Capital scaled. Deliverability became the filter.
Romania closed September with a striking combination: a €561 million financing package pushed a 1.3 GWp solar project toward construction; public-sector battery requests outran their dedicated funding envelope within hours; another Sunday produced negative quarter-hour prices; and the power system moved between multi-gigawatt daytime exports and large evening imports while Cernavodă remained unavailable. The commercial signal is no longer simply that Romania needs more solar or more batteries. It is that revenue structure, connection quality, controllable delivery and execution discipline are becoming the variables that decide which megawatts are economically useful.
The bankable asset is becoming a controlled delivery profile, not a panel count.
The week's most important signals came from different layers of the market but point in the same direction: merchant price shape is volatile, the system's physical need changes sharply by hour, grid reinforcement is becoming explicit infrastructure policy, and the largest financed PV project pairs scale with contracted revenue support rather than relying on the headline spot price.
Solar capacity remains financeable at very large scale when revenue and execution risks are sufficiently structured. Dama Solar is the week's clearest evidence: up to €561 million of financing from a 14-lender group, with 520 MW covered by CfDs and an additional corporate PPA disclosed by the sponsor. [S06] [S07]
That should not be read as validation of an unhedged merchant-PV case. It is evidence that institutional capital can fund Romanian solar when price, offtake, construction and counterparty risks are packaged appropriately.
Stop using the weekly base price as the central revenue case.
Run PT15 shape sensitivities, weekend cases, curtailment, imbalance and location-specific export constraints. The reviewed solar-window proxy was only 57.8% of the weekly base average.
Buy connection quality before buying merchant-spread optimism.
The market clearly rewards time flexibility, but gross price shape is not EBITDA. Qualification, grid rights, charging access, round-trip losses, degradation, optimiser terms and ancillary-market saturation remain asset-specific.
Revenue de-risking is scaling faster than pure merchant underwriting.
Dama's financing structure, public support mechanisms and C&I PPAs all point toward contracting becoming a larger part of bankability while merchant exposure is managed rather than ignored.
The weekly average improved. The Sunday problem returned.
The price console uses the seven delivery days from 24 through 30 Sep. The “solar window” is an unweighted average of available OPCOM PT15 prices from 08:00–17:00 local time. It is a price-shape proxy and explicitly not a national solar capture-price calculation. [S01]
Hover, tap or focus the chart for exact values. Keyboard: left/right arrows. Source: OPCOM delivery-day outcomes retained by SolarIndustry.ro. No missing market interval is substituted. [S01]
View accessible price data table
| Date | Base | 08–17 proxy | Proxy/base | Low PT15 | Negative PT15 | Volume |
|---|---|---|---|---|---|---|
| 24 Sep | 1,224.06 | 925.95 | 75.6% | 764.59 | 0 | 34,171 MWh |
| 25 Sep | 1,160.52 | 933.23 | 80.4% | 712.68 | 0 | 34,765 MWh |
| 26 Sep | 845.10 | 425.16 | 50.3% | 3.62 | 0 | 39,367 MWh |
| 27 Sep | 603.19 | 93.71 | 15.5% | −16.01 | 21 | 46,679 MWh |
| 28 Sep | 813.64 | 305.59 | 37.6% | 9.35 | 0 | 62,552 MWh |
| 29 Sep | 895.80 | 620.29 | 69.2% | 316.71 | 0 | 54,804 MWh |
| 30 Sep | 761.74 | 338.34 | 44.4% | 50.10 | 0 | 51,437 MWh |
30 Sep maximum retained solar observation. Operational feed; not installed capacity. [S01]
12:00 retained hourly mean; negative means net export. Solar + wind were 5,587 MW against 4,487 MW of load. [S01]
From −2,685 MW net export at 13:00 to +1,813 MW net import at 18:00. Derived from retained system observations. [S01]
Up to €561m from 14 lenders for the 1.3 GWp project; EIB is an anchor institution. [S06] [S07]
Execution replaced pipeline size as the useful unit of comparison.
The timeline deliberately distinguishes financing, construction, market outcomes, consultations and funding applications. A project being authorised, financed, physically under construction or commercially operating are different evidence states and are not aggregated into one “pipeline” number.
Econergy added up to €160m of sponsor equity around its Romanian portfolio.
A Phoenix Financial investment is intended to capitalise a new UK holding structure consolidating 16 Romanian projects. Econergy says approximately 1.1 GW of Romanian solar, wind and BESS is under construction or ready for connection. The number is issuer-originated and mixes technologies and maturity states, so it is treated here as sponsor financing evidence rather than verified operating capacity. [S16]
ANRE moved prosumer multi-location settlement into a public hearing, not into final law.
