Data Centre & Sovereign Compute
Distributed 1–5 MW data centre capacity on sites where power, fibre and land are already in place — deliberately below the thresholds that make large projects slow.
Small is deliberate
Hyperscalers and large customers deal directly with large investors and utilities. For smaller distributed requirements there is almost nobody: technology buyers have no on-the-ground energy capability, and the people who hold suitable sites have no route to a compute tenant. We are the combination.
No queue
A 1–5 MW requirement is an ordinary commercial new-service request to the distribution utility. No cluster study, no ISO process.
No CIAC exposure for you
We work prepared sites. Where utility contribution is required, it is funded by the tenant or the financier — never carried by the project sponsor.
Speed is the product
In the busiest territories, pending large-load requests now exceed what the utility has ever delivered at peak. Small and ready beats large and queued.
What we do
Capacity sourcing
For buyers who need megawatts now: prepared sites with existing service, fibre and land, matched to a real requirement.
Site qualification
We verify what a site actually has. An existing interconnection is usually adjacency, not capacity — we ask the utility in writing.
Campus & institutional hosting
Universities and institutions hold land, power and fibre but no route to revenue. We structure the hosting model.
Distributed design
Modular 1–5 MW nodes — rack density, liquid cooling and thermal integration specified with engineers of record.
Offtake structuring
Lease terms, credit and guarantees. Tenant credit kills more deals at this scale than siting does.
Multi-site programmes
Portfolios across several locations, kept genuinely separate so loads are not aggregated into a large-load tariff.
Institutional hosting
Universities, colleges and public institutions often hold exactly what a distributed data centre needs — land, an existing electrical service, fibre and a stable counterparty — with no route to turning it into revenue. We structure that: the hosting model, the commercial terms, and the workforce programme that can run alongside it on the same campus.
How we work
Define the requirement
Megawatts, term, density, latency and the entity that will sign.
Match a prepared site
Existing service, fibre path and land. We confirm capacity with the utility rather than assuming it.
Structure
Lease, credit support and the works needed to energise, with funding responsibility explicit.
Hand to delivery
Design and build executed by specialist firms. We stay on the advisory, site, power and commercial side.
Sovereign compute
Modular, grid-parallel compute nodes deployed on campus and institutional sites. For organisations running AI workloads, that means capacity inside their own state on infrastructure with a known location and a stable long-term cost — rather than a hyperscale subscription that reprices. It is the infrastructure layer beneath our managed AI services.
Two sides of the same conversation
If you need compute capacity, tell us the megawatts and the timeline. If you hold land, power or a campus, tell us what is already there.
Get in touch