Serbia’s transition from a country that regularly generated an electricity surplus into one increasingly exposed to imports is becoming one of the most important structural risks facing its energy system. Domestic consumption has risen by roughly 45% since 2000, while investment in large new generating capacity failed for much of the following two decades to keep pace. At the same time, the lignite mines supporting Elektroprivreda Srbije’s thermal fleet have become more difficult to operate, hydropower is increasingly exposed to drought and low river levels, and the electricity system is entering a period in which decarbonisation requirements will make its traditional coal-heavy model progressively more expensive.
The deterioration did not happen suddenly. Serbia’s power balance gradually weakened as economic activity and electricity demand recovered after the disruptions of the 1990s. Domestic electricity consumption increased from approximately 28.3 TWh in 2000 to 41.2 TWh in 2024, an increase of about 45%. Generating capacity, particularly within the state-controlled EPS portfolio, did not expand at anything approaching the same rate.
For much of the period between 2005 and 2020, Serbia nevertheless remained a net electricity exporter in annual terms, with 2018 the principal exception. Hydropower played a particularly important balancing role, allowing the country to generate surpluses during favourable hydrological periods and export electricity while importing more heavily during winter or periods of weak domestic production.
That model became increasingly fragile because the underlying generation portfolio remained concentrated around lignite-fired thermal plants built decades earlier.
Coal historically supplied around 70% of Serbia’s electricity needs, with the Nikola Tesla and Kostolac thermal complexes forming the backbone of baseload production. Hydroelectric plants provided most of the remainder, while wind, solar and other renewable sources only began to reach material scale much later.
The system’s vulnerability became unmistakable in December 2021, when severe operational problems at the Nikola Tesla thermal power complex exposed weaknesses in both coal quality and mine management. Poor-quality material reaching the boilers contributed to a collapse in generating availability at precisely the point when winter electricity demand was high.
The consequences extended beyond an isolated technical failure.
Serbia was forced into the regional electricity market to replace lost domestic production, contributing to a sharp rise in imports during 2021 and 2022. The latter became the most difficult year in Serbia’s modern electricity trade balance, while expensive emergency energy purchases placed substantial financial pressure on EPS.
The same period exposed a second weakness: Serbia was increasingly importing not only electricity, but also the coal needed to run the thermal plants that were supposed to provide domestic energy security.
Coal imports had already been rising from relatively low levels around the beginning of the century. Serbia imported roughly 2.1mn tonnes in 2000, but the figure reached approximately 6.8mn tonnes in 2022. Although imports subsequently declined, they remained around 5.4mn tonnes in 2024, still far above the levels typical of the previous decade.
This creates an uncomfortable contradiction in Serbia’s traditional energy-security argument.
Coal-fired generation is usually presented as the domestic component of the power system because Serbia possesses significant lignite reserves. But a thermal system dependent on increasingly difficult domestic mining conditions and millions of tonnes of imported coal provides less energy sovereignty than the headline installed capacity suggests.
The Kolubara mining basin remains particularly important because it feeds the Nikola Tesla plants. Yet lignite is neither unlimited nor homogeneous. Existing deposits become progressively more difficult and expensive to exploit as mining moves into new areas, while geological conditions affect both calorific value and the amount of overburden that must be removed before usable coal can be extracted.
The challenge is therefore not simply whether Serbia possesses coal underground. It is whether EPS can open and operate new mining fields rapidly enough, at acceptable cost and with sufficient fuel quality, to sustain thermal generation during the period before lower-carbon replacement capacity is available.
Recent production figures suggest some improvement. EPS reported that coal production increased by around 2mn tonnes in 2025 compared with 2024, while the Drmno mine supplying the Kostolac thermal complex exceeded 10mn tonnes, its highest annual production since operations began. EPS has also started investment in additional mining systems intended to secure future coal supply.
That improvement helps stabilise the immediate power balance, but it does not solve the strategic problem.
