One percent.

That’s how many Romanian renewable projects that posted the old grid connection guarantee actually cleared all three connection milestones. Not ninety. Not fifty. One.

Every buyer running BESS due diligence in Romania budgets for the obvious fights — DSCR stress tests, revenue stack assumptions, the augmentation reserve nobody wants to fund early. Those get checked. Those show up on every template.

What doesn’t show up is the five-item list below, and every one of them has killed a real deal timeline somewhere in the last eighteen months. This piece covers both halves: what each risk actually costs, and — because flagging a problem without a fix is just a more expensive way of scaring people — exactly what to do about it before it becomes a renegotiation.

Why the Deals That Die in Romania Rarely Die on the Model

Somewhere in a data room right now, an analyst is checking a box that says “building permit: obtained” and moving to the next tab. The box isn’t lying. It’s just answering a narrower question than the one the buyer thinks they’re asking.

Romanian permitting and grid-connection law is built around specific, narrow instruments — a permit tied to one parcel, an approval scoped to one road authority, a technical opinion that expires on a clock nobody’s tracking.

A due diligence checklist built for a different market — or even a checklist built for Romania two regulatory cycles ago — reads these instruments as boxes to tick. It misses that each one covers exactly what it says and nothing more.

Five gaps show up over and over. Here’s what each one actually is, what it costs when it’s missed, and how to close it before it’s a crisis instead of a line item.

Risk 1: The Building Permit That Doesn’t Cover the Whole Project 🏗️

Romania issues construction authorizations (autorizație de construire) per cadastral parcel. Your substation footprint sits on one parcel. Your cable route to the connection point almost never does — it crosses other landowners’ plots, easements, sometimes multiple parcels in sequence, and each one is its own permitting scope, not an extension of the first.

The legal review window is 30 days per complete file. Fine, for one file. A different situation entirely when the cable route needs its own file for every parcel it crosses, filed and reviewed one at a time rather than as a single package. “Building permit: obtained” can be completely accurate and still describe a project that’s only permitted at the substation and nowhere along the 4km of cable route between there and the grid.

How to de-risk it: Commission a full route survey — every parcel the cable crosses, named and mapped — before signing an LOI, not after. Treat “permitting status” as a per-parcel table, not a single yes/no field, and require the seller to produce it as part of the data room rather than accepting a summary statement.

Where possible, file applications for every parcel in parallel through a dedicated permitting consultant rather than waiting for one to clear before starting the next; nothing in the process requires them to run sequentially, only the seller’s own project management sometimes does. Build permitting float — at minimum 60-90 days beyond the sum of individual review windows — into the SPA’s conditions precedent and long-stop date.

Risk 2: Grid Works That Run on the TSO’s Calendar, Not Yours 🔌

An ATR (Aviz Tehnic de Racordare — the technical connection approval) promises capacity. It says nothing about a delivery date, because the physical work behind it — reconductoring existing lines, upgrading substations — sits on Transelectrica’s own capital investment schedule, and that schedule doesn’t move because your FID date does.

The scale of the mismatch is the headline number here: Romania currently has a 66–70GW renewable pipeline in development, and industry estimates put realistically connectable capacity at roughly 10% of that.

Since the 2026 grid connection reform, the obligation to secure land ownership for the reinforcement works themselves has shifted to the developer too, starting from the date the connection contract is signed. The “waiting on the TSO” risk didn’t disappear. It picked up a property acquisition obligation on top of it.

How to de-risk it: Cross-check the specific reinforcement or reconductoring work your connection point depends on against Transelectrica’s published network development plan (Planul de Dezvoltare a Rețelei Electrice de Transport, covering 2024–2033) before treating a connection date as reliable — a project sitting on funded, scheduled works is a fundamentally different risk than one waiting on work that’s still aspirational.

Get written confirmation from the seller of exactly which works the connection depends on, not just the ATR’s headline capacity figure. And since land acquisition for reinforcement works is now a developer obligation, confirm during DD whether that land has actually been secured, optioned, or merely identified — those are three very different states with three very different timelines to close the gap.

Risk 3: The Approval Your Urbanism Certificate Never Mentions 🛣️

The certificat de urbanism lists the approvals tied to the site parcel. It does not list CNAIR (Compania Națională de Administrare a Infrastructurii Rutiere — Romania’s national road authority) approval for crossing a national road with an underground cable, because that approval was never a site question. It’s governed separately, under CNAIR’s own technical norm on placing electrical cables in the vicinity of national roads, and it answers to a different authority entirely.

A DD team working strictly from the urbanism certificate as “the list of everything required” has no reason to suspect CNAIR sign-off is missing — it was structurally never going to appear on that document. It surfaces only if someone specifically asks whether the cable route crosses a national road, and specifically follows up on the answer.

How to de-risk it: Overlay the cable route against Romania’s national road network before signing anything — this is a straightforward GIS exercise, not a legal mystery, and it tells you immediately whether CNAIR is even in scope. If it is, request a preliminary CNAIR opinion early rather than waiting for construction-readiness, since this approval runs on its own timeline independent of every other permit in the project.

Add “CNAIR sign-off status, if applicable” as its own explicit line in every DD checklist template going forward — not a sub-bullet under building permits, a separate line, because it’s a separate authority with a separate process.

Risk 4: The Committee at the Prime Minister’s Office 🏛️

Capital from outside the EU into a sensitive sector — energy infrastructure included — can trigger mandatory review by CEISD (Comisia pentru Examinarea Investițiilor Străine Directe), Romania’s foreign direct investment screening authority, created under Government Emergency Ordinance 46/2022 implementing EU Regulation 2019/452.

It is not a formality: the committee includes representatives from Romania’s intelligence services, and the Ministry of Energy has already referred at least one energy-sector acquisition — MVM Grup’s proposed purchase of E.ON Energie România — into this exact process.

The threshold is €5 million for non-EU investors in a covered sector. The clock: up to 45 days for a standard review once the file is complete, escalating to a 90-day in-depth assessment extendable by a further 45 days if the case is flagged — and that’s before the 30 days (plus a possible 15-day extension) an investor may need to answer information requests just to get the file declared complete.

Call it up to 135 days of committee review, layered on top of however long it takes to get a complete file in front of them in the first place.

How to de-risk it: Determine non-EU investor status and sector sensitivity at term-sheet stage, not after signing — this is a five-minute classification exercise that changes the entire deal timeline if the answer is yes. Build the full review window into the SPA’s long-stop date explicitly, rather than treating it as background regulatory risk that “shouldn’t be a problem.”

Where the ownership structure genuinely allows for it, some sponsors route non-EU capital through an EU-domiciled holding vehicle — whether that’s appropriate for a specific deal is a structuring question for qualified Romanian and EU counsel, not something to decide off a Substack article, but it’s worth raising with counsel early enough that it’s actually still an option.

Risk 5: ATR in Hand Isn’t a Grid Connection 💸

These get treated as interchangeable in casual deal conversation. They are not. An ATR is a 12-month offer — it becomes contractually binding only once accepted and the connection tariff is paid, and it lapses automatically if that payment doesn’t happen within 3 months.

Under the current reform, even after the connection contract is signed, the developer has 12 months from signing (18 months from the original ATR date, whichever comes first) to complete the setting-up authorization — miss that window, and both the ATR and the connection agreement terminate automatically, with no extension by default.

“The project has ATR” and “the project has a grid connection” are two sentences that sound like they mean the same thing. They don’t. One is a time-limited option a missed deadline can void without warning. The other is the asset itself.

How to de-risk it: Never underwrite a deal on ATR status alone — confirm, in writing, whether the connection contract is signed, whether the connection tariff has been paid, and the exact date both clocks (12-month contract, 18-month ATR) started running.

Make current connection-contract status, not ATR issuance, the condition precedent that actually matters in the SPA. If a target is still on ATR only, price the deal — or the timeline — for the possibility that the window closes before the deal does.

Building This Into How You Actually Run DD

None of these five risks require a specialist nobody has access to. They require asking a slightly more specific question than the standard checklist asks, and asking it early enough that the answer still has time to change the outcome.

We’ve written before about what banks actually require before they’ll move a BESS project toward financial close — the lender-facing version of this same discipline — and about what happens once a project reaches the debt committee stage, where several of these same gaps show up again from the other side of the table.

The one percent this piece opened with isn’t a talent problem, and it isn’t a market problem. It’s five specific documents that most DD templates were never built to catch — plus, now, five specific ways to catch them anyway. ⚡

No Romanian bank wants to carry €300 million of exposure to a single battery storage project on its own balance sheet. That isn’t a battery-specific problem — it’s what happens to any energy financing once the ticket size crosses a threshold no single lender’s risk appetite will absorb alone. There’s a standard answer to it, and it’s older than renewables: syndication.

BESS syndicated loan financing works the same way syndication works for any large energy asset. A lead bank — the arranger — structures the credit package, sets the terms, and then sells down pieces of it to other banks and international financial institutions (IFIs) until the full ticket is covered.

The borrower signs one credit agreement. The lenders each hold a slice of the same exposure, priced and governed identically. What changes project to project isn’t the mechanic — it’s who’s in the room and how the risk gets sliced.

What Is Loan Syndication, and How Does the Structure Actually Work?

Loan syndication is the process of multiple lenders jointly funding a single borrower under one shared credit agreement, each holding a defined portion of the total exposure. Two structures show up most often in Romanian and broader European energy finance:

Club deals, where a small, pre-selected group of banks — usually the ones already relationship-lending to the sponsor — split a ticket relatively evenly and negotiate terms together from the outset. These suit mid-size financings where the sponsor already has established banking relationships and doesn’t need to go to a broad, unfamiliar lender base.

