Co-located BESS merits selective expansion. Not a blanket curtailment hedge.
Romania now presents a credible structural case for solar-plus-storage: photovoltaic penetration is reshaping short-term prices, storage deployment is accelerating and Transelectrica identifies binding renewable-integration constraints in key regions. The investment case is nevertheless narrower than the headline narrative. Public evidence proves grid pressure and solar capture-price risk; it does not yet quantify national or regional solar curtailment at a level sufficient to underwrite avoided-curtailment revenue. The recommended strategy is therefore node-led hybridisation with site-level investment gates, not fleet-wide BESS deployment.
Approve development-stage expansion. Gate construction capital site by site.
The decision is not whether batteries are strategically relevant to Romania. They are. The decision is whether ownership of co-located storage can produce a sufficiently defensible return after connection constraints, degradation, market competition, charging costs and merchant-revenue compression are recognised.
Board action
Approve origination, engineering, interconnection, legal and market-model work through a pre-FID stage. Construction capital returns to the Investment Risk Committee asset by asset.
Confidence
Medium-high on the strategic direction; low-to-medium on asset economics until site telemetry, connection terms, capex, degradation and contractual revenue data are available.
Principal risk
The market can support storage while individual co-located projects still fail their hurdle rate because value is captured by competing standalone BESS, grid reinforcement or connection-contract restrictions.
Storage is already material, but the official stock still describes an early-stage market.
Transelectrica's Q1 2026 report is the latest official TSO installed-capacity stock identified at the research cut-off. The H1 2026 report is scheduled for 14 Aug 2026, two days after this study's cut-off. [S01] [S09]
Four value pools exist. Only one is immediately visible in public market data.
An investment model should not collapse all storage value into a single “arbitrage revenue” line. The four pools have different evidence strength, bankability and site dependence.
Captured-price uplift
Romania's short-term market shows pronounced intraday shape, while Transelectrica explicitly links rising solar/wind output to more very-low and negative-price intervals. The signal is real; project value still requires dispatch simulation against the site's actual solar profile. [S01] [S04]
Avoided solar curtailment
Regional integration constraints are documented, but current public evidence located does not quantify recurring 2025–2026 solar curtailment by node or asset. Treat this value as zero until dispatch instructions, export caps or site telemetry prove it. [S01] [S03]
Balancing and ancillary value
JRC documents storage participation in Romanian balancing/congestion-management routes and an expected larger storage role. Qualification, delivered revenue, availability and saturation must be evidenced for each project rather than inferred from technical eligibility. [S03]
Connection-capacity value
In constrained nodes, an existing site, land position, transformer bay or connection right can be economically valuable. The benefit is real only to the extent that the relevant ATR and network studies allow charging and discharging without a disproportionate reinforcement requirement. [S01] [S05]
The curve is storage-shaped. The observed spread is not battery margin.
OPCOM's DAM page retrieved on 12 Aug 2026 showed a deep middle-of-day trough followed by a sharp evening rise in its then-selected 60-minute Romanian price series. The exhibit is a one-day market-shape observation, not a revenue forecast. [S04]
Price falls to €32.83/MWh at interval 13 and reaches €300/MWh at interval 20.
Unit: EUR/MWh / 24 hourly intervals / OPCOM page retrieved 12 Aug 2026
Source: OPCOM ROPEX DAM. Geography: Romania. Observation: dynamic page retrieved 12 Aug 2026. Interpretation: intraday timing value is visible, but a storage model must deduct round-trip losses, solar charging opportunity cost, degradation, availability constraints, market/optimizer fees, balancing exposure and applicable charges. [S04]
The grid problem is geographically real, but connection scarcity should be valued node by node.
Transelectrica's Q1 2026 report explicitly identifies Dobrogea and Banat as already congested for integrating additional renewable generation, while investments in progress or planned are expected to make approximately 5,000 MW of additional integration capacity available across those two regions by 2027. [S01]
Current integration constraint is explicit.
Reported fact: Transelectrica describes Dobrogea as already congested and unable to accommodate new capacity under existing grid conditions. [S01]
Decision implication: existing connection position and hybridisation optionality deserve higher diligence here than a national-average BESS model.
The same connection-headroom constraint is reported.
Reported fact: Banat is included alongside Dobrogea in Transelectrica's current congestion statement. [S01]
Decision implication: a project should establish whether storage improves utilisation of an existing connection or instead triggers a new import/export requirement.
