SolarIndustry.ro / Case Study / Investment & Project Finance
Romanian utility-scale solar: bankability is shifting from capex efficiency to delivery and revenue quality.
Romania's solar market is not becoming unbankable; it is becoming more selective. The evidence now requires investors and lenders to distinguish a low-cost PV plant from a financeable power asset. Grid deliverability, project-weighted capture price, route-to-market structure, balancing allocation, counterparty quality and storage optionality increasingly determine debt capacity and equity resilience. A higher-capex project with firmer grid rights and better-shaped contracted cash flow can therefore be economically superior to a cheaper project exposed to an uncertain connection and an undifferentiated merchant-price assumption.
The asset class remains financeable. The generic underwriting approach does not.
Romanian PV has moved into a phase where revenue timing and deliverability deserve at least as much diligence as irradiation, EPC cost and nominal connection capacity.
Merchant exposure is no longer one variable.
Base-price level, solar shape, nodal or contractual basis, negative-price exposure, forecast error, imbalance settlement, curtailment treatment and route-to-market fees should be modelled separately. Compressing them into one “merchant discount” obscures where cash flow actually fails.
The connection queue is large; the advanced subset is much smaller.
ANRE reported 113,652 MW of maximum approved export power across 1,672 projects with valid ATRs at 1 July 2026. Only 11,153 MW sat at the combined connection-contract, construction-authorisation and establishment-authorisation stage: 9.8% of the ATR MW denominator. That ratio measures administrative progression, not probability of commissioning. [S02]
Development optionality now consumes more financial capacity.
ANRE's May 2026 reforms increased the financial guarantee associated with ATR issuance for relevant projects above 1 MW from 5% to 20% of the connection tariff excluding VAT, introduced an establishment-authorisation guarantee of €30/kW and specified a €20,000/MW capacity-allocation-auction guarantee for 2026. [S03]
Storage is valuable where it changes a specific exposure.
Romania already reported 924.43 MW / 1,762.50 MWh of storage by 1 July 2026. That increases the strategic relevance of BESS, but it also means a merchant-storage case must anticipate more competing flexibility. Avoided curtailment, capture uplift and ancillary-market revenue should each be evidenced independently. [S01]
The market is scaling faster than a single “installed solar” number can describe.
The finance question is not whether Romania has room for more PV. It is which projects can convert a growing market into deliverable, contractable and financeable MWh.
Gross photovoltaic capacity in Transelectrica's national generation-park breakdown; 3.737 GW net. This population is presented separately from prosumers. [S01]
Installed prosumer power reported separately by Transelectrica. It should not be mechanically added to the generation-park PV figure without reconciling source definitions. [S01]
Installed storage power, paired with 1,762.50 MWh reported energy capacity. The implied fleet E/P ratio is approximately 1.91 hours; this is descriptive, not a recommended BESS duration. [S01]
Maximum approved export power represented by renewable projects with valid ATRs. It is a connection-process stock, not an executable construction pipeline. [S02]
The first underwriting error is mixing stocks.
The generation fleet, prosumer population, ATR queue, construction-authorised pipeline and advanced authorised pipeline answer different questions. A lender evaluating a 150 MW project needs to know what capacity can compete for the same grid corridor and market hours, but it should not treat every ATR MW as a future commissioned MW. Likewise, distributed prosumer capacity can affect residual daytime demand without being interchangeable with transmission-connected utility assets.
That distinction matters because solar's economic externality is increasingly temporal. Transelectrica explicitly linked the growing share of solar and wind production in H1 2026 with a greater share of short-term-market intervals exhibiting very low and even negative prices. [S01] The relevant risk is therefore not simply “more solar capacity”; it is correlated generation entering overlapping intervals faster than demand, interconnection, flexibility and storage can absorb it.
The daytime price problem is visible before a project-specific capture model is built.
SolarIndustry's retained OPCOM evidence gives a screening signal. It must not be confused with a solar capture price because the 08:00–17:00 proxy is unweighted by actual project generation.
What this evidence establishes—and what it does not.
Observed: daytime PZU prices can diverge materially from the all-day base price, and negative 15-minute intervals occurred during the selected 29-day period. Transelectrica separately reports that a greater solar and wind share has coincided with more very-low and negative short-term price intervals. [S01]
Derived: the simple average of the 29 daily solar-window proxies was approximately 48.1% below the corresponding simple average of daily PZU base prices. This is useful for screening the direction and potential magnitude of shape pressure.
Not established: a Romanian utility-scale PV plant would have captured 48.1% less than baseload. A real capture price depends on the project's quarter-hourly generation profile, outages, clipping, curtailment, settlement route, PPA mechanics, imbalance, basis and actual delivered MWh.
Bankability begins with the correct denominator and the correct clock.
The investor needs to know what each generated MWh is worth when it arrives, not what an average MWh was worth during the year.
Separate market capture from net realised revenue.
These equations should not be collapsed into one haircut. The capture factor addresses timing. It does not automatically include balancing, PPA basis, route-to-market cost, curtailment or BESS dispatch. A project-finance model that applies a generic capture discount to an already solar-shaped forecast can also double-count the same risk.
The minimum investment-grade capture study.
- Quarter-hourly or finer generation profile. Use an independent energy-yield study and reconcile DC/AC ratio, clipping, degradation, availability and seasonal distribution.
- Market-price chronology. Historic and forecast prices need consistent settlement intervals, timezone treatment and daylight-saving handling.
- Forward cannibalisation mechanism. The forecast should explicitly respond to new PV, load growth, thermal retirements, hydro conditions, interconnection and flexibility rather than extrapolating one historic capture factor.
