The daylight discount reopened.
Flexibility moved from pipeline to operating need.
Romania’s market did not spend the week in simple renewable surplus. Daylight electricity was sharply devalued on the weekend even while retained system observations showed positive average net imports on every review day. By midweek, prices tightened again, exposing the underlying problem more clearly: generation is increasingly abundant at the wrong hours and scarce at others. At the same time, storage moved forward on several different maturity tracks — one major BESS entered operation, ANRE approved establishment authorisations, developers disclosed multi-hundred-MWh projects, optimisers entered the route-to-market layer and public-sector solar-plus-storage procurement became more concrete. The commercial issue is therefore no longer whether Romania needs flexibility. It is which flexibility can actually connect, operate and monetise across a volatile 15-minute market.
The market priced time more aggressively than energy.
The relevant comparison is not “solar versus storage.” It is uncontrolled production versus controllable delivery. This week made that distinction unusually visible in both price and physical-flow evidence.
Romania’s seven-day PZU base mean eased only 4.7% week on week to 946.00 RON/MWh, while the unweighted 08:00–17:00 price proxy fell 32.0% to 465.17 RON/MWh. The daylight/base ratio therefore dropped from 68.9% to 49.2%, and 20 Sep recorded 16 negative PT15 intervals. Yet retained Transelectrica observations show positive average net imports on every review day; on 23 Sep the system moved from 696 MW net export at 12:00 to 1,962 MW net import at 19:00. Meanwhile the 120 MWh Teiuș BESS entered operation and ANRE approved several storage establishment-authorisation decisions. The decision implication is direct: underwrite timing, connection maturity and controllability before headline MW or MWh. [S01] [S02] [S03] [S04]
For PV: the base PZU price is no longer an adequate revenue proxy. A project can face attractive daily averages and still experience severe daylight compression.
For BESS: this week supports the value of time-shifting but does not prove a battery margin. Round-trip losses, degradation, imbalance exposure, grid charges, cycling constraints and ancillary-service revenues remain asset-specific.
For grid delivery: approvals accelerated, but announced MWh remain economically irrelevant until permits, connection, equipment, commissioning and market access converge.
A weak daylight weekend reset the weekly profile.
Seven delivery days are shown exactly as retained. The “solar window” is the unweighted mean of available OPCOM PT15 results from 08:00 through 17:00 local time. It is a price-shape proxy, not a solar capture-price calculation.
Derived sum of Borzești 918 MWh, Lebăda 816 MWh, Buciumi 176 MWh and Zephyr 120 MWh. Maturities differ materially; this is not a national pipeline total. [S05–S08]
Eurowind’s Teiuș BESS announced as completed and commissioned on 23 Sep after testing, integration and verification. [S09]
Storage-related establishment-authorisation decisions identified in ANRE’s 22 Sep approved-document list, plus a separate storage operating licence for Dacia Green Stoc. [S04]
AFIR confirmed €500m for new solar with integrated storage and €150m for renewable-energy storage, opening 28 Sep. [S11]
Storage dominated the headlines, but maturity was the real story.
The timeline distinguishes operating assets, regulatory decisions, company pipeline disclosures and policy consultations. It intentionally does not treat an announcement, an authorisation and commercial operation as equivalent evidence.
Borzești put 918 MWh behind an EPC disclosure.
Renovatio Solar disclosed a full-EPC role for the 918 MWh Borzești Energy Storage project, developed for RNV Infrastructure and Greenvolt Power, with CATL as technology partner and an estimated 12-month execution period. No MW power rating was disclosed in the source reviewed, so no duration is inferred. The key signal is commercial procurement depth, not yet operating capacity. [S08]
Capalo AI added another route-to-market layer.
Capalo AI said it had formally entered Romania, registered as a Balance Responsible Party / PRE with Transelectrica and already held a supply licence. This expands the optimisation and trading-services set available to standalone and hybrid assets. It is evidence of service-market formation, not evidence of realised BESS revenue. [S10]
Daylight prices briefly collapsed into negative territory.
The 20 Sep PZU base averaged 731.27 RON/MWh, while the unweighted 08:00–17:00 proxy was only 70.18 RON/MWh. Sixteen PT15 intervals were negative and the daily low was −2.47 RON/MWh. The daylight/base ratio fell to 9.6%. This was the week’s clearest price-cannibalisation event — but it was a one-day extreme rather than evidence that every subsequent daylight hour had become structurally negative. [S02]
ANRE moved two prosumer rule sets into phase-II consultation.
