Summary
- Danish colocation occupancy has risen from 42% to 76%, while self-built facilities account for 89% of live IT capacity.
- Grid applications reached 60GW in March 2026, with data centres representing more than half of proposed new electricity demand.
- Stricter connection assessment is making credible power delivery increasingly important to the viability of proposed campuses.
Denmark’s data centre market is entering a more constrained phase as rising occupancy and hyperscale development meet an electricity connection queue measured at several times the country’s existing peak demand.
New research from DC Byte puts Danish colocation occupancy at 76%, up from 42%, while self-built facilities account for 89% of live IT capacity. DC Byte describes that as the highest self-built share of any national market in EMEA.
Google, Apple, Meta, and Microsoft have established substantial self-built infrastructure in Denmark, leaving the market unusually dependent on hyperscale campuses rather than conventional third-party colocation.
Western Denmark is attracting another wave of development. DC Byte identifies land availability, electricity infrastructure, proactive municipalities, and improving subsea connectivity as factors behind the region’s growth, with Prime Data Centers’ planned 550MW campus around Esbjerg and Varde among the largest projects in the pipeline.
The constraint is that a large pipeline of electricity demand is seeking access to the same network.
Applications for new consumption across Danish transmission and distribution systems reached around 60GW in March 2026, according to the market analysis, with data centres accounting for more than half.
For comparison, Denmark’s existing peak electricity demand is only a fraction of that queue. The applications also include other large consumers and energy projects, meaning data centres are competing with batteries, electrification, and Power-to-X developments for connection capacity.
Connection readiness becomes part of site selection
Energinet temporarily paused new transmission connection agreements in March as applications accelerated beyond the assumptions used in existing network planning.
The Danish transmission system operator subsequently introduced a revised approach to major connections, grouping projects and applying stricter tests around their maturity and progress rather than allowing an expanding speculative queue to dictate network investment.
The 60GW figure should not be read as a forecast that all of that capacity will be built. Grid queues can include projects at very different stages of development, competing applications, and schemes that never secure financing or planning permission.
Its scale nevertheless shows why connection applications have become a poor proxy for deliverable capacity. The physical network requires substations, transformers, transmission routes, switchgear, and permitting processes that can take years to design and construct.
For developers, this shifts part of the site-selection calculation. Industrial land, fibre connectivity, and planning support have limited value if the project cannot establish when sufficient electricity will actually be available.
The strongest projects are therefore increasingly likely to be those that can demonstrate a credible construction programme, customer demand, financing, and a technically realistic connection requirement rather than simply reserving a large position in the queue.
The Danish market has strong underlying demand. The rise in colocation occupancy from 42% to 76% shows existing third-party capacity being absorbed, while the scale of hyperscaler investment demonstrates continued appetite for large self-built sites.
That structure can itself make power planning more difficult. A retail colocation building may fill progressively over several years, while a hyperscale campus can require very large blocks of connection capacity as successive data halls are commissioned.
Western Denmark provides opportunities because it sits closer to substantial renewable generation and offers larger development locations than Copenhagen. Neither characteristic guarantees firm power on a commercially useful timetable.
Energinet has accelerated network investment, but transmission construction cannot expand as quickly as developers can file applications for hundreds of megawatts.
Denmark therefore provides a useful example of the change taking place across European data centre markets. Customer demand remains strong, but the ability to translate demand into operational capacity is increasingly governed by grid readiness.
The next generation of Danish projects will be judged not only on how many megawatts developers propose, but on how much of that capacity has a credible route through planning, network reinforcement, financing, and ultimately energisation.