Serbia’s official 2026 Energy Balance still assumes a system overwhelmingly dependent on thermal production. Gross electricity generation is planned at approximately 39.3 TWh, about 5% above the estimated 2025 level of 37.36 TWh. Conventional thermal power plants are expected to generate approximately 24.5 TWh, equivalent to 62.35% of gross production, while combined heat-and-power plants add another 1.38 TWh. Hydropower is expected to contribute approximately 9.83 TWh, or around 25%.
Wind generation is projected at approximately 2.11 TWh, equivalent to 5.37%, while solar generation is expected to reach about 465 GWh. The difference in scale remains substantial: coal-based generation is still measured in tens of terawatt-hours, while utility-scale solar remains below half a terawatt-hour in the official annual balance.
The direction is changing, however.
EPS commissioned Kostolac B3, adding 350 MW of thermal capacity, and completed its first wind farm and the Petka solar facility during 2025. Those renewable projects contributed another 76 MW, bringing the increase in the EPS generation portfolio over two years to approximately 426 MW.
The more consequential projects are still ahead.
EPS plans investment of around €1bn during 2026, with a large share directed towards new renewable capacity and improvements in system reliability. The central project is a portfolio of 1 GW of self-balancing solar plants combined with 200 MW of battery storage, being developed with Hyundai Engineering and UGT Renewables.
The solar portfolio is expected to generate around 1.6 TWh annually once completed. The largest individual development is planned across Negotin and Zaječar with approximately 460 MW, while another 302 MW is planned near Bošnjace in the municipality of Lebane. Initial capacity is expected to enter the system from 2027, with completion targeted around mid-2028.
For Serbia’s electricity balance, 1.6 TWh of additional annual generation is meaningful but not transformative on its own.
Against final electricity demand approaching 30 TWh and gross domestic system consumption above that level once network losses and the energy sector’s own use are included, the project would cover only part of the structural gap created by rising demand and declining reliability of older thermal assets.
Solar also cannot directly replace coal megawatt-for-megawatt.
A lignite unit can theoretically generate around the clock, subject to fuel supply and outages. Solar production is concentrated into daylight hours and varies strongly by season and weather. Large-scale photovoltaic deployment therefore reduces fuel consumption and imports during solar hours but increases the value of flexible generation, storage and interconnection.
This is why Serbia’s planned Bistrica pumped-storage hydropower plant may ultimately have greater strategic importance than its installed capacity alone suggests.
The project is planned at around 660 MW in the Nova Varoš and Priboj area. Technical documentation is approaching completion, the first tender for preparatory infrastructure has been launched and expropriation procedures are progressing. Authorities estimate that Bistrica could support integration of as much as 1,500 MW of additional renewable generating capacity.
Bistrica would operate differently from a conventional power station.
When electricity is abundant and prices are low — increasingly during periods of strong solar or wind generation — the plant would use electricity to pump water into an upper reservoir. That stored water could later be released through turbines during periods of higher demand or lower renewable production.
The economic value is therefore concentrated in flexibility, balancing and peak generation, rather than simply annual MWh output.
EPS has indicated that around 25–30% of the energy available to Bistrica could come from renewable natural inflows from the upstream Uvac hydroelectric system, while modern variable-speed units would allow much more flexible pumping than the technology used at Serbia’s existing Bajina Bašta pumped-storage plant.
That flexibility becomes increasingly valuable as the generation mix shifts.
A system with large volumes of solar does not necessarily suffer from an annual shortage of electricity. It can instead experience surpluses at noon and shortages during winter evenings. Pumped storage and batteries convert part of that temporal mismatch into usable system capacity.
Serbia therefore needs to distinguish between energy capacity and firm capacity in its investment programme.
Solar can be built relatively quickly and at increasingly competitive costs. Wind provides a higher capacity factor and often generates during different hours. Hydropower provides flexibility. Batteries provide short-duration balancing. Pumped storage provides longer-duration system support.
Coal still supplies firm generation today, but maintaining that role indefinitely becomes increasingly difficult economically and environmentally.
