Price and capture: the week improved the proxy, not the structural equation.

The seven delivery days split into two regimes. From 13 through 16 August, the daylight proxy still displayed substantial compression: it was 68.3% of PZU base on 13 August, 58.1% on 14 August, 29.8% on 15 August and 55.7% on 16 August. Then system scarcity dominated more of the curve. The proxy/base ratio rose to 81.1% on 17 August, 87.1% on 18 August and 87.3% on 19 August. [S01]

That is a meaningful short-run improvement for merchant solar exposure. It is not yet evidence of structural repricing. Both nuclear units were unavailable, hydro conditions were weak enough to be central to the system discussion, and Transelectrica explicitly described imports as part of the adequacy solution. Under those conditions, scarcity can lift daylight hours as well as evening hours. If firm supply returns, the price curve can flatten by the evening falling, the midday falling, or both. Which part moves determines solar's relative capture outcome.

True capture price remains intentionally unpublished.

A defensible solar capture price is generation-weighted: Σ(price_t × solar generation_t) / Σ(solar generation_t). The retained public evidence layer contains OPCOM PT15 prices and a separate Transelectrica-derived operational solar series, but the public generation scope and settlement alignment are not assumed to be identical. Mixing inconsistent scopes would produce a precise-looking number that the evidence does not justify. The 08:00–17:00 measure therefore remains explicitly labelled a price proxy, not capture price. [S01]

This limitation matters for underwriting. A utility-scale plant's realised capture will depend on its own location, AC export cap, clipping, curtailment, availability, PPA or CfD structure, imbalance route and dispatch constraints. National aggregate solar output cannot substitute for asset-level generation weighting. A lender or sponsor using 69.1% as an assumed future capture rate would therefore be converting a one-week system diagnostic into a long-term asset forecast without the evidence required to do so.

The physical profile is now more important than the daily balance.

Sunday 16 August is the clearest example. At 12:00, retained solar output averaged 2,503 MW against only 2,961 MW of load. Romania was exporting 1,593 MW because the rest of the generation fleet continued to produce alongside solar. By 20:00, solar had fallen to 17 MW and the same system was importing 1,558 MW. The within-day exchange movement was therefore 3,151 MW. [S01]

The pattern repeated on working days. On 18 August, the system moved from 1,793 MW of net exports at 15:00 to 938 MW of net imports at 20:00. On 19 August it moved from 686 MW of net exports at noon to 1,874 MW of imports at 20:00. These are not battery-profit calculations. They are physical indicators of how much the net system position can change over the hours in which solar output falls and evening demand rises.

This is why national installed-solar capacity is an incomplete adequacy metric. Solar can materially suppress daytime net demand, reduce thermal dispatch and create exports; none of those benefits automatically survives into 20:00–22:00. The system needs resources that can shift energy across time, change demand, retain dispatchable output, or move electricity across borders at the required hour.

Părău 2 changes the financing conversation more than the capacity conversation.

The 342 MW solar / 150 MW / 300 MWh battery configuration implies a two-hour battery and a BESS power rating equal to about 43.9% of the disclosed solar MW figure. Those ratios are derived from EBRD's project disclosure; they should not be generalised into an optimal Romanian hybrid design. They do, however, show the scale at which a major lender is willing to underwrite co-located flexibility alongside PV. [S06]

The financing structure is more instructive than the headline battery size. EBRD is providing up to €120 million within an aggregate debt package of up to €229 million. The EU supports the EBRD A/B loan through an InvestEU first-loss guarantee of up to €115 million because the project retains substantial merchant-revenue exposure. A 125 MWac part of the solar plant has a 15-year CfD at €49.4/MWh; the remainder is merchant. This means the financing precedent is not a pure merchant-battery bet. It is a blended project in which contracted renewable revenue, merchant exposure, storage optionality and credit enhancement coexist.

For future hybrids, that architecture suggests a more rigorous question than “what is the arbitrage spread?” A lender needs to know which cash-flow risk each component is solving. Storage can shift output, reduce imbalance exposure, support PPA shaping, participate in balancing or ancillary services and potentially help operate within an export envelope. It cannot prudently be credited with the full value of every stack simultaneously. Revenue streams compete for state of charge, cycling budget and availability.

The €150 million standalone-storage scheme rewards capital efficiency, not only eligibility.

The state-aid scheme has a superficially simple headline: €150 million for standalone batteries. The competitive mechanics are more important. The scheme targets a minimum 2,174 MWh, accepts new standalone storage connected at transmission or distribution level, sets a minimum power of 1 MW, requires at least two hours of duration, caps aid at €69,000/MWh and €15 million per undertaking, and can cover up to 100% of eligible investment cost subject to those limits. Costs outside the eligible perimeter and the balance of total project cost remain the beneficiary's responsibility. [S08]

Dividing the full €150 million budget by the 2,174 MWh minimum target gives approximately €68,997/MWh. That is effectively the scheme's €69,000/MWh ceiling. The arithmetic reveals the policy design: if every winning MWh were supported at the maximum, the budget would only just cover the minimum target. If competition drives requested aid materially below the ceiling, the same budget can support more than 2,174 MWh, subject to eligible applications and awards.

