The Danish Ministry of Climate, Energy and Utilities plans to give grid operators the power to rank connection requests according to their contribution to critical services, electrification, and the energy system, replacing an allocation model based largely on who entered the queue first.

Under an emergency plan backed by a broad parliamentary agreement, protected needs and socially critical functions would receive the highest priority. Direct electrification, renewable energy, carbon capture, hydrogen, and other industrial projects would follow, with storage and grid supporting facilities placed in a third category. Certain large electricity consumers would occupy the fourth and lowest group.

Datacentres are not prohibited, nor will every proposed facility necessarily fall to the bottom. A project may improve its position by locating near generation, accepting flexible consumption, supporting the Danish economy, or reducing pressure in a congested part of the network. Political statements accompanying the agreement nevertheless identify large, inflexible, foreign owned datacentres as developments that should not displace hospitals, heating, transport, households, or industrial electrification.

Denmark’s government says capacity is close to exhaustion in several areas, while demand for new connections is growing more quickly than planned network expansion. Legislation is expected after the summer following consultation, with the intention that Energinet and distribution companies can apply the rules when processing large connection requests in the autumn.

The ministry’s published emergency plan allows projects in the lower three categories to be reordered according to their effect on the grid. Draft legislation would also permit some applications to be refused in exceptional circumstances.

Such discretion may direct scarce infrastructure towards projects with greater economic or social value, but it also introduces uncertainty for investors. Datacentre developers make decisions years ahead, arranging land, fibre, electricity, planning consent, finance, equipment, and customers around an expected energisation date. A connection regime whose criteria are still being defined may prompt them to add storage, accept interruptible supply, redesign campuses, or move investment elsewhere.

Digital growth runs into physical scarcity

The Danish proposal reflects a wider European collision between computing demand and electrification. Governments want more renewable energy, electric transport, heat pumps, low carbon industry, hydrogen production, and AI infrastructure, although every ambition draws on networks that take many years to reinforce. Policy commitments can multiply far more quickly than substations, transmission lines, and engineering teams.

Connection queues also contain speculative demand. Developers may apply for capacity at several sites before deciding which campus to build, while some projects never obtain finance or customers. Denmark will need to distinguish credible proposals from reservations that tie up capacity without creating a process that automatically favours large companies able to assemble more elaborate applications.

Assessing societal value will prove equally difficult. A facility supporting hospitals or public administration presents a clearer case than a generic cloud region, but the boundary soon blurs. Danish manufacturers, retailers, banks, and government bodies may run essential systems in a foreign owned datacentre, while an AI campus serving overseas customers could still support construction work, tax receipts, district heating, grid balancing, and demand for local renewable generation.

Economic benefits can also be overstated when facilities employ relatively few people after construction, import much of their equipment, and consume electricity that could support more labour intensive industrial activity. A credible assessment must examine employment, heat reuse, flexibility, tax, water, resilience, carbon effects, and the opportunity cost of the grid connection rather than accepting the description attached to the building.

Grid friendly design offers the most constructive response. Batteries, flexible computing loads, lower carbon backup generation, direct renewable connections, and the ability to reduce demand during constrained periods can turn a datacentre from a fixed consumption block into a more responsive participant in the electricity system.

Those capabilities carry costs and will not suit every workload. Financial transactions, public services, and other systems sold with continuous availability cannot be interrupted merely because the grid is congested. Operators may instead have to separate flexible AI workloads from services that require uninterrupted power.

The plan also needs to sit alongside faster network expansion. Prioritisation can allocate scarcity more deliberately, but it does not produce another megawatt of transmission capacity. If queue management becomes a substitute for investment, Denmark may simply preside more efficiently over a shortage that restricts industry as well as datacentres.

Connection policy is consequently becoming part of industrial strategy. Datacentre developers will have to show that their campuses can support the power system, contribute to the domestic economy, and compare favourably with factories, transport, heating, and other digital projects seeking capacity at the same connection point.