ANRE republished phase-II drafts covering transitional quantitative compensation and multi-location commercialisation, billing, settlement and value allocation, with comments due by 1 Oct. These were consultations at the cut-off, not final rules. The commercial relevance is settlement design and the ability to allocate prosumer value across consumption points. [S14] [S13]
ANRE converted several storage projects from agenda items into decisions.
ANRE’s approved-document list records establishment-authorisation decisions for multiple storage projects, including Buciumi and Zephyr 1, alongside Isaccea, Baba Ana 2, Berghia, Poiana-Turburea, Arad 2 and a mixed generation/storage project at Mehedința. A separate commercial-operation licence for energy storage was approved for Dacia Green Stoc. For Buciumi, EBRD describes an 88 MW / 176 MWh standalone merchant asset. For Zephyr, PPC’s own project disclosure is 60 MW / 120 MWh. [S04] [S06] [S05]
Grenergy made Lebăda visible in its formal investor pipeline.
Grenergy’s September investor presentation identifies Lebăda in Romania at 204 MW / 816 MWh, classifies the project within backlog, gives an expected ready-to-build date of Q1 2027 and operation in Q3 2028, says it expects to apply to the EU Modernisation Fund and reports PPA negotiations as advanced. The same presentation shows 4,719 MWh of Romanian storage pipeline across company-defined maturity categories. Those figures are developer pipeline, not operating national stock. [S07]
AFIR turned the public-sector funding theme into a dated procurement window.
AFIR confirmed online submissions from 28 Sep at 10:00 through 20 Nov at 23:59, with €500m allocated to new solar with integrated storage for public entities and €150m to renewable-energy storage. Storage support can cover up to 100% of eligible expenditure, capped at €10m per beneficiary and €200,000/MWh. Evaluation starts in chronological order. For EPCs and public buyers, this converts a funding concept discussed last week into a near-term execution timetable. [S11]
Solar scale expectations rose, but the headline remains an industry estimate.
Industry representatives cited roughly 9.1 GW of Romanian solar across parks and prosumers and expected the market to exceed 10 GW around the beginning of 2027, with possible year-end figures around 11 GW. Eurowind separately said it expected to begin its 220 MW Vișina project within weeks. These are useful construction-pipeline signals, but the 9.1–11 GW figures are not treated here as an independently reconciled official operating stock. [S18] [S17]
Teiuș provided the week’s strongest delivery evidence.
Eurowind announced completion and commissioning of its 120 MWh BESS at Teiuș, Alba, following testing, integration and verification. The disclosed project investment is €21m with NRRP support. This is qualitatively different from the much larger pipeline announcements: it is an operating reference rather than future flexibility. [S09]
Green Breeze added 99.2 MW of actual wind supply.
OX2 handed over the 99.2 MW Green Breeze wind farm in Galați County to Nala Renewables and said the 16-turbine project was fully operational and supplying the grid under a long-term corporate PPA. The project matters to solar and storage economics because new wind changes the same residual-load and balancing environment into which PV and batteries sell. [S16]
Behind-the-meter solar-plus-storage broadened beyond industrial plants.
Sibiu Airport announced financing for approximately 1.700 MWp of PV plus 2.088 MWh of storage, targeting 71.06% use of forecast output for own consumption. A day earlier, Bacău Airport disclosed a signed design/build contract for 1.25 MW PV and 2.502 MWh of storage, targeting more than 70% of its own electricity consumption. These are small beside utility-scale BESS, but commercially important because their value case begins with avoided grid purchases and autoconsumption rather than merchant arbitrage. [S20] [S19]
The flexibility deficit is temporal, locational and commercial.
Four separate questions need to remain separated: what electricity cleared for, what solar could actually capture, what the power system physically needed by hour, and which storage projects are mature enough to respond.
The base market softened. Daylight value fell much faster.