Arranged syndicates, where a lead bank or IFI structures the deal first, then distributes participations more widely — often including an A/B loan structure. In an A/B loan, the anchor institution (commonly a development bank like the EBRD) holds the “A” tranche directly on its own books, then syndicates a “B” tranche out to commercial banks and other IFIs, who lend under the anchor institution’s preferred-creditor status and due-diligence work rather than running the full credit process from scratch themselves.

That second structure matters most for BESS. It’s the mechanism that lets a technology class without a long Romanian repayment track record borrow at scale anyway — because commercial banks aren’t underwriting the battery in isolation, they’re underwriting alongside an institution that already has.

Why Energy Projects Reach for Syndication in the First Place

The trigger is almost always concentration limits, not the borrower’s creditworthiness. Every bank caps how much exposure it will hold against one borrower, one sector, or one technology class still building its track record — regardless of how strong an individual project’s numbers look.

Read more

Paks nuclear in Hungary came within a day of shutting down.

Not a controlled maintenance pause — a full, unplanned shutdown of Hungary’s only nuclear power plant, the one that supplies nearly half the country’s electricity.

Prime Minister Péter Magyar said it out loud over the weekend: the Danube’s cooling water intake at Paks had fallen to –138 cm, the lowest ever recorded there, and the plant’s fixed-elevation pumps were running out of river to pull from. Output was already cut 90%. In its 44-year operating history, Paks has never fully shut down. This was the closest it’s come.

Four hundred kilometres downstream, Romania was running its own version of the same emergency. Cernavodă had one reactor down and the other clinging on at half output, cooled by water that authorities were, quite literally, trying to redirect by sinking rock-filled barges into the riverbed. It worked — barely.

Eight centimetres of extra flow bought the plant nine more days instead of four or five. The government legislated emergency powers to disconnect industrial consumers with 24 hours’ notice, and Dacia and Ford moved up their August maintenance shutdowns to shave load off the grid.

Serbia’s Đerdap dam — the largest hydropower plant on the entire river — was producing a third of its normal output. Slovenia’s Krško nuclear plant was curtailed 20%. Regional day-ahead prices blew past €700/MWh.

This wasn’t a Romania story. It was never a Romania story. It’s a Danube story, and the Danube touches ten countries.

The one country that wasn’t in the news this week

Here’s the detail that didn’t make the wire copy: while Hungary was calculating how many hours Paks had left, Bulgaria was exporting power into the crisis.

Bulgaria has 4.8 GW and roughly 14.7 GWh of battery storage online right now — more than double what Romania has, running for an hour longer per cycle. It’s on pace to close the year near 7 GW / 20.8 GWh, built substantially on the back of nearly €600 million in EU grant funding through a national storage auction programme. By relative size to its own grid, Bulgaria is the number one battery storage market on Earth. Not in Europe. Earth.

None of that was an accident, and none of it happened this week. Bulgaria placed this bet years before the Danube hit record lows. And when the region’s nuclear and hydro fleets started failing in near-identical ways this summer, Bulgaria wasn’t the country scrambling to sink barges — it was the country selling power to the neighbors who were.

That’s not a coincidence worth glossing over. That’s the entire argument for battery storage, delivered by the market itself, in real time, with an invoice attached.

Is this actually getting worse, or does it just feel that way?

It’s getting worse, and there’s a peer-reviewed answer, not just a vibe.

Read more

Reactor 1 at Cernavodă went dark at 11:00 on a Tuesday morning. Not a fault. Not a leak. A river.

That’s not a metaphor — and it’s also the clearest battery storage bankability argument Romania’s market has produced all year.

By July 28, the Danube’s flow into Romania at Baziaș had fallen to roughly 1,650 cubic metres per second — about a third of what should be moving through in a normal July. Days later it was still falling, toward 1,600. The multiannual July average is 4,750.

Do that math twice, because it’s worth sitting with: Europe’s second-longest river was running at barely a third of normal, and Romania’s only nuclear plant needs that water to cool itself.

So Nuclearelectrica shut Reactor 1 down in a controlled sequence. Reactor 2 was scheduled to follow the next morning.

It didn’t — an overnight technical review found a way to keep it running — but for about twelve hours, close to a fifth of Romania’s electricity supply was one bad number away from zero.

This Wasn’t Supposed to Happen Twice in One Decade

Romania’s reservoir levels are sitting at 71% — the lowest recorded for this point in the year since 2018, per Transelectrica’s own published data. Hidroelectrica, the state hydro operator, told investors its output was down a third in the first half of 2026 alone.

Water levels this low aren’t a Romania problem. They’re a river problem, and the river runs through five countries.

Serbia’s Đerdap plant — the shared hydro facility at the Iron Gate, one of the largest in Europe — is running somewhere between a fifth and a third of capacity depending on which day you check.

The government held an emergency energy session. Two towns declared local states of emergency because the water near Bezdan measured minus 131 centimetres.

Hungary’s Paks plant supplies close to half the country’s electricity. It just cut 491 megawatts because there isn’t enough cooling water to run at full output, and the state utility says a complete shutdown is “inevitable” if conditions don’t improve. The Prime Minister has already floated the possibility of outages for large industrial users.

Bulgaria’s only nuclear plant, Kozloduy, is on the same warning list.

Four countries. One river. One failure mode, hitting hydro and nuclear at the same time, because both of them need water and the water isn’t there.

I keep coming back to one detail buried in the forecast. Water levels aren’t expected to recover in the coming days, or even weeks — no significant rain is forecast, and a fresh heatwave is moving in behind this one, with daytime highs above 38°C. This isn’t a bad week. It’s a bad season sitting on top of a bad decade.

Here’s the Uncomfortable Part

The industry’s go-to answer for renewable intermittency has always been more water. Pumped hydro storage — the so-called “water battery” — just crossed 200 GW of installed capacity globally, and Spain and Portugal responded to last year’s blackout by announcing 3.5 GW more of it.

Water storage solving a water problem sounds fine until the water itself is the thing that’s missing.

A pumped hydro plant needs two reservoirs and enough flow to cycle between them. A nuclear plant needs a river to cool its core. Right now, across four countries, both categories of asset are constrained by exactly the same cause.

The “water battery” doesn’t just fail to help during a drought — in some configurations, it’s part of what’s failing.

A lithium BESS doesn’t touch the river. It charges off whatever’s on the grid, discharges into whatever’s short, and the Danube’s flow rate is irrelevant to every part of that cycle.

That’s not a marginal advantage. During a correlated hydro-and-nuclear shortfall, it’s the only dispatchable asset class in the regional stack that is structurally immune to the failure mode currently taking down a fifth of Romania’s nuclear capacity and a third of its hydro output.

There’s a second layer to this that gets less attention than the energy shortfall itself: inertia. Large synchronous units — the nuclear reactors, the hydro turbines — don’t just supply energy, they supply grid stability, the physical momentum that keeps frequency steady when something trips.

Pull several of those units offline at once across four interconnected countries, and the grid loses stability reserves at the exact moment it can least afford to.

Battery storage is one of the few assets that can supply frequency regulation (FCR/aFRR — frequency containment reserve and automatic frequency restoration reserve) within seconds, without needing a drop of water to do it.

What Is Driving the Bankability Case for BESS Right Now?

The short answer: price. Romania’s day-ahead market hit €224/MWh on June 29 — an EU high for that day, roughly two and a half times what Bulgaria (€90) and Greece (€87) were trading at on the same market, same hour. That gap is what happens when a grid loses dispatchable capacity faster than it can replace it.

That matters on the debt side of the table too. A project stress-tested to a 1.35x DSCR (Debt Service Coverage Ratio) floor across the full loan tenor — not just year one — is explicitly built to survive a bad year. A drought that spikes balancing-market prices doesn’t threaten that coverage ratio. It pads it.

I’ll say the quiet part. Most BESS pitches lean on renewables growth and grid congestion to make the bankability case. Those are real, but they’re slow variables — multi-year trends that ask an investment committee for patience.

A four-country nuclear-and-hydro drought making front-page news this week is not a slow variable. It’s the fastest, most legible argument for battery storage this market has produced in years, and it’s unfolding while the ink on this piece is still wet.

Why This Keeps Coming Back

2018, 2022, now 2026. Three Danube droughts severe enough to hit power generation inside eight years isn’t noise — it’s a pattern with a shrinking return period.

A 2026 study tracking EU-wide drought years found fossil generation had to backfill an extra 180 TWh over seven years to cover hydro shortfalls — 7% of the entire EU’s generation in the worst year — at a cost of 141 million extra tonnes of CO2. Somebody pays for that backfill. Right now it’s gas plants and consumers.

It could be batteries, at a fraction of the emissions and, increasingly, a fraction of the cost per MWh delivered during the exact hours it matters.

None of this means BESS replaces nuclear or hydro. It means the grid needs an asset in the stack that doesn’t share their single point of failure. That’s not a hypothetical anymore. It’s what Cernavodă proved on a Tuesday morning in July.

Before the Webinar — the Last Word on This

I’m walking through the Romanian market’s financing mechanics — grid connection, DSCR structuring, revenue stacking, what “bankable” actually means under exactly this kind of volatility — live this Wednesday, August 5, 6:00–7:30pm Romania Time (EEST). Register here — it’s two days out, and given what the Danube is doing to this market right now, I’d rather you hear the mechanics from someone working them than from a headline.

For the regulatory side of why this window is tighter than it looks, I broke down the ATR guarantee deadline squeeze in Romania’s BESS Buyer’s Market Has an Expiration Date.