The ranking rule is therefore commercially material. The project asking for the lowest amount of aid per MWh receives the maximum technical-economic score for that criterion, while the project at the highest requested aid level receives the lowest score, with intermediate values scored linearly. Developers that optimise capex, financing, connection cost and merchant revenue expectations can improve both economics and award probability by asking the state for less.

Another important detail is connection maturity. The legal scheme does not require the technical connection approval, ATR, at the moment of the funding application, but requires it before the first payment request. That widens the pool of applications while preserving a later grid-deliverability gate. It also creates execution risk: a funding award is not a substitute for securing a workable connection.

The scheme is explicitly for standalone storage. Developers should not silently transpose its economics onto co-located PV+BESS projects such as Părău 2. Hybrid storage and standalone storage can compete in many of the same market intervals, but they sit inside different development, connection, funding and dispatch constraints.

Demand response adds a new competitor to the flexibility stack.

ANRE's 19 August regulation is strategically important because it changes the supply side of flexibility. Until now, much of the investment discussion has framed evening scarcity as a contest between batteries, hydro, thermal generation and imports. The new mechanism allows eligible consumption itself to bid a reduction quantity and price, either directly or through suppliers and aggregators. Transelectrica selects the economically most advantageous offers up to the required volume, and delivered reduction is checked against the relevant consumption baseline. [S10]

For a battery developer, the immediate implication is not that demand response destroys BESS economics. The initial market size, frequency of activation, clearing prices, baseline quality, participation depth and interaction with balancing products remain unknown. The implication is that scarcity value cannot be assumed to belong exclusively to storage. Flexible industrial load can sometimes provide a cheaper MWh of system relief by not consuming it in the first place.

For C&I consumers and aggregators, the logic is reversed. A load profile that was previously only a procurement liability can become a market asset if production processes, thermal inertia, backup systems or scheduling allow a verified temporary reduction without excessive operational cost. The commercial threshold is the customer's real cost of interruption or rescheduling, not the wholesale price alone.

ANRE also explicitly separates the mechanism from rationing. That distinction became important within hours because public discussion suggested industrial disconnections had already been ordered. The Ministry of Energy denied that interpretation. A system under stress can therefore produce two very different instruments: voluntary paid flexibility and compulsory emergency limitation. The evidence reviewed for 19 August supports the first as an approved market mechanism and does not establish that broad compulsory industrial curtailment had already been activated. [S15]

Residential batteries moved closer to a real market, but the programme is still draft policy.

The proposed 400 million lei programme would support batteries for existing physical-person prosumers regardless of whether the original photovoltaic system came through an AFM scheme or private funding. The draft proposes up to 15,000 lei including VAT, no more than 75% of eligible cost and no more than 1,250 lei for each installed kWh, with a minimum 12 kWh storage capacity. [S09]

The system rationale is consistent with this week's operating profile: distributed PV generates when the system can already be long, while evening load persists after solar disappears. Behind-the-meter storage can reduce household exports at low-value hours and grid withdrawals later. But residential BESS value is not identical to utility BESS value. A home battery primarily optimises self-consumption and retail exposure; it does not automatically have access to wholesale arbitrage, ancillary-service or balancing revenue.

The commercial warning is maturity. The guide is in consultation. Installers and distributors can prepare product matrices, procurement options and compliance pathways, but should avoid presenting the 15,000 lei level, 12 kWh minimum or final application process as a guaranteed live subsidy until the final guide and opening terms are published.

Grid deliverability has two different problems: MW transfer and electrical strength.

Last week's briefing concentrated on the enormous gap between projects holding valid ATRs and projects that had accumulated later development gates. This week adds a different grid layer. The two ±150 MVAr STATCOM installations at Sibiu Sud and Bradu provide dynamic reactive-power support and voltage control. Those functions become more valuable as synchronous generation is displaced and variable inverter-based generation grows. [S11]

But a STATCOM is not an extra 150 MW or 300 MW of grid export capacity. MVAr measures reactive capability; MW measures active power. A solar developer should therefore ask two separate connection questions: whether the network has enough active-power transfer capacity at the relevant node and contingency state, and whether voltage/reactive/stability conditions can support the proposed inverter-based generation. Improving one does not automatically solve the other.

This distinction matters for storage as well. A battery can be electrically valuable at a constrained node even if national energy-arbitrage spreads compress. Conversely, a battery located where connection rights, charging conditions or network limits prevent useful operation may not monetise the national price spread visible on OPCOM. “Romania needs storage” is a system statement; it is not a project-specific revenue guarantee.

The Dama headline did not change again this week.

The 18 August Balkan Green Energy News article described the establishment authorisation as the project's “final permit” and cited 1.24 GWp / roughly 1.07 GW grid connection. The underlying authorisation was already the central project-development item in Solar Industry Romania's 6–12 August briefing. Treating the later publication date as a second permitting milestone would inflate apparent weekly project progress. [S14] [S13]

The next Dama evidence that should alter a decision model remains execution evidence: financial close, EPC notice to proceed, final AC/DC and export definitions, current BESS MW/MWh design, connection works and contractual treatment of the output outside the supported CfD tranche. Repeated media characterisations of an existing approval do not change those unknowns.