Reinforcement projects show where pressure is expected to propagate.
Transelectrica's 2024–2033 development plan identifies overload-driven reconductoring and flow-control requirements including Brașov–Gutinaș and București Sud–Fundeni, with high renewable output in the south-east and southern PV contributing to studied overloads. [S02]
Not proven: these planning studies do not establish the future curtailment hours of any specific solar plant.
Transelectrica's reference scenario embeds substantially more solar through 2033.
The grid development plan's reference scenario is based on the then-draft updated NECP. It is a planning input, not a commissioned-project pipeline or a SolarIndustry forecast. [S02]
Solar rises from 2.90 GW in 2023 to 9.70 GW in the 2033 planning reference case.
Unit: GW / selected planning years / Transelectrica 2024–2033 RET development plan
Data summary: 2023 wind 2.97 GW / solar 2.90 GW; 2024 3.05 / 3.90; 2025 3.70 / 4.70; 2028 5.50 / 6.90; 2030 7.00 / 8.30; 2033 7.30 / 9.70. Source: Transelectrica RET Development Plan 2024–2033, reference scenario based on draft updated PNIESC. [S02]
Romania's storage policy signals are supportive, but they are not one coherent project-sizing instruction.
The JRC records both an NECP battery target and a larger storage-needs assessment referenced by Romania's National Energy Strategy. These constructs use different purposes and should not be merged into a single “required BESS market size”. [S03]
Policy target
At least 1,200 MW / 2,400 MWh of battery storage by 2030 is reported by JRC from Romania's final updated NECP, alongside 2,000 MW by 2035 and 4,500 MW by 2040. [S03]
This is a national policy target. It does not specify which share should be co-located, standalone, merchant, contracted or located in congestion zones.
System-needs reference
Romania's National Energy Strategy is reported by JRC as referring to a Transelectrica assessment of 2–4 GW of storage needs for 2025–2030 with an average duration of eight hours. [S03]
The larger need assessment is not equivalent to the NECP battery target and does not justify specifying an eight-hour lithium-ion project at any individual site without dispatch and system studies.
Storage economics improved on network charging while connection commitments became more capital-intensive.
Regulation is moving in two directions at once: stored electricity reinjected into the grid receives relief from specified regulated charges, while connection and establishment guarantees have become more demanding. [S05] [S06]
| Measure | Reported requirement | Decision implication | Evidence class |
|---|---|---|---|
| ATR connection guarantee | 20% of connection tariff excluding VAT for production / consumption-and-production sites with approved export power above 1 MW. | Development-stage cash and guarantee capacity become material; co-location does not automatically make connection changes capital-light. | ANRE / 21 May 2026 [S05] |
| 2026 capacity-allocation participation guarantee | €20,000/MW of requested allocation capacity. | Connection strategy can create meaningful pre-construction capital exposure. | ANRE / 21 May 2026 [S05] |
| Establishment authorization guarantee | €30/kW of installed power, held through works reception; for hybrid/repowering projects with increased installed power, ANRE states the guarantee is calculated on the incremental power versus the original ATR. | Hybridisation can preserve some value from the original site configuration, but the exact updated ATR remains decisive. | ANRE / 21 May 2026 [S05] |
| Stored-energy network charges | ANRE exempts stored and subsequently reinjected electricity from specified transmission extraction, distribution and system-service tariffs and green-certificate obligations. | Removes a material double-charging problem from storage cycling economics. | ANRE / 8 Jul 2025 [S06] |
| Own use and technological losses | Network tariffs remain applicable to storage own consumption and technological losses. | Do not model charging and conversion losses as tariff-free by default. | ANRE / 8 Jul 2025 [S06] |
Subsidised standalone BESS is a competitive risk to merchant revenue pools.
The European Commission authorised Romanian state aid SA.121308 in March 2026 for stand-alone battery storage financed from the Modernisation Fund. The scheme has an overall budget of RON 764.295 million and runs to 31 Dec 2030. [S07]
Scope discipline
The cited European Commission authorisation is explicitly for stand-alone battery storage. This report does not assume that co-located projects qualify.
Second-order effect
A successful national BESS build-out can compress the very day-ahead and ancillary spreads that attract new capital. Long-run underwriting must therefore include market cannibalisation.
Research gap
This report does not rely on unverified assumptions about subsequent grant awards or implementation status beyond the cited Commission authorisation.
No responsible project IRR can be produced from public market data alone.
