Author: Derek Michalski, Chief Editor.
For most of the liberalised era, European network tariffs and connection policy operated on a simple, largely unexamined premise: new demand connects in the order it applies, and the resulting network costs are socialised across the customer base broadly in proportion to consumption. That premise is now under sustained pressure. Data centre electricity demand in the EU stood at roughly 96 TWh in 2024 — about 3% of total EU electricity consumption — and Ember Energy projects this will reach 168 TWh by 2030, a 75% increase in six years, extending to as much as 5.7% of EU consumption by 2035 on some trajectories. Unlike most demand growth regulators have historically planned around, this load is lumpy, concentrated in specific corridors, arrives faster than transmission and distribution reinforcement can be delivered, and — critically — is now large enough that how its network costs are allocated materially affects what households and industrial consumers pay. The result is a live redesign of both connection policy and tariff methodology across several major European markets, running under three distinct but converging approaches.
Approach one: replacing first-come-first-served with strategic curation of the connection queue
The most immediate pressure point has been the connection queue itself. Great Britain’s experience is illustrative: by late 2025, the contracted demand queue had grown to a scale regulators judged unsustainable, driven substantially by speculative applications — developers reserving capacity for projects, including data centres, that may never proceed, while genuinely ready projects wait behind them. The National Energy System Operator’s connections reform, implemented following Ofgem approval and roughly eighteen months of design work with industry and government, moved the system away from pure chronological ordering toward assessing projects against strategic alignment and delivery readiness. NESO’s own account of the reform frames it as unlocking up to £40 billion in annual investment by allowing 283 GW of generation and storage capacity, and 99 GW of transmission-connected demand, to progress on a basis of demonstrated readiness rather than queue position alone.
Demand connections specifically remain a live regulatory front. Ofgem’s February 2026 “Demand Connections Reform” call for input sets out a three-pillar approach — Curate, Plan and Connect — explicitly designed to prevent speculative demand applications, including data centre projects, from crowding out strategically important connections. This followed a government commitment in November 2025 to address speculation in the demand queue directly, and it sits inside the broader Clean Power 2030 framework, under which grid connection policy has been formally aligned to national strategic energy planning rather than left to bilateral, first-come-first-served negotiation between applicant and system operator.
Approach two: locational and constraint-based connection policy
A second and more structurally significant approach has emerged in Ireland, where data centre demand has grown from roughly 5% of national electricity consumption in 2015 to approximately 22% in 2024 — concentrated overwhelmingly around Dublin, where data centres now account for close to half of regional demand. The Commission for Regulation of Utilities’ final Large Energy Users Connection Policy, published after an extended consultation, replaces the blanket regional moratorium that had effectively frozen new data centre connections in the Dublin area with a constraint-based framework: system operators must assess each application against the specific capacity and congestion conditions of the transmission or distribution location proposed, rather than applying an area-wide prohibition or an undifferentiated queue position.
The policy’s more consequential feature, from a market design perspective, is that it converts renewable energy and flexibility commitments from a policy aspiration into a connection precondition. Applicants seeking a defined Maximum Import Capacity must demonstrate a credible plan — subject to ongoing reporting — for meeting a substantial share of demand from renewable sources or on-site/autoproducer generation, and system operators are required to publish standardised, transparent locational data on available network capacity and constraints by 31 March 2026. This shifts the connection process from a rationing exercise to a genuine price- and risk-allocation mechanism: large users internalise more of the cost and planning burden their load imposes on the system, rather than that burden falling by default on network-wide tariffs.
The political salience of this shift is worth noting for context, even where the underlying figures originate from advocacy rather than regulatory sources. Analysis circulated by opposition politicians in early 2026, drawing on CRU network charge data, calculated that Irish households pay roughly 7.6 cents per kWh in network charges compared with approximately 0.7 cents per kWh attributed to data centre connections — a disparity that, whatever one’s view of the calculation methodology, has become a central reference point in the domestic debate over cost allocation, and helps explain the urgency behind the CRU’s policy reset.
Approach three: cost-reflective network tariff design at the EU level
The third approach operates above the level of individual connection decisions, targeting the tariff methodology itself. Network tariffs currently account for roughly 20–22% of household electricity bills and 11–12% of industrial bills across the EU, according to Eurelectric — a share the association argues must be more clearly correlated with actual network usage and the costs individual connections impose, rather than allocated through broad, undifferentiated volumetric charges. Eurelectric frames tariffs that are not cost-reflective as a direct source of cross-subsidy between consumer classes, and the scale of investment involved raises the stakes: Eurelectric estimates around €730 billion is required for European distribution networks and €477 billion for transmission networks by 2040, with annual distribution investment needing to roughly double by 2050.
The European Commission has signalled that network tariff design will be a formal workstream in 2026, building on its Action Plan for the Future of Grids, with the explicit aim of ensuring tariff structures send accurate price signals to large and flexible consumers rather than diffusing costs uniformly across the customer base. This is a materially different lever from either of the connection-policy approaches above: rather than deciding who gets to connect and where, it addresses what any connected user — data centre or otherwise — actually pays once connected, and whether that payment reflects the network reinforcement their specific load pattern requires.
A shared design principle, three different points of intervention
What unites these approaches, despite their procedural differences, is a single underlying shift: European regulators are moving from treating network cost recovery as a background allocation exercise to treating it as an active instrument of demand management and distributional fairness. Great Britain is intervening at the point of queue entry, filtering out speculative demand before it can distort planning assumptions. Ireland is intervening at the point of connection design, converting renewable and flexibility commitments into a binding condition of access in constrained locations. The European Commission and Eurelectric are intervening at the point of tariff calculation, arguing that the price signal itself needs to become genuinely cost-reflective rather than a case of averaging.
For energy sector participants — developers, financiers and network operators alike — the practical implication is that data centre demand can no longer be modelled as broadly analogous to other large industrial load in terms of network cost treatment. Each of the three levers described above is capable of independently reshaping the economics of a hyperscale connection: a locational rejection or conditional approval under an Irish-style policy, a queue de-prioritisation under a British-style readiness filter, or a tariff redesign under an EU-style cost-reflectivity mandate. The markets currently moving fastest — Ireland and Great Britain in particular — offer the clearest early evidence of how these mechanisms will be calibrated in practice, and are likely to inform how slower-moving jurisdictions eventually approach the same underlying problem.
