Serbia’s Đerdap 3 pumped-storage hydropower project has entered a new procurement phase after a U.S.-linked public call for American companies to express interest in development and construction. The indicative project scale is 2,400 MW, with an investment value estimated at roughly €2.63bn–€3bn. The call closed on 25 June 2026, and Serbia’s Ministry of Mining and Energy confirmed that six U.S. companies submitted expressions of interest, with qualification review now under way.
The call was issued through the U.S. Embassy in Belgrade under the Serbia–U.S. strategic energy cooperation framework. Serbia said the initiative reflects a shift from diplomatic engagement toward a hard-infrastructure position within its power sector. An underlying agreement was signed in Washington on 18 September 2024 by Serbian Foreign Minister Marko Đurić and U.S. Under Secretary Jose W. Fernandez. Belgrade presented the agreement as a platform for cleaner energy, long-term supply security and deeper Serbia–U.S. cooperation.
Grid flexibility needs and project configuration details
The project is being positioned for a period when Serbia remains heavily dependent on coal while renewables expansion continues and grid flexibility requirements increase. Serbia’s transmission system faces growing needs related to managing intermittent generation and operational stress conditions. The pumped-storage concept is designed to support flexibility in periods when the system requires balancing capability.
Đerdap 3 is described by EPS as a reversible pumped-storage facility rather than a run-of-river scheme. The project concept places the facility at the 1,007th kilometre of the Danube, using water from Đerdap Lake at around 68 metres elevation. During surplus electricity periods, water would be pumped into planned upper reservoirs Pesača and Brodica, then released through turbines during peak-load hours to generate electricity.
EPS outlines a three-stage development approach with total reservoir water volume of 578 million cubic metres. In the final phase, the energy equivalent is described as around 484 GWh. Total installed capacity in the concept is stated at 2,400 MW.
Indicative costs and competing investment priorities
The reported scale is linked to both potential system value and cost exposure. At an estimated CAPEX of €2.63bn, the headline figure implies roughly €1.1mn per MW if the full 2,400 MW configuration is realised. At an estimated CAPEX of €3bn, the implied level rises to about €1.25mn per MW, before financing costs, contingencies, grid works, environmental mitigation, land, access roads, tunnels, electromechanical packages and potential cross-border constraints are included.
The project would represent a multi-cycle capital allocation decision for EPS and Serbia’s public-sector balance sheet. It would compete with projects including RHE Bistrica, grid reinforcement, solar and wind development, battery storage additions, coal-mine transition costs and EPS balance-sheet restructuring. The procurement outcome therefore depends on how the project’s system role translates into investable revenue arrangements.
Status of procurement process and FEED timeline
The current procedure indicates that the project remains at an early stage despite figures already circulating publicly. The public-call materials envisage subsequent work to define final configuration elements including number of units, installed capacity and reservoir size through feasibility studies and FEED documentation .
The initial development package is expected to include conceptual design; preliminary design with feasibility study; geotechnical and hydrological reports; construction planning; and environmental studies including EIA and ESIA . The FEED phase is expected to last around 36 months, with the selected supplier potentially moving into construction later subject to final agreement.
This timing suggests that the scheme should not yet be treated as a ready-to-build plan for full capacity at once. A first-stage configuration of 600 MW–1,000 MW is described as potentially more financeable against domestic system needs, while full build-out at 2,400 MW would require a regional market case. The FEED work is expected to cover more than engineering optimisation by defining revenue model elements such as market role, dispatch logic, cross-border value and cost allocation structure.
Romanian coordination on Danube impacts and permitting requirements
Romania is identified as a key non-financial constraint due to shared Danube hydropower assets including . The existing systems include Đerdap 1 and Đerdap 2, which are shared Serbian-Romanian assets on the Danube near the Iron Gates complex. Romanian authorities have moved toward negotiations with Serbia on a memorandum of understanding for information exchange regarding Đerdap 3.