Project capex, augmentation, degradation, warranty terms, connection cost, dispatch restrictions, financing, tax, site-specific solar production and commercial contracts were not supplied. An invented NPV or IRR would create false precision.
Avoided spill or curtailment
Model only the MWh demonstrably unavailable for sale or exported below an enforceable limit. Base-case treatment before site proof: zero.
Captured-price uplift
Simulate PV and BESS dispatch interval by interval against actual market settlement prices. Use foregone solar sale value as a charging cost where applicable.
Balancing / ancillary net revenue
Include only revenues compatible with qualification, availability, SoC constraints and stacking rules. Deduct optimizer and imbalance exposure.
Connection and infrastructure value
Quantify avoided network works, shared substation/land/SCADA costs and schedule value only where engineering and the ATR substantiate the saving.
Availability, tolling or floor
Recognise only when a credible counterparty term sheet or contract defines payment, performance, indexation, penalties and dispatch control.
Do not stop at round-trip efficiency
Capex, augmentation, fixed O&M, degradation, losses, insurance, market access, guarantees, financing, tax, land, EMS/SCADA, cyber/fire compliance and connection works belong in the model.
The formula deliberately does not convert the €267.17/MWh single-day OPCOM max-minus-min range into margin. A battery cannot assume perfect charging/discharging at the extrema, zero losses, zero degradation, unlimited cycle availability or perfect foresight.
Six gates should separate strategic approval from construction FID.
These are proposed Investment Risk Committee governance rules, not regulatory requirements. Their purpose is to prevent a structurally attractive market from becoming an excuse for weak asset underwriting.
One full seasonal operating dataset.
Obtain at least 12 months of interval-level solar output, export, curtailment/dispatch flags, captured prices and outages where available. The 12-month period is an analyst governance recommendation to span seasonality.
ATR and connection contract survive legal/technical review.
Confirm import, export, simultaneous operating limits, protection, metering, reinforcement works, milestones, guarantees and treatment of modified installed power. [S05]
Do not rely on one merchant value pool.
Preferred governance standard: at least two economically meaningful value pools should survive diligence, with at least one sufficiently observable or contractable to support downside underwriting.
Duration follows use case.
Do not select a two-, three- or four-hour system because it is a market convention. Power, energy and cycling capability should follow the binding dispatch requirement and warranty economics.
Return clears the company's own risk-adjusted hurdle.
No hurdle rate was supplied. The IC should reject a project that requires an artificially low discount rate, unsupported terminal value or optimistic merchant extrapolation to pass.
Guarantees and liquidity remain survivable.
Model connection and authorization guarantees, supplier payment milestones, construction contingencies and working-capital timing alongside economic return. [S05]
Selective co-location dominates a blanket rollout because it preserves option value.
Qualitative decision matrix below is analyst judgement based on the evidence in this report. It is not a substitute for asset-level financial returns.
The preferred target is not “a solar project”. It is a solar project with monetisable timing and connection asymmetry.
Screening should rank sites before engaging in full technical design. The strongest projects combine measurable exposure with an existing asset position that reduces incremental execution risk.
| Screen | Preferred evidence | Positive signal | Reject / deprioritise when |
|---|---|---|---|
| Node / grid | ATR, network studies, operator correspondence, export-limit history | Scarce connection position or recurring operating constraint with workable hybrid terms | Battery requires disproportionate reinforcement or loses connection advantage |
| Solar economics | Interval generation and realised captured price | Persistent discount versus benchmark or demonstrable spill | Solar capture remains strong and curtailment is immaterial |
| Infrastructure | Single-line diagram, substation, protection, land, SCADA | Material reusable infrastructure and space | Retrofit complexity removes co-location cost/schedule benefit |
| Revenue diversity | Optimizer bids, offtake/tolling term sheets, qualification assessment | Multiple independent value pools | Project return rests on one volatile merchant stream |
| Contract control | PPA/CfD, route-to-market, connection and land agreements | Owner controls dispatch economics and has sufficient asset tenure | Counterparty retains value while project owner bears degradation or imbalance risk |
Poland is a trajectory warning, not a Romanian curtailment proxy.
PSE operates a formal process for non-market redispatch of renewables when market resources are insufficient for system balancing and publishes compensation procedures. That demonstrates how curtailment risk can mature operationally in a high-renewables CEE market. [S08]
The relevant lesson is contractual.