Last week’s briefing recorded a seven-day PZU base mean of 992.62 RON/MWh and an 08:00–17:00 proxy of 683.76 RON/MWh, placing the proxy at 68.9% of the base. This week the base mean fell to 946.00 RON/MWh, only a 4.7% decline, but the daylight proxy fell to 465.17 RON/MWh — a 32.0% contraction. The ratio therefore lost 19.7 percentage points. [S03]
The economic interpretation is stronger than “power prices were lower.” Most of the deterioration occurred in the hours most exposed to solar production. The 19–20 Sep weekend was particularly severe: the daylight/base ratio fell to 24.4% on Saturday and 9.6% on Sunday. By 22 Sep it had recovered to 71.0%, showing that the extreme compression is not yet a stable regime. For underwriting, the correct conclusion is therefore higher intraday shape risk, not a claim that daylight power has permanently lost value.
| Delivery date | PZU base RON/MWh |
08–17 proxy RON/MWh |
Proxy / base | Low PT15 RON/MWh |
Negative PT15 | Avg exchange MW |
|---|---|---|---|---|---|---|
| 17 Sep 2026 | 890.53 | 496.96 | 55.8% | 215.84 | 0 | +505 |
| 18 Sep 2026 | 928.91 | 474.47 | 51.1% | 79.25 | 0 | +1,269 |
| 19 Sep 2026 | 844.56 | 205.68 | 24.4% | 5.01 | 0 | +1,385 |
| 20 Sep 2026 | 731.27 | 70.18 | 9.6% | −2.47 | 16 | +778 |
| 21 Sep 2026 | 964.83 | 450.90 | 46.7% | 97.18 | 0 | +791 |
| 22 Sep 2026 | 1,219.99 | 865.55 | 71.0% | 614.64 | 0 | +1,510 |
| 23 Sep 2026 | 1,041.88 | 692.43 | 66.5% | 290.76 | 0 | +875 |
Exchange is the arithmetic mean of retained system observations by local day; positive = net import, negative = net export. Price fields are OPCOM PZU references. No missing interval is filled or interpolated. [S02]
465.17 RON/MWh is not a solar capture price.
It is an unweighted clock-window proxy. It does not weight prices by actual PV generation and therefore cannot be inserted directly into an asset revenue model.
True capture needs temporally matched price and generation.
A solar capture price requires the market price in each relevant settlement interval multiplied by actual solar production in that same interval. The retained SolarIndustry dataset deliberately keeps operational observations and OPCOM outcomes separate, and the current public dataset does not provide a fully harmonised national PT15 solar-generation series suitable for a decision-grade national capture calculation.
Solar capture price = Σ(Pt × Gsolar,t) / ΣGsolar,tCheap daylight and import dependence coexisted.
The retained operational evidence is important because it prevents an easy but incorrect inference: low daylight prices do not mean Romania was simply long energy all week. The daily average exchange balance remained positive — net import — on every review day, ranging from +505 MW on 17 Sep to +1,510 MW on 22 Sep. The simple arithmetic mean of the seven daily exchange means was about +1.02 GW. [S02]
What changed inside the day was more revealing. On 23 Sep, retained observations show 2,036 MW of solar and a 696 MW net export at 12:00. By 19:00, reported solar had fallen to 5 MW and the exchange balance had moved to 1,962 MW net import — a 2,658 MW reversal in seven hours. Similar export-to-import rotations appeared on 19 and 20 Sep. The system problem is therefore not only annual energy adequacy. It is the ability to reposition energy over hours while respecting local and national network constraints.
Reported solar versus net exchange
The largest number was not the most important number.
Four high-profile non-operating projects highlighted this week sum to 2.03 GWh: Borzești 918 MWh, Lebăda 816 MWh, Buciumi 176 MWh and Zephyr 120 MWh. That sum is useful only if its evidence boundary stays visible. One is disclosed through an EPC announcement, one is in a developer-defined backlog with a 2027 RTB target, and two obtained ANRE establishment-authorisation decisions. None of those 2.03 GWh should be treated as operating flexibility at the research cut-off.
By contrast, Teiuș’s 120 MWh is smaller but was announced as commissioned. For near-term system impact, the operating 120 MWh is more material than a gigawatt-hour-scale project with an unresolved construction or connection timeline. For market-entry decisions, therefore, a maturity-adjusted project ledger is more useful than a raw MWh league table.
Selected BESS references active in this week’s evidence
Bar length represents disclosed MWh only. It does not encode probability of completion or commercial quality. Hover or focus each row for the maturity evidence.
Wholesale time-shifting
The 20 Sep negative PT15 cluster and subsequent high evening values create a visible gross price-shape signal. Do not convert that spread directly into battery EBITDA: efficiency losses, cycling limits, degradation, market fees, charging constraints and dispatch optimisation all sit between spread and cash flow.
