in a closed loop, using about as much water as 18 average UK homes.

Its cooling needs only around 7% extra power, against around 80% for a typical supercomputer, and it ranks among the greenest supercomputers in the world on the Green500 list. It hosts around 4,000 users and more than 1,000 research projects, including work on climate and air pollution.

Judges' analysis

The judges recognised a national-scale system with sustainability designed in from the start, backed by clear figures on energy, water and footprint. It shows how large AI infrastructure can be built responsibly.

Literal Labs

Nominated by Paul Southall, Commplicated

Literal Labs, a Newcastle University spin-out, has built a new kind of AI model from logic rather than the heavy multiplication behind neural networks. Its Logic-Based Networks run on the ordinary processors and microcontrollers already in use, with no specialist chips required.

Independent MLPerf benchmarks measured them as 54 times faster and 52 times more energy-efficient than equivalent neural networks. They are in live use in water utilities, solar energy and food supply, where running on existing hardware avoids replacing devices in the field.

Judges' analysis

The judges saw Literal Labs as the most original approach in the category, with energy savings that are both large and independently verified. Avoiding new hardware altogether adds a further environmental benefit.

Locai Labs

Locai One

Nominated by James Drayson, Locai Labs

Locai Labs has built Locai One, a UK-designed AI computer that runs advanced AI locally on around 500 watts from a standard plug. It is air-cooled and uses no water, and is sized to the job rather than relying on large, permanently running data centres.

Its SPACE technique trims unneeded parts of large AI models so they can run on much smaller systems, and has already reduced a leading open-source model by 25%. Locai One systems are manufactured in Bolton and deployed across UK public and private organisations.

Judges' analysis

The judges recognised a credible and distinctive approach that combines right-sized hardware with model compression, alongside strong UK design and manufacture.

Live results

Votes so far. Results are updated regularly while voting is open.

Isambard-AI
0
Literal Labs
0
Locai Labs
0
0 votes castLast updated: awaiting first votes

Cast your vote

Choose one finalist and tell us who you are.

You can vote once in this category, and you are welcome to vote in every category. If the same email address votes more than once in a category, only the most recent vote is counted. The People's Vote runs alongside the independent judging panel's assessment.