And if you want to see how Romania’s revenue numbers stacked up against Hungary, Bulgaria, and Greece before this drought even hit, that comparison is in Romanian Energy Storage Market vs. Three European Peers.

The river will come back up eventually. The next drought will too. The only real question is whether the grid has an asset in place by then that doesn’t care either way.

Note on the use of AI

Yes, I use AI in researching & drafting the content that I put out. That doesn’t mean I just post what Claude spits out.

A reliability protocol is embedded in the way the content is generated:

1. Research from trusted sources (you will find it below);
2. A cannoic benchmark of what is being generated vs. a set of benchmarks that are meant to reduce any slop that may come out;
3. HiTL (human in the loop) - I always check & edit what comes out to make sure there’s no BS slipping in between the lines.

Sources:

• Record low Danube levels push some countries in Eastern Europe to the brink of energy emergency

• Water level in Romanian dams at lowest level since 2018 — Romania Insider

• Romania shuts down both nuclear power reactors as Danube levels plunge — Romania Insider

• Danube’s record low leads to Hungary, Romania nuclear shutdowns — World Nuclear News

• Romania Declares State of Alert as Power Output Falls on Danube Drought — Bloomberg

• Extreme heat pushes Romania’s electricity prices to EU record high — Caliber.Az

• Low Danube levels force Serbia to cut hydro power generation — Daily Sabah

• Historic Danube water lows signal blackout risks — Balkan Green Energy News

• Europe’s power grid has a big drought problem — TechXplore

Transelectrica publishes the validated bidder list for October’s grid capacity auction. Nine days ago, a much quieter deadline passed — one that matters more, and almost nobody outside a small circle of specialists noticed it.

Both dates are the same story from two different angles: Romania’s BESS pipeline is being cut down to size, on a clock that already started.

I’m hosting a live briefing on exactly this on August 5 — where the pipeline actually stands right now, what’s about to get filtered out of it, and what that does to pricing for whatever survives. If you’re a fund, family office, or strategic investor evaluating Romania for BESS exposure, registration is here.

The full session details are at the end of this piece — for now, here’s why the timing question matters more than the thesis question.

1. ❓ How Much of Romania’s BESS Pipeline Is Actually Real?

Romania’s approved BESS and renewables pipeline sits at 81,322 MW across 1,431 projects holding valid ATRs, according to ANRE’s April 2026 snapshot. Of that, only 107 projects — 3,843 MW — were on track to reach commercial operation in 2026. That’s roughly 4.7% of the approved capacity.

The 81,322 MW figure is the number that gets quoted in every pipeline overview. The 3,843 MW figure is the number that actually matters to an investor deciding where to deploy capital this year.

Everything between those two numbers — the other 95.3% — is paper: real permits, mostly unreal projects behind them.

That gap has existed for a while. What changed in the last two months is that ANRE built a mechanism specifically designed to close it — and started the clock on it before most of the market had read the order.

2. 🕐 The Clock That Already Started

On 25 May 2026, ANRE published Order no. 15/2026. It raised the standard grid-connection financial guarantee — the capital a developer has to post just to keep an ATR alive — from 5% of the connection tariff to 20%. A four-fold increase.

It also attached teeth: projects with evacuation capacity above 1 MW now have 12 months from signing their grid connection contract, and no later than 18 months from ATR issuance, to obtain a setting-up authorization, or the position lapses.

Here’s the detail that should reframe how you read this market today rather than next year. Anyone who had registered for capacity allocation before the order took effect had until 14 July 2026 to post the new, higher guarantee.

That deadline was nine days ago.

This isn’t a filter that starts running at some point in the future. It’s already running. The market just hasn’t caught up to what that means for pricing yet.

ANRE’s own historical data tells you roughly what the filtered population looks like. Under the old, cheaper 5% guarantee, only 12% of projects ever signed a connection agreement. Only 3% obtained a building permit.

Only 1% made it through all three milestones — signed agreement, building permit, setting-up authorization. If that conversion pattern holds even loosely under a guarantee that just quadrupled, the number of ATRs surviving this filter is a small fraction of 81,322 MW, and it’s about to become visible to every fund evaluating this market at roughly the same time.

3. 🔨 The Separate Squeeze on New Capacity

The obvious response — if the existing pipeline is thinning, just file for new capacity — runs into a different wall entirely.

Romania’s competitive grid-capacity auction (Ordinul ANRE 53/2024) replaced first-come-first-served allocation for any new generation or storage request of 5 MW or more.

This is a distinct mechanism from Order 15/2026 above — worth keeping separate, because both are compounding the same squeeze from different directions. Registration for the first auction round closed 14 July 2026.

Transelectrica publishes the validated bidder list tomorrow, 24 July. The first daily auction sessions don’t begin until 30 October 2026, and the capacity they allocate covers a 10-year window starting in 2028.

I covered the mechanics of this auction — the EUR 20,000/MW bid bond, the reinforcement-priced starting bids, the sub-zones currently publishing zero megawatts available without reinforcement — in detail in the earlier grid-auction deep dive.

The short version for this piece: a greenfield developer filing today has no realistic path to commercial operation before 2028.

Greenfield isn’t a near-term option anymore. It’s a 2028-and-later option, at a materially higher entry cost than most current pipeline overviews have priced in.

4. 📊 Old Regime vs. New Regime, Side by Side

Pre–May 2026 Post–May 2026 (Order 15/2026)

5. 📈 Buyer’s Market Today, Seller’s Market Tomorrow

Put the two mechanisms together — Order 15/2026 thinning the existing pipeline, Ordinul 53/2024 closing greenfield until 2028 — and the market-timing question answers itself.

Right now, capital allocators can still survey a pipeline that looks wide, most of which nobody has correctly sorted into real versus non-viable yet. Diligence is cheap relative to the asset precisely because sellers themselves often don’t know which of their ATRs survive the guarantee deadline and the setting-up authorization clock.

A buyer willing to do that diligence properly is pricing against the real 4.7%-and-shrinking population, while the rest of the market is still anchored to the 81,322 MW headline.

That mismatch is the buyer’s market. It exists because information is still asymmetric and the filter hasn’t finished running.

It closes the moment that changes. Once the guarantee deadlines force the non-serious majority of that backlog to lapse — and once the validated auction list and a visibly thinner ATR pool make it obvious to every fund in the market how small the real, near-term-viable pipeline actually is — every serious investor is competing for the same shrinking set of assets. That’s a seller’s market, and on this timeline, it’s a matter of months.

None of this changes the underlying economics once you find a real asset. A 150MW/300MWh Romanian project still needs to clear a 1.35× DSCR floor stress-tested across the full debt tenor, still needs an equity case built on a 12% IRR floor, still produces its revenue from the same DAM arbitrage, FCR/aFRR, and balancing mix.

What’s changing is which projects get to use those numbers at all, and how much competition there is to find out which ones qualify.

6. 🔍 The Only Real Option Also Has a Clock

Here’s the detail that should sharpen the urgency rather than soften it. With greenfield effectively closed until 2028, the existing ATR pool isn’t just the best near-term option for Romanian BESS exposure — it’s the only one for any investor who wants revenue-generating capacity before the end of this decade.

But that pool is exactly what Order 15/2026 is cutting down, on a clock that started nine days ago and runs through the 12-to-18-month setting-up authorization window. The asset class an investor needs to buy into is shrinking at the same time its price floor is about to move upward.

Finding the genuinely real projects inside that pool — credible sponsor, valid land control, a defensible path through the setting-up authorization deadline — before the filter finishes running, is the opportunity in this market right now. Once it’s obvious to everyone, it’s already priced in.

Join the Live Briefing — August 5

I’m walking through this pipeline in detail — what’s already been filtered, what’s still viable, and what a defensible entry into the surviving pool looks like — live, for investors, on August 5.

Date Wednesday, August 5, 2026 Audience Investors and capital allocators evaluating BESS market entry in Romania Topics State of the ATR pipeline post-filter · auction mechanics and the 2028 commissioning timeline · why this is a closing buyer’s market · live Q&A Format Live, ~45 minutes + open Q&A Registration Register here

No pitch deck, no sales angle — bring your hardest question about whether this window is really as short as it looks.

Register for the August 5 briefing →

For how Romania compares to its regional peers on the underlying economics, see the four-market BESS comparison.

RENEWD advises developers, lenders, and investors navigating BESS project finance, grid connection strategy, and bankability in Romania and CEE.

Hungary €800. Bulgaria €798. Greece €797. Romania €792.

Eight euros separates first place from fourth on Europe’s newest battery storage revenue ranking. If you’re choosing where to put capital based on that number alone, you’re choosing on noise, not signal.

The real question isn’t which of these four markets pays the most. It’s which one actually gets your project built, financed, and generating that revenue on schedule. Those turn out to be very different rankings.

1. ❓ Which European Market Pays BESS Investors the Most in 2026?

Hungary currently posts the highest day-ahead arbitrage spread in Southeastern Europe at €800/MW/day, according to ENTSO-E’s latest Market Report covering the twelve months to June 2026 — narrowly ahead of Bulgaria (€798), Greece (€797), and Romania (€792). All four sit far above the European average floor of roughly €215/MW/day.

Market ENTSO-E Spread (EUR/MW/day) vs. EU Floor Hungary €800 ~3.7× Bulgaria €798 ~3.7× Greece €797 ~3.7× Romania €792 ~3.7× EU average floor €215 1×

Worth saying plainly: these four numbers are statistically a rounding error apart. ENTSO-E’s figures are realised day-ahead prices, not volume-weighted — a price-structure indicator, not a promise of what any individual battery actually captures.

