Most Romanian power models built in 2024 contain a nuclear availability assumption somewhere between 88 and 92 percent. That number is not wrong because of an engineering fault, a fuel problem, a strike or a regulatory action. It is wrong because of a hydrograph.<\/span><\/p>\n On 28 July 2026, Unit 1 of the Cernavod\u0103 nuclear power plant was taken offline. On 13 August, at 10:53, the controlled shutdown of Unit 2 was completed. Both CANDU-6 units, roughly 700 MW each and together about a fifth of Romania’s electricity generation, have been in safe shutdown since. As of the first week of September, the National Institute of Hydrology and Water Management projects Danube inflows at the Romanian border will stay below multiannual averages through September, October and November.<\/span><\/p>\n The variable that determined whether Romania had baseload in August was not a plant outage schedule. It was the water level in a canal.<\/span><\/p>\n Key figures, at a glance:<\/b><\/p>\n \u00a0<\/span><\/p>\n The Danube did not fall gently. Copernicus analysis found that nearly two-thirds of the river recorded the lowest July flow rates in 34 years. Cernavod\u0103 draws cooling water from the Danube through the Danube to Black Sea Canal, and the plant has a hard operational minimum below which it cannot run.<\/span><\/p>\n The sequence was compressed:<\/span><\/p>\n The state of alert has been extended through 30 September.<\/span><\/p>\n \u00a0<\/span><\/p>\n The failure here is not a failure of forecasting weather. It is a failure of category.<\/span><\/p>\n Firmness in Romanian power models has conventionally been treated as a <\/span>technical availability<\/b> problem: forced outage rates, planned maintenance windows, fuel supply. Cernavod\u0103 performed brilliantly on all three. Its capacity factor history is among the best in Europe. Nothing in the plant’s engineering record would have flagged a 40-day zero-output event.<\/span><\/p>\n What was missing was a <\/span>hydrological availability<\/b> layer, and it was missing across the entire thermal and hydro fleet at once. Consider what the Danube basin did in the same weeks:<\/span><\/p>\n This is the part that matters for portfolio construction. <\/span>The correlation you assumed you had is not there.<\/b> A Romanian offtaker whose hedge relies on imports from Hungary or Bulgaria during a Romanian supply shock has discovered that the shock is regional and the counterparty is short at exactly the same hour, for exactly the same reason. Cross-border capacity is real, around 4,000 MW today, but capacity is not energy when everyone upstream is also short.<\/span><\/p>\n \u00a0<\/span><\/p>\n Three things filled the hole, in order of volume: <\/span>imports, thermal generation, and demand reduction.<\/b><\/p>\n Evening import flows moved from roughly 1,700 MW before the Unit 2 shutdown to above 2,500 MW after. Lignite units that were scheduled for closure under Romania’s recovery plan stayed online instead. Romania negotiated with the European Commission over Milestone 119a, the coal decommissioning commitment tied to EUR 771 million in recovery funds, and on 14 August secured a tenfold reduction in the penalty coefficient, from 5 to 0.5. Turceni and Rovinari units were retained. Around 710 MW of decommissioning remains owed.<\/span><\/p>\n On the demand side, Dacia in Mioveni and Ford Otosan in Craiova were both offline from 3 August, framed by both companies as pre-planned annual maintenance and holiday shutdowns rather than a crisis response, saving roughly 200 MW. Ford Otosan restarted on 20 August, Dacia on 25 August. Kronospan volunteered load reduction during the 21:00 to 23:00 evening peak. Mandatory industrial restriction was prepared but, according to the government, most likely not to be applied. Transelectrica retained a 24-hour notice mechanism for disconnecting large industrial consumers, with households, hospitals and critical infrastructure excluded.<\/span><\/p>\n And then there is the storage number, which deserves its own line.<\/span><\/p>\n Romania had around <\/span>2,300 MWh of battery storage<\/b> available in August. It was not enough to cover the evening peak, which meant expensive imports through the night hours alongside heavier gas and coal dispatch. That is not a criticism of the build-out, which has been fast: roughly 663 MW of storage was connected in the first eight months of 2026 alone, and Transelectrica reported passing 2,514 MW of total new generation and storage connections in 2026. It is an observation about duration. A four-hour evening gap across a 2,000 MW nuclear shortfall is an 8,000 MWh problem. Romania had a quarter of that.