- Contract mapping. Identify which MWh settle against spot, CfD reference price, PPA fixed price, floor, collar or other formula.
- Physical-to-commercial reconciliation. Gross generation, auxiliary consumption, transformer losses, metered export, curtailed MWh and nominated volume should reconcile through one audit trail.
- Downside correlation. Low prices, high solar output, curtailment and battery charging opportunities are not independent shocks. Stress cases should preserve plausible relationships.
“Merchant risk” should be disaggregated before anyone prices it.
Each layer has a different hedge, different diligence owner and different failure mode.
| Risk layer | What can move | Evidence required | Primary mitigant | Bankability consequence |
|---|---|---|---|---|
| Base-price level | Long-run wholesale price curve | Independent market study, forward-market evidence, scenario assumptions | CfD, fixed-price PPA, floor, conservative debt case | Controls the merchant tail's absolute revenue level |
| Solar shape / capture | Project price relative to baseload | Generation-weighted interval model | Shaped offtake, storage, diversified generation profile | Can impair revenue even if annual baseload forecast is unchanged |
| Basis | Difference between project settlement and hedge/reference point | Contract and market settlement mapping | Matched settlement point/reference | Can leave an apparently “fixed” hedge economically open |
| Volume / profile | Generation differs from contracted shape | P50/P90 profile, PPA volume clauses | Pay-as-produced structure, tolerance bands, portfolio management | Baseload PPAs may create buy-back exposure |
| Imbalance | Actual export differs from nomination | Forecast accuracy, BRP terms, historic imbalance cost | Competent BRP, intraday optimisation, BESS where proven | Cash-flow volatility can increase despite a headline fixed price |
| Negative-price provisions | Contract payment may change or cease in specified intervals | Exact CfD/PPA drafting and settlement rules | Dispatch, storage, contractual protections | Requires contract-specific modelling; cannot be inferred from “fixed price” label |
| Curtailment / redispatch | Metered export may be below available generation | Connection terms, historical site/node evidence, compensation regime | Grid rights, compensation, BESS only where charging path is available | Unproven curtailment value should not support base-case debt |
| Counterparty | Offtaker fails, terminates or weakens | Credit analysis, collateral, parent support, termination economics | Investment-grade counterparty, LC/guarantee, replacement rights | A strong fixed price from a weak obligor can remain weak collateral |
The balance between contracted and merchant cash flow is a debt-sizing choice.
A partially hedged project is not intrinsically weaker than a fully contracted one if the residual merchant tail is conservatively valued and debt amortisation is aligned with robust cash flow. Conversely, a long-tenor PPA does not eliminate risk if the volume obligation creates large profile settlements or the counterparty can terminate on weak security.
The correct investment question is therefore not “merchant or contracted?” It is: which exposures remain open after the contract, who carries them, and can debt service survive when those exposures move together?
Revenue certainty is contractual engineering, not a label.
Romania now provides real precedents across CfD-backed, PPA-backed and wholesale-exposed solar. The public disclosures prove financing pathways exist; they do not establish universal leverage for every merchant project.
| Structure | Primary benefit | Residual exposure | Critical lender diligence |
|---|---|---|---|
| Two-way CfD | Long-tenor revenue stabilisation under the scheme | Contract-specific reference price, availability, settlement, compliance, excluded or merchant volume | Eligibility, commissioning deadlines, settlement mechanics, security and termination |
| Pay-as-produced corporate PPA | Can fix or formula-price delivered renewable output | Counterparty credit, basis, imbalance, contract exclusions, merchant tail | Credit support, tenor, change in law, curtailment, negative-price and termination clauses |
| Baseload / profiled PPA | Can produce predictable contracted volume | Shape mismatch and buy-back risk when output differs | Hourly/quarter-hourly volume obligation and replacement-power economics |
| Floor / collar | Protects a portion of downside while preserving defined upside | Counterparty, basis, settlement and residual tail | Whether lenders recognise the floor as durable CFADS support |
| Partial hedge + merchant tail | Balances contracted debt support and equity upside | Unhedged volume and post-hedge tenor | Debt sizing to the contracted period and stressed merchant tail |
| Pure merchant | Full market upside and contracting flexibility | All price, shape and route-to-market risk | Independent market case, leverage tolerance, liquidity and refinancing assumptions |
Interpretation: this transaction shows that institutional financing can accommodate different revenue profiles inside a Romanian solar financing package. Public disclosure does not provide enough detail on per-project ring-fencing, leverage, covenants, cross-support or downside assumptions to conclude that a standalone 100% merchant PV project would receive equivalent non-recourse terms.
The financing market already distinguishes revenue structures.
Two public cases are more useful as evidence of lender behaviour than generic claims that “PPAs are bankable” or “merchant projects can finance.”
Three solar assets with €192 million financing announced. One plant benefited from a 15-year CfD while two were expected to sell on the Day-Ahead Market. The public source demonstrates mixed revenue architecture inside institutional financing, not a generic leverage benchmark. [S07]
The Theia Solar Green Loan project covers three PV plants in Teleorman. EIB states that revenues will rely on commercial PPA(s) and wholesale-market sales. It also identifies environmental, cumulative-impact and promoter-capability appraisal. [S08]
CfD competition adds a separate benchmark—but not a merchant forecast.
