Grids for data centres꞉ ambitious grid planning can win Europe’s AI race | Ember

In this chapter

Recommendations
3.1
Read

Planning for the future

Trends in data centre investment in Europe make it clear that grid capacity has become a prized resource and a pull factor for major investors. If Europe is to have a place in the global AI race, grid planners need to step up their game.

The rapid development of AI can provide a much needed boost for European economies. But it will not happen unless the energy infrastructure is prepared to accommodate the necessary data centre expansion. Forward-looking and anticipatory grid planning is a key solution, but there are also short-term measures such as data centre flexibility and alternative connection contracts that can tackle grid congestion and unlock millions in investments. Multiple tactics need to be deployed in parallel to give Europe a chance of hitting its data centre deployment target and avoid further downgrading of forecasts. Countries that become leaders in innovative grid planning will likely emerge as Europe’s new data centre hubs.

3.1

Recommendations

3.1.1 Governments and regulators need to recognise grids as strategic infrastructure for economic growth and competitiveness

Grid planning is no longer just about managing the power system – it is a strategic lever to attract high-value industries and unlock national economic ambitions. Countries seeking to attract such investments should adapt their grid planning practices and investment strategies accordingly. This requires more ambitious and forward-looking approaches, which are also crucial to enable anticipatory investments.

National grid development plans should recognise the opportunities coming with AI, but also electrification, flexibility and digitalisation. Planning for technological innovations ahead of time can become a pull factor for investments, bringing benefits for the economy and energy consumers. This is even more critical in sectors developing as quickly as data centres, where investment cycles are much shorter than in the case of grids.

 

3.1.2 Governments need to consider priority AI zones to tackle congestion

To keep AI investments in the country, governments and grid operators can direct AI projects into areas of better grid availability – e.g. outside of capital cities. This tactic is already being tested as a short-term solution to grid congestion in largely concentrated data centre hubs like Paris and London.

 

3.1.3 Spatial planning offers optimisation benefits in the long term

Joint spatial planning between grid operators, clean power developers and data centre investors can offer multiple benefits. Locating large electricity demand nodes close to clean energy supply infrastructure lowers grid investment needs. It also allows data centres to stabilise their electricity costs through near-site PPAs, which can deliver network charge savings as well.

 

3.1.4 Smarter grid connection deals can fast-track data centre deployment

System operators should offer phased or non-firm grid connection options to enable new data centres to connect faster and reduce growing queues. These solutions can be implemented quickly, even in highly congested areas, and can bring the additional benefit of improved utilisation of existing infrastructure.

 

3.1.5 Data centre flexibility has vast potential, but requires more research

Pilot projects on data centre flexibility should be initiated, given the potential to unlock capacity in grid constrained areas and shorten connection wait times. These projects might require research and development funding from governments and data centre companies, but can lead to savings in grid balancing and expansion costs.

 

3.1.6 Better data availability is needed to unlock opportunities

The abovementioned recommendations cannot be implemented without good availability of data on data centre locations, investment prospects, data centre demand profiles and utilisation scenarios. Grid operators need to provide grid capacity maps and demand-side connection queues as well.

These datasets are required for grid planning and modelling, spatial planning, designing policies and incentives, as well as research and development activities.

Previous Chapter
2: Rewiring for growth
Next Chapter
Supporting materials
Share