The 2026 balance illustrates how narrow Serbia’s margin has become.
The government plans electricity imports of around 6.93 TWh during the year and exports of approximately 7.41 TWh, both figures including transit. That compares with estimated 2025 imports of 8.09 TWh and exports of 7.29 TWh. On that measure, 2025 was effectively a net-import year, while the 2026 plan assumes a modest improvement.
Final electricity consumption is expected to rise again in 2026, reaching approximately 29.97 TWh, about 1% above 2025. Industry and construction are projected to consume around 10.05 TWh, while households remain the single largest category, representing about 44.6% of final demand.
That demand structure limits the government’s room to solve shortages simply through industrial curtailment.
Electricity is not only an input for steel mills, mines or manufacturing plants. It is a politically sensitive household commodity consumed by millions of Serbian residents, many of whom also rely heavily on electricity for heating during winter.
Every structural generation shortfall therefore eventually has to be met through additional domestic capacity, energy efficiency, imports or demand management.
Imports are not inherently a sign of system failure.
An interconnected European electricity market is designed around cross-border trading. A country should import when neighbouring electricity is cheaper and export when domestic generation has an advantage.
The problem arises when imports cease to be an optimisation tool and become a requirement for maintaining supply.
That distinction matters financially because Serbia cannot control regional wholesale prices.
During periods of cold weather, drought or generation outages across Southeast Europe, neighbouring countries may need electricity at the same time. Import prices can therefore rise precisely when Serbia’s demand is highest and its domestic generation weakest.
The 2021–2022 crisis provided an extreme demonstration of that exposure.
The strategic objective should consequently not be electricity autarky. It should be maintaining enough reliable domestic capacity and flexibility that Serbia can choose when to import rather than being forced to buy regardless of price.
Hydrology is making that balance more difficult.
Serbia’s hydropower system has historically provided its strongest electricity surpluses during periods of favourable river flows. Extremely low Danube levels reduce production from the Đerdap plants and remove part of the summer surplus that previously supported electricity exports. Weak river levels therefore compound rather than replace the structural problem in coal generation.
Climate variability increases the risk of relying on hydropower as the automatic counterweight to thermal generation.
A dry year can simultaneously reduce hydro output, increase air-conditioning demand and disrupt river-based cooling or transport. It can also affect agricultural production and wider economic activity.
Serbia consequently needs a more diversified power portfolio rather than simply replacing one dominant source with another.
Wind should play a materially different role from solar in that portfolio.
Wind projects typically generate at higher annual capacity factors and their output is less concentrated around midday. They remain intermittent, but their production profile can complement solar and reduce the amount of storage required for a given renewable share.
EPS’s entry into wind generation is therefore important even though the initial capacity remains small relative to the thermal fleet.
Private investors are likely to remain equally important because Serbia’s decarbonisation requirement is larger than EPS can finance alone.
The company itself has set a goal for 45% of electricity generation to come from renewable sources by 2030, implying a major change from a system where fossil generation still contributes more than 60%.
Reaching that target requires not only wind and solar generation but also new transmission lines, substations, balancing assets and market reforms.
Grid investment increasingly becomes part of the generation investment case.
A renewable project without transmission capacity has limited economic value. As the number of wind and solar connection applications increases, EMS must determine where new capacity can be integrated without creating congestion or destabilising the system.
This raises the risk of connection delays becoming one of the largest hidden costs in Serbia’s renewable transition.
A wind or solar project delayed 12–18 months after construction capital has been committed suffers a measurable reduction in equity returns because interest continues to accrue while the plant generates no revenue. Depending on leverage and project economics, such a delay can remove several percentage points from the equity IRR even when the underlying generation asset remains technically sound.
Solar faces an additional capture-price risk.
As several gigawatts of photovoltaic capacity enter the Serbian and neighbouring markets, midday electricity prices can fall sharply during sunny periods. A project may therefore generate electricity when the wholesale market is least valuable.
Battery storage helps reduce that exposure but does not eliminate it.