Romania's second CfD auction, completed in August 2025 with EBRD support, awarded 2,751 MW. Combined first- and second-auction awards reached 4.2 GW. The second auction attracted more than 5.5 GW of solar and wind bids, with solar bids reported as low as €35/MWh. [S06]
The €35/MWh observation is a competitive bid result under a specific support structure. It should not be substituted for a wholesale-price forecast, universal PV LCOE or minimum price required by every project. What it does show is that contracted projects can compete aggressively for long-duration revenue stabilisation and that investors evaluating merchant projects should compare their risk-adjusted return against a growing body of state-supported contracted capacity.
113.7 GW of valid ATR capacity is a headline. 11.2 GW at the deepest reported stage is the underwriting signal.
Administrative progression is not a commissioning probability, but it is a far better indicator of project maturity than the ATR headline alone.
Maximum approved export power with connection contracts.
Connection contracts plus construction authorisations.
Connection contracts, construction authorisations and ANRE establishment authorisations.
An ATR is evidence of a connection process. It is not a substitute for a deliverability model.
Investment-grade diligence has to identify what physically stands between the project and unrestricted, timely export.
The relevant asset is not merely “150 MW with an ATR.” It is a defined export entitlement attached to a defined substation, network topology, reinforcement programme, construction sequence, technical-compliance package and commissioning path.
Verify: ATR validity, approved export/import power, point of connection, connection agreement, security package, milestones, change-control rights.
Failure mode: the project owns land and permits but cannot preserve or exercise the connection on the intended schedule.
Verify: exact network works, responsible party, procurement status, permitting, cost allocation, dependencies on third-party projects and long-stop date.
Failure mode: PV construction finishes before the grid work required for export.
Verify: transformer rating, line capacity, substation configuration, N-1 implications, reactive-power requirements, protection and SCADA.
Failure mode: nominal approved MW cannot translate into expected operating profile under actual network conditions.
Verify: dispatch instructions, redispatch/curtailment provisions, compensation, outage coordination, metering and telemetry.
Failure mode: the model assumes all technically available generation is economically deliverable.
Verify: energisation, testing, compliance certification, grid-code studies, punch-list and final acceptance.
Failure mode: EPC completion is mistaken for revenue commencement.
Speculative grid optionality now has a more visible liquidity cost.
ANRE's 2026 reform makes development-stage financial capacity part of the project-selection test.
Financial guarantee for relevant production / consumption-and-production projects with approved export power above 1 MW increased from 5% to 20% of the connection tariff excluding VAT. [S03]
Guarantee specified for requested capacity in the 2026 allocation process. Application and transition details require project-specific legal confirmation. [S03]
Financial guarantee based on installed power, valid until works reception and subject to the conditions set out in the reform. [S03]
Illustrative liquidity exposure for a 100 MW project.
At the stated 2026 capacity-allocation level, 100 MW corresponds to a €2.0 million auction guarantee. At €30/kW, a 100 MW establishment-authorisation guarantee corresponds to €3.0 million. The ATR-related amount cannot be calculated without the project's connection tariff.
These amounts should not automatically be added together: applicability, sequence, release and overlap depend on the project pathway and implementing rules. They are also not equivalent to permanently spent capex. They are nevertheless relevant to sponsor liquidity, guarantee-line availability, acquisition pricing and the cost of warehousing a development portfolio.
Milestone asymmetry remains important.
ANRE's reform also set deadlines for obtaining establishment authorisations and provided for extensions in justified cases. Separately, the framework allows the network operator to seek successive 12-month extensions where connection or reinforcement works for which it is responsible require more time. [S03]
Connection quality is locational, not national.
A Romania-wide grid narrative is inadequate for underwriting a project tied to one substation and one reinforcement sequence.
Transelectrica's H1 2026 report explicitly identifies Dobrogea in the south-east and Banat in the south-west as regions with strong wind and solar potential that are already congested and, under existing conditions, do not allow integration of additional capacities. The company states that investments under way and planned for those regions are expected to provide approximately 5,000 MW of additional integration capacity by 2027. [S01]
The 5,000 MW figure is planned additional integration capacity, not commissioned renewable generation, not an allocation to any named developer and not a guarantee that each current ATR will become deliverable. The schedule, technical scope and dependency of the exact project on those reinforcements remain diligence items.
What a locational grid memorandum should contain.
- single-line diagram from project transformer to the relevant upstream network;
- existing and planned generation connected to the same electrical area;
- all network upgrades embedded in the project's ATR and connection agreement;
- network-owner responsibility and evidence of procurement or construction status;
- planned outage and energisation sequence;
- import capability if BESS is contemplated;
- export cap treatment when PV and BESS share the connection;
- reactive power, fault ride-through, protection, SCADA and telemetry requirements;
- operational restriction, redispatch and curtailment language;
- contractual remedies if grid completion misses the project long-stop date.
Storage optionality is becoming part of the asset design. Merchant storage scarcity should not be assumed to persist.
The current fleet and policy pipeline support a stronger flexibility market. They also create a competitive response that storage investors must model.
ANRE reported that, among the 673 projects with connection contracts and construction authorisations in its 1 July 2026 dataset, 155 projects representing 11,268.7 MW of maximum approved export power had been reported by network operators as plants with storage installations or standalone storage. Forty of those projects, representing 3,143.2 MW of approved export power, were reported as expected to commission in 2026. [S02]
Those MW figures describe project approved export power, not the energy capacity or discharge power of the batteries themselves. Treating 11.27 GW as a BESS-capacity pipeline would be a denominator error.
Policy is also supportive. The European Commission's published state-aid record for SA.121308 describes a Romanian Modernisation Fund scheme for standalone battery-storage investment, with an overall budget of RON 764.295 million and duration through 31 December 2030. [S09] ANRE had already removed specified double network charging for electricity stored and subsequently reinjected, while retaining network tariffs for the storage facility's own consumption including technological losses. [S10]
Storage creates value only when the project can name the exposure it changes.
