Serbia’s Hydropower Shortfall Highlights Delays in Bistrica and Đerdap 3

Serbia’s worsening hydrological conditions are putting greater pressure on the country’s electricity system while two major pumped-storage projects remain years away from operation. RHE Bistrica, with planned capacity of 661 MW, is currently targeted for completion in 2032, while the 1,800 MW RHE Đerdap 3 is scheduled for 2036. Neither project has yet secured a fully defined financing structure.

The immediate pressure is reflected in hydropower production. Over the past three months, Đerdap 1 has reportedly generated at roughly 20% of its normal production level, while Đerdap 2 has operated at around one-third of capacity. Energy Minister described Đerdap’s output as the lowest recorded, while the government has continued to state that Serbia’s electricity supply remains stable.

System stability, however, does not eliminate the financial consequences of constrained domestic generation. Elektroprivreda Srbije (EPS) has had to rely on electricity imports, maintain available lignite generation and manage remaining hydropower reserves selectively.

On 4 August, Serbia’s average electricity consumption was forecast at 4,030 MW, compared with domestic production of approximately 3,534 MW. The resulting average import requirement was 496 MW, increasing to almost 797 MW during peak hours. Coal accounted for approximately 79% of reported generation on the preceding day, while hydropower averaged only 621 MW and wind supplied around 110 MW.

Electricity prices expose the cost of limited flexibility

The wholesale market has also reflected the pressure on supply. The SEEPEX day-ahead baseload price reached €174.64/MWh, while the evening maximum climbed to €496.80/MWh at H20. Serbia’s average baseload price remained below those of Hungary, Romania, Croatia and Slovenia, but the daily average concealed a significant evening scarcity premium. Electricity purchased during the tightest hours can therefore carry a substantially higher cost than the baseload figure indicates.

On 4 August, SEEPEX prices fell to approximately €70/MWh at H13 before approaching €497/MWh at H20, creating a theoretical hourly spread of more than €426/MWh. The drought has consequently increased the value of electricity storage, but neither Bistrica nor Đerdap 3 can contribute additional storage capacity during the current hydrological cycle. Serbia must continue managing the gap between its need for greater system flexibility and the delivery of the infrastructure intended to provide it.

Bistrica remains the more advanced storage project

RHE Bistrica is further advanced than Đerdap 3. The project is designed around four reversible generating units with total installed capacity of 661 MW and expected annual electricity production of approximately 1,600 GWh. The government’s energy infrastructure development plan places the investment at €962.5 million, while more recent public estimates have risen to approximately €1.2 billion. On those figures, the implied investment cost ranges from around €1.46 million/MW to €1.82 million/MW.

The final cost will depend on the defined project scope, including underground construction, reservoirs, water-transfer systems, turbines, generators, grid connection, access infrastructure, environmental mitigation, land acquisition, financing costs and contingencies. Changes to the technical baseline could therefore materially affect the final capital requirement. Serbia has been negotiating financing with the Japan International Cooperation Agency (JICA). The current timetable envisages a formal financing agreement during EXPO 2027, with Japanese banks and the selected delivery structure expected to provide funding. The amount, tenor, sovereign support arrangements and allocation of construction risk have not yet been publicly fixed.

A financing agreement in 2027 would not by itself constitute financial close. Before debt could be drawn, Bistrica would need an agreed feasibility case, environmental and social approvals, a bankable EPC structure, confirmed grid arrangements, available land, verified costs and a complete package of conditions precedent.

The government currently targets completion in 2032, with construction expected to start in 2027. That would leave approximately five years for construction, a demanding but potentially achievable schedule if the project enters the construction phase with mature design, completed expropriation, secured financing and sufficient geological investigation.

Land, permits and design remain critical to the schedule

Bistrica has not yet reached that stage. Existing documentation reportedly includes a preliminary feasibility study and main design, while the construction permit is expected in 2026. Land discussions have only recently begun in Priboj and Nova Varoš, where property must be acquired for both preparatory activities and permanent infrastructure.

