ACER asks Romania to recalculate more realistically the risk that the electricity system will not meet demand in 2027

Romania needs to revise some assumptions used in assessing the risk that the electric system will not be able to meet demand in the coming years, according to the EU Agency for the Cooperation of Energy Regulators. ACER states that the analysis by Transelectrica identifies a high risk for 2027, but some calculations do not sufficiently reflect the likely situation of production, storage, and flexible consumption capacities.

In short Transelectrica estimated a LOLE indicator of 143.55 hours for 2027, above the reliability standard of 13 hours. LOLE does not mean guaranteed power outage hours, but the expected number of hours per year in which the available capacity may not be sufficient to cover demand. ACER states that the risk for 2027 is mainly determined by the slower installation of wind capacities and the lack of thermal capacities, particularly the combined effect of the accelerated exit from lignite and the delay of the 1.7 GW gas plant at Mintia. The agency considers that some of Romania's assumptions are justified, including the lower hydro and wind capacity compared to the European scenario and the delay of new gas plants. However, it does not justify the assumptions regarding too few batteries, the lack of consumption flexibility, the too rapid exit from lignite, and the omission of new nuclear capacities by 2035. An alternative scenario included even in the Romanian assessment shows that maintaining 854 MW of available lignite capacity in 2027 would reduce LOLE to 11.34 hours, below the 13-hour standard. ACER states that this scenario seems to better reflect the likely situation, considering the delays in the energy transition. ACER recommends that Romania fully include the analysis of the economic viability of capacities, take into account batteries and flexible consumption, correlate higher demand with production resources, and update information for the next European assessments of the adequacy of the electric system.

ACER's opinion targets the national assessment of resource adequacy carried out by Transelectrica and transmitted to the agency by the Ministry of Energy on April 27, 2026. Such an assessment checks whether an electric system has sufficient available capacity, production, imports, storage, and flexibility to cover demand under normal conditions and during difficult periods.

The central indicator used is LOLE, Loss of Load Expectation. In simple terms, it shows how many hours per year there could statistically be situations where the available resources are insufficient for the estimated demand. It is not a forecast of power outages, but a risk indicator used in electric system planning.

Transelectrica analyzed the years 2027, 2030, and 2035. In the central scenario, the national assessment estimates a LOLE of 143.55 hours for 2027, well above the reliability standard of 13 hours. For 2030, the indicator is 0.69 hours, below the 12-hour standard. For 2035, the assessment estimates 6.69 hours, above the 2-hour standard.

The difference compared to the European assessment is significant. In the ERAA 2025, the European assessment conducted by ENTSO-E, Romania had 7.59 hours in 2028, 1.37 hours in 2030, and 0.20 hours in 2035. ACER states that the Romanian assessment uses different data and assumptions about available capacities, the pace of investments, and the structure of demand.

For 2027, ACER identifies two main causes of the high risk: the slower installation of wind capacities than anticipated and the lack of available thermal capacities. In this second category falls the accelerated closure of lignite capacities and the delay in the commissioning of the 1.7 GW gas plant at Mintia.

The Mintia plant is relevant because it could nearly double Romania's gas production capacity if it were to come online on time. ACER notes that the Romanian government has supported the acceleration of the project and that there was information regarding possible production from September 2026, but the Ministry of Energy later accepted to postpone the completion deadline until the end of 2027. For this reason, the agency considers the assumption of a delay in new gas capacities justified.

Two other gas projects, a 460 MW CCGT plant in Turceni and an 825 MW one in Ișalnița, are also delayed. ACER notes that, for Ișalnița, the second construction tender was canceled on January 30, 2026, because no eligible bids were received before the deadline. According to the European Commission, the initially planned commissioning date for 2026 has been moved to 2029.

The issue of lignite is treated differently by ACER compared to Transelectrica's central scenario. The Romanian assessment assumes that, in 2027, only 155 MW of lignite capacity will remain available, based on an accelerated exit from coal and lignite. ACER states that this assumption is not consistent with the delays of new gas capacities and the slower pace of wind development.