A BESS business case should be decomposed into independently testable value pools.
| Value pool | Economic mechanism | Evidence needed before base-case inclusion | Key failure mode |
|---|---|---|---|
| Capture-price uplift | Move energy from low-value PV intervals into higher-value intervals | Interval PV profile, price forecast, charging constraints, efficiency, degradation, cycling policy | Spread compression or insufficient discharge headroom |
| Clipping recovery | Capture DC generation that would otherwise be clipped | DC/AC design, clipping study, battery topology and charging limits | Clipping volume too small to justify battery cost |
| Curtailment recovery | Charge energy that would otherwise be curtailed and discharge later | Verified site/node curtailment pattern and technical ability to charge during restriction | Curtailment is assumed rather than observed, or restriction also prevents charging/export |
| Imbalance reduction | Use battery flexibility to reduce deviation between nomination and delivery | BRP settlement data, forecast error distribution, dispatch rights | Battery is committed to other services when imbalance occurs |
| Ancillary / balancing services | Provide reserve or balancing products | Qualification, market rules, price history, availability requirement, optimizer capability | Market depth and prices decline as flexible capacity enters |
| Connection utilisation | Share an export cap between PV and battery across time | Connection agreement, metering, import/export rights and dispatch study | Contractual or technical restrictions prevent expected stacking |
| Contract shaping | Use storage to better match a PPA or other delivery profile | Contract nomination, tolerance and settlement terms | Battery degradation cost exceeds avoided shape penalty |
Storage-ready can be the rational answer when BESS is not yet finance-positive.
The decision should separate irreversible construction from reversible design optionality.
| Configuration | Best fit | Advantages | Primary diligence concern |
|---|---|---|---|
| PV only | Strong capture economics, contracted revenue or weak BESS value case | Lower capex and operational complexity | Leaves future timing risk unhedged |
| PV storage-ready | Storage value plausible but timing/economics uncertain | Preserves future option through land, electrical design, control architecture and permitting strategy | Option must be real: reserved space without grid/import rights may have little value |
| AC-coupled BESS | Independent dispatch and retrofit flexibility matter | Operational separation from PV conversion system | Additional conversion losses, connection/import-export treatment and metering |
| DC-coupled BESS | Clipping recovery and integrated PV charging are important | Potentially efficient use of shared equipment and clipped energy | Design complexity, availability interaction and support-scheme/PPA metering |
| Standalone BESS | Storage economics do not depend on a PV asset | Independent siting and market optimisation | Separate grid access and stronger dependence on standalone merchant value pools |
The investment committee should demand the counterfactual.
A co-located battery should be compared with PV-only and PV-storage-ready alternatives using the same market path. If BESS generates its return primarily by assuming historically exceptional price spreads, unverified curtailment or persistent ancillary scarcity, the value case is fragile. If the battery materially improves a contracted delivery obligation, preserves generation otherwise demonstrably lost, or creates a durable revenue floor, the case is stronger.
Storage also introduces new liabilities: degradation, augmentation, warranty throughput limits, availability guarantees, fire-safety requirements, EMS/optimizer dependence, auxiliary consumption, replacement capex and residual-value uncertainty. The project should not treat battery capex as the only incremental cost.
A battery cannot be assumed to “shift contracted solar” until the contract says it can.
Commercial eligibility, metering and dispatch rights have to be mapped before storage optimisation is valued.
Questions that must be answered contract by contract.
- Which meter determines qualifying generation and contractual settlement?
- Can the BESS charge from the grid, PV only, or both?
- If grid-charged energy is later exported, how is it distinguished from renewable generation?
- Who controls BESS dispatch: project company, offtaker, BRP or optimizer?
- Can battery dispatch cause the project to miss PPA delivery obligations?
- How are negative-price intervals treated by the offtake contract or support arrangement?
- Does BESS cycling change the contractual availability calculation?
- Does charging during a grid restriction remain physically and contractually possible?
- How are storage losses, auxiliary consumption and metered imports allocated?
- Who receives ancillary-service revenue and who carries associated imbalance?
The answers determine whether storage is a hedge, a separate merchant business or both. They also determine whether the battery improves debt service or merely moves value between equity revenue streams.
The model should be auditable from irradiance to debt service.
A finance model that starts with annual MWh and annual price has already aggregated away several of the risks that now matter most.
| Layer | Core inputs | Required output | Control test |
|---|---|---|---|
| 1. Resource | Irradiance, temperature, interannual variability | Interval resource series | Independent source and long-term correction |
| 2. Energy yield | Module/inverter model, losses, degradation, availability | P50/P90 generation profile | Independent engineer reconciliation |
| 3. Grid export | Export cap, outages, auxiliary load, verified curtailment assumptions | Metered export MWh | Connection agreement and single-line consistency |
| 4. Market | Quarter-hourly prices, capture model, basis | Market value of each interval | No direct use of annual baseload for PV output |
| 5. Contract | PPA/CfD formula, volume, tenor, floors/collars | Contract settlement | Financial model matches executed legal drafting |
| 6. Balancing | Forecast error, BRP terms, intraday correction | Imbalance and route-to-market cost | Evidence from BRP offer or operating history |
| 7. BESS | Power, energy, efficiency, degradation, augmentation, dispatch | Incremental cash flow by value pool | No double counting of PV revenue or avoided curtailment |
| 8. Opex / tax | O&M, land, insurance, network charges, tax, inflation | EBITDA and CFADS | Contracts and advisor tax model |
| 9. Financing | Drawdown, fees, interest, reserve accounts, amortisation | DSCR, LLCR, debt balance | Term-sheet mechanics reproduced exactly |
| 10. Equity | Acquisition price, development spend, distributions, exit assumptions | Equity cash flow and return metrics | No refinancing or terminal-value assumption hidden in base case |
Debt capacity should emerge from cash-flow resilience, not from a target leverage ratio.