On revenue potential alone, an investor has no real basis to prefer one of these four markets over the other three.

So the ranking that actually matters starts here — not with the price, but with what stands between that price and a commissioned, revenue-generating asset.

2. 🇭🇺 Hungary — The Highest Number, the Least-Tested Market

Hungary’s headline is real, but the market behind it is early-stage. Roughly 60–70 MW of storage is currently operational, with another 550 MW licensed and expected online by mid-2026. Trade coverage describes the market itself as “underdeveloped,” with limited competition — which cuts both ways for an investor.

The bigger open question is regulatory: Hungary’s new competitive grid-capacity allocation mechanism — national-level tenders, evaluated by a five-member committee — was expected to launch in Q2 2026 and, as of this writing, is a market still waiting on its first real test run.

A mechanism with no completed cycle yet is a mechanism with no track record. That’s not disqualifying. It’s a different risk profile than a market where the process has already run.

3. 🇧🇬 Bulgaria — The Most Capital, the Nearest Cliff

Bulgaria is, by a wide margin, the most capital-proven market on this list. Roughly €2 billion has flowed in over two years, ~10 GWh has been awarded under the EU-funded RESTORE program, and the country now holds close to 3 GW of operational and contracted BESS capacity — the highest in Europe relative to grid and economic size.

Bulgaria also did something structural that matters: it revised its Energy Law to formally recognize storage as an independent market participant, removing a regulatory bottleneck that still exists in some neighboring markets.

The catch sits in the timeline. RESTORE-funded projects face a hard commercial-operation deadline of 31 July 2026 — which means a large share of that €2 billion is racing a subsidy clock, not necessarily building for the long-term merchant economics the ENTSO-E number implies.

Grid constraints and curtailment are also already being reported in southern Bulgaria during peak solar hours.

4. 🇬🇷 Greece — Third on Revenue, Flagged on Execution

Greece’s grid-scale batteries are now generating real revenue for the first time — genuine progress. But the same coverage that reported it included a direct caution: grid connection delays and cautious bank financing are “slowing deployment compared with Romania and Bulgaria.”

That’s not RENEWD’s read — it’s the market’s own trade press drawing the comparison unprompted, and it’s the clearest third-party signal on this list that a strong revenue number and a smooth deployment path aren’t the same thing.

5. 🇷🇴 Romania — Fourth on Revenue, and the Only One Not on the Backlog List

Here’s the comparison that actually moves the needle. A June 2026 AFRY study for Beyond Fossil Fuels found 455 GW of European battery storage capacity — roughly €100 billion in value — sitting in distribution grid connection queues across eight named markets: Bulgaria, Czech Republic, UK, Germany, Greece, Italy, Poland, and Spain. Germany and the UK are worst-hit, each with backlogs exceeding 100 GW.

Romania is not on that list. Neither is Hungary, though Hungary’s exclusion is more a function of market size than resolved queue management.

Romania’s is a function of policy. Ordinul președintelui ANRE nr. 53/2024 replaced first-come-first-served grid access with a competitive, price-based auction specifically designed to clear the zombie pipeline of low quality projects with no backing.

It hasn’t solved the problem instantly — plenty of Romanian sub-zones are still publishing zero MW available without reinforcement while the backlog works through the new milestone system — but it’s an active mechanism attacking the problem, not a queue quietly growing for lack of one.

And capital is already acting on that distinction. Electrica, a regulated Romanian utility, secured grid connection permits for 17 battery storage projects totaling 700 MWh in June 2026, using its own balance sheet as part of a 2030 Strategy — not chasing a subsidy deadline.

Nuvve Holding Corp, a US-listed storage and V2G company, signed an agreement the same month to acquire a 42 MW Romanian project (BESS Sibiu SRL), with deal terms — a modest closing payment plus a COD-linked earn-out — that read like a buyer pricing real execution risk carefully, not chasing a headline spread.

6. 🧮 One Discipline That Applies to All Four

None of this changes the underwriting math, wherever you land. A €792–800/MW/day spread is a market condition, not a bankable input.

A bankable model in any of these four markets runs power prices 20–30% below base case, cuts ancillary-service volume 15–20% for market saturation, and still clears a 1.35× DSCR floor in the worst single year across the full debt tenor — not the average year, the worst one.

That discipline is what separates a project that survives a spread compression from one that was never more than a bet with financing attached.

🎙️ See This Live — Investor Briefing, August 5

I’m hosting a working session for investors currently evaluating, or considering, BESS exposure in Romania specifically — where the market actually stands against its regional peers, what the new legislative changes open up or close off, and where the real constraints sit right now.

Date Wednesday, August 5, 2026 Audience Investors and capital allocators evaluating BESS in Romania Topics State of the Romanian market vs. regional peers · the grid auction and what it’s actually clearing · opportunities and constraints from the new legislative changes · live Q&A Format Live, ~45 minutes + open Q&A Registration Register here

No pitch deck, no sales angle — bring your hardest question about whether this market holds up under real diligence.

Register for the August 5 briefing →

For the mechanics behind Romania’s auction and what it’s clearing, see what the 2026 grid auction changed for ATR-holding pipelines. For the full underwriting framework behind the numbers above, see the five things banks actually require before financing a BESS project.

RENEWD advises investors, developers, lenders, and navigating BESS project finance, grid connection strategy, and bankability in Romania and Europe.

A colleague of mine sat in a room in late June and heard a number that should have been the biggest story in Romanian energy that week.

Zero.

Not “constrained.” Not “tightening.” In many of the network areas Transelectrica reviewed for its new capacity auction, the megawatts published as available — without reinforcement — came back at zero.

Nobody outside a small circle of specialists noticed. That’s the gap this piece is trying to close.

1. 🔌 What Actually Changed: From First-Come-First-Served to Auctions

For most of the last decade, getting grid access in Romania worked on a simple principle: whoever filed the paperwork first got the queue position. File early, get an Aviz Tehnic de Racordare (ATR — technical connection approval), wait your turn.

That principle is now dead. Ordinul președintelui ANRE nr. 53/2024 — the order from Romania’s energy regulator, introduced in 2024 and amended again in early 2026 — replaces first-come-first-served with a competitive auction mechanism for any new generation or storage capacity of 5 MW or more.

The timeline, confirmed directly from Transelectrica’s own procedural documents:

• By 30 June 2026: Transelectrica finalizes a “global study” and publishes the capacity available for connection — without additional network reinforcement — for every network zone and every year of the allocation period

• 1–14 July 2026: Applicants submit allocation requests through the new online platform (licitatii.transelectrica.ro) — this window is open as of this writing

• By 24 July 2026: The validated participant list is finalized

• By 23 October 2026: Transelectrica updates the study with the reinforcement works required to connect the submitted requests, and publishes the participant list for the live auction

• From 30 October 2026: Daily auction sessions begin, one per working day, 08:00–13:00, allocating capacity year by year in chronological order across a 10-year allocation period starting in 2028

If your capacity request exceeds what’s available without reinforcement, the starting auction price is calculated by spreading the estimated cost of the required network reinforcement across the capacity being auctioned — meaning the price you pay is directly tied to how much grid upgrade your connection requires.

Why build a whole auction apparatus for this? Because the queue it replaces had stopped meaning anything.

2. 📊 The 80,000 MW Problem

Here’s the number that explains everything else in this piece.

As of an April 2026 government audit, Romania has issued over 1,400 ATRs for more than 80,000 MW of combined capacity — close to ten times the country’s actual electricity demand. Under the old rules, holding an ATR cost almost nothing and required almost no proof of intent to build.

The result was predictable: speculative filings piled up faster than the grid could ever accommodate them, with paper capacity crowding out capacity.

The government’s own sample audit is the clearest evidence of how much of that 80,000 MW was never going anywhere. Of 722 ATR-holding companies examined, 318 are pure SPV’s, with shareholding capital under 1,000 RON.

This is the backlog the new auction mechanism exists to clear. And it’s the backlog that explains why “available capacity” and “capacity you can actually get” have become two very different questions.

3. ❓ Is There Still Grid Capacity Available Without Reinforcement in Romania?

Read more

On Friday, 26 June 2026, Cristian Bușoi — Romania’s State Secretary for Energy — stood in Luxembourg and signed the EU’s first-ever tripartite agreement on energy storage. AccelerateEU battery storage just went from a policy communication to a signed sectoral commitment, and every trade outlet covering it ran some version of the same headline: Europe locks in the storage buildout.

Here’s the sentence none of them led with. Printed in the European Commission’s own account of the agreement: the pledges member states are making under it “do not create legal obligations for signatories.”

Sit with that for a second.

1. 🖊️ What the EU Tripartite Storage Agreement Actually Is

The tripartite agreement on energy storage was signed in the margins of the Council of EU Energy Ministers meeting in Luxembourg. Twenty-two EU member states joined. Signatories also include national promotional banks, storage and renewable energy developers, storage manufacturers, and large industrial energy consumers — a genuinely cross-sector coalition, which is what makes it a “tripartite” instrument rather than a standard intergovernmental accord.

The agreement runs 2026–2028. Its stated aims: accelerate storage technology development, manufacturing, and deployment; raise annual installation rates; widen the use of storage in industry; and — this is the part that matters for anyone financing a project in Romania — build a stronger project pipeline and improve the conditions under which storage gets financed at scale.

Seventeen of the 22 signatory countries have already submitted concrete pledges, expressed either as installed power (MW) or total energy capacity (MWh). The remaining five have until the end of 2026. Romania is among the 22.