Bucharest’s stated interest includes assurance that Đerdap 3 will not negatively affect production at Iron Gates I and Iron Gates II or Danube navigation . Because pumped storage changes outcomes when water is moved as well as stored, hydrological modelling would need to show that pumping and generation cycles do not disturb operating regimes of existing assets or create downstream impacts during sensitive periods . Any lender or export-credit agency reviewing Đerdap 3 would be expected to require a formal Serbian-Romanian framework rather than only political assurances.
The permitting process is also described as decisive given that the project sits within the wider Đerdap landscape in one of Serbia’s most sensitive protected natural areas . An EIA alone is described as insufficient where lenders apply strict environmental and social standards. A credible ESIA would need to include public consultation processes, biodiversity assessment, transboundary review and mitigation planning .
Storage competition, revenue stack elements and financing structure
The storage economics are described as shifting as batteries increasingly serve short-duration flexibility needs such as frequency response and intraday arbitrage . Battery characteristics including modularity, shorter construction time and declining costs make them competitors for many grid services. In contrast, Đerdap 3’s stated advantage would be duration, scale and system resilience based on reservoir-based storage with hundreds of GWh energy equivalent.
The project’s revenue stack would therefore need to reflect services beyond wholesale arbitrage alone since arbitrage revenues are described as insufficient for covering construction through debt service over operating life . Serbia would require frameworks covering capacity remuneration, reserve procurement, ancillary services, balancing-market access and potentially cross-border flexibility products . If Đerdap 3 were intended to serve wider Balkan and Central European markets, its value would be measured against avoided curtailment, reduced thermal reserve costs, lower import exposure during scarcity hours and improved system security during hydro-poor or wind-poor periods .
The financing package is described as likely hybrid rather than single-source funding. Potential components include state support; export-credit backing; development-finance participation; commercial debt; and contractor-linked finance . U.S.-linked company participation could help open access to American export-credit or development-finance tools but does not equate to committed U.S. financing by itself . Lenders would still seek clarity on contracted revenues; whether EPS or the state carries payment obligations; whether regulated capacity payments apply; and whether cross-border services can be monetised under existing market rules .
EPC risk profile and regional market positioning requirements
Pumped-storage EPC structures are described as requiring attention due to high geological and civil-works risk across tunnels, caverns, reservoirs, slopes, hydraulic transients and electromechanical interfaces . Cost escalation risk can increase if early studies are weak or if risk allocation mechanisms are unrealistic . Even where turnkey EPC models create a single accountable contractor politically, banks are expected to assess subsurface risk exposure alongside variation mechanisms force majeure language delay damages performance testing and interfaces between FEED findings and final construction obligations .
The project’s strategic case depends on how it fits into Serbia’s future industrial electricity demand alongside renewables integration needs . Serbia is described as seeking more renewables; more manufacturing investment; more mining processing; more data-centre-type load; and more resilient electricity supply . Industrial users exposed to European carbon rules are described as focusing on credibility traceability and firmness of lower-carbon power rather than average electricity mix alone . Pumped storage itself does not make electricity green but can make renewable-heavy supply more usable less volatile and more bankable when paired with metering guarantees of origin PPAs dispatch rules and transparent carbon accounting .
Southeast Europe’s market context is also cited as relevant given rising solar penetration coal plant ageing less reliable hydrology and cross-border price signalling driven by market coupling . Storage assets located in Serbia could eventually influence price spreads among Hungary Romania Bulgaria Croatia Bosnia and Herzegovina Montenegro and Greece . The value proposition would depend on market design that rewards cross-border flexibility rather than limiting dispatch value within narrow domestic logic .
The procurement documents describe credibility being built through next steps including FEED work permitting Romanian coordination financial structuring over an expected period of around 36 months . With six U.S.-linked expressions of interest received following qualification review initiation after the 25 June 2026 deadline, momentum exists but further evaluation remains part of the process .
