Summary
- Portugal’s national assessment identifies electricity adequacy risks in every target year from 2028 to 2035.
- ACER says the modelling understates possible investment in batteries, demand response, and other resources as scarcity increases.
- Rapid data-centre demand is widening the gap between national and European assumptions about Portugal’s future electricity needs.
ACER has challenged Portugal’s assessment of future electricity supply risks, arguing that the country may be overstating its projected adequacy gap even as rapid data-centre growth places substantially greater pressure on the power system.
The EU Agency for the Cooperation of Energy Regulators issued its opinion on Portugal’s National Resource Adequacy Assessment after the national analysis reached a markedly different conclusion from the European Resource Adequacy Assessment for 2025.
Portugal’s assessment identifies adequacy risks in every target year between 2028 and 2035. The European assessment, by contrast, found no adequacy concern for Portugal.
The national model projects that Portugal’s loss of load expectation — the number of hours in which demand is expected to exceed available supply — will move above the country’s reliability standard of 1.46 hours a year. ACER said the projected exceedance is slight in 2028, rises to about five times the standard in 2030, and reaches nearly 50 times the threshold by 2035.
Data-centre demand sits directly inside that divergence. Compared with the European reference case, the Portuguese assessment assumes higher demand from new energy-intensive consumers, particularly data centres, alongside slower renewable deployment, lower availability of pumped-hydro storage, conservative battery assumptions, and the movement of the Tapada do Outeiro gas-fired plant into an out-of-market reserve.
The resulting capacity assumptions differ substantially. ACER said Portugal’s national assessment assumes around 15GW less installed electricity capacity than the European case by 2030 and around 12GW less by 2035, largely because it expects slower wind and solar deployment.
The regulator is not arguing that accelerating electricity demand presents no problem. Its criticism is instead aimed at how the model treats the market response once scarcity begins pushing power prices higher.
ACER said Portugal’s analysis does not adequately examine whether those price signals could attract new batteries, demand response, or other electricity resources. It also considers the contribution assigned to demand-side flexibility too limited.
Cross-border supply creates another difference. ACER said the national assessment does not fully reflect factors such as Spain’s capacity remuneration mechanism, which could influence how much electricity is available for import during periods of scarcity.
The dispute illustrates a difficult problem for electricity planners as large data-centre connection requests multiply across Europe. A proposed campus can represent hundreds of megawatts of future demand, but requested capacity, contracted capacity, installed IT equipment, and actual electricity use can materialise on different timetables.
Underestimating those loads risks leaving generation and network capacity short when facilities are ready to connect. Overestimating them can encourage investment or intervention against demand that is delayed, reduced, or never built.
The problem becomes more complicated when the supply side is also changing quickly. Battery economics, renewable deployment, demand response, interconnection, market prices, and the commercial position of existing generators can all shift during the development period for a large data-centre project.
Portugal is trying to attract digital infrastructure partly through access to renewable electricity and international connectivity. Large additions of compute capacity therefore feed directly into national assumptions about generation, storage, transmission, flexibility, and security of supply.
ACER has recommended that future Portuguese assessments use the European economic viability methodology more consistently, model market-led investment in batteries and demand response, account appropriately for out-of-market resources, and apply more realistic assumptions about demand-side flexibility.
The outcome does not establish that Portugal faces no future electricity constraint. It shows how strongly the answer depends on assumptions about both data-centre demand and the resources that higher prices could bring into the power system.
As European connection queues fill with increasingly large loads, those modelling choices will influence how much generation, storage, and network reinforcement governments conclude they need to deliver.