None of that is nothing. It’s the EU’s first attempt to coordinate storage deployment the way it once coordinated vaccine procurement — pooling political will across governments, financiers, and industry into one signed document.

But “first attempt at coordination” and “binding obligation” are two different sentences, and the gap between them is exactly where this piece lives.

2. 🎓 Who Signed for Romania — and Why the Signature Isn’t Symbolic

I’ve written before about how a signature can carry more weight than the paper it’s on. This is one of those cases.

Bușoi didn’t arrive at energy policy sideways. He chaired the European Parliament’s ITRE Committee — Industry, Research and Energy, the body that shapes EU energy legislation before it ever reaches Council — for the second half of the 2019–2024 term, re-elected to the post in January 2022. He sat across the table from the Clean Energy Package negotiations. He knows which clauses in an EU energy text have teeth and which are aspirational filler, because he spent years drafting the distinction.

He left Brussels in 2024. He’s now the person signing Romania’s name to instruments built by the institution he used to help run.

I’m not going to tell you that changes what the agreement legally requires — it doesn’t, and we’ll get to why in a minute. But it changes something else: the credibility of Romania’s seat at the table when the next round of storage-related EU legislation gets negotiated.

That’s a different kind of value than a legal obligation. It’s slower. It’s harder to put in a financial model. It’s also real.

3. 🔢 The Number Everyone’s Repeating

200 GW.

That’s what Bușoi himself cited in his own statement announcing the signature: “It is estimated that by 2030, storage capacities of approximately 200 GW will be needed to meet the needs of the European energy system” — compared with roughly 55 GW installed at the start of 2026.

Read more

Professor X wasn’t the most powerful X-Man because of his telepathy.

He was the most powerful because he built Cerebro — a machine that amplified what he already knew, extended his reach to places he couldn’t physically be, and ran continuously while he was doing everything else.

That’s what I built with Claude. Not an AI assistant. Not a chatbot I talk to when I’m stuck. A domain-expert intelligence system running my knowledge — my deal parameters, my market patterns, my voice, my brand — at a scale and speed that no team I could hire would match.

This is the story of how I got there. With the actual numbers. And an honest answer to the question I get asked constantly: “Can’t I just do this myself?”

🤯 The Moment I Understood What I Was Actually Building

I was on a call with a developer in Bucharest — normal project finance conversation, revenue stack, grid connection status, the usual.

The other half of my system was already working. In parallel. Without me.

It was running a global BESS demand intelligence scan across eight geographies simultaneously — the US, MENA, Australia, India, Latin America, Germany, Spain, Romania. Pulling signals from LinkedIn posts published that week by practitioners on the ground.

Extracting deal announcements, regulatory shifts, market saturation signals, names of people who were explicitly asking to meet at Intersolar. Cross-referencing everything against the financial benchmarks I’d spent years developing in actual transactions. Producing a structured intelligence brief while I was on the phone talking about one project.

When I finished the call, I had a 250-data-point global market scan sitting in my folder. Written to my brand. Structured the way I’d structure it. With the right flags highlighted — not the ones a generic AI would flag, but the ones that matter in this specific market.

That was the moment. Not when I first used Claude. When I first understood what I’d built with it.

💸 Let’s Talk About the Numbers Nobody Publishes

I’m going to do something unusual and share what this system would cost if I’d bought it the traditional way.

Global BESS demand intelligence report — 8 geographies, ~250 signals, formatted and structured: Traditional route: specialist research firm, 3–4 weeks, €15,000–25,000. What I spent: one session, roughly 4 hours of system runtime, cost of my Claude subscription.

Branded client proposal — full design, RENEWD brand system, 15+ pages, print-ready PDF: Traditional route: design agency briefing + production, 2–3 weeks, €3,000–6,000 per document. What I spent: a few hours. It knows my fonts, my hex values, my logo placement rules, my voice. Down to #00391E — not #003300, not #002200. Exactly #00391E. Because I told it once, it wrote it into its permanent memory, and it has never used the wrong colour since.

BESS × data centre market research — LinkedIn chatter scan, gap analysis, 5 productizable opportunities identified, articles drafted: Traditional route: market research firm + specialist energy copywriter, minimum 4 weeks, €20,000–35,000. What I spent: one session.

CO2 and carbon accounting regulatory analysis — four EU frameworks mapped, banking implications synthesised, advisory angle developed: Traditional route: energy finance lawyer, minimum 2 days at €350–500/hour, €3,500–8,000. And they’d still miss the PCAF avoided emissions update. What I spent: 45 minutes and a cup of coffee.

Add it up. In a single extended working session — the kind of session I now run regularly — I produce deliverables that would have cost me €40,000–60,000 through traditional channels and taken 6–10 weeks of elapsed time.

I’m not running a hundred-person firm. I’m running RENEWD. This system is why that’s enough.

🧠 What “My Genius Embedded” Actually Means

Here’s the thing people misunderstand when I describe this.

They hear “Claude” and they think: ChatGPT, but Anthropic. A smart chatbot. Something anyone can access for $20 a month.

That’s not what this is.

What makes the system valuable isn’t Claude. It’s what I put inside Claude. Years of deal experience, encoded as permanent knowledge. My specific financial parameters — not the generic “DSCR should be above 1.0x” that any MBA could tell you, but the actual 1.35x minimum stress-tested across the full debt tenor that Romanian and European banks require for BESS project finance right now.

The LFP degradation curve that distinguishes warranty-period performance (0.7% annually) from post-warranty reality (~1.0%) and knows why that gap matters for augmentation CAPEX modelling in Year 12–15. The revenue stack benchmarks — EUR 14–20M annually for a 150MW/300MWh system in Romania — broken down by DAM arbitrage, FCR/aFRR, and balancing market, with the current spread ranges I’m seeing in actual transactions.

That knowledge doesn’t come from training data. It comes from deals. From hours in front of lenders’ credit committees. From arguments about degradation assumptions at 10 PM before a model deadline. From watching a revenue stack that looked bulletproof in 2023 compress in 2024 as more capacity flooded the ancillary market.

I encoded all of it. Into a skill file that loads every time the system touches anything related to BESS finance. So when it reads an acquisition target’s financial model and says “the FCR revenue assumption is optimistic given current market saturation in Romania,” it’s not guessing. It’s running my judgment. At machine speed. While I’m on the phone.

That is not replicable by someone who subscribes to Claude tomorrow. They have the tool. They don’t have the 10 years of context I put inside it.

🔌 The Architecture, Briefly

For the people who want to understand how the pieces fit:

The system runs on Claude’s Agent SDK — a framework for building AI agents that use tools autonomously across multiple steps. Not a single question-and-answer. A loop of: observe → reason → act → observe again → reason again → act again — until the task is complete.

Connected to that loop via the Model Context Protocol (Anthropic’s open standard for AI tool integration) are several critical connectors. A browser automation layer that injects JavaScript directly into LinkedIn’s React component tree — bypassing the rate limits and API restrictions that would otherwise make systematic intelligence gathering impossible.

A prospecting API that enriches companies and people flagged during scans, returning registration data, estimated revenue, key personnel, and recent news in a single call. Web search and web fetch for deep-dive research. A file system connector that writes finished deliverables directly to my workspace.

Skills — structured knowledge packages — load into the system’s working memory when tasks require them. The BESS finance skill. The brand production skill with the complete RENEWD identity system. The creative entropy engine that makes content sound like me, not like a press release. The reliability protocol that runs confidence checks before anything is delivered.

And underneath all of it: a persistent memory system. Files that survive session boundaries. So when the system encounters a developer it scanned three weeks ago, it pulls the prior context automatically. It remembers that the founding team changed. It remembers the ATR status from the last check. It remembers the revenue assumption that didn’t hold up on closer examination.

It doesn’t start from zero every time. Neither do I.

⚡ What This Looks Like in a Real BESS Deal Intelligence Run

A concrete example — the kind of thing we run regularly.

I want to know which BESS developers in CEE are approaching a deal inflection point. Pre-marketed. Not on any broker’s radar. The ones sitting on ATRs they’re running out of time to execute on.

The agent loop opens LinkedIn via the Chrome MCP and scans posts from the past 30 days — practitioners in Romania, Poland, Hungary, Bulgaria. It’s not reading for keywords. It’s reading for signals: a CEO talking about “a difficult but clarifying quarter.” A project company that filed an SPV 18 months ago with no subsequent EPC announcement. An engineering team connection disappearing from a developer’s profile. These are patterns I’ve learned to recognise. The system recognises them now too.

Flagged developers get enriched automatically through the prospecting API. I’m looking at company registration dates, estimated size, last known financing round, key personnel. For each one, the system cross-references against its memory: have we seen this name before? What was the context? What’s changed since?

The output is a ranked list — not a data dump. Developers ordered by deal likelihood, with a summary of the signals that flagged them and a recommended first action. It lands in my folder, written in the format I’d have written it myself.

From query to brief: typically under 30 minutes for a CEE-wide scan. Without anyone touching it after I start the session.

Then I pick up the phone. Already knowing more than the broker who’ll call me about the same project next month.

🌍 Why This Matters — and Who Should Be Paying Attention

Two groups.

Energy developers and project owners. If you’re sitting on a BESS project in CEE — permitted, gridded, partially developed — and wondering whether it’s a viable acquisition candidate and at what value: I have a system that can give you a preliminary answer in hours, not months. Based on real comparable signals, not a spreadsheet model built on assumptions.