ANRE held the announced debate on the methodology for multi-location commercialisation, billing, settlement and value allocation and on transitional quantitative compensation. The consultation remained open through 1 Oct, so these texts were still proposals at the research cut-off. The commercial issue is potentially significant for distributed generation because settlement design determines whether surplus value can follow a customer across multiple consumption points. [S09]
Corbii Mari showed hybridisation moving into civil and electrical execution.
Contractor AJ Brand describes the Nofar project at Corbii Mari as 281 MWp of solar plus 434 MWh of batteries and two high-voltage substations, including 400 kV infrastructure, with civil and electrical work running in parallel. This is stronger maturity evidence than a development announcement, but it remains a construction-stage project rather than operating capacity. [S13]
Negative quarter-hours returned on the next Sunday.
PZU base averaged 603.19 RON/MWh, but the unweighted 08:00–17:00 proxy collapsed to 93.71 RON/MWh. Twenty-one PT15 intervals were negative and the minimum was −16.01 RON/MWh. A similar event had occurred on Sunday 20 Sep. The recurrence weakens the argument that the earlier event was purely isolated, although two Sundays are still insufficient evidence for a stable seasonal regime. [S01] [S03]
The power system became long by gigawatts at noon while public storage demand accelerated.
At 12:00, retained observations show 2,804 MW solar and 2,783 MW wind against 4,487 MW of load, with Romania exporting a net 2,935 MW. The same delivery date set an OPCOM PZU volume benchmark of 62,552 MWh, which OPCOM described as the highest level of 2026 outside the winter maxima. Separately, AFIR opened €500m of solar-plus-storage support and €150m of storage support for public entities. [S01] [S02] [S05]
Dobrogea grid reinforcement became a concrete capacity project.
Rețele Electrice România announced approximately 86m lei including VAT for three 110 kV lines in Constanța and one in Tulcea. The company states that conductor and adjacent-equipment upgrades will approximately double the lines' transport capacity and permit higher renewable injections. This is direct evidence that network capacity, not only generation capacity, is now an investment category in the renewable build-out. [S08]
A 55 MW / 220 MWh BESS moved from construction evidence into governance uncertainty.
GEN-I ordered an audit of the Gheorgheni storage project, appointed interim management at GEN-I Sonce and resolved to assess whether the project should be put on hold while circumstances connected to an investigation are clarified. GEN-I states that it had not received official or detailed information on the proceedings at the time of its announcement. The underlying project had entered construction with a 55 MW / 220 MWh disclosed configuration. No conclusion about alleged wrongdoing is drawn here; the market implication is project-continuity and counterparty-governance risk. [S11] [S12]
AFIR's storage envelope was already over-requested within the first day.
AFIR reported 690 storage requests worth €241.44m and 760 solar/autoconsumption requests worth €213.51m by 16:00 on opening day. The storage amount was therefore approximately 161% of the dedicated €150m allocation, while the solar-plus-storage request value represented about 42.7% of its €500m envelope. Submission is not award and application value is not disbursement, but the signal on public-sector demand is unusually clear. [S04]
Construction depth broadened across utility solar.
China Huadian began construction at Coțofenii din Dos on a project reported at 154.7 MWac / approximately 180 MWp, with first-quarter 2028 commissioning targeted. Separately, Parapet reported construction progress at ENGIE's 215 MWp Cornățelu project. Both are construction signals rather than proof of future commissioning dates; nevertheless, they show that Romania's solar build-out is moving through physical procurement and civil works, not merely development-company announcements. [S14] [S15]
Dama Solar provided the strongest financing evidence of the week.
Rezolv Energy announced up to €561m of project financing for the 1.3 GWp Dama Solar project in Arad from a 14-lender consortium. Reuters independently reported the financing; EIB's project record shows Dama Solar approved on 28 Sep with approximately €100m of proposed EIB financing and an approximate total project cost of €728m. The sponsor says 520 MW is supported by two Romanian CfDs and that a corporate PPA covers additional output. Construction is the next stated phase, with operation targeted for H2 2028. [S06] [S07]
Alba Iulia turned grid digitalisation into an awarded infrastructure programme.