PSE states that compensation for non-market redispatch depends on the applicable circumstances and documentation. It also states that an RES owner is not entitled to compensation where its connection agreement contains specified non-guaranteed-supply provisions related to system balancing. [S08]
Implication for Romania
Do not forecast Polish curtailment volumes into Romania. Instead, treat the comparator as evidence that connection-contract wording and compensation architecture can determine who economically bears curtailment once the operational problem becomes material.
The preferred strategy survives more futures than a pure arbitrage thesis.
Scenarios are directional analyst constructs. No probabilities are assigned because the available evidence does not support calibrated likelihoods.
Renewables and BESS grow while grid reinforcement advances unevenly.
Solar timing pressure remains commercially relevant. Storage revenues gradually become more competitive. Strong projects combine capture-price value, connection advantage and additional revenue streams rather than relying on one spread.
Renewable build-out outruns local grid reinforcement.
Selected sites develop recurring export limits, increasingly weak captured prices or demonstrable spill while evening/flexibility value remains robust. Existing connection positions become more valuable.
Grid and standalone BESS scale faster than scarcity value.
Reinforcements release connection headroom and subsidised/merchant storage compresses intraday and ancillary opportunities. Co-location works only at unusually advantaged sites or with contracted floors.
Connection or market design changes the value stack.
Flexible connection agreements, compensation rules, market redesign, support-scheme changes or different contracting structures alter the allocation of congestion and dispatch risk. Re-underwrite rather than extrapolate historic spreads.
Assume the expansion failed in 2029. These are the most plausible reasons.
Downside is prioritised over upside because BESS development combines irreversible construction capital with merchant and regulatory variables that the owner does not fully control.
Failure modes and early indicators
Early indicator: declining realised cycling value and ancillary clearing value as competing storage commissions. Response: stop treating historic spreads as forward margins; contract a floor where economically rational.
Early indicator: site export remains unconstrained through high-solar seasons. Kill trigger: the project requires undocumented avoided-curtailment revenue to clear the hurdle rate.
Early indicator: reinforcement scope, guarantees or import/export restrictions expand during ATR work. Kill trigger: incremental grid works erase the shared-infrastructure and schedule benefit.
Early indicator: optimizer dispatch consumes warranty throughput faster than revenue compensates. Response: align optimizer incentives with net asset contribution, not gross traded value.
Strongest case against the recommendation
Do not expand into owned co-located storage if the company's candidate sites show little or no export constraint, only modest capture-price discount, expensive connection modifications and no contractable revenue floor, while standalone BESS deployment continues to compress merchant flexibility value. Under those conditions, the rational strategy is to remain storage-ready, use contractual flexibility where available and preserve capital for later entry.
Build the evidence before building the batteries.
The programme should deliberately convert a partially irreversible capital decision into a sequence of reversible evidence gates.
Create the site-level data room.
Collect interval production, realised prices, dispatch/export-limit records, ATRs, connection contracts, single-line diagrams, land rights, grid studies, PPA/CfD terms and outage history. Rank missing evidence by decision impact.
Shortlist and model the highest-exposure nodes.
Run interval dispatch simulations, connection/legal review and preliminary system sizing. Request binding or budgetary EPC/BESS and optimizer inputs with explicit degradation and augmentation assumptions.
Convert market value into investable terms.
Negotiate route-to-market, tolling/floor or optimizer terms where available. Complete network studies and define the exact guarantee, reinforcement and commissioning cash profile. Return only qualified sites to the IC.
Scale only after the first assets prove net contribution.
Track realised captured-price uplift, curtailment recovery, ancillary contribution, degradation cost, availability and connection performance against underwriting. New projects use observed portfolio evidence rather than the original market thesis.
The programme should be governed by evidence that can falsify the thesis.
These KPIs are intended to tell the Board early when the expansion case is strengthening or weakening.
| KPI | Measure | Why it matters | Decision use |
|---|---|---|---|
| Solar capture discount | Site realised €/MWh versus defined market benchmark | Measures timing erosion directly. | Higher persistent discount strengthens shifting value. |
| Verified constrained energy | MWh with documented export limitation / dispatch instruction | Separates curtailment fact from congestion narrative. | No recurring constrained MWh weakens the hedge thesis. |
| Net BESS contribution | Revenue less energy cost, losses, degradation, fees and variable O&M | Prevents gross trading revenue from masking asset cost. | Primary operating-economic KPI after commissioning. |
| Revenue concentration | Share of net contribution from largest value pool | Identifies dependence on a saturating merchant market. | High concentration requires stronger stress testing. |
| Connection capital | Guarantees + reinforcement + connection works per project | Captures liquidity as well as economic cost. | Unexpected escalation can stop FID. |
| Warranty throughput consumption | Actual cycle/energy throughput versus underwritten degradation path | Links dispatch strategy to physical asset life. | Optimizer incentives adjusted if throughput outruns contribution. |
The recommendation is stronger than the available project economics, so the two are kept separate.