Investors and buyers building BESS pipelines in Europe. The deals that matter are not the ones that hit your inbox through brokers. They’re the ones you find three months earlier, when the developer is still thinking about whether to run a process or find a direct partner. The intelligence to find them systematically? That’s now a solved problem on my end.

Somewhere in Washington in March 2026, the executives of three of the world’s most valuable companies sat across a conference table from the President of the United States and were told: pay your own way.

Google. Meta. Amazon. The message was not subtle. You want the data centers, you pay for the power. The infrastructure you need is no longer a public good.

That meeting got covered as a political story. It was a BESS project finance story.

Because buried inside that directive — inside the regulatory pressure building across PJM, Ireland, the Netherlands, and increasingly CEE — is a structural shift in who the customer for battery energy storage actually is. The offtaker is changing. And the new offtaker is, in almost every respect, better for bankability than the old one.

This is what that means for how you structure, finance, and underwrite BESS projects over the next three years.

⚡ Why the Grid Is Breaking Under AI’s Weight

Data centers consumed an estimated 447TWh of electricity globally in 2025. By end of 2026, Gartner projects that figure reaches 565TWh — a 26% increase in twelve months.

In PJM, the largest grid region in the United States, electricity prices spiked 76% directly attributable to data center load growth. The federal regulator, FERC, is now forcing the question the industry had been deferring: who pays for the transmission and generation infrastructure required to keep these facilities running?

The answer, increasingly, is: the companies that built them.

Ireland is a different version of the same problem. Data centers are projected to account for up to 32% of national electricity consumption by 2026. The grid wasn’t sized for this. The planning system wasn’t built for this. And the political tolerance for socialising infrastructure costs across residential and commercial ratepayers is running out.

This is the energy problem everyone is talking about.

Here’s what almost nobody is talking about yet.

🔄 The Offtaker Revolution

When hyperscalers are forced to build their own power infrastructure — their own generation, their own storage, their own grid independence — the BESS market acquires a new class of customer.

Not a TSO. Not a grid operator running a capacity auction with a 60-page prequalification process and 18-month eligibility timeline.

Amazon.

Amazon’s procurement process looks nothing like a TSO’s. They sign at scale. They move fast. They have balance sheets that make development finance look like rounding errors. And — critically — they don’t need a government revenue support mechanism to make the investment worthwhile. They have their own margin. Their own capex programme. Their own incentive to own the infrastructure rather than depend on a grid that can’t scale as fast as their compute.

The standard feasibility framework for grid-connected BESS in a market like Romania depends on stacking revenues across DAM arbitrage, FCR/aFRR, and the balancing market. The bankability analysis models DSCR stress-tested at 1.35× across the full loan tenor against conservative revenue assumptions. It’s sophisticated. It works. But it depends on regulatory access: OPCOM prequalification, ANRE eligibility, Transelectrica dispatch protocols, PICASSO integration timelines.

A hyperscaler corporate PPA bypasses most of that. The revenue certainty comes from a contract with a single creditworthy counterparty. The financial model gets shorter. The lender conversation starts from a different place.

📊 Grid-Connected vs. Hyperscaler-Facing BESS: The Financing Comparison

The structural advantages are significant. A hyperscaler PPA removes the regulatory eligibility risk that makes lenders nervous in emerging markets. It replaces stacked, uncertain revenue streams with a single contracted counterparty. It simplifies the financial model at the point where most BESS deals in CEE still require DFI credit enhancement to get across the line.

That doesn’t mean it’s simpler to execute. It means the complexity moves somewhere else.

🔋 What the Technology Requirements Actually Change

Behind-the-meter BESS for hyperscalers is not just grid-connected BESS with a different customer.

A grid-connected BESS in the current Romanian market optimises for ancillary market access. That means 2-hour duration, fast-ramping capability, OPCOM prequalification, and revenue modelling that stacks DAM arbitrage (30–40 EUR/MWh spread), FCR/aFRR (60–70 EUR/MWh equivalent), and balancing market revenues (80–120 EUR/MWh, highest margin, most volatile).

A hyperscaler-facing BESS optimises for load following and peak shaving. Longer duration — 4, 6, 8 hours or more. Different cycling profiles. Different degradation exposure. The LFP warranty degradation assumption (0.7%/yr under warranty, approximately 1.0%/yr post-warranty) holds at the chemistry level, but the augmentation event timing and replacement CAPEX modelling shift because the cycling regime is heavier and less predictable than ancillary dispatch.

At 8+ hours of duration, you’re no longer in well-trodden LFP territory. Google has already placed its bet: a 30GWh reservation with Form Energy for iron-air long-duration storage. Meta has committed 100GWh to Noon Energy’s reversible solid oxide fuel cells. These aren’t announcements for the press release. They’re technology procurement decisions made by organisations that ran the economics and decided merchant duration risk is worth owning.

Iron-air and solid oxide fuel cells are not yet commercially de-risked at scale. That matters for how lenders underwrite these projects and what technical advisor consensus looks like.

For 2–4 hour LFP behind-the-meter BESS — which is where the near-term hyperscaler opportunity realistically sits — the technology is understood. The debt is available. The question is deal structure and finding the right counterparties.

🌍 The European Lag Is Not a Warning. It’s a Window.

The United States is roughly three years ahead on this curve.

The regulatory pressure hit there first because the data center concentration hit first — Virginia’s corridor alone reshaped how PJM thinks about capacity planning. FERC moved because the political cost of 76% electricity price spikes, attributable to load that benefits a handful of companies, became impossible to absorb.

Europe is structurally different but directionally the same.

The hyperscalers building in Ireland, the Netherlands, Denmark, and increasingly in Central and Eastern Europe are facing identical pressure: grids not sized for their load, interconnectors at capacity, planning timelines measured in years.

The political and regulatory impulse to require these companies to internalise their infrastructure costs will follow the US pattern. It already is, in Ireland. It will come to Romania, Poland, and Hungary.

For CEE BESS developers, the timeline matters. The hyperscaler build-out in the region is 2–4 years behind the US coast. Which means the regulatory pressure to internalise power infrastructure is still ahead of the market — not behind it. That’s the window.

The EBRD-backed R.Power project at Scornicești — Romania’s first large-scale standalone BESS financial close — demonstrates that grid-connected project finance is now viable in Romania.

What comes next, as the data center footprint grows across Bucharest, Cluj, and the industrial corridors, is the behind-the-meter layer. The developers who understand both models will own both pipelines.

⚠️ The Risk Register

This isn’t a clean story.

Counterparty risk is real. A hyperscaler PPA looks excellent until the hyperscaler pauses its capex programme. They’ve done it before. AI investment cycles are volatile — data center builds have been delayed and repriced as compute demand shifts. Lenders will price this into covenants.

Developers need to think carefully about termination provisions, CAPEX guarantees, and what happens to a behind-the-meter BESS asset if the data center it was built for goes dark.

Long-duration technology isn’t commercially de-risked. Google’s iron-air bet and Meta’s solid oxide fuel cell commitment are technology wagers, not proven commercial deployments at utility scale. Lenders operating in European markets will not underwrite these technology stacks without significantly more project precedent than currently exists. The near-term opportunity is LFP at 2–4 hours, not iron-air at 12.

Regulatory frameworks for behind-the-meter BESS in CEE are nascent. Behind-the-meter doesn’t mean off-grid. These assets still interact with local distribution networks. ANRE in Romania, URE in Poland, and MEKH in Hungary have not yet developed comprehensive frameworks for large-scale behind-the-meter storage at hyperscaler campuses. Regulatory ambiguity is a bankability constraint even when the offtake counterparty is excellent.

Neither of these risks invalidates the thesis. They define where the work is.

🧭 What This Means for How You Position Your BESS Pipeline

The immediate implication is not to pivot your entire pipeline to hyperscaler offtake. The grid-connected ancillary services model in CEE is still the right first-mover position for most developers. The OPCOM prequalification, the regulatory framework, and the DFI appetite are there.

The strategic implication is to understand both models well enough to execute either one.

That means:

Knowing what your asset’s duration and dispatch profile needs to be for each offtake structure — and how that changes the capex basis and degradation modelling.

Knowing how lender conversations start from a different place when your offtaker is a hyperscaler versus a TSO — and what due diligence looks like in each case.

Knowing which DFIs and commercial lenders are beginning to develop hyperscaler BESS financing frameworks, because those relationships will matter when the pipelines arrive.

The developers in CEE who figure out who their real customer is — in 2024 terms and in 2027 terms — are the ones who will own the next generation of project flow.

Follow RENEWD on Substack for the next piece in this series: how PICASSO aFRR integration changes the revenue stack for Romanian BESS and what it means for projects currently in development.

Same developer. Same LFP chemistry. Similar scale.

One deal needed a development bank, a European Union first-loss guarantee, and a regulatory de-risking framework to close. The other closed on purely commercial terms — two commercial banks, no public instruments, no DFI backstop.

That’s a precise map of where Romania and Poland stand on BESS project finance in June 2026. And the gap between the two structures is not a technology problem, a developer problem, or a lender problem.

It’s a market structure problem.

🔋 The Two Deals

In December 2025, the EBRD approved financing for the Scornicești BESS project in Olt County, Romania — 127MW / 254MWh, developed by R.Power S.A. (50.01%) and Eiffel Investment Group (49.99%) through a Romanian SPV.

Total project cost: €67 million. Debt: up to €35 million in non-recourse senior loan from the EBRD, plus a €15 million bridge facility covering capex ahead of an anticipated EU grant disbursement. Bank Gospodarstwa Krajowego (BGK), Poland’s state development bank, also contributed, bringing total debt to approximately €46 million.