The EU public-procurement record shows a final contract value of approximately 476.07m lei excluding VAT for Transelectrica's 220/110/20 kV Alba Iulia digital-substation pilot, including digital twin, private cloud, GIS and photovoltaic components. The procurement record indicates the contract was concluded on 16 Sep and the award published on 24 Sep; the 30 Sep disclosure should therefore be read as an implementation/public-reporting milestone rather than a new contract date. [S20]
The constraint is no longer one-dimensional.
Price cannibalisation, national energy adequacy, local connection capacity, battery economics and project bankability are related, but they are not interchangeable. The most useful decision model separates them and asks where controllability has economic value.
Daylight value recovered week on week, but its distribution became the more important fact.
The seven-day PZU base mean fell from 946.00 RON/MWh in the previous briefing to 900.58 RON/MWh, a 4.8% reduction. In contrast, the unweighted 08:00–17:00 proxy rose from 465.17 to 520.32 RON/MWh, an 11.9% increase. Its ratio to the base therefore recovered from 49.2% to 57.8%. [S01] [S03]
That sounds constructive for PV until the daily distribution is inspected. Thursday and Friday were relatively strong: the daylight proxy captured 75.6% and 80.4% of the daily base. Saturday fell to 50.3%. Sunday collapsed to 15.5%, with 21 negative quarter-hours. Monday remained weak at 37.6%; Tuesday rebounded to 69.2%; Wednesday fell to 44.4%.
The negative-price cluster on 27 Sep is especially relevant because it followed a 20 Sep Sunday with 16 negative PT15 intervals. Repetition on consecutive Sundays provides stronger evidence of weekend oversupply risk than either event in isolation. It still does not justify assuming every future Sunday will behave the same way: weather, outages, cross-border flows, load and bidding conditions remain variable.
520.32 RON/MWh is useful. Calling it the national solar capture price would be analytically wrong.
A solar capture price weights each settlement interval by actual solar production in that same interval. The retained SolarIndustry.ro dataset intentionally stores the public operational feed and OPCOM outcomes as separate evidence layers, and it does not currently provide a fully harmonised national PT15 solar-generation series that can be matched to every market settlement interval without introducing interpolation or timing assumptions. [S01]
The 08:00–17:00 proxy instead answers a narrower question: “What did electricity cost on average during the clock hours most exposed to solar output?” It is valuable for identifying shape compression, comparing weeks and detecting downside regimes. It should not be inserted directly into a project revenue model.
Severe clock-hour compression exists.
27 Sep produced a 93.71 RON/MWh daylight proxy against a 603.19 RON/MWh base, while 21 quarter-hours cleared below zero.
The fleet-weighted revenue effect.
Without temporally matched national solar production, the precise capture price and capture rate cannot be stated to institutional standard.
For individual projects the better solution is project-specific: use metered or P50/P90 interval generation, actual node or bidding-zone settlement prices, curtailment assumptions, imbalance exposure, PPA/CfD terms and any storage dispatch constraints. National averages can frame the market; they cannot replace asset modelling.
30 Sep moved from 2.7 GW of export to 1.8 GW of import in five hours.
Retained Transelectrica observations provide the physical counterpart to the price signal. At 13:00 on 30 Sep, the system averaged 3,021 MW of reported solar, 6,713 MW of total production and only 4,028 MW of load, with net exchange at −2,685 MW: a large export position. By 18:00, solar had fallen to 295 MW, load had risen to 6,317 MW and net exchange had moved to +1,813 MW of imports. The exchange rotation was 4,498 MW in five hours. [S01]
Exhibit 02 / 30 Sep operating profile
Reported solar and net exchange · MW · negative exchange = exportYellow: reported solar. Grey: net exchange. Hover, tap or use left/right arrows. Values are local-hour arithmetic means of retained public operational observations; missing readings are not interpolated. [S01]
Net export with 3,021 MW solar and 4,028 MW load.
Net import with solar down to 295 MW and load at 6,317 MW.
Derived exchange swing. It measures system rotation, not an addressable battery market by itself.
The daily average hides both the deep midday export and evening import states.