There is no defensible universal Romanian solar DSCR threshold that can be imposed across lenders, revenue structures and project stages.
Those formulas are standard; the thresholds are deal-specific. A lender may distinguish contracted and merchant periods, P50 and downside generation, different price forecasts, reserve requirements and amortisation profiles. The report therefore does not manufacture a “market standard” minimum ratio unsupported by a current term-sheet dataset.
A stronger debt-sizing sequence.
- Build a sponsor base case from fully reconciled physical and commercial assumptions.
- Create a lender energy case using the independent yield methodology required by the financing process.
- Replace sponsor merchant assumptions with the lender-approved market case or agreed haircut methodology.
- Apply the executed offtake terms, including termination, negative-price and volume mechanics.
- Apply project-specific grid availability and curtailment treatment rather than a generic national assumption.
- Model imbalance and route-to-market costs explicitly.
- Include BESS revenue only to the extent accepted by the lender and supported by contracts or conservative evidence.
- Size and sculpt debt to the required coverage profile.
- Run combined downside cases before determining distributable equity.
Single-variable sensitivities are useful diagnostics. They are not a downside case.
The major Romanian solar risks can reinforce one another. The model should show the combined cash-flow effect.
| Stress | Primary variable | Correlated variables to consider | Early-warning evidence |
|---|---|---|---|
| Solar cannibalisation | Capture factor weakens | More negative intervals, charging competition, PPA repricing | Project-weighted capture trend vs baseload |
| Low wholesale case | Merchant price curve declines | Lower BESS arbitrage revenue | Forward curve and independent forecast revisions |
| Grid delay | COD moves later | IDC, EPC claims, PPA start mismatch, guarantee extension | Network works critical-path slippage |
| Export restriction | Delivered MWh falls | Capture mix, PPA volume, BESS charging opportunity | Operator notices and actual node restrictions |
| Balancing shock | Imbalance cost rises | Forecast error and intraday liquidity | BRP settlement history |
| PPA counterparty event | Contract cash flow lost or delayed | Replacement price, collateral draw and merchant exposure | Credit rating, financial statements, collateral status |
| EPC / capex overrun | Construction cost increases | Contingency, sponsor equity, delay | Procurement packages and schedule contingency |
| BESS revenue compression | Arbitrage/ancillary value declines | Cycle count, degradation and optimizer economics | Spread distribution and qualified flexible capacity |
| BESS degradation | Usable MWh declines faster | Augmentation capex and warranty claims | SOH and throughput versus warranty curve |
| Interest / refinancing | Debt cost or terminal refinancing worsens | Coverage and equity distributions | Hedge coverage and maturity profile |
Every model assumption needs a contract owner—or it remains equity risk.
Bankability improves when construction, operational and revenue risks are placed with parties capable of controlling them and backed by enforceable remedies.
| Contract | Risk to allocate | Bankability provisions | Residual diligence |
|---|---|---|---|
| EPC | Cost, schedule, performance | Fixed-price scope where achievable, LDs, performance tests, security, interface matrix | Cap, exclusions, relief events, contractor credit |
| Module / inverter supply | Equipment performance and defects | Bankable warranty, degradation/performance terms, assignability | Manufacturer credit and claims process |
| O&M | Availability and response | Availability KPI, response times, spare-parts strategy | Exclusions and interface with OEM warranties |
| Connection | Grid scope, timing and cost | Defined milestones, responsibilities, security and remedies | Operator extension rights and reinforcement dependencies |
| PPA / CfD | Price, volume and counterparty | Clear settlement, credit support, change-in-law and termination regime | Residual market/basis exposure |
| BRP / route-to-market | Nomination and imbalance | Transparent fee, data access, settlement and termination | Incentive alignment and historic performance |
| BESS OEM / integrator | Availability, degradation, throughput | Capacity/efficiency guarantees, augmentation and warranty regime | SOH definition, excluded operating modes, warranty security |
| Optimizer | Dispatch and market access | Rights, revenue definition, benchmark, data ownership, termination | Merchant forecast risk remains with project unless guaranteed |
The independent engineer now needs to test commercial consequences, not only plant performance.
Technical diligence should explain how design decisions alter capture, export and contractual settlement.
Yield and profile
P50/P90, interannual variability, hourly/quarter-hourly generation, degradation, clipping, availability and loss stack.
Connection and dynamic compliance
Studies, protection, reactive power, fault ride-through, SCADA, metering, export/import limits and commissioning sequence.
Constructability
Geotechnical, drainage, flood, access, civil scope, cable routes, substation delivery and critical-path procurement.
Technology bankability
Track record, warranties, degradation, supply-chain quality, spares, service capability and replacement assumptions.
BESS duty-cycle validation
Power/energy sizing, SOH, round-trip efficiency, thermal management, augmentation, safety and warranty throughput.
Meter-to-model reconciliation
Demonstrate how the technical export profile maps into the revenue model and contracted settlement.
Investment-grade solar is a rights package, not merely an engineered plant.
Land, permits, grid rights, environmental obligations and revenue contracts have to survive financing and enforcement.
Legal diligence priorities.
