Night aerial view of the Dutch glasshouse belt glowing under a dark sky

Energy-producing AI data centers

Co-generating power, flexible compute, and heat.

As energy makers, we access power and permits in Europe's grid-constrained metro regions.

Partners & supporters

NVIDIAVertivSchneider ElectricArmadaRaffayEfficientJohnson ControlsFrigelUnica DatacentersRoyal HaskoningDHVJenbacher

Why we exist

From energy taker to energy maker.

Building the neo-utility for energy and compute.

AI compute can be the new heating engine of the world. Heating is almost half of the world's energy end use (IEA), and most of it is still fossil. For AI and decarbonization to grow together, AI infrastructure has to become an energy producer, and the same watts have to benefit more than one party. Not a facility that only takes power and exports tokens, but local value: power for the grid, heat for the community, sovereign capacity for startups, enterprises, neoclouds and AI labs to build local and go global. When all three win, capacity comes fast. A Digital Energy Center gives the place it stands more than it takes.

What we do

One platform. Three outputs. Three beneficiaries.

Every Digital Energy Center produces AI capacity (tokens), heat and power for the community it vertically integrates within.

Power already connectedLiquid-cooled AI FactoryTokens, heat and grid flexibility

AI

compute capacity
  • Colocation for AI, tailor-built to your workloads and requirements
  • Sites of 5-25 MW, interconnected into large clusters
  • Liquid-cooled native, for GPU racks of 40-600 kW

Who benefits from this wattNeoclouds, AI labs, enterprises, startups

Secure AI capacity

Heat

usable heat
  • Baseload heat, delivered year-round
  • 30-year heat supply agreements
  • Agriculture, residential, industrial: all can benefit from a clean, carbon-free heating alternative

Who benefits from this wattGrowers, industry, homes, heat networks

Secure 30-year heat supply

Power

flexible power
  • Flexible grid draw that turns down when the network is tight
  • On-site generation that supports the grid, not just the site
  • Grid relief for congested regions, so homes and businesses connect sooner

Who benefits from this wattGrid operators, municipalities, the region

Talk to us about grid flexibility

Why Digital Energy

Speed. Scale. Purpose-built.

01

Speed

AI capacity live in months, not years. While others queue for the grid to allocate new power, we use what is already there. RFS as fast as 30 weeks from financial close, indicative.

02

Scale

Join a platform where you can grow, instead of a site you outgrow. We advance 18 secured sites in parallel, so your next megawatts are already in development.

03

Purpose-built

We build natively for liquid-cooled GPU workloads. No cloud Frankenstein retrofit. Built for the workloads you need: sovereign and local, with your data under your control, in your own dedicated AI Factory.

The platform

A global platform, built metro by metro.

AI inference belongs where the world's tech and people already are: the metro regions. They concentrate connected power, heat demand, fiber and talent. We use that density to build many sites at once, and we integrate each one with the industries around it. Europe first, then the metros of the world. Local value in every place we build, global impact from the network.

Europe first: 300 distributed nodes by 2030, powering and generating 5 GW across Europe.

Measured in operation at Kirkenes, Norway, north of the Arctic Circle: PUE 1.03, and 1.03-1.17 in production.

Map of the Netherlands in Digital Energy blue

Where

The metro regions where the world's tech and people are: Amsterdam first, Europe next, then the global hubs.

How

Density as the advantage: connected power, heat users, fiber and talent on one site, built many at a time.

What

Compute integrated with local industry: heat for growers and homes, flexibility for the grid, inference capacity for the region.

18sites secured
300 MWin development, of which 150 MW is Phase 1
600 MWunderlying energy infrastructure run by our partners today, to be made multipurpose
5 GWidentified today

The product

Build-to-suit AI Factories.

Optimized for NVIDIA NVL72 and beyond. Liquid-cooled native.

  • N+1 redundancy on power and cooling
  • Tier 3 equivalent design
  • 99.99% availability SLA
  • PUE 1.03 measured at Kirkenes (1.03-1.17 in production)
  • Rack densities of 40-600 kW
  • Quiet by design: under 40 dB(A) at the nearest home at night
  • Designed to the NVIDIA Cloud Partner (NCP) reference architecture
  • Turnkey deployment and operations available

Full technical specification and site dossiers under NDA: ask when you book.

Row of NVIDIA Vera Rubin class GPU racks: the systems our liquid-cooled halls are built for

Built for your AI, not for everyone's compute.

A hall built for every kind of compute is optimal for none. We build each AI Factory once, for one tenant: your layout, your densities, your access.

More tokens from less energy.

Infrastructure for inference

Inference, close to the people it serves.

Training tolerates distance. Inference belongs next to its users.

Europe's metros: Amsterdam first

  • 350 million+ people under 10 ms from Amsterdam
  • AMS-IX, one of the world's largest internet exchanges, and NL-IX on the doorstep
  • Existing grid connections and heat users across the Amsterdam Metro region

What's next for Europe:

London · Paris · Frankfurt · Madrid · Oslo · Helsinki

What's next? The global frontiers

The same model will follow: energy-producing infrastructure that vertically integrates with and benefits the communities it sits within.

San Francisco · Dallas · New York · Tel Aviv · Riyadh · Dubai · Mumbai · Singapore · Tokyo

On-prem, without the burden of on-prem

Your own dedicated AI Factory, your hardware, your access, without building or running it yourself. That is why location matters: get to your racks, fine-tune them, keep your data a drive away.

The neo-utility

Energy infrastructure that computes.

Engines improve when they use their heat instead of venting it. AI is the next engine. A neo-utility runs it for everyone: power that is already connected becomes AI capacity, and the heat and flexibility go back to the grid and the community. We are not building data centers that only consume energy. We are building energy infrastructure that computes.

Talk to the founders

Recognition

What the industry says.

European Data Centre Association
“The next generation of European infrastructure won't be built by companies optimising the old model. … Digital Energy demonstrates how AI infrastructure can contribute beyond compute, by reusing heat, supporting food production and delivering sovereign AI capacity. This is exactly the kind of structural innovation our industry needs.”
Michael Winterson, Secretary General, European Data Centre Association  ·  Read the press release →

In the news and on stage

AI buyers

Need AI capacity in Europe, fast?

You bring the GPUs and the workload. We deliver a dedicated, liquid-cooled AI Factory on power that is already connected, and operate it to a 99.99% SLA. Tell us the megawatts and the date.

Secure AI capacity

Heat users

Want heat you can plan on for 30 years?

If you run a greenhouse, a factory or a heat network in the Amsterdam Metro region, we supply baseload heat from AI compute into your system, year-round, replacing gas. Tell us where you are and how much heat you use.

Secure 30-year heat supply

Talent

Want to build the energy layer of AI?

We are hiring engineers, developers and operators who would rather build the foundation than rent it: energy-producing AI Factories, from permit to power-on, across Europe and beyond.

Join the team
NVIDIAVertivSchneider ElectricArmadaRaffayEfficientJohnson ControlsFrigelUnica DatacentersRoyal HaskoningDHVJenbacher

Next step

Power intelligence, fast.

Tell us what you're building or heating. We'll tell you how fast we can deliver.

jonathan@digital-energy.group