View 30 Sep hourly system data used in exhibit
| Hour | Load MW | Solar MW | Exchange MW | Reported storage MW |
|---|---|---|---|---|
| 00 | 5,181 | 0 | +259 | 61 |
| 01 | 4,983 | 0 | +174 | 12 |
| 02 | 4,944 | 0 | +252 | 37 |
| 03 | 5,035 | 0 | +415 | 29 |
| 04 | 5,114 | 0 | +644 | 14 |
| 05 | 5,294 | 0 | +849 | 21 |
| 06 | 5,747 | 0 | +1,078 | 169 |
| 07 | 6,241 | 59 | +986 | 335 |
| 08 | 6,225 | 656 | +681 | 271 |
| 09 | 5,696 | 1,594 | −52 | 56 |
| 10 | 5,076 | 2,397 | −1,247 | 2 |
| 11 | 4,613 | 2,677 | −1,843 | 0 |
| 12 | 4,065 | 2,854 | −2,553 | 41 |
| 13 | 4,028 | 3,021 | −2,685 | 105 |
| 14 | 4,392 | 2,894 | −2,062 | 0 |
| 15 | 4,459 | 2,550 | −1,647 | 20 |
| 16 | 4,914 | 2,157 | −832 | 48 |
| 17 | 5,627 | 1,214 | +533 | 103 |
| 18 | 6,317 | 295 | +1,813 | 266 |
| 19 | 6,911 | 2 | +1,529 | 277 |
| 20 | 6,990 | 0 | +1,480 | 409 |
| 21 | 6,675 | 0 | +1,467 | 254 |
| 22 | 5,969 | 0 | +1,687 | 69 |
| 23 | 5,531 | 0 | +1,297 | 47 |
28 Sep showed the opposite extreme even more clearly. At 12:00, wind and solar together averaged 5,587 MW against 4,487 MW of load and accounted for approximately 75.3% of total reported production. Net export averaged 2,935 MW. A daily statement such as “Romania was a net importer” or “Romania had excess renewables” is therefore inadequate without an hour attached.
There are now repeated observable reasons to value flexibility, but none justifies multiplying a spread by MWh and calling the result profit.
Three distinct signals appeared this week. First, negative PT15 pricing returned on 27 Sep. Second, the system produced several-gigawatt intraday exchange reversals as solar output fell and load rose. Third, the retained operational feed on 30 Sep showed the reported storage category rising from low daytime values to 266 MW at 18:00 and 409 MW at 20:00. These are directionally consistent with the system needing flexible output later in the day. [S01]
27 Sep created low-price charging opportunities in the day-ahead market. Availability of cheap charging energy does not establish realised battery revenue.
30 Sep exchange rotation from 13:00 to 18:00. Batteries are one possible flexibility resource among imports, hydro, thermal generation, demand response and network actions.
Corbii Mari BESS disclosed in active hybrid construction. Physical execution matters more than nominal pipeline size. [S13]
A decision-grade BESS revenue case needs, at minimum, charging and discharge prices at settlement granularity, round-trip efficiency, usable energy, cycling constraints, degradation cost, availability, grid charges, imbalance exposure, trading/optimisation fees, collateral requirements, ancillary-service eligibility and realistic market-share assumptions. Charging may also be physically constrained precisely when wholesale prices are most attractive.
The stronger strategic question is therefore “Which value pools can this exact connection and asset access?” A standalone battery with strong grid rights, fast controls, qualified market access and an experienced optimiser may be better positioned than a larger nominal project constrained by connection or route-to-market limitations. Conversely, widespread battery entry can arbitrage away part of today's visible price shape; the strategy must survive its own success.
Dobrogea's 110 kV reinforcement is more commercially meaningful than another unqualified national pipeline total.
Rețele Electrice România's four-line modernisation programme is explicit about the problem it is intended to solve: doubling transport capacity on the targeted lines and enabling greater renewable injection. The project is approximately 86m lei including VAT, with more than 46.3m lei excluding VAT co-financed through the relevant development programme. [S08]
This matters because a solar project's economic capacity is not its module nameplate. The commercially relevant capacity is what can be injected, when, under its connection conditions and network constraints. As more PV reaches construction, connection studies, reinforcement schedules, curtailment exposure, substation maturity and transformer availability become more important due-diligence items.
Move the ATR and reinforcement schedule into the investment committee pack.
Require project-specific connection rights, milestones, export limits, reinforcement dependencies, commissioning prerequisites and change-of-control implications. “Grid secured” should never be accepted without the underlying document and scope.
Hybrid design should begin with the connection, not be appended later.
Battery duration, inverter sizing and operational rules should respond to export caps, curtailment, ramping and market access. A BESS can be a deliverability asset as well as a trading asset.
Transelectrica's Alba Iulia digital-substation pilot adds a different layer: operational observability and control. The approximately 476.07m lei awarded contract includes digital-twin, cloud and related digital infrastructure. It will not solve near-term renewable congestion nationally, but it illustrates where transmission-system modernisation is heading. [S20]
Solar abundance at noon did not remove Romania's dependence on firm supply and imports after sunset.