- land ownership or lease chain, cadastral status, access and easements;
- urbanism, construction and establishment authorisations and their expiry/milestone conditions;
- ATR and connection-contract transferability and change-of-control implications;
- security package, lender step-in and direct agreements;
- PPA/CfD assignability, termination payments and lender cure rights;
- corporate approvals, shareholder funding commitments and guarantee availability;
- BESS permitting, fire-safety, hazardous-material and emergency-response obligations where applicable.
Environmental and social diligence is finance diligence.
EIB's public Theia project summary states that its appraisal addresses the EIA screening process, potential cumulative impacts, impacts on nature-conservation sites and the promoter's capacity to implement the project. [S08] For lenders and institutional investors, environmental or biodiversity defects can become permitting delay, capex, operating restriction, litigation or reputational exposure. They therefore belong in the project model and conditions-precedent schedule rather than an annex completed after commercial underwriting.
Tax diligence should follow cash flow.
Model reviews should reconcile tax assumptions with SPV structure, financing, depreciation, VAT timing, land arrangements, revenue contracts, BESS charging and any applicable support scheme. This report does not prescribe project-specific Romanian tax treatment; current advice should be obtained for the actual ownership and financing structure.
The future Romanian PV market will bifurcate by evidence quality.
These classifications are strategic judgements, not lender ratings. They identify what would normally need to be true before a project can move toward investment-grade financing.
| Archetype | Revenue | Grid | Storage | Financing view |
|---|---|---|---|---|
| A / Contracted + deliverable | Long-tenor credible offtake with controlled residual exposures | Mature rights, defined works and credible COD | Only if additive | Strongest project-finance candidate, subject to full diligence |
| B / Partially contracted + mature grid | Contract supports debt period; merchant tail conservatively modelled | Deliverability substantially evidenced | Optional or targeted | Financeable with disciplined debt sizing |
| C / Merchant + strong grid | Project-weighted merchant case and liquidity support | Mature and deliverable | Potential hedge | Possible for suitable capital; leverage likely more sensitive to market assumptions |
| D / Contracted + weak grid | Strong headline offtake | Major unresolved reinforcement or COD dependency | Cannot automatically solve | Conditional; contract value may not be realisable on time |
| E / Merchant + speculative grid | High open price/capture exposure | Immature or uncertain | Often presented as a remedy | Development optionality rather than finance-ready asset |
Replace one blended project score with seven evidence gates.
A project should be able to fail one gate without a strong score elsewhere disguising the problem.
Land, permits, establishment authorisation and corporate rights are valid, transferable where needed and aligned with financing.
Stop condition: critical rights expire before realistic construction or cannot be secured by lenders.
Connection terms, reinforcement scope, schedule, export cap and operator dependencies are independently mapped.
Stop condition: COD depends on an unverified network assumption.
Independent yield, degradation, availability and interval profile support the revenue case.
Stop condition: project economics require sponsor yield materially above independent evidence.
Capture, PPA/CfD, merchant tail, basis, imbalance and negative-price exposure are transparent.
Stop condition: baseload forecast is used as project realised price without an interval model.
EPC scope, capex, schedule, equipment and contractor credit are financeable.
Stop condition: residual interfaces exceed sponsor contingency or critical procurement is unsecured.
Each BESS value pool is evidenced and contractually available; degradation and augmentation are funded.
Stop condition: equity case depends mainly on unverified curtailment or persistent extreme spreads.
Debt case survives combined downside with adequate liquidity and covenant headroom.
Stop condition: required leverage exists only in sponsor base case.
Development-stage M&A should price evidence, not MW.
The connection queue makes apparent pipeline scale easy to acquire on paper. The value lies in de-risked progression.
Purchase-price allocation should distinguish four things.
Recoverable project spend. Land, studies, applications and engineering may have evidence value but do not automatically equal market value.
Scarce rights. A durable grid position, permits and land package can justify a development premium when independently verified.
Execution work completed. Advanced engineering, procurement and network works reduce remaining schedule risk only to the extent they are usable by the buyer.
Speculative option value. Pipeline projects dependent on future auctions, reinforcement or unresolved permits should be valued as options with explicit failure pathways rather than at a uniform €/MW benchmark.
Earn-outs can align value with evidence.
For immature projects, consideration can be linked to objective milestones: preservation of grid rights, construction authorisation, establishment authorisation, financing-ready offtake, notice to proceed, energisation or COD. The exact mechanism is transaction-specific, but the principle prevents the acquirer from paying full finance-ready value before the project has crossed finance-ready gates.
The direction is investable. The dispersion between projects should widen.
These are scenario frameworks rather than probability-weighted forecasts. They identify what would change the bankability conclusion.
Base case
Conditions: Romanian PV and BESS deployment continues; grid reinforcements progress unevenly; capture pressure becomes a permanent underwriting variable; CfD/PPA financing coexists with merchant exposure.
Implication: contracted and grid-mature assets retain strong financing access, while merchant projects require lower leverage, stronger sponsor liquidity or differentiated capture/storage economics.
Upside
Conditions: transmission reinforcement, interconnection, demand growth and flexible load/storage absorb solar output faster than expected; corporate PPA depth expands; balancing/route-to-market capability improves.
Implication: capture erosion moderates and well-sited merchant projects regain debt capacity. Storage value shifts from scarcity trading toward contracted shaping and grid services.
Risk case
Conditions: PV additions outpace flexible demand and grid reinforcement; low/negative daytime intervals become more frequent; network works slip; storage merchant revenues compress as competition increases.
Implication: merchant PV valuations fall first at weak nodes. Debt migrates toward CfD/PPA-backed projects and acquisitions place larger discounts on immature connections.
