Summary
- A European Commission proposal would let national regulators move beyond first-come, first-served grid queues and rank projects using maturity, congestion, economic, social and environmental criteria.
- Data centres could face a separate network-charge regime, but the charge could be higher or lower only where their consumption profile has a proportionately different impact on network costs.
- The proposal is not law, and its cited €75 billion-€100 billion investment need and potential 60% grid-cost increase apply to Europe’s electricity networks as a whole, not specifically to data centres.
Europe’s grid queue is becoming an economic allocation system. The scarce asset is no longer just electricity; it is the right to connect a large new load at a useful place and time. Under a European Commission proposal published on 17 July, regulators could decide that position in line is less important than whether a project is real, how it changes congestion and what benefits it brings.
For data-centre developers, that shifts the contest from securing an early application to proving that a campus deserves capacity. For regulators, it creates a more difficult question: which users should pay more for the network they require, and which should advance because they make the system—or the wider economy—work better?
Scarcity gets a ranking
The Commission’s COM(2026) 600 proposal says connection queues were present in at least 16 member states in evidence collected during 2025. Its answer is not to give data centres an express fast lane. Proposed Article 18d would allow national regulators, where grid capacity is scarce, to deter speculative requests, test project maturity and prioritise categories of users. The illustrative list runs from public services, households and small businesses to data centres, energy-intensive industry and transport.
Regulators could also rank users within a category. The criteria would have to be objective, transparent and non-discriminatory, but could include congestion relief and economic, environmental or social benefits. That makes “queue jumping” less a matter of buying a better position than of satisfying a policy scorecard.
A data centre with committed customers, a credible construction schedule and a design that can moderate demand may therefore fare better than a speculative reservation. Yet it could still lose to housing, a hospital or an industrial project if the regulator assigns greater value to those uses. The proposal preserves national discretion, so two otherwise similar campuses in different member states may not receive the same answer.
Who pays depends on the load
The second allocation mechanism sits in the tariff. Proposed Article 18 would permit special charge regimes for categories including data centres only if regulators can demonstrate that their consumption profile has a proportionately lower or higher effect on total network costs. Cost-reflectivity is the constraint: this is neither an automatic data-centre discount nor a blanket penalty on digital infrastructure.
The wider tariff design points to what operators may be asked to prove. The proposal calls for locational signals, incentives to cut peaks and time-of-use elements. Its recitals say data-centre location, timing and consumption can support efficient connection, particularly where a facility provides flexibility, storage or additional clean generation. A campus that can shift part of its load, accept a flexible connection or support the grid may lower its case-specific burden. A rigid project that creates a new peak in a constrained zone may face the opposite result.
This turns engineering choices into financing variables. Batteries, on-site generation, workload flexibility and the credibility of a power schedule could affect both the waiting time for a connection and the recurring charge after it is energised. Grid-ready land should become more valuable, while sites whose economics assume an unconditional firm connection become harder to underwrite.
The big numbers belong to the whole grid
The proposal cites an ACER estimate that annual European transmission and distribution investment may need to reach €75 billion to €100 billion by 2050. It also says total grid costs could rise by as much as 60% by 2050 compared with 2022. Those are system-wide figures covering Europe’s electricity networks and the demands of electrification, renewables and new industrial and business loads. They are not a forecast of data-centre investment, a data-centre tariff increase or the cost caused by data centres.
The data-centre-specific effect is narrower and still unpriced. The proposal recognises that their electricity use is expected to grow and pressure grids, while also allowing that well-located and flexible facilities can reduce system strain. No EU-wide surcharge, discount or connection timetable is specified. The eventual commercial effect would depend on the final law, later tariff methodology and each national regulator’s evidence.
A proposal, not a rulebook
COM(2026) 600 is a Commission legislative proposal under file 2026/0203 (COD). It does not itself change a developer’s bill or queue position. Under the EU’s ordinary legislative procedure, the European Parliament and Council must agree on a final text; either can amend the proposal along the way.
Even if adopted substantially as drafted, important detail would follow. The Commission would still develop a common tariff structure and harmonised methodology, while national regulators would choose congestion measures and apply them to individual markets. The immediate watchpoints are therefore political and regulatory: whether data centres remain an explicit tariff category, how public-benefit tests are defined, what proof establishes project maturity and how regulators compare flexibility claims.
The economic direction is nevertheless clear. A connection request would no longer function simply as a ticket with a timestamp. It would become a bid for scarce infrastructure, judged on readiness, cost and public value. The winners will be operators able to show not only that they can consume power, but that their location and behaviour justify getting it sooner.