On 24 Sep, a Ministry of Energy state secretary said the next ten days did not offer conditions for restarting Cernavodă given low Danube flows and that imports remained the principal replacement source for the missing nuclear output. On 28 Sep, CNSU Decision 23/2026 extended the national state of alert for another 30 days, citing continued hydrological risk and providing for available generation, imports or consumption reduction as measures needed to protect system operation. [S18] [S19]
Retained operational data are consistent with that stress: the nuclear category remained at zero in the reviewed hourly records visible in the source layer, while daily average exchange was positive—net import—on six of the seven review days. The simple arithmetic mean of the seven daily exchange means was approximately +490 MW, lower than the prior week's approximately +1.02 GW but still an import position overall. Only 28 Sep had a negative daily average exchange balance. [S01]
Institutional lenders are willing to fund extraordinary solar scale when the revenue architecture is sufficiently structured.
Dama Solar is the week's most consequential project-finance event. Rezolv announced up to €561m from a 14-lender consortium for a 1.3 GWp solar project. EIB's project page records an approximately €100m proposed EIB contribution and approximate €728m project cost, while Reuters confirms the larger syndicated package. The sponsor says two CfDs cover 520 MW and that a corporate PPA provides another contracted revenue layer. [S06] [S07]
The inference should be precise. The deal demonstrates depth of lender appetite for Romanian renewable infrastructure and validates the ability of the CfD/PPA ecosystem to support very large projects. It does not demonstrate that lenders are comfortable extrapolating a 900 RON/MWh weekly PZU average into a long-term merchant revenue case.
Econergy's additional sponsor-equity arrangement reinforces the capital-availability signal from another angle. But capital abundance can increase competition for the genuinely scarce inputs: viable grid positions, bankable offtake, transformers, experienced EPC labour, commissioning resources and high-quality operating assets. [S16]
1.3 GWp / financed for construction
Up to €561m project-finance package · 14 lenders · 520 MW CfD-backed capacity plus disclosed corporate PPA.
Maturity: financing signed; construction next according to sponsor. Target operation H2 2028 is a forecast, not commissioned capacity.
281 MWp + 434 MWh BESS
Civil and electrical work described as progressing in parallel; two high-voltage substations disclosed.
Maturity: contractor-reported construction. Energisation remains future.
215 MWp / construction progress
Parapet disclosed approximately 340,000 modules and 23 central inverters within its ENGIE project scope.
Maturity: physical construction evidence; commissioning not inferred.
154.7 MWac / ~180 MWp
China Huadian construction start reported; approximately 240 GWh/year and Q1 2028 operation are project forecasts.
Maturity: construction reported by specialist press; lower evidence grade than regulator or lender documents.
Public-sector storage demand exceeded its dedicated AFIR budget before the first day was over.
AFIR opened two Modernisation Fund calls on 28 Sep: €500m for new solar generation with integrated storage for public entities and €150m for renewable-energy storage. By 16:00, AFIR reported 760 applications worth €213.51m under the production call and 690 applications worth €241.44m under storage. [S04] [S05]
Yellow = requests reported by AFIR. Grey outline = call allocation. Applications are not awards and may not all prove eligible. [S04] [S05]
€241.44m requested against a €150m storage envelope by 16:00 on opening day. AFIR said it was examining the possibility of additional allocations subject to available Modernisation Fund resources and the Ministry's decision.
For EPCs and equipment suppliers this is more useful than a general statement that “storage demand is strong.” It creates a near-term procurement cohort with defined public beneficiaries, funding rules and chronological evaluation. It also creates execution risk: many small or mid-sized public projects can compete simultaneously for engineering, PCS/BMS/EMS integration, transformers, permitting support and commissioning resources.
Tulcea and Făgăraș provide local examples of the same direction. Tulcea approved feasibility and technical-economic documentation around a roughly 9.92m lei storage project intended to work with its municipal PV asset, while Făgăraș approved a storage project of approximately 13.84m lei alongside a minimum 2 MW solar project. These are project-preparation signals, not operating batteries. [S21] [S22]
In C&I, Greenvolt Next reported a Romanian backlog of approximately 24 MW and described its local distributed-generation portfolio as PPA-led, with rising storage interest. Its stated 30 MW contracted figure refers to 2025 and should not be misread as a new 2026 capacity addition. Because these are vendor-reported figures, they are useful as demand texture rather than a market-size estimate. [S17]
Multi-location settlement is potentially material, but investors should not price draft rights into contracts yet.
