Future Proof Data Centers

by Svahl

Norway is having an honest argument about data centres. Power is finite, the grid queue is long, and communities are entitled to ask what they receive in return for hosting infrastructure that consumes a great deal and, in some cases, gives little back.

We think that argument is healthy. We also think it has already produced an answer, and that the answer is being written into rules faster than much of the sector has adjusted to.

That answer has a name in our vocabulary: the future proof data centre. It is the standard we are building to, in a gigawatt-scale cluster in Central Norway, developed in phases over the next fifteen years.

A future proof data centre is not simply a newer one. It is a facility that would still be permitted, still be welcome, and still be economic under the rules and public expectations of the 2040s, not only those of the year it was approved.

In practice that means five things.

It is efficient by design, not by retrofit. The EU Energy Efficiency Directive brings heat reuse and energy performance requirements for data centres into force from 2026. Facilities designed around those requirements will meet them comfortably. Facilities designed to yesterday's assumptions will spend the next decade retrofitting toward a moving target.

It gives its energy back. A data centre converts almost all the electricity it draws into low grade heat. Venting that heat to the air is a choice, not a law of physics. A future proof site is located and engineered so the heat has somewhere useful to go, typically district heating, industrial process heat or food production, with the connection built as part of the project rather than promised as a future option.

It cools without competing for scarce local resources. Cooling is where many projects create conflict, because evaporative systems consume fresh water and air cooling consumes additional power. Coastal Central Norway offers a third path: stable, cold deep water used in a closed loop, where pumping energy covers friction rather than refrigeration. Designed with the ecological status of the receiving water as the binding constraint, this removes an entire category of local objection before it arises.

It adds capacity to the energy system rather than only drawing from it. This is the point most often missed. A project that signs a power contract and waits depends on investment decisions it does not influence. A project that co-finances the assets which unlock capacity, in hydro refurbishment, wind, storage, substations and transmission, is a contributor to the system rather than a claimant on it. That distinction is increasingly what regulators and host municipalities are actually screening for. Our team has built several hundreds of MW of renewables over the last decade.

It hosts work the host country values. Capacity in itself is not a public benefit. Capacity serving research, healthcare, public administration and regulated industry, under European jurisdiction and European data governance, is. "What is this power being used for" is a legitimate question for a democracy to ask, and a future proof facility has a specific answer.

Why Central Norway

The region is, in our assessment, the most attractive place in Europe to build to this standard, and the reasons are structural rather than promotional.

It has grid depth, with high voltage transmission capacity and substation infrastructure at a scale that supports gigawatt-class development in phases rather than in a single speculative leap. It has cold, deep fjord water adjacent to industrial sites, which makes efficient cooling an engineering question rather than a political one. It has existing district heating and industrial heat demand, which turns surplus heat from a disposal problem into a delivered public good. It has a hydro-anchored, low carbon power system with a genuine industrial tradition of building large facilities responsibly.

And it has access to a large tech competence center. Central Norway is one of the few places where large scale compute can sit next to the research community that will use it. That proximity is why we think this region can host not only capacity but the work that justifies capacity.

None of this is a claim to entitlement. It is a claim that the region's natural advantages align unusually well with where regulation is heading.

Why this matters now

Three things changed at roughly the same time.

Policy moved from guidance to instruments. Authorities and political stakeholders have openly set out alternatives for how data centre load should be handled, including a concession-based path that would screen projects against criteria rather than admit them in order of arrival. Some have been explicit that it is legitimate for politics to question the societal value of particular industries. Whichever alternative is chosen, the direction is selection rather than open access.

The numbers became impossible to ignore. Data centre power consumption in Norway rose sharply over the past year, more than half the Statnett connection queue is now data centres, and NVE's load forecast points to a multi-terawatt-hour increase before 2030. When one sector dominates the queue for a scarce national resource, prioritisation follows. That is not hostility toward the industry. It is arithmetic.

Capital started pricing permitting risk. In the United States, local opposition has begun to slow projects, and investors have started discounting announced build schedules accordingly. Permitting durability has moved from an appendix footnote to a valuation input. Once that happens, projects that are straightforward to permit carry a real premium over projects that are merely large.

Our position

We are not making a moral argument. We are making a deliverability argument, and the two happen to point the same way.

If the rules are tightening, compliance is not a cost centre. It is the moat. Efficiency requirements, heat reuse obligations, concession criteria and sovereignty expectations all reward the same design choices. A project that already meets the standard of the 2030s does not need to be defended, renegotiated or rebuilt. It proceeds.

So our development sequence starts from the constraints, not from the megawatts. We select for grid maturity and heat offtake before we discuss capacity. We treat fjord ecology as a design input rather than a hurdle to be argued down. We intend to invest in the generation and grid assets that make our load possible. We build in phases, each one energised and useful before the next is committed, so that growth toward gigawatt scale is earned rather than assumed. And we anchor the workload in European research and regulated sectors, where sovereign capacity is genuinely needed and the public benefit case is straightforward.

We accept the corollary. This is slower to start and more capital intensive in the early phases, and it disqualifies sites that look attractive on paper. We think it is the right trade. The alternative is building assets whose licence to operate depends on the rules never changing.

An invitation

Norway can host significant AI infrastructure without trading away industrial electrification, community consent or environmental standards. But not every project can, and the country is entitled to choose which ones do.

We would rather compete on that basis than lobby against it. If the sector is to be filtered, we intend to be on the right side of the filter, and we would rather help define the criteria in public than discover them in a rejection letter.

To politicians and regulators: we support a criteria-based approach, and we are ready to be measured against it. To host communities: heat, jobs and grid investment should be commitments in the permit, not goodwill in a press release. To researchers and industry: the capacity is being built for work like yours. To investors: in this market, the scarce asset is not megawatts announced. It is megawatts that will actually be allowed to energise.

Future proof is not a marketing phrase for us. It is the specification.