<\/span><\/p>\n \u00a0<\/span><\/p>\n For anyone running an offtake book, a PPA desk or a merchant-exposed asset in Romania, four things move:<\/span><\/p>\n \u00a0<\/span><\/p>\n The obvious read is that Romania has become a climate-fragile market and capital should discount it. That read is wrong, and it is wrong for a specific and checkable reason: <\/span>the alternative locations are not less exposed, they are less advanced in pricing the exposure.<\/b><\/p>\n Romania’s response to this event is the most informative thing about it.<\/span><\/p>\n Cooling architecture is changing in the build pipeline, not just the operating fleet.<\/b> The Doice\u0219ti small modular reactor project, which reached final investment decision in February 2026, is a 462 MWe plant made of six NuScale modules of 77 MWe each, sited on a former coal plant roughly 90 km northwest of Bucharest. Critically, it is not designed around continuous river abstraction. Each module sits in a large reinforced concrete basin holding a fixed water volume; normal heat rejection runs through a closed secondary circuit to cooling towers, and passive decay heat removal transfers into the basin without pumps or external water, transitioning to natural air circulation once residual heat drops below roughly 400 kW per module. This is vendor-stated design rather than independently audited performance, and the project carries a roughly 20-month schedule slip and leadership change at Nuclearelectrica that investors should price. But the direction is unambiguous: Romania’s next nuclear MW are not Danube-dependent in the way its current nuclear MW are. Target commercial operation is 2033.<\/span><\/p>\n The firm capacity pipeline is real and dated.<\/b> Cernavod\u0103 Units 3 and 4, 700 MW each, target commercial operation in 2030 and 2031, backed by an EPCM consortium of Fluor, AtkinsR\u00e9alis, Ansaldo Nucleare and Sargent & Lundy and by export credit commitments from US EXIM, Canada’s EDC and Italy’s SACE. Unit 1’s refurbishment carries a EUR 1.9 billion EPC contract and a EUR 800 million European Investment Bank loan approved in July 2026, with the reactor offline for refurbishment across 2027 to 2029. Neptun Deep, a EUR 4 billion project between OMV Petrom and Romgaz, targets first gas in 2027 at roughly 8 bcm per year against approximately 100 bcm of recoverable reserves, which materially changes the domestic fuel position of Romania’s gas fleet.<\/span><\/p>\n Storage economics have moved from thesis to policy.<\/b> Romania received EUR 636.9 million from the EU Modernisation Fund on 6 July 2026 specifically for battery storage, the largest single national allocation in a EUR 2.5 billion tranche covering eleven member states. A further EUR 150 million national round opened for applications on 1 September 2026 and closes 30 October, capped at EUR 15 million per company and EUR 69,000\/MWh, targeting 2,174 MWh of new capacity, with contracts expected January to February 2027 and a 48-month build deadline.<\/span><\/p>\n Renewable price discovery is working.<\/b> Across three CfD rounds, Romania has awarded roughly 4.6 GW. Solar strike prices fell from an average of EUR 51\/MWh in the December 2024 round to EUR 40.35\/MWh in August 2025. Wind cleared at EUR 65\/MWh, then EUR 73.89\/MWh, then EUR 59.95 to 74.90\/MWh in December 2025. Solar capacity passed 7 GW at the end of 2025 after 2.2 GW of additions in a single year, wind sits above 3.1 GW, and prosumer capacity reached 3.35 GW, up 43 percent year on year.<\/span><\/p>\n The grid is being repriced too.<\/b> ANRE approved Transelectrica’s 2026 to 2035 investment plan on 4 September 2026, covering 117 projects including 25 entirely new ones, and specifically required an accelerated schedule for the Hungary and Bulgaria interconnection reinforcements. Cross-border capacity is currently around 4,000 MW with a stated target of 7,000 MW within two to three years. Romania’s interconnection ratio was 18.5 percent in 2024, above the EU’s 15 percent minimum.<\/span><\/p>\n And demand is arriving regardless. Sany International announced a EUR 1 billion investment in April 2026 combining a 70 MW data centre in Timi\u0219oara with 1,600 MW of solar. Romania is projected to become the second-largest data centre hub in Central and Eastern Europe after Poland.<\/span><\/p>\n \u00a0<\/span><\/p>\n We do not think August 2026 is a reason to reduce Romanian exposure. We think it is a reason to restructure it.<\/span><\/p>\n Here is our position, stated plainly.<\/span><\/p>\n Climate risk to baseload generation is not a Romanian problem. It is a thermal-generation-on-rivers problem<\/b>, and Romania is simply the first market in the region where it produced a headline number instead of a footnote. Hungary, Bulgaria, Serbia and Austria all took the same hit in the same weeks. Any investor who reallocates from Romania to a neighbouring Danube-basin market has not reduced the exposure. They have renamed it.