ANRE's consultation concerns rules for commercialisation, billing, settlement and multi-location value allocation as well as transitional quantitative compensation. The immediate implication is procedural rather than economic: the final text and implementation mechanics need to be known before suppliers, aggregators or multi-site consumers can treat the new settlement structure as certain. [S09]
Separate media reporting described Parliamentary discussions around condominium rooftop PV and small balcony systems. No adopted rule is assumed in this briefing. Until a bill number, final legal text, technical connection rules, metering treatment and supplier obligations are reconciled, this remains a regulatory watch item rather than a bankable demand forecast. [S23]
What could make the central thesis wrong?
Weekend compression may remain episodic.
Two consecutive negative-price Sundays are stronger evidence than one, but autumn outages, weather and unusual system conditions may be amplifying the pattern. A durable structural claim requires a longer sample.
New batteries can destroy part of their own arbitrage opportunity.
If BESS deployment accelerates, charging demand can lift low-price intervals and discharge can reduce evening scarcity. Current gross spreads should not be capitalised indefinitely.
Cernavodă is distorting the observed balance.
Nuclear unavailability materially increases import need. A restart would change evening residual demand and could reduce some of the currently visible flexibility value without removing solar shape risk.
Funding demand is not project delivery.
AFIR applications can fail eligibility, procurement, permitting, connection or implementation. €241m of storage requests must not be converted into installed MWh without project-level evidence.
Expect continued shape volatility while adequacy remains weather and hydrology sensitive.
These are directional scenarios, not calibrated probabilities. The first delivery day is already partly observable through the 1 Oct PZU result; later outcomes depend on solar, wind, load, cross-border conditions, hydrology and unit availability. The forward framework therefore uses observable triggers rather than false probability precision.
Volatile daylight discount; evening imports remain relevant.
Solar continues to create low residual-load periods on favourable days while the loss of nuclear output and evening demand preserve scarcity after sunset. Day-ahead value should remain more sensitive to hour and weather than to the weekly base average.
- Assumption: Cernavodă is not materially restored in the first part of the window.
- Assumption: solar and wind alternate between surplus and tighter-output days.
- Expected implication: no stable one-directional merchant-price regime.
Monitor: daylight/base ratio, evening net import, wind output, Cernavodă status and PT15 negative counts.
Lower renewable coincidence lifts solar-hour relative pricing.
If wind is weaker during high-solar periods, load is firmer and cross-border export capacity remains available, the solar-window ratio could improve even without a higher base price. This would benefit unhedged PV capture relative to the reviewed Sunday regime.
- Trigger: daylight/base ratio sustainably above roughly 65% across several delivery days.
- Trigger: no new negative PT15 cluster.
- Trigger: midday export stays manageable despite high PV.
Decision response: do not relax downside underwriting; treat improvement as evidence to update the distribution, not to erase the downside tail.
High renewable coincidence deepens midday surplus while evening adequacy remains tight.
Strong simultaneous wind and solar can recreate low or negative daylight prices, while the post-sunset system can still rotate into high imports if firm generation remains unavailable. This is the most demanding configuration for unshifted merchant solar and the most visible gross value signal for flexibility.
- Trigger: daylight/base ratio below 25%.
- Trigger: another multi-hour cluster of negative PT15 prices.
- Trigger: evening net imports move above roughly 2 GW in retained observations.
Decision response: preserve curtailment and low-capture sensitivities; confirm storage charging rights and collateral before assuming arbitrage access.
The next decision should be different for each participant.
These are research conclusions from the observed market and project evidence. They are intended as decision-support directions, not personalised investment, legal, trading or technical advice.
Underwrite interval revenue and connection constraints before adding more merchant MW.
Build revenue cases from PT15 output and price profiles, not PZU base averages. Run explicit weekend and high-renewable coincidence cases. For hybrid projects, optimise battery design against the actual ATR/export envelope and PPA/CfD settlement mechanics before final equipment sizing.
Minimum test: can the project still meet its financing case if a meaningful share of weekend daylight resembles a 15–40% proxy/base regime rather than the weekly average?
Prioritise assets with controllable grid access and multiple qualified revenue channels.
The market supplies compelling timing signals, but a single-spread merchant thesis is vulnerable to saturation. Secure charging/discharging rights, optimiser/PRE capability, ancillary qualification, warranty-compatible dispatch rules and contractual treatment of imbalance and collateral.
Minimum test: remove 30% of assumed arbitrage gross value and verify that the project can still service fixed costs through the remaining credible stack.