Balancing and ancillary access
Capalo’s PRE/BRP entry and EBRD’s Buciumi description both reinforce the route-to-market layer. Bankability depends on qualification, optimiser performance, settlement rules and future service-market saturation — not only on day-ahead arbitrage. [S10]
Connection and curtailment value
Craiova’s Șimnic documentation is especially instructive: the planned 5.3 MW / 21.2 MWh usable BESS is tied directly to the PV project’s ATR conditions, with the reviewed feasibility material stating the plant cannot be commissioned under that ATR without the storage component. Storage can therefore be a deliverability asset, not merely a trading asset. [S21]
Behind-the-meter optimisation
The Bacău and Sibiu airport projects start from autoconsumption, not merchant spread. Greenvolt Next separately reported 209 MW of Romanian storage projects and increasing interest from industrial and commercial customers adding batteries to existing PV. The 209 MW figure is company-reported and is not converted to MWh because no comparable energy capacity was disclosed. [S24]
Storage is becoming part of the connection solution.
The strongest grid signal this week was not a national congestion statistic. It was the way storage appeared inside project-specific delivery evidence. ANRE’s 22 Sep committee approved establishment-authorisation decisions across several storage projects; Buciumi connects by an approximately 1.5 km underground line to the existing Gutâi 400/220/110 kV substation according to EBRD; and the Craiova project explicitly ties its battery to ATR compliance. [S06]
This changes how a developer should define “grid-ready.” An ATR is necessary evidence, but it is not the final commercial state. The underwriting sequence should distinguish grid study and ATR, connection contract, establishment authorisation where applicable, land and environmental readiness, equipment procurement, construction, energisation/testing, final connection certification and market registration. Treating those stages as a single “ready” category hides the most expensive execution risk.
Price the connection solution before pricing the battery.
A lower-cost battery package can be economically inferior if its design fails the ATR, reactive-power, EMS, protection, telemetry or grid-code requirements needed for commissioning. For storage attached to a constrained renewable project, connection certainty is part of the battery’s value.
AFIR creates a near-term bid market, not automatic demand.
The €650m public allocation is substantial, but eligible budgets still require technically compliant projects, procurement capacity, design maturity and implementation. Because evaluation begins chronologically, bid readiness matters; because compliance conditions remain binding, speed without documentation quality creates rejection risk. [S11]
Government rhetoric is converging with the observed ramp.
A Ministry of Energy state secretary argued on 22 Sep that more storage would reduce dependence on imports during peak hours and improve use of renewable generation. That is a policy position, not a permitting reform or revenue guarantee; the physical 23 Sep exchange reversal nevertheless shows why the issue has become operationally salient. [S25]
Prosumers already represent a multi-gigawatt system layer.
ANRE’s latest reliable prosumer stock identified at the cut-off is for 31 Jul 2026: 373,249 prosumers and 4,148.86 MW of installed capacity. The same ANRE report records 139,967 prosumers with storage installations. That is 37.5% of the prosumer count, derived from ANRE’s site counts — but it is emphatically not 37.5% of prosumer MW and does not provide aggregate battery MW or MWh. [S12]
The distinction matters because distributed flexibility is often discussed as though an installed battery count were dispatchable system capacity. Without aggregate power, usable energy, inverter configuration, state-of-charge management and market/aggregator access, the system contribution cannot be inferred from installation count alone. The current policy direction can increase the installed base, but turning residential storage into grid value requires control, incentives and market integration.
Registered prosumers in ANRE’s July situation report. [S12]
Prosumer installed generation capacity; 4,148.86 MW reported. [S12]
Prosumer sites reported with storage installations; no aggregate storage MW/MWh in this report. [S12]
139,967 ÷ 373,249. A count-based adoption ratio only; not a capacity or energy share.
Do not silently reconcile incompatible solar totals.
The industry’s approximately 9.1 GW current-solar estimate includes parks and prosumers under its own market framing. ANRE’s 4.149 GW figure above is prosumers only. No attempt is made here to subtract or combine these into a “more precise” national total without a same-date, same-definition operating dataset.
Latest primary national reference identified is older than the review week.
