Within EPS power utility company the story of repeated feasibility does not stop with classical hydropower. It becomes even more pronounced when looking at projects explicitly designed to solve Serbia’s most visible system weaknesses: flexibility, balancing and security of supply. Reversible hydropower and gas-fired generation have been identified for more than a decade as strategic answers to those challenges. Yet they, too, remain largely confined to studies, announcements and updated concepts.
The reversible hydropower plant Bistrica is the clearest example. For EPS, Bistrica is not a marginal project. It is repeatedly framed as the backbone of future system balancing, especially as wind and solar penetration increases. In most mature variants, Bistrica is designed as a pumped-storage facility with installed capacity in the range of 600–700 MW, with storage capability sufficient for 8–10 hours of full-load operation. That implies usable storage energy of roughly 5–6 GWh, a scale that would materially change Serbia’s flexibility profile. Capital expenditure estimates have evolved over time but generally sit between €900 million and €1.2 billion, reflecting civil works dominance, electromechanical equipment costs and grid integration.
Feasibility studies for Bistrica have confirmed the same core conclusions again and again. From a system perspective, the project is highly valuable. It would reduce curtailment risk, support peak shaving, stabilise frequency and lower dependence on imports during stress periods. Financially, however, the project is structurally hard to close under current market rules. Pumped storage monetises value through arbitrage, ancillary services and system support, yet Serbia lacks a mature capacity or flexibility market that would guarantee predictable revenue. Every feasibility therefore relies on forward-looking assumptions about market reform, regional coupling and price volatility. EPS is left facing a project that is system-critical but revenue-uncertain, capital-intensive and exposed to 10–12 years from first excavation to full payback. Under those conditions, repeated feasibility becomes a rational holding pattern. The project is too important to cancel and too risky to approve.
The same equilibrium appears in EPS-linked gas power plant concepts. Over the past decade, several gas-fired generation projects have been announced, re-announced or hinted at, often in the 300–500 MW combined-cycle range. Capital costs are typically modelled between €250 million and €400 million, depending on configuration and gas connection scope. These plants are usually justified as transitional assets, providing flexible backup to renewables and replacing aging lignite units in system balancing roles.
Here again, feasibility studies are not the problem. Technical layouts, heat rates, emissions profiles and grid connection points are well understood. What blocks execution is structural exposure. Long-term gas supply pricing remains uncertain, especially in a post-crisis European market. Power offtake is predominantly merchant, with no domestic capacity market to stabilise revenues. Carbon pricing adds another layer of long-term uncertainty. Every feasibility therefore produces acceptable returns only under scenarios that assume future regulatory clarity and market depth that do not yet exist. EPS, tasked with maintaining tariff stability and financial solvency, has little incentive to lock in fuel price risk and merchant exposure for assets with 25–30 year lifetimes.
What links Bistrica and gas power plants inside EPS is that both are flexibility assets whose value is collective rather than easily monetised. They stabilise the system, but the market does not fully pay for that stability. Feasibility studies repeatedly quantify system benefits in avoided imports, reduced curtailment and improved reliability, often amounting to €50–100 million per year in avoided system costs under stress scenarios. Yet those benefits accrue diffusely to consumers and the state, while the capital risk sits squarely on EPS’s balance sheet.
From an institutional perspective, this creates a trap. EPS is expected to behave commercially, but also to deliver national energy security and transition objectives. Approving Bistrica or a large gas plant would mean committing close to €1–1.5 billion in combined capital for assets whose revenues are not contractually secured. Repeated feasibility allows EPS to demonstrate strategic alignment with government policy and EU transition narratives without crossing that exposure threshold.
There is also a sequencing issue. Both reversible hydro and gas plants only become fully rational investments once market reforms are in place: capacity remuneration mechanisms, ancillary service markets with depth, long-term balancing contracts, or explicit state support frameworks. In Serbia, these reforms are discussed in parallel with the projects themselves. Feasibility therefore becomes a bridge between a present system that cannot yet reward flexibility and a future system that is assumed to do so. As long as that future remains undefined, the bridge never ends.
In this sense, Bistrica and gas power plants are not exceptions within EPS. They are the logical extension of the same institutional behaviour seen in classical hydropower. Projects that would materially improve the system remain suspended in analysis because they concentrate long-term risk in an environment that diffuses responsibility and shortens decision horizons.
The paradox is that these are precisely the assets Serbia needs most as coal units age and renewable penetration rises. Without large-scale storage and flexible thermal backup, the system’s dependence on imports and emergency measures increases. Yet inside EPS, the rational response to that paradox has been to keep refining feasibility rather than to force commitment.
Until Serbia explicitly decides who pays for flexibility and how it is remunerated — through tariffs, capacity payments or direct state backing — reversible hydropower like Bistrica and EPS-linked gas plants will remain flagship projects in strategy documents, not in operation. The studies will continue to improve. The numbers will be sharpened. And the system will continue to wait for someone to own the risk that feasibility so carefully avoids.