Of that senior loan, €29 million sits behind an InvestEU first-loss guarantee — deployed under EBRD’s RenewEU De-Risking Framework. Revenue structure: partially merchant, drawing on Romania’s ancillary services and balancing market. No long-term contracted revenue floor.

In June 2026, R.Power closed the Jedwabno BESS deal in northeastern Poland — 150MW / 300MWh. €64 million. Non-recourse project finance. Club deal: Siemens Bank and Erste Group. No public instruments. No DFI involvement. Revenue structure built on a 17-year capacity market contract plus a long-term optimisation agreement with Axpo.

Same developer. Entirely different financing architecture.

🏗️ What Made Each Deal Bankable

Jedwabno’s bankability rests on two contracts.

A 17-year contract from Poland’s capacity market auction — government-backed minimum revenue certainty across the full debt tenor. This is the instrument that converted a merchant BESS asset into something a debt committee could model with precision.

Stacked on top: a long-term optimisation agreement with Axpo — locking in route-to-market and incremental revenue flow back to the project.

Together, these gave Siemens Bank and Erste Group enough cashflow visibility to underwrite €64 million without any first-loss protection. Michal Swół, R.Power’s CIO, was direct about what the deal represents: “large-scale BESS is now fully bankable in CEE.” He’s right. But the mechanism matters.

Scornicești doesn’t have either of those instruments available.

Romania doesn’t yet have a functioning capacity market delivering 17-year contracts. Revenue comes from ancillary services, frequency containment, and balancing — real markets, real cashflows, but volatile and uncontracted. Without a revenue floor, commercial banks face an underwriting problem: what does the downside case look like? How do you stress-test cashflows across a 10-year debt tenor when nothing is contracted?

The InvestEU first-loss guarantee answers that question. It doesn’t replace contracted revenue — it absorbs the first layer of loss if merchant revenues disappoint, making the senior debt position safe enough for EBRD to underwrite.

The EBRD project documentation states the additionality plainly: structuring “complex, limited recourse project financing for one of the first large-scale, standalone, partially merchant battery energy storage systems in Romania, enabling investment in a market where financing for merchant-based storage projects is not readily available.”

📊 Side-by-Side: Two Structures, One Market Gap

🔄 Two Stages, One Trajectory

These aren’t two different philosophies. They’re the same structure applied at two different stages of market development.

Stage 1 is market creation. The asset class is new in the local context. Commercial lenders have no track record to underwrite, revenue mechanisms are forming, and regulatory frameworks haven’t yet stabilised.

Development banks take the first-mover risk — not to subsidise the project, but because their mandate is to invest precisely where markets can’t yet function efficiently. The InvestEU first-loss guarantee at Scornicești is that mandate deployed exactly as designed. However, stage 2 looks different.

Read more

Eight of the top ten global BESS system integrators are Chinese companies.

In the first half of 2025, every single one of the top ten battery cell suppliers by shipment volume was Chinese. In Europe specifically, approximately 80% of newly installed inverters came from Chinese manufacturers — Huawei and Sungrow leading the ranking, according to Wood Mackenzie.

This is not a criticism of Chinese manufacturing quality. Chinese BESS and inverter technology is, in many configurations, technically excellent. The LFP chemistry that European lenders now treat as the bankable standard was developed and scaled in China.

It is, however, a description of a structural dependency that the European Commission has decided — with the April 2026 EIB financing ban — is incompatible with EU-funded energy infrastructure.

What happens next is not a question of if. It is a question of how fast, how far, and whether your project is positioned to survive the transition.

1. What the April 2026 Ban Actually Says 📋

The European Commission’s guidance, communicated to financial institutions on May 1, 2026, restricts EIB and European Investment Fund financing for solar, wind, and energy storage projects using inverters and Power Conversion Systems from high-risk country suppliers. The designated countries: China, Russia, Iran, and North Korea.

Three things in that paragraph are worth unpacking.

First: the ban covers PCS explicitly. Power Conversion Systems — the component that manages energy flow between the battery DC architecture and the AC grid — are named alongside inverters. This is not an accidental overlap. A BESS PCS from Sungrow, Huawei, or any other Chinese-origin supplier is directly in scope. The initial reporting focused on solar inverters; the updated guidance made clear that BESS projects are equally affected.

Second: the ban applies to companies owned or controlled by entities or persons from those countries. This is not purely a country-of-manufacture designation. A European-registered entity that is majority-owned by a Chinese parent company falls within scope. Supply chain due diligence now extends to corporate ownership structures, not just product origin.

Third: the transition window is specific and short. Financial institutions could notify pipeline projects by May 1, 2026; only projects sufficiently mature to be approved by November 1, 2026 qualify for grandfathering. By July 1, 2026, all EU services must review ongoing activities and propose integration of the restrictions. From April 2027, the regime becomes fully binding for new projects with limited derogations.

If you are currently developing a BESS project expecting to access EIB, EBRD, or InvestEU financing — the instruments that sit at the center of every blended finance structure for Romanian BESS — and your PCS selection has not been reviewed against this guidance, that review is now urgent.

2. The Regulatory Stack That Made This Inevitable ⚖️

The April ban did not arrive without warning. The signals ran for over a year, each one more specific than the last.

March 2025. Commissioner Jørgensen at the Solar Power Europe Summit announces a formal cybersecurity review of critical elements in the EU solar value chain. The same review mechanism that resulted in Huawei and ZTE being designated high-risk in 5G networks was now being pointed at energy hardware.

December 2025. The EU Economic Security Strategy is published. Among its explicit actions: the Commission will prevent access by high-risk entities to Union-supported actions, including those supported by public financial institutions. Solar inverters are cited as the illustrative example case — in a box, in the text, not in a footnote.

January 2026. The proposed revision of the Cybersecurity Act (CSA2) is published, introducing for the first time a horizontal EU framework for ICT supply chain security. Article 99: security risk assessments. Article 100: designation of third countries posing cybersecurity concerns. Article 104: identification of high-risk suppliers, with binding phase-out requirements across NIS2-covered sectors.

April 2026. The EIB financing restriction is announced, grounded in the Economic Security Strategy and preliminary findings of the cybersecurity review.

Thirteen months. Four escalating milestones. The 5G precedent was visible the entire time. In that sector, the same commission review mechanism resulted in formal designations that — while implemented unevenly across member states — permanently restructured the European telecommunications supply chain.

The BESS and solar industry is at the March 2025 moment of the 5G sequence. The April 2026 ban is not the end of this process. It is the beginning of the enforcement phase.

3. NIS2 Romania: The Legal Layer That’s Already Active 🔒

I’ve covered NIS2 and its implications for BESS operators in detail previously. The short version for the analysis here: Romania transposed NIS2 through Law 124/2025 (from GEO 155/2024), in force July 10, 2025. DNSC implementing orders entered into force August 20, 2025. The registration deadline for in-scope entities was September 19, 2025.

Grid-connected battery energy storage systems in Romania are classified as critical infrastructure operators under this framework. What that classification carries:

Read more

Most BESS project funding application never make it to the credit committee.

They get there eventually — the slide deck, the model, the teaser — and the bank asks three questions that the developer hasn’t answered. Not because the project is bad. Because the preparation was wrong. The model showed the upside case. The degradation curve was linear. The augmentation event was a footnote.

I’ve reviewed enough Romanian BESS financing attempts to say this clearly: projects that fail to raise debt don’t usually fail because of bad fundamentals. They fail because of presentation gaps that, to a bank’s credit team, look like thinking gaps.

Here are the five things that matter. Not a checklist — a framework for how banks actually think about BESS credit risk.

1. DSCR 1.35×, Stress-Tested Across the Full Tenor — Not Year One 📊

The number you’ll hear most is 1.25× minimum. Some lenders go higher; some go lower for highly contracted projects. For Romanian standalone BESS, the relevant threshold is 1.35× minimum (and even this number might fall short sometimes), and that’s not the average across the loan life — it’s the floor in the stress case, in the worst single year, at any point across the full debt tenor.

That distinction matters enormously.

A project that comfortably covers 1.5× DSCR in years 1–5 and then struggles to reach 1.1× in years 12–14 is not a bankable project. It’s a project with a hidden problem that will surface at exactly the wrong moment — when degradation accelerates, when the augmentation CAPEX draws down, when market revenues compress. The lender’s stress case is specifically designed to find that moment.

Banks will run their own model. Not yours. And their assumptions will be more conservative than yours on every dimension that matters: power price forward curves, revenue stack volume, degradation rate, augmentation timing and cost. If your project only works on your numbers, it won’t pass their model.

The stress-test inputs that matter most:

• Power prices: 20–30% below base case

• FCR/aFRR volume: reduced by 15–20% to reflect market saturation

• Degradation: 1.5–2× the warranty rate in the post-warranty years

• Augmentation CAPEX: modeled at year 12, not deferred

• Curtailment risk: 5–10% revenue haircut on merchant tranches

A project that still covers 1.35× under all of those simultaneously is a bankable project. That’s the bar.

2. A Revenue Stack with a Contracted Anchor — But Not 100% Tolling ⚡

Banks want contracted revenue. Enough to anchor the DSCR. Not so much that the project is delivering 100% of its capacity to a single counterparty who may or may not be around in year 8.

The balance that European lenders have converged on in 2026: roughly 50–80% of revenue contracted through a tolling agreement or firm capacity payments, with the remainder left as merchant. NORD/LB’s Marcus Starke put it directly at BBDF 2026: “Typical structures are 12-month merchant phases, then around 80% tolling for five to seven years. This works for projects with high gearing.”