Wind has different economics. Higher capacity factors and a less concentrated production profile reduce the severity of solar-style price cannibalisation, although wind projects remain vulnerable to grid constraints and periods of correlated regional output.
Pumped storage such as Bistrica has the opposite economic exposure: greater hourly price volatility can improve its value because the plant earns money from moving electricity from cheap periods into expensive ones.
That is why Serbia’s energy investment programme increasingly needs to be viewed as an integrated system rather than a list of separate projects.
The 1 GW solar portfolio, 200 MW of battery storage, approximately 660 MW Bistrica pumped-storage plant, hydro modernisation, new wind farms and additional transmission capacity are complementary investments.
Their combined value is considerably greater than treating each project as an isolated source of electricity.
The remaining question is the role of coal.
Serbia cannot close its thermal fleet quickly without creating a substantial security-of-supply problem. Thermal plants are still expected to produce almost two-thirds of gross electricity in 2026. Removing that generation before replacement capacity exists would increase imports dramatically.
Keeping the entire existing lignite system operating indefinitely is equally difficult.
Mining costs, environmental investments, ageing equipment and European carbon policy will progressively increase the economic burden.
The introduction of the EU’s Carbon Border Adjustment Mechanism adds a direct commercial dimension. Serbian electricity exports are becoming increasingly sensitive to the carbon intensity of the generation mix, meaning coal-heavy output may be physically available but commercially less attractive in European markets.
The implication is significant.
Serbia can rebuild a physical electricity surplus and still find that coal-generated exports are economically unattractive once their embedded carbon cost is recognised.
New low-carbon generation therefore has two values: it reduces dependence on imports and preserves the competitiveness of Serbian electricity in increasingly carbon-sensitive European markets.
The same applies indirectly to Serbian industry.
Exporters of steel, aluminium, fertilisers and other carbon-intensive goods increasingly need access to lower-carbon electricity if they are to remain competitive in the EU market. The electricity transition is therefore becoming part of Serbia’s wider industrial policy rather than remaining an issue confined to EPS.
The country has lost valuable time.
For almost two decades after Serbia returned to a relatively stable electricity surplus, demand expanded while major new generation projects progressed slowly. The result was a system capable of functioning comfortably during favourable hydrology and stable coal production but increasingly vulnerable when either component failed.
The 2021–2022 crisis exposed that vulnerability. Low Danube levels expose it from another direction.
The investment cycle now beginning is substantially larger than anything EPS has implemented in recent years, but much of the capacity capable of changing the structural balance will not be available immediately.
The 1 GW solar portfolio is targeted towards 2027–2028, while Bistrica remains in the preparatory phase. Meanwhile, demand continues rising and coal remains responsible for most domestic production.
Serbia therefore faces several years in which the old system must remain reliable while the new one is built around it.
That transition period is financially and operationally the most difficult part of the process.
Too little investment in coal and existing hydro risks outages before renewable replacements are ready. Too much investment in extending old lignite assets risks creating stranded costs once carbon pricing and environmental requirements tighten.
The rational path is increasingly concentrated on reliability expenditure for existing thermal units, accelerated renewable construction, hydro rehabilitation, grid reinforcement and large-scale flexibility rather than attempting to recreate the electricity system Serbia operated twenty years ago.
Serbia is unlikely to return to the old model in which abundant lignite and favourable hydrology automatically produced large export surpluses.
The economically stronger objective is a different one: a system where domestic wind, solar, hydro, storage and a declining thermal fleet provide enough reliable electricity that imports become a commercial choice rather than an emergency necessity.
The shift from exporter to importer was therefore not caused by a single failure at Nikola Tesla, one dry summer or rising consumption alone. It reflects two decades in which electricity demand grew faster than Serbia replaced and diversified its generating base.
Reversing that imbalance will require several gigawatts of new capacity, billions of euros of investment and a generation portfolio fundamentally different from the one that created Serbia’s electricity surpluses in the first place.