Use a maturity-adjusted project ledger rather than a MW league table.
Dama shows that large Romanian PV can attract deep institutional financing when revenues are structured. Gheorgheni shows that non-technical governance events can interrupt even construction-stage storage. Investment committees should score connection, offtake, EPC, permits, governance, commissioning and route-to-market separately.
Minimum test: distinguish “announced”, “authorised”, “financed”, “under construction”, “energised” and “commercially operating” in every capacity table.
Prepare for a public-sector procurement cluster, but do not build inventory against applications alone.
Storage requests exceeded the dedicated AFIR allocation on opening day. That can create a substantial bid and delivery pipeline, but funding approval, procurement and connection still sit between application and equipment order. Secure framework pricing and engineering capacity without treating 1,450 applications as confirmed projects.
Minimum test: track funded award, tender publication, contract award and notice-to-proceed as separate conversion stages.
Value solar-plus-storage against the site's load curve, not the national wholesale average.
Behind-the-meter economics begin with avoided purchases, demand timing, production continuity and contract structure. A national daylight discount can strengthen storage logic but does not determine a factory's optimal battery size. Obtain interval consumption data before accepting generic “solar + battery” sizing.
Minimum test: model self-consumption, peak demand, outage value and PPA terms separately from merchant export revenue.
Treat flexibility and reinforcement as one infrastructure portfolio.
The reviewed system can be long by almost 3 GW at noon and import heavily only hours later. Reinforcement, storage, demand response and controllable generation therefore solve different parts of the same timing-and-location problem. Funding programmes should preserve evidence of grid impact rather than rewarding nameplate capacity alone.
Minimum test: require each supported project to state the relevant connection bottleneck, expected operating profile and measurable system or self-consumption benefit.
Re-price the downside.
- Refresh PV price-shape sensitivities with September PT15 evidence.
- Audit project connection documentation and export constraints.
- Reconcile BESS revenue models with actual optimiser and grid terms.
- Track AFIR applications through eligibility and procurement.
Convert pipeline into maturity evidence.
- Track construction-to-energisation conversion by project.
- Monitor actual battery commissioning, not announced MWh.
- Map transformer, PCS and grid-equipment lead times.
- Update weekend capture-risk distributions as autumn data accumulate.
Change the thesis when the data change.
- Multiple weeks with daylight/base above 70% would weaken the near-term cannibalisation thesis.
- Frequent sub-25% weekends would strengthen it materially.
- Cernavodă restart would reduce one current adequacy distortion.
- Rapid operating BESS additions should trigger lower long-run arbitrage-spread assumptions.
Evidence first. Unknowns remain visible.
Material market values were reconstructed from retained public operational and OPCOM evidence. Company announcements are used for their own projects and claims, not as independent proof of national market size. Secondary sources are retained where no stronger directly accessible source was identified and are labelled accordingly.
How the key derived numbers were built.
- 7-day PZU base: arithmetic mean of OPCOM delivery-day base prices for 24–30 Sep.
- 08:00–17:00 proxy: arithmetic mean of each day's unweighted retained PT15 daylight proxy; not production weighted.
- Daylight/base: weekly daylight proxy divided by weekly base mean.
- Cleared volume: seven-day sum of retained OPCOM delivery-day volume.
- Daily system fields: arithmetic means of all stored Transelectrica readings for each local day; sample counts differ and missing observations are not filled.
- Exchange sign: positive = net import into Romania; negative = net export.
- 30 Sep reversal: +1,813 MW at 18:00 minus −2,685 MW at 13:00 = +4,498 MW.
- AFIR storage pressure: €241.436622m requested / €150m allocation = approximately 161%.
What this report does not claim.
- No national solar capture price is stated because a fully harmonised PT15 national solar-generation series was not available in the retained public dataset.
- No battery EBITDA is inferred from PZU spreads.
- No announced, authorised or financed project is counted as operating unless commissioning evidence supports that classification.
- Transelectrica's reported storage category is used as an operational reference, not as a complete inventory of Romanian battery dispatch.
- Project output forecasts and commissioning dates remain sponsor or media forecasts until realised.
- AFIR applications are demand evidence, not awards or installed capacity.
- The two consecutive negative-price Sundays are a material signal but still a short sample for structural seasonal forecasting.