Structural break
Conditions: material market-design, network-tariff, support-scheme, balancing or connection-allocation changes alter the economics faster than historical models can capture.
Implication: existing long-term assumptions require re-underwriting. Projects with flexible contractual structures, strong grid rights and low fixed leverage preserve more strategic options.
Assume the investment disappoints. What probably went wrong?
The most useful red team attacks the project's hidden dependencies rather than repeating generic risk factors.
| Failure | How it becomes visible | Why the original case missed it | Contingency |
|---|---|---|---|
| Capture deteriorates faster than forecast | Project-weighted price falls relative to reference | Model relied on annual baseload or static capture factor | Re-contract, optimise availability, add storage only if incremental case clears |
| Grid COD slips | Reinforcement milestones repeatedly move | Operator work was treated as a date rather than a critical path | Long-stop protections, staged NTP, sponsor liquidity buffer |
| PPA hedge performs poorly | Shape or imbalance settlements consume fixed-price benefit | Headline price was analysed without settlement mechanics | Renegotiate shape, improve forecasting/BRP, use BESS selectively |
| BESS merchant case compresses | Spread and ancillary revenue fall as competing storage grows | Historical scarcity was extrapolated | Preserve multiple value pools, contracted optimisation, lower leverage |
| Guarantee / collateral burden grows | Development portfolio consumes bank lines | Guarantees treated as immaterial because not capex | Stage portfolio, recycle weak projects, allocate liquidity by milestone quality |
| Debt case depends on refinancing | Coverage is adequate only with optimistic terminal debt | Refinancing was embedded as a certainty | Amortise to resilient cash flow and treat refinancing as upside |
The portfolio should diversify electrical and contractual risk, not just geography.
Ten projects in different counties can still be one economic exposure if they generate into the same hours under the same market structure.
Portfolio construction should track five concentrations.
Node and network concentration. Projects should be mapped by electrical dependency and shared reinforcement, not only map coordinates.
COD concentration. A portfolio with the same delivery year can be exposed to the same equipment, financing, auction and connection bottlenecks.
Revenue concentration. Diversify CfD, corporate PPA and merchant exposure deliberately rather than inheriting a portfolio-wide merchant tail.
Counterparty concentration. Several PPAs with one buyer do not diversify credit risk.
Shape concentration. South-facing PV assets produce highly correlated hours. Tracking configuration, geographic weather diversity, storage and complementary generation can change the aggregate capture profile, but the effect should be modelled rather than presumed.
Preferred expansion sequence.
- Prioritise projects with durable connection rights and independently evidenced delivery schedules.
- Secure or design credible revenue protection before maximising leverage.
- Preserve storage-ready optionality at sites where flexibility may become valuable but BESS does not yet clear the investment hurdle.
- Acquire immature connection positions only at prices consistent with milestone failure risk.
- Measure portfolio capture and balancing exposure at interval level.
- Allocate guarantee-line capacity to projects with the highest probability-adjusted strategic value rather than largest headline MW.
Move from market thesis to bankable evidence in four workstreams.
The sequence is designed to kill weak assumptions before expensive financing work is committed.
Grid and rights reconstruction. Build the connection critical path, permit matrix, land chain and guarantee schedule. Obtain underlying operator correspondence and technical studies.
Decision: confirm whether the project deserves further diligence.
Interval revenue model. Rebuild yield and settlement at quarter-hourly level; calculate project capture, merchant tail, PPA/CfD settlement, imbalance and downside.
Decision: establish an investable revenue case independent of sponsor summary outputs.
Technical and commercial diligence. Validate EPC, equipment, O&M, BRP, offtake and BESS design; reconcile model assumptions to contracts.
Decision: determine residual risks that remain with SPV equity.
Financing market test. Issue lender information package based on a reconciled model, not sponsor headline returns. Obtain term-sheet feedback on merchant tail and storage revenue treatment.
Decision: establish actual debt capacity and lender conditions.
Investment committee and documentation. Run combined downside, finalise conditions precedent, price residual risk and negotiate acquisition/EPC/PPA protections.
Decision: approve, reprice, defer or reject.
Bankability now changes what each capital provider should ask first.
The diligence order matters because each audience is exposed to a different failure point.
Pay for delivery quality.
Value grid maturity, contract durability and project-weighted capture before optimising acquisition €/MW. Use milestone-linked consideration where execution evidence is incomplete.
Underwrite residual exposures.
Ask what remains open after the PPA/CfD, size debt to stressed CFADS and require explicit treatment of capture, imbalance, grid delay and merchant tail.
Portfolio-manage correlation.
Monitor node, capture, counterparty, COD and merchant concentrations. Project count and geographic spread alone can materially overstate diversification.
Reconcile every layer.
Make the model, technical report, connection documents and revenue contracts tell the same story at the same settlement granularity.
The winning Romanian PV asset will not necessarily be the cheapest plant.
Its advantage will be the quality and durability of the cash-flow system around the plant.
Romanian utility-scale solar remains an investable asset class. The current evidence does not support a thesis that merchant exposure, grid congestion or storage competition make new PV structurally unfinanceable. It does support a more demanding conclusion: financeability is becoming increasingly project-specific.
The market now contains enough solar for same-hour price effects to matter, enough connection demand for grid progression to carry economic value, enough storage for flexibility to become both a hedge and a competitive market, and enough financing precedent to show that strong projects can still attract institutional capital under different revenue structures.
The resulting underwriting standard is clear. A future PV investment should not be approved because the sponsor can show an attractive annual yield, a valid ATR and a long-run average power price. It should be approved when the investment committee can trace the physical MWh through the grid, the settlement MWh through the revenue contract and the stressed cash flow through debt service.