<\/span><\/p>\n The repricing is happening in the right direction, and that is the investable part.<\/b> Romania’s structural answer to an evening-peak, low-water, low-wind failure mode is more storage, more firm non-river-dependent capacity, stronger interconnection and a deeper CfD market. Every one of those four is funded, dated and in motion right now, not aspirational. The EUR 636.9 million Modernisation Fund allocation and the EUR 150 million national round open until 30 October are the clearest signal available: the country is buying exactly the asset class that August proved short.<\/span><\/p>\n For offtakers and PPA desks, the correct response is three specific changes<\/b>, not a country-level decision. First, write hydrological events explicitly into force majeure and into availability definitions rather than relying on generic language. Second, price evening shape separately from annual volume, because August broke the shape while the annual balance was improving. Third, stop treating Romanian, Hungarian and Bulgarian counterparties as uncorrelated firmness, because in a Danube low-flow event they are the same trade.<\/span><\/p>\n For merchant-exposed investors, this event improves the case for storage and for firm flexible capacity in Romania, and weakens the case for unhedged baseload-shaped assumptions.<\/b> A market that just demonstrated a four-hour evening scarcity window, at EUR 150\/MWh monthly averages with intraday peaks well above that, and which is simultaneously subsidising the technology that captures that window, is a market where flexibility is being paid to show up. That is not a market to exit. It is a market where the product mix needs to change.<\/span><\/p>\n Where we would be honest about the risk:<\/b> the restart date is genuinely unknown, hydrological forecasts point below average through November, Unit 1 goes offline again for refurbishment in 2027, and the SMR programme carries schedule and governance risk that should be underwritten rather than assumed. Anyone modelling Romanian nuclear at 90 percent availability for 2027 and 2028 should redo that work this quarter.<\/span><\/p>\n Romania remains, in our view, the right location in Central and Eastern Europe for renewable and storage capital. The reason is not that it avoided this shock. The reason is that it is the market in the region currently paying the most, and moving the fastest, to fix the thing the shock exposed.<\/span><\/p>\n \u00a0<\/span><\/p>\n What would change our view:<\/b> a second consecutive drought year without a corresponding acceleration in storage commissioning, or a slip in the Cernavod\u0103 Units 3 and 4 timeline beyond 2031.<\/span><\/p>\n What is your desk assuming for Romanian nuclear availability in 2027? We would genuinely like to know if anyone has already moved it.<\/span><\/p>\n <\/p>\n","protected":false},"excerpt":{"rendered":" The number you are not modelling Most Romanian power models built in 2024 contain a nuclear availability assumption somewhere between 88 and 92 percent. That number is not wrong because of an engineering fault, a fuel problem, a strike or a regulatory action. It is wrong because of a hydrograph. On 28 July 2026, Unit… Romania Has Spent Over a Month Without Nuclear Power. The Number Repricing Your Portfolio Is a River Flow Forecast<\/span><\/a><\/p>\n","protected":false},"author":6,"featured_media":2933,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"off","neve_meta_content_width":70,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"categories":[43],"tags":[],"class_list":["post-2932","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"acf":[],"_links":{"self":[{"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/posts\/2932","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/comments?post=2932"}],"version-history":[{"count":1,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/posts\/2932\/revisions"}],"predecessor-version":[{"id":2937,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/posts\/2932\/revisions\/2937"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/media\/2933"}],"wp:attachment":[{"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/media?parent=2932"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/categories?post=2932"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/momentumgroup.ro\/en\/wp-json\/wp\/v2\/tags?post=2932"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n
What actually happened, in sequence<\/b><\/h2>\n
\n
Where 2024 assumptions broke<\/b><\/h2>\n
\n
What actually covered the gap, and what it cost<\/b><\/h2>\n
What this actually changes in a contract<\/b><\/h2>\n
\n
The uncomfortable conclusion is not the obvious one<\/b><\/h2>\n
Momentum Energy’s View<\/b><\/h2>\n