- The current Cernavodă outage changes market conditions; post-restart price and import behaviour may differ materially.
| ID | Organisation / source | Use in report | Evidence class | Link |
|---|---|---|---|---|
| S01 | Solar Industry Romania / retained public data history | Transelectrica hourly observations, daily system means, OPCOM PZU base/daylight/low/negative intervals/volume; review and 1 Oct forward indicator. | Retained primary-source evidence + transparent derivation | Open source |
| S02 | OPCOM | 28 Sep PZU record: 62,552 MWh traded; official market-operator context. | Primary / market operator | Open source |
| S03 | Solar Industry Romania / previous weekly briefing | 17–23 Sep comparison baseline and continuity control to avoid repeating prior developments as new. | Prior research product / primary-source synthesis | Open source |
| S04 | AFIR | First-day public-entity applications: 760 production requests / €213.51m and 690 storage requests / €241.44m. | Primary / funding agency | Open source |
| S05 | AFIR | Call timetable, first-submitted/first-evaluated process and €500m production / €150m storage allocations. | Primary / funding agency | Open source |
| S06 | European Investment Bank | Dama Solar approval, approximately €100m proposed EIB finance and approximately €728m total project cost. | Primary / institutional lender | Open source |
| S07 | Reuters / Rezolv Energy announcement | Up to €561m Dama financing, 14-lender consortium, 1.3 GWp project and financing structure; sponsor disclosure adds 520 MW CfD and PPA detail. | High-quality secondary + issuer disclosure | Open source |
| S08 | Rețele Electrice România | 86m lei Constanța/Tulcea 110 kV reinforcement, four lines, stated doubling of transport capacity and renewable-integration purpose. | Primary / DSO disclosure | Open source |
| S09 | ANRE | Prosumer multi-location and transitional quantitative-compensation public-hearing notice; consultation status through 1 Oct. | Primary / regulator | Open source |
| S10 | ANRE statement reported by News.ro | Exact 22 Sep establishment-authorisation breakdown: 66.247 MW PV, 132.65 MW added storage, 394.568 MW standalone storage. | Regulator-originated / secondary carrier | Open source |
| S11 | GEN-I | Gheorgheni project audit, interim management and assessment of possible project hold. | Primary / issuer | Open source |
| S12 | GEN-I | Gheorgheni project baseline: 55 MW / 220 MWh, EPC agreement and construction commencement. | Primary / issuer | Open source |
| S13 | AJ Brand | Corbii Mari project reference: 281 MW solar, 434 MWh storage; construction context corroborated through contractor disclosure. | Primary commercial / contractor | Open source |
| S14 | Parapet / InvesTenergy | Cornățelu 215 MWp construction progress and disclosed equipment quantities. | Contractor-originated / specialist press | Open source |
| S15 | Profit.ro | China Huadian Coțofenii din Dos construction start; 154.7 MWac / approximately 180 MWp and project forecast. | Specialist secondary reporting | Open source |
| S16 | Econergy / Energy Global | Phoenix Financial equity commitment up to €160m, 16-project holding structure and company-reported Romanian portfolio status. | Issuer-originated / industry publication | Open source |
| S17 | Greenvolt Next / Agenda Construcțiilor | Company-reported C&I backlog, historical 2025 contracted capacity and PPA/storage demand commentary. | Management interview / self-reported | Open source |
| S18 | CNSU Decision 23/2026 / Monitorul Oficial text mirror | Extension of state of alert from 29 Sep, hydrological-risk basis and measures available to protect electricity-system operation. | Official legal text reproduced by third party | Open source |
| S19 | Ministry of Energy state-secretary statement / News.ro | 24 Sep statement on Cernavodă restart conditions, Danube flow and imports as principal replacement source. | Official statement via press | Open source |
| S20 | Publications Office of the European Union | Alba Iulia 220/110/20 kV digital-substation procurement, final contract value and technical scope. | Primary / public procurement | Open source |
| S21 | Radio România Constanța | Tulcea municipal storage feasibility decision and approximately 9.92m lei project value. | Public broadcaster / secondary local | Open source |
| S22 | Salut Făgăraș | Făgăraș municipal storage feasibility approval and minimum 2 MW PV project context. | Local secondary reporting | Open source |
| S23 | Știrile ProTV | Reported Parliamentary discussion of condominium and balcony PV proposals; used only as watchlist evidence, not adopted law. | Secondary / legislative watch | Open source |
Research cut-off: 30 Sep 2026, 23:59 EEST. Prices are market outcomes, not forecasts. Operational values are retained public observations and may differ in sample density by hour/day. Company project capacities and schedules remain company disclosures unless separately verified. This report is independent market research for decision support and does not constitute investment, trading, legal, tax or engineering advice.