For a pumped-storage project, land acquisition extends well beyond the main plant site. Reservoirs, water-transfer infrastructure, tunnels, shafts, roads, transmission facilities and construction compounds all require access and property rights. Delays in expropriation can prevent contractors from opening construction fronts even after financing and equipment contracts have been arranged. Environmental and water-management approvals are similarly important. The project must address impacts on river systems, biodiversity, local communities, landscape, water users and sediment management. Any major design changes introduced after excavation begins could increase both construction costs and financing requirements.

The 2032 completion target consequently depends on several workstreams advancing simultaneously through 2026 and 2027. A 12- to 18-month delay in permitting, financing or land acquisition could push commercial operation into 2033 or 2034, while increasing capitalised interest, owner’s costs and contractor escalation.

At an investment value of approximately €1.2 billion, even a one-year delay could generate tens of millions of euros in additional financing charges, inflation, extended project-management costs and possible contractor remobilisation.

Storage would shift Serbia’s electricity purchases

Bistrica would not create primary energy in the same way as a conventional generating plant. A pumped-storage facility uses electricity to move water into an upper reservoir and subsequently generates electricity by releasing that water through reversible turbines.

Because the process incurs energy losses, the plant consumes more electricity during pumping than it later produces. Its economic function is instead to shift electricity availability between different periods, absorbing power when prices are low and generating when demand or market prices are higher. That function is becoming more relevant as solar generation contributes to lower midday prices across the region while electricity demand and prices can rise after sunset.

A pumped-storage facility would not capture the entire difference between low and high market prices because pumping losses, transmission charges, market liquidity and operating constraints reduce the effective spread. Nevertheless, Bistrica could potentially generate revenues from day-ahead arbitrage, intraday trading, balancing energy, frequency services, reserve capacity, congestion management and emergency system support.

With 661 MW of installed capacity, Bistrica would be large enough to materially alter Serbia’s evening electricity balance. Compared with the approximately 797 MW peak import requirement recorded on 4 August, its capacity could theoretically cover most of that shortfall for a limited period, depending on stored water and transmission conditions.

Its financial case therefore cannot be assessed solely against the projected 1,600 GWh of annual output. A bankable model would need distinct assumptions for pumping costs, generation prices, balancing revenues, reserve payments and avoided electricity imports, together with scenarios for hydrology, renewable deployment, market coupling and the future availability of Serbia’s coal fleet.

The existing Bajina Bašta pumped-storage plant illustrates the operating model. It can pump during periods of lower prices and generate electricity during morning and evening peaks. One existing storage facility is insufficient for a power system that is adding variable renewable generation, continuing to depend on ageing lignite assets and facing increasingly difficult summer hydrological conditions.

Đerdap 3 remains at an earlier development stage

The proposed RHE Đerdap 3 is substantially larger than Bistrica. Current plans envisage six reversible units with combined capacity of 1,800 MW and an estimated investment of €2.63 billion, equivalent to approximately €1.46 million/MW. The project has a target completion date of 2036, placing it at least a decade away under the current schedule. Its financing structure has not been closed and the required funding has not been secured.

The Ministry of Mining and Energy launched an expression-of-interest process limited to US companies under the Serbia-US energy cooperation framework. Six companies reportedly responded, but Bechtel was the only participant to submit a compliant offer and qualify for the second stage. Bechtel has previously prepared preliminary work associated with Đerdap 3 and has expressed interest in the project since 2022. It is therefore positioned as the probable strategic delivery partner, although the government working group has not published a deadline for completing the selection procedure.

The contracting model combines the potential selection of a construction partner outside the standard public-procurement procedure with a separate competitive process for planning and part of the technical documentation.

That documentation is expected to cover the general design, preliminary feasibility study, strategic environmental assessment and other development documents. The structure confirms that Đerdap 3 remains in an earlier project-definition phase than Bistrica. The final technical configuration, geological conditions, reservoir arrangement, grid interfaces, environmental constraints and construction timetable still need to be developed into an investable technical baseline.

The €2.63 billion investment estimate should therefore be viewed as a planning figure rather than a fixed construction cost. For a large underground pumped-storage facility, geological conditions can significantly affect tunnelling costs, excavation methods, support systems, water sealing and the overall construction schedule.