Research prioritised Transelectrica, ANRE, OPCOM, European Commission/JRC and PSE evidence. Public statements, derived calculations, analyst judgement and unknowns are explicitly distinguished.
What is established
Romanian storage capacity is material and growing; short-term solar timing pressure is observable; Dobrogea and Banat face stated renewable-integration congestion; grid reinforcements are planned; storage receives specified network-charge relief; connection guarantees have tightened; and a large standalone-BESS support scheme has been authorised.
What remains unknown
No decision-grade public dataset was identified here for current solar curtailment MWh by Romanian node/site; no user project capex, WACC/hurdle, financing, degradation curve, connection scope, optimizer terms or site telemetry were supplied. Those unknowns prevent a responsible project NPV/IRR.
Source-date limitation
Research cut-off is 12 Aug 2026. Transelectrica's H1 2026 report is scheduled for 14 Aug; therefore 1 Apr 2026 is used as the latest official TSO installed-storage stock identified at cut-off. [S09]
OPCOM reproducibility limitation
The cited OPCOM graph page is dynamic rather than a fixed dated permalink. The plotted 24-point 60-minute series and reported indices are the values observed on retrieval. The exhibit should be archived internally before formal IC circulation.
-
S01
Transelectrica — Q1 2026 report
Tier A / published 28 May 2026 / Romania / latest official TSO quarterly stock used.
Reported 599 MW / 1,129.7 MWh storage as of 1 Apr 2026; 3,339 MW gross PV; 3,616 MW prosumer capacity as of 1 Mar; short-term low/negative-price observation; Dobrogea/Banat congestion and approximately 5,000 MW additional integration headroom expected by 2027. -
S02
Transelectrica — RET Development Plan 2024–2033
Tier A / network-planning document / scenario values are planning inputs, not commissioned pipeline.
Reference wind/solar scenario and network studies identifying south-east transfer overloads, reconductoring and flow-control requirements. -
S03
European Commission Joint Research Centre — Implementation of the Commission Recommendation on Energy Storage
Tier A/B institutional evidence / 2026 publication; some regulatory observations use information current to April 2025 and are dated accordingly.
Romanian storage policy, NECP targets, referenced TSO needs, ancillary/congestion-management framework, historical 2023 redispatch observation and support instruments. -
S04
OPCOM — ROPEX DAM price and traded-volume report
Tier A market operator / retrieved 12 Aug 2026 / dynamic page; archive recommended.
60-minute Romanian DAM series observed on retrieval: minimum €32.83/MWh, maximum €300/MWh, base index €137.91/MWh and base volume 39,261.1 MWh. -
S05
ANRE — connection and licensing amendments, 21 May 2026
Tier A regulator / Romania / current rule-change announcement cited directly.
20% ATR-related connection guarantee for relevant sites; €20,000/MW 2026 allocation guarantee; €30/kW establishment-authorization guarantee; hybrid incremental-power treatment. -
S06
ANRE — stored-energy regulated-tariff exemption, 8 Jul 2025
Tier A regulator / Romania.
Specified network/system tariff and green-certificate exemptions for stored electricity subsequently reinjected; own consumption and technological losses remain subject to network tariffs. -
S07
European Commission / Official Journal — State aid SA.121308
Tier A / decision adopted 6 Mar 2026 / Official Journal publication 1 Apr 2026.
Romanian Modernisation Fund scheme for stand-alone battery storage; direct grant; RON 764,295,000 overall budget; duration through 31 Dec 2030. -
S08
PSE — non-market redispatch of renewable generation
Tier A CEE comparator / Poland / used qualitatively, not as a quantitative Romanian proxy.
Polish TSO operating process, system-balancing rationale, compensation documentation and connection-contract condition relevant to compensation eligibility. -
S09
Transelectrica — Financial Calendar
Tier A / retrieval 12 Aug 2026 / used to define latest-available-data limitation.
H1 2026 report scheduled for 14 Aug 2026.