Transelectrica’s Q1 2026 report states that storage facilities totalled 599 MW / 1,129.7 MWh as of 1 Apr 2026. It is useful as a historical reference, not as a 23 Sep operating total. Teiuș and other additions after April mean that simply repeating the April figure as “current” would be wrong; equally, adding selected press announcements to it would not produce an auditable national stock. [S15]
What could make the storage thesis look too easy?
- 20 Sep may be an extreme rather than a new normal. The daylight/base ratio rebounded to 71.0% on 22 Sep, 66.5% on 23 Sep and 75.6% for 24 Sep delivery. One negative-price Sunday cannot support a permanent merchant spread assumption.
- More batteries can arbitrage away their own opportunity. The faster operating storage grows, the more midday charging and evening discharge can compress the simple energy spread and intensify competition in balancing services.
- Authorisation is not construction. ANRE’s 22 Sep decisions are important maturity evidence, but permitting, procurement, financing, connection works and commissioning remain separate failure points.
- Imports are not a battery-only problem. Cross-border economics, wind, hydro, thermal and nuclear availability, internal congestion and outage patterns affect exchange. The +1.02 GW simple average of daily retained exchange means cannot be attributed to insufficient storage alone.
- Pipeline totals can mislead. Company “pipeline,” “backlog,” “advanced development,” “EPC awarded” and “operational” labels describe materially different states. Adding all announced MWh and calling them capacity is not decision-grade analysis.
- Behind-the-meter economics are different. Airport, municipal and industrial systems may monetise avoided purchases, self-consumption and connection value more than wholesale spread. A merchant model cannot be copied into a C&I case without a load profile.
Expect volatility to matter more than the weekly average.
The strongest observable leading indicator available at the research cut-off was already the 24 Sep PZU result: daylight value had tightened materially from the weekend trough. That argues against extrapolating 20 Sep mechanically, but not against continuing to underwrite renewed daylight stress.
Price shape stays unstable, not permanently collapsed.
Weekday demand and changing renewable output keep daylight/base ratios materially above the 20 Sep extreme on some days, while high-solar/low-load intervals remain capable of producing sharp localised discounts. Project execution, funding applications and ANRE authorisation flow remain the dominant non-price themes.
- Assumptions
- No structural market-rule break; no assumption of persistent negative prices; 24 Sep’s tighter next-day signal is directionally relevant but not extrapolated through the week.
- Indicators
- Daily daylight/base ratio, PT15 negative count, midday export/evening import rotations, AFIR application start and new ANRE decisions.
- Decision implication
- Keep asset models at PT15 or at minimum hourly resolution. Use the base price only as context, not as PV revenue.
Daylight value remains closer to the 22–24 Sep range.
Lower renewable availability, stronger working-day load or tighter system conditions could sustain a higher daylight/base ratio and reduce immediate merchant-PV pressure. This would improve near-term solar price realisation but can narrow the simplest energy-arbitrage opportunity for batteries.
- Confirmation
- Several delivery days with positive low PT15 values and daylight/base ratios remaining materially above the weekend trough.
- Who benefits
- Unhedged or partially hedged solar is less exposed to cannibalisation; BESS remains valuable through balancing, system services, connection and contractual revenue rather than relying on one spread.
- What changes
- Do not remove storage from hybrid designs merely because one week’s energy spread tightens. Test revenue-stack substitution instead.
Another sunny low-load block recreates the weekend stress.
If negative PT15 intervals reappear on multiple days while evening scarcity remains pronounced, merchant solar capture pressure would intensify and the operational value of controllability would become more visible. The immediate opportunity for BESS increases, but so does the risk that developers over-capitalise a temporary spread.
- Confirmation
- Repeated negative-price clusters, very low daylight/base ratios and recurring multi-gigawatt midday-to-evening exchange reversals.
- Failure mode
- Financing assumes today’s gross price spread survives large storage commissioning volumes and increasing optimiser competition.
- Response
- Stress-test revenue against spread compression, efficiency, degradation and lower ancillary-service prices before committing irreversible capex.
Decisions for the next underwriting cycle.
These are commercial decision rules derived from the week’s evidence. They are not asset-return forecasts and should be applied only after project-specific technical, legal and financial diligence.
Replace annual-price thinking with interval-level revenue design.
For PV, model PT15 or at least hourly generation against market prices and curtailment constraints. Run PV-only, constrained-PV, hybrid and contracted cases separately. A 946 RON/MWh weekly base price tells you almost nothing about the 20 Sep daylight outcome. Do not label the unweighted 08–17 proxy a capture price; calculate asset-specific capture from matched generation and prices.