The agency does not say that Romania must abandon the goal of exiting lignite, but it calls for the central scenario to reflect the likely situation. The Romanian assessment includes an alternative scenario in which 854 MW of available lignite capacity remains operational in 2027. In that scenario, LOLE drops to 11.34 hours, below the reliability standard of 13 hours. ACER considers that this scenario seems more suitable for the central scenario, as it takes into account the delays in the energy transition.

ACER, however, accepts other more cautious assumptions from Romania. The lower hydro capacity than in the European assessment is considered justified, as it reflects historically lower levels of hydro production. The agency still recommends that Romania maintain the actual hydro capacity in the model and limit production by availability factors, so as not to create the impression that hydropower plants are being withdrawn from the system.

The assumption regarding lower wind capacity is also considered justified. The Romanian assessment assumes 4.4 GW of wind in 2027, starting from about 3 GW in 2026. ACER states that this trajectory is closer to the current pace of wind development in Romania and the target of 7.4 GW for 2030 from the National Energy and Climate Plan.

In contrast, ACER criticizes the assumptions regarding batteries. The Romanian assessment estimates 0.5 GW of batteries in 2027, compared to 1.4 GW in the European assessment for 2028, and 1.2 GW in 2030, compared to 2 GW in ERAA. The agency states that these assumptions are too conservative, especially in the context of the state aid of 150 million euros for storage and the new ANRE rules that eliminate double charging of stored energy and reinjected into the grid.

Double charging was a problem for investments in batteries because energy could be taxed upon entering the storage system and again upon reinjection into the grid. ACER states that the removal of this barrier, together with the support scheme, sends a favorable signal to investors for the development of storage in Romania. For this reason, the agency calls for battery projections to be aligned with other studies, including ENTSO-E's Seasonal Outlooks.

Another major gap is the flexibility of consumption. Transelectrica introduced in the assessment a higher electricity demand than in the European scenario, against the backdrop of increasing consumption from electric vehicles, heat pumps, and data centers. In 2027, demand is 4% higher than in ERAA, in 2030 by 11%, and in 2035 by 20%.

ACER states that these consumption projections may better reflect electrification and the emergence of large consumers, but the criticism is that the new demand is treated as inflexible. In the model, the consumption of electric vehicles, heat pumps, and data centers does not decrease during periods of stress for the system, although Romania is preparing rules that allow Transelectrica to purchase consumption flexibility through daily auctions.

For a system with many renewables, consumption flexibility matters because certain activities can be shifted to hours with more abundant or cheaper energy. Without this flexibility, the model may overestimate risks and indicate the need for additional plants that would not be necessary to the same extent if part of the demand could respond to price or system stress.

ACER also criticizes the treatment of new nuclear capacities. The European assessment assumes about 1.8 GW of new nuclear capacity in Romania by 2035, including units 3 and 4 at Cernavodă and the small modular reactor at Doicești. The Romanian assessment excludes them from the central scenario for 2035 due to uncertainties regarding the timeline.

The agency states that the complete exclusion of units 3 and 4 at Cernavodă from 2035 is not justified. According to Nuclearelectrica, unit 3 is planned for 2030 and unit 4 for 2031, and the Ministry of Energy anticipates possible delays until mid-2035. ACER recommends that the assessment take into account the availability of capacity from mid-2035, not its absence throughout the year.

For 2035, ACER also notes that half of the LOLE indicator in the Romanian assessment is caused by internal bottlenecks in the transmission network, not by a lack of production capacities. The agency states that such restrictions do not indicate a market problem that should be solved through a capacity mechanism, but structural issues that need to be addressed through network development and more efficient use of existing infrastructure.

ACER's opinion has a procedural consequence for Romania. The national assessment must take into account the agency's opinion and, if necessary, be modified. If the responsible authority does not fully include ACER's observations, it must publish a report with detailed reasons.

Romania is rapidly changing the structure of its electric system. Coal and lignite are being gradually phased out, gas plants need to come online as a transition solution, renewables are increasing, and consumption could rise through the electrification of transport, heating, and the emergence of data centers. The adequacy assessment checks whether all these changes fit in time.

ACER's opinion shows that the risk for 2027 mainly depends on the concrete succession of projects: how quickly lignite exits, when Mintia comes online, how much wind is installed, and how much reserve capacity remains available. For 2030 and 2035, the focus shifts to batteries, flexible consumption, the grid, and clarifying the timeline for new nuclear capacities.