What is known, what is derived and what still belongs in the data room.
The report prioritises current primary and institutional evidence. Public market data are used to frame underwriting questions, not to replace project-specific technical, legal or lender diligence.
Current reported market structure
Installed fleet, storage stock, grid-connection stages, current guarantee reform and cited public financing transactions are directly supported by primary or institutional sources.
Structural interpretation
Capture-price pressure, storage competition and likely financing bifurcation are analytical conclusions supported by current observations but remain sensitive to future grid, demand, policy and market development.
Bankability outcome
Node-level curtailment, exact reinforcement schedule, project capture factor, PPA settlement, lender leverage, DSCR thresholds and BESS revenue are unknown until the actual project and contracts are diligenced.
Method.
Primary-source evidence was prioritised from Transelectrica, ANRE, Romania's Ministry of Energy and EU institutional records. EBRD and EIB disclosures were used for financing precedents. SolarIndustry.ro's retained operational history was used as a traceable market-observation layer. The SolarIndustry hourly history averages all retained public Transelectrica readings recorded during each local Europe/Bucharest hour; missing hours are left absent rather than interpolated. PZU data are maintained separately because they are delivery-day market results. [S04]
Price-screen methodology.
The 20 July–17 August 2026 exhibit uses 29 OPCOM PZU delivery days retained by SolarIndustry. “Solar window” is the platform's unweighted 08:00–17:00 price proxy. The report calculates simple means across those daily observations and compares the resulting solar-window mean with the mean daily PZU base price. This does not weight prices by project output and must not be used as a project capture factor.
Grid-pipeline methodology.
ANRE's 1 July 2026 maximum approved export power is shown by administrative stage. Derived percentages divide the capacity at each stage by 113,652 MW of valid-ATR approved export power. The ratios describe stage progression only. They do not estimate a project's probability of construction or imply that capacity falling outside a later stage has been cancelled.
Material limitations.
No current public source identified in this research provides the site-specific quarter-hourly generation, curtailment, PPA settlement, BRP cost, nodal reinforcement schedule, debt terms or BESS operating data required to make an investment decision on a named project. The 29-day price screen is deliberately short relative to an asset life and is used only to demonstrate current market shape. Historic observations should not be extrapolated mechanically over a financing tenor.
Regulation can change. The May 2026 ANRE communication is used as evidence of the adopted connection/licensing reform, but a transaction should verify the final implementing order, effective dates, transitional treatment and applicability to the project with Romanian counsel.
Auditable evidence ledger.
Research cut-off: 17 Aug 2026. Source dates and observation dates are kept distinct where material.
- S01Transelectrica — Half-yearly Report January–June 2026. Report dated 14 Aug 2026. Used for national generation-park capacity at 1 Jul 2026, storage capacity, prosumer stock, short-term low/negative-price commentary, balancing-price volatility context and grid-development/congestion statements. Primary public-system/operator evidence.
- S02ANRE — Status of renewable electricity-generation projects with approved export capacity ≥1 MW at 1 Jul 2026. Used for valid ATR count and MW, connection-contract stages, construction/establishment-authorisation stages, projects reported with storage and expected commissioning distribution. Primary regulator evidence.
- S03ANRE — Connection and licensing regulation reform communication, 21 May 2026. Used for 20% ATR-related financial guarantee, establishment-authorisation deadlines and €30/kW guarantee, 2026 allocation-auction guarantee and grid-operator extension provisions. Project applicability requires legal verification.
- S04SolarIndustry.ro — Romania operational electricity data history. Retrieved through research cut-off 17 Aug 2026. Used for retained OPCOM PZU daily evidence and methodology. Original operational source is Transelectrica; original PZU source is OPCOM. The displayed solar-window series is explicitly an unweighted screening proxy, not a solar capture-price calculation.
- S05OPCOM — Romanian electricity market operator. Original public market source for PZU delivery-day and PT15 evidence retained and analysed in the SolarIndustry data-history layer.
- S06EBRD — Romania's second renewable CfD auction, 21 Aug 2025. Used for 2,751 MW second-auction award, 4.2 GW combined first/second awards, >5.5 GW bids and reported solar bid prices as low as €35/MWh. Institutional secondary / transaction-framework evidence.
- S07EBRD — €192 million Romanian solar financing package, 27 Nov 2025. Used for the 531 MW Slobozia, Corbii Mari and Iepuresti II package, EBRD/commercial lender participation, Slobozia CfD and disclosed Day-Ahead-Market exposure for the other projects.
- S08European Investment Bank — Theia Solar Green Loan. Project summary released 6 Mar 2025; approved 19 Jun 2025. Used for 710 MWp Teleorman portfolio, commercial PPA(s) plus wholesale-market revenue architecture and EIB environmental/social appraisal topics. Public financing-project evidence; financing amount not disclosed.
- S09Official Journal of the European Union — State aid SA.121308, published 1 Apr 2026; Commission decision adopted 6 Mar 2026. Romanian Modernisation Fund scheme for standalone battery-storage installations; overall budget RON 764.295 million, direct-grant form, duration through 31 Dec 2030.
- S10ANRE — Removal of duplicate regulated charges for stored and reinjected electricity, 8 Jul 2025. Used for specified exemptions on stored/reinjected electricity and the continuing tariff treatment of storage own consumption and technological losses.
- S11Romanian Ministry of Energy — Order No. 1120/26 July 2024 approving the CfD state-aid scheme. Primary scheme reference used to anchor the Romanian CfD framework; executed project contracts and later auction documentation remain controlling for project-specific settlement.