Secure route-to-market and grid maturity before optimising the headline spread.
Build the revenue case in layers: wholesale arbitrage, intraday, balancing/ancillary services, connection value and contracted/tolling revenue where available. Then discount each layer for qualification, saturation and execution. Capalo’s market entry demonstrates growing service infrastructure, while Grenergy’s tolling focus demonstrates the search for contracted bankability; neither removes merchant-risk diligence.
Sell compliance and commissioning capability, not containers.
The bid differentiator is moving toward guaranteed usable MWh, power response, degradation terms, EMS integration, fire and safety design, grid-code compliance, protection, telemetry, transformer/substation scope and credible commissioning schedules. Public calls opening 28 Sep add volume, but project owners should prefer executable system integration over lowest equipment-only pricing.
Apply a maturity haircut to every pipeline number.
Keep separate columns for land, ATR/grid access, connection agreement, environmental approvals, ANRE establishment authorisation, EPC, equipment deposits, construction, energisation, commercial-operation status and contracted revenue. A disclosed 918 MWh EPC project, an 816 MWh backlog project, a 176 MWh authorised merchant BESS and a commissioned 120 MWh system are not interchangeable assets.
Size storage against the load curve and connection limit, not the subsidy ceiling.
Bacău and Sibiu airports illustrate a behind-the-meter logic where solar and storage are designed around own consumption. Craiova illustrates a second value source: connection compliance. Buyers should obtain at least 12 months of interval load data, model PV coincidence, verify maximum import/export limits and only then select MW/MWh. Maximising grant-supported MWh is not the same as maximising lifecycle value.
Measure controllable ramp capability by location and hour.
The 23 Sep noon-to-evening exchange reversal demonstrates why aggregate installed renewable MW is insufficient as an adequacy metric. Track local congestion, export capability, dispatchable storage power, usable duration, state-of-charge constraints and evening deliverability. The network needs controllability where and when the imbalance appears, not simply more nominal energy capacity.
Evidence is kept separate from interpretation.
Research cut-off: 23 Sep 2026, 23:59 EEST. Primary public sources were preferred. Company disclosures are accepted for company-specific project claims and labelled accordingly; secondary publications are used where a current primary announcement was not independently accessible.
Market and system calculations
- PZU review-week means are simple arithmetic means of the seven delivery-day references from 17 through 23 Sep 2026.
- The 08:00–17:00 figure is an unweighted price-window proxy calculated from available OPCOM PT15 results; it is not generation-weighted.
- Negative-price duration on 20 Sep is 16 PT15 intervals × 15 minutes = 4.0 hours. This is interval duration, not a claim of continuous battery opportunity after operational constraints.
- Retained Transelectrica hourly values are arithmetic means of all stored public readings within each Europe/Bucharest local hour. Empty hours remain empty.
- Exchange sign convention: positive = net import into Romania; negative = net export.
- The simple +1.02 GW review-week exchange figure is the arithmetic mean of seven daily average-exchange values, not an energy-weighted cross-border settlement figure.
What this report deliberately does not claim
- No national solar capture price is claimed because a harmonised national PT15 solar-generation series was not established in the retained dataset used for this issue.
- No battery gross spread is presented as battery margin. Efficiency, losses, degradation, cycling, fees, grid constraints, imbalance and ancillary-service opportunity cost are excluded from a simple price spread.
- No current national operating BESS MW/MWh total is constructed by adding press announcements to Transelectrica’s older 1 Apr reference.
- No developer pipeline is treated as commissioned capacity. Project maturity is shown explicitly.
- No causal claim is made that storage alone explains Romania’s imports or could alone remove them.
- Industry estimates of total solar capacity are attributed rather than silently reconciled with official prosumer or TSO datasets using different dates and definitions.