Đerdap 3 will require major grid investment

The scale of Đerdap 3 also creates substantial transmission requirements. An 1,800 MW pumped-storage facility would operate both as a major electricity consumer during pumping and as a major generator during discharge. The transmission system would need to absorb the pumping load without creating congestion and accommodate the full generating output without compromising system stability. This requires detailed load-flow, short-circuit, dynamic-stability, frequency-response and N-1 security studies.

Associated substations and high-voltage transmission lines could require several years of separate development and construction.

Serbia’s transmission network is already subject to congestion and connection constraints. The timetable for Đerdap 3 must therefore account not only for construction of the pumped-storage plant itself but also for the grid infrastructure needed to make its 1,800 MW capacity fully usable. Completing the storage facility before the required transmission infrastructure would create a mismatch between installed generation capacity and the system’s ability to use it.

Combined projects represent more than €3.8 billion

Using the higher estimate for Bistrica, the two pumped-storage projects represent approximately €3.83 billion of combined investment. Once development contingencies, grid reinforcement, owner’s costs and construction-period financing are included, the overall programme could exceed €4 billion. Such an investment scale requires Serbia to determine how individual risks will be allocated among the state, EPS, contractors and lenders.

Hydrological and electricity-market risks cannot simply be transferred to an EPC contractor. Construction and performance risks can be addressed through contractual guarantees, while long-term revenue depends on electricity-price spreads, balancing rules, renewable penetration and the operating strategy adopted by EPS or another project company.

A state-supported financing model could provide access to lower-cost debt but would create contingent liabilities for the sovereign balance sheet. Project finance would require more predictable revenues, potentially through availability payments, regulated storage remuneration, capacity payments or long-term contracts for ancillary services.

Pure merchant financing would be challenging for projects of this scale. Day-ahead arbitrage can produce substantial revenues during scarcity events, but those revenues are volatile and could decline as batteries, interconnections and flexible demand expand. Lenders would therefore require a dependable contracted revenue floor.

The two projects also need to be assessed as a portfolio. Bistrica could provide additional flexibility in the early 2030s, while Đerdap 3 could substantially expand Serbia’s storage capacity later in the decade. Their operating strategies, grid requirements and revenue streams could overlap during the same low-price and high-price periods. Their economics will therefore need to be tested against Serbia’s planned solar, wind, battery and interconnection pipeline. Bistrica cannot be valued solely using current electricity-price spreads, while Đerdap 3 cannot assume that the scarcity conditions recorded in 2026 will continue unchanged into the 2040s.

Existing assets must cover the near-term shortfall

Neither pumped-storage project can resolve Serbia’s immediate hydrological constraints. The current shortfall will have to be managed through the existing system, including greater reliance on lignite generation, selective use of hydropower reserves, the rehabilitated Bajina Bašta storage facility and electricity imports, alongside possible demand-management measures during the tightest periods.

Regional electricity availability also cannot be assumed. Romania is protecting the remaining operating unit at the Cernavoda nuclear plant against extremely low Danube flows. Hungary has experienced reduced nuclear availability, while Croatia has been importing close to half of its summer electricity requirements. Serbia has recently relied heavily on electricity flows from Bulgaria and North Macedonia.

The exposure therefore extends beyond the direct cost of imported power. Serbia also faces the risk that neighbouring markets require additional electricity at the same time. Evening prices approaching €500/MWh illustrate the financial impact when flexible generation and transmission capacity become constrained simultaneously.

Bistrica and Đerdap 3 could reduce that exposure once completed, but their schedules remain dependent on project preparation and financing. Bistrica still requires a closed financing package, land acquisition, permits and a construction-ready design. Đerdap 3 requires a bankable technical concept, a strategic-partner decision, an environmental pathway, a grid-development programme and a financing model. Serbia’s hydrological stress has emerged in 2026, while its first major new pumped-storage capacity is not scheduled before 2032 and the larger Đerdap 3 project not before 2036. The resulting gap between current system pressure and future storage capacity is shaping the country’s electricity-security investment timetable.