Carbon3.ai, a newly formed UK firm, has unveiled a £1 billion ($1.3 billion) project to construct a nationwide network of data centres powered entirely by domestic renewable energy. This ambitious initiative, if successful, could serve as a blueprint for sovereign AI infrastructure in the UK, benefiting enterprises, research institutions, and public bodies alike.
The company plans to transform existing industrial and energy-generation assets across the UK into high-performance compute hubs. The £1 billion investment is estimated to cover the cost of converting 50 MW of secured renewable baseload generation into data-centre capacity. Carbon3.ai is collaborating with HPE, Vast Data, and WWT to develop this “national grid for AI,” drawing inspiration from HPE Private Cloud AI.
According to its founders, the project aims to deliver a fully sustainable “AI mesh” that is owned, powered, and secured domestically. This approach seeks to reduce reliance on foreign-controlled cloud infrastructure while providing low-carbon compute capacity. CEO Tom Humphreys emphasised, “The UK’s competitiveness in AI depends on infrastructure that is truly sovereign, sustainable, and resilient.” He added, “It’s not enough to invest in data centres. We need a national backbone for AI that’s owned, powered, and secured right here at home.”
The rising energy demand from data centres sharpens the case for renewable compute. Globally, data centres are responsible for about 1.7% of global electricity consumption, with the UK seeing approximately 2.5% of its national electricity consumed by these facilities. Expanding this infrastructure with renewable energy will help meet increasing demand while addressing sustainability and emissions goals. Data centres could also provide flexibility for power systems, supporting broader transitions toward net-zero energy grids.
One of the announced projects involves a modular “AI data centre” next to a landfill-gas-powered energy site in Derbyshire. The plan is to draw power from the adjacent facility to create an off-grid, renewable-powered compute hub. The finished network could host up to 100,000 GPUs across multiple sites, offering scalable, low-latency, AI-ready compute for enterprises, public institutions, and researchers.
This initiative could reshape how UK operators approach data centre projects. By building new facilities based on renewable power from the start, Carbon3.ai is setting a precedent. Owning power generation and infrastructure may better position operators to comply with evolving ESG expectations, government policies, and energy-supply constraints. A domestic sovereign network could also reduce dependence on foreign cloud providers and improve data governance for sensitive workloads.
Moreover, this project offers a potential solution to the increasing stress on the power grid. As data-centre demand grows, integrating them into national energy planning will become increasingly important. This initiative could set a new standard for sustainable and resilient AI infrastructure, influencing global markets and policies.
The implications for the sector are profound. If successful, Carbon3.ai’s model could inspire similar projects worldwide, accelerating the transition to renewable-powered data centres. It could also spur innovation in energy storage and grid management technologies, as well as influence policy decisions regarding data sovereignty and sustainability. The project’s success will hinge on its ability to demonstrate the viability of renewable-powered AI infrastructure, setting a new benchmark for the industry.