What was not repeated from last week
Ford Otosan’s 5.1 MW Craiova PV investment, the 14 Sep final AFM household-battery guide, the initial AFIR programme architecture, Dacia’s Mioveni self-consumption project, the Ișalnița battery, the Porțile de Fier II storage award, PPC’s earlier retail time-of-use signal and the prior ANRE connection-pipeline baseline were already covered or referenced in the 10–16 Sep briefing. They are not repackaged as new developments here. [S03]
The household-battery programme remains relevant, but the important formal milestone — final guide publication — occurred on 14 Sep. A later ministerial reaffirmation is therefore treated as continuity rather than a new programme launch. [S22]
Two supplied headlines required adjustment
Zephyr: PPC’s primary disclosure states 60 MW / 120 MWh. The approximately 129.9 figure refers to project cost in million lei, not battery energy. This report therefore uses 120 MWh. [S05]
VIFOR: 23 Sep reporting stated that the 192 MW phase one was in testing and that its formal inauguration was planned for early October. It is therefore not counted here as a 23 Sep commercial-operation addition. [S23]
Official PZU/PT15 delivery-day market outcomes and 23 Sep market reference.
Primary market operator OpenRetained OPCOM references and Transelectrica observations; hourly arithmetic-mean methodology, no interpolation.
Retained primary-derived dataset OpenPrior-week comparison baseline and continuity control for 10–16 Sep 2026.
Prior research issue Open22 Sep Regulatory Committee agenda and approved-document set; storage licences and establishment-authorisation decisions.
Primary regulator OpenZephyr 1 project specification: 60 MW / 120 MWh, project cost and Modernisation Fund support.
Issuer disclosure OpenBuciumi BESS project: 88 MW / 176 MWh, standalone merchant model, financing status and connection description.
Institutional project disclosure OpenSeptember 2026 investor presentation: Lebăda 204 MW / 816 MWh, RTB and operating targets, Romanian company pipeline.
Investor disclosure OpenCompany-originated Borzești Energy Storage disclosure: 918 MWh, EPC role, CATL partner, estimated execution period.
Commercial self-report Open23 Sep company announcement distributed to press: Teiuș 120 MWh BESS completed and commissioned; €21m investment.
Company announcement / secondary host OpenRomanian market entry, PRE/BRP registration and supply-licence statement.
Issuer disclosure Open22 Sep public-entity solar/storage session notice: opening date, closing date, allocations, aid caps and evaluation sequence.
Primary funding authority OpenProsumer situation at 31 Jul 2026: site count, installed MW and number of prosumers reported with storage installations.
Primary regulator dataset OpenPhase-II multi-location prosumer commercialisation, billing, settlement and value-allocation consultation.
Primary regulatory consultation OpenPhase-II transitional quantitative-compensation consultation for prosumer contracts.
Primary regulatory consultation OpenQ1 2026 report: 599 MW / 1,129.7 MWh storage facilities as of 1 Apr 2026; used strictly as a dated reference.
Primary TSO disclosure Open23 Sep handover of 99.2 MW Green Breeze wind farm; fully operational and supplying the grid.
Issuer disclosure Open22 Sep management update on 220 MW Vișina PV project and expected near-term construction start.
Management statement / secondary OpenIndustry estimates for approximately 9.1 GW current solar and >10 GW around early 2027; explicitly treated as industry estimates.
Industry forecast / secondary OpenSigned design/build contract: 1.25 MW PV, 2.502 MWh storage, >70% own-consumption target, maximum nine-month implementation.
Project disclosure / secondary OpenFinancing contract: 1.70038 MWp PV, 2.088 MWh storage and 71.06% forecast own-use share.
Beneficiary press release / AGERPRES OpenȘimnic proposal: 5.3 MW / 21.2 MWh usable BESS, project cost and reported ATR requirement. Council vote was scheduled after the research cut-off.
Municipal-document reporting Open14 Sep final household battery-programme guide; used for continuity control, not counted as a new 17–23 Sep development.
Primary ministry disclosure Open23 Sep VIFOR verification: 192 MW phase one in testing; formal inauguration reported for early October rather than 23 Sep commercial commissioning.
Secondary verification OpenCompany-reported 209 MW Romanian storage-project portfolio and increasing C&I interest in battery retrofits to PV.
Commercial self-report / secondary Open22 Sep policy statement on accelerating storage and reducing peak-hour import dependence; treated as policy commentary, not a regulatory decision.
Attributed official statement Open© Solar Industry Romania · Weekly research briefing · Week ended 23 Sep 2026. This report distinguishes reported facts, derived metrics, analyst interpretation and forward scenarios. Forward scenarios are directional research cases, not guarantees, securities recommendations or trading instructions. Project figures should be re-verified against current permits, grid documents, financing and issuer disclosures before capital commitment.