ABN AMRO has financed a 100% merchant project — but at materially lower gearing and with additional protections. Santander CIB: “You inevitably need tolling agreements, in part, to spread the risk.”

The 100% tolling structure has its own problems. A tolling agreement transfers market risk to the toller, but it creates operational risk for the project: delivery obligations, availability thresholds, sanctions for underperformance. Terralayr’s Mikko Preuss at BBDF 2026: “A tolling agreement doesn’t mean risk-free mode for a battery. It mitigates market exposure risk but increases operational risk — delivering megawatts, efficiency. There’ll be sanctions, most likely.”

Read more

The grant covers up to €15M.

A mid-range utility-scale BESS project in Romania costs €20–40M depending on size. Sometimes more. The arithmetic is obvious: the Modernization Fund grant is meaningful — it’s not sufficient. What bridges the gap, and how you structure it, determines whether your project gets built or sits on a shelf next to a beautifully prepared feasibility study.

This article is about stacking. How to combine the €150M Modernization Fund grant with EBRD instruments, InvestEU guarantees flowing through Romanian local banks, in-natura security structures, and private capital — without violating the no-double-funding rule or the grant’s contractual conditions.

One caveat before we start. I’ve written the clause-by-clause breakdown of what the Ghid actually says — the assignment prohibition, the 5-year ownership claw-back, the lender standstill conditions. This article builds on that analysis. If you haven’t read it, the structuring decisions below will make less sense.

1. The Constraint That Shapes Everything: What the Grant Actually Covers 📐

Start here, because the structuring logic flows from the cost split.

The Modernization Fund grant covers up to 100% of eligible costs — construction, equipment, installation, fittings, intangible assets. It does not cover: grid connection (branșament), feasibility study, design fees, permitting, project management, financial audit, VAT, contingency reserves, or price adjustment.

For a typical Romanian standalone BESS project, the eligible/ineligible split looks roughly like this:

Two things stand out. Grid connection is the largest single ineligible item — for a 50–150 MW project connecting at HV level, this alone can be €3–10M. And VAT on the full eligible cost base is a separate cash requirement that must come from equity or debt.

The no-double-funding rule (confirmed in the Ghid’s Article 5 and the model contract) says: no cost can be funded from two public sources simultaneously. It does NOT say you cannot use a bank loan alongside a grant. It says the same euro of cost cannot be claimed from both the grant AND another public funding program.

The practical structure is a cost split: grant covers the eligible capex bucket, bank debt covers the ineligible bucket plus any eligible costs above the grant cap, private equity covers the equity layer. As long as each euro of cost is allocated to exactly one source, no double-funding issue arises.

2. The Instruments Available Right Now 🏦

Five categories of financing can be layered with the Modernization Fund grant. Each has different access conditions, pricing, and compatibility with the grant’s contractual structure.

Let me go through each one in detail.

3. EBRD Direct Lending: The IFI Path 🌍

EBRD has demonstrated an active appetite for Romanian BESS. The clearest signal: in May 2026, EBRD closed a €70M loan to NGEN Energetske Rešitve for five BESS projects across Latvia, Poland, Romania, and Slovenia — backed by an InvestEU first-loss guarantee and technical assistance from the EU Advisory Hub. The Romanian project in that portfolio operates on a pure merchant model.

EBRD direct lending for BESS projects in Romania typically looks at:

• ATR confirmed and valid

• Revenue model: merchant is acceptable (EBRD confirmed it), contracted (tolling or capacity payment) is preferred

• Security package: land mortgage, equipment pledge, revenue assignment — share pledge is possible but structurally more complex with grant conditions attached

• Environmental and social standards compliance (ESAP — Environmental and Social Action Plan required)

• DSCR: 1.35× across the full loan tenor under stress case

When combined with the Modernization Fund grant, the EBRD debt sits on the non-grant portion of the capital stack. EBRD does not lend against eligible costs already covered by a grant (that would be double-funding). It finances: the ineligible cost bucket, any eligible costs above the €15M cap, and working capital for early operations.

EBRD pricing for Romanian BESS: floating rate over EURIBOR, typically EURIBOR + 2.5–3.5% depending on project risk, with the InvestEU backstop compressing the margin vs. commercial terms. Tenor: 10–15 years, consistent with BESS project life and grant durability period.

4. InvestEU via Romanian Banks: The Local Path 🇷🇴

This is the most significant recent development for mid-size Romanian BESS developers — and almost no one is talking about it yet.

In May 2026, EBRD launched its InvestEU Risk-Sharing Facility in Romania’s financial sector: €270M in guarantees, flowing through Romanian commercial banks, covering 80% of each green loan’s default risk. Libra Internet Bank was the first to join — a €40M risk-sharing agreement enabling €50M in green lending over three years. Other Romanian banks are expected to follow.

What this means in practice: a Romanian developer approaches Banca Transilvania, Exim Banca Românească, or another local bank. That bank, backed by an 80% EBRD InvestEU guarantee, can now lend against a BESS project at meaningfully reduced collateral requirements and better pricing than would otherwise be available for a merchant project with limited operating history.

This is particularly relevant for standalone BESS developers who don’t have the scale or track record to access EBRD directly, but who have a bankable project — confirmed ATR, solid feasibility study, realistic revenue model.

The local bank path has two additional advantages the IFI path doesn’t. Romanian credit committees understand the domestic grid, the Transelectrica connection process, and the ANRE regulatory environment. And Romanian bank execution is faster — a domestic credit process runs on weeks, not the months that EBRD due diligence requires.

The market has already validated this model for standalone BESS without the grant. Aukera Energy closed a €60M debt facility through Kommunalkredit Austria for its 250 MW / 500 MWh standalone project in Gura Ialomitei, Ialomita County. Kommunalkredit acted as sole mandated lead arranger. Now that InvestEU guarantees are flowing through domestic banks, the Kommunalkredit-equivalent transaction is increasingly replicable through Romanian institutions.

5. In-Natura Structuring: Using the Project as Security 🏗️

Here is where the grant conditions and standard project finance collateral structures need to be reconciled carefully.

The Ghid’s model financing contract (Annex 11) contains a template bank letter for land mortgages. The bank commits to a standstill — it will not enforce against the land and the financed investment for at least 5 years from commissioning. The standstill is conditional on loan performance: if the borrower defaults, the bank’s commitment terminates with notice to the Ministry.

This is the in-natura security structure the Ghid contemplates. The project assets — land, structures, and the BESS system itself — serve as the primary collateral. Not the SPV shares. The physical asset.

For project finance purposes, this actually works reasonably well if structured correctly:

The asset-based security package:

• Land mortgage (ipotecă imobiliară) — permitted, subject to bank standstill commitment

• Equipment pledge (gaj pe echipamente) — not explicitly addressed in the Ghid, but not prohibited; the standstill commitment covers the “investment financed from the Modernisation Fund”

• Revenue assignment (cesiunea creanțelor din venituri) — assignment of future revenues from ANRE-regulated markets (FCR/aFRR payments, balancing market settlements) to the lender; this is contractual, not an asset pledge, and likely falls outside the Ghid’s assignment prohibition which covers the financing contract itself, not revenue streams

• ATR document assignment — subject to legal analysis; the Ghid’s prohibition targets the financing contract, but a lender will want ATR document rights in an enforcement scenario

The security structure that is most uncertain is the share pledge (gaj pe acțiuni). The Ghid is silent on it. EBRD and local banks routinely take share pledges as first-line security in project finance transactions. But the Ghid’s 5-year ownership change provision (Article 5 of the model contract) creates a risk that share pledge enforcement — where the lender takes ownership of the SPV — could trigger the claw-back mechanism.

Until the Ministry provides a clarification (which should be requested during the current consultation window), the safest structuring approach is: asset-based security as primary collateral, with share pledge as secondary or deferred security that requires Ministry approval before enforcement. Some lenders will accept this structure. Others will not — particularly for smaller facilities where share pledge is standard.

This is a gap that needs a legal opinion before financial close. Not after.

6. Private Capital: Structuring for the Long Hold 💼

The grant’s durability and assignment conditions make it a natural fit for infrastructure equity — long-horizon investors who want contracted exposure to Romanian grid infrastructure and are not planning a 3–5 year exit.

The relevant profile: a professional infrastructure fund with a 10–15 year hold mandate, comfortable with the 5-year durability window and willing to model the ownership continuity requirement into the investment thesis.

The combination that works:

Grant (capex subsidy) + IFI/local bank senior debt + infrastructure equity

In this structure, the grant reduces the equity required at construction. It functions not as a revenue source but as a cost-reduction mechanism — eliminating €10–15M of equity that would otherwise need to be deployed at WACC. The equity IRR improvement is significant: on a 150MW/300MWh project generating EUR 14–20M annually across the revenue stack, the grant converts a 9–10% unlevered return into the 12%+ equity IRR threshold without changing the operational model.

The key requirement for infrastructure equity in this structure: confirm that a future fund-to-fund transfer (e.g., from Fund I to Fund II of the same manager) does not trigger the “nature of ownership” change provision. This is another consultation question. A transfer between vehicles managed by the same team, with continuity of operational purpose, should not constitute a change that “offers illicit advantages” — but the Ministry needs to say so in writing.

What doesn’t work: venture or growth equity with a 5–7 year exit horizon, any structure requiring a change-of-control to an unrelated third party within the durability window, and any equity instrument that converts to debt at a trigger point (convertible instruments that could change the “nature of ownership” of the financed infrastructure).

7. The GoO Revenue Layer: January 2027 and What It Changes 🌱

One structuring dimension that almost no financing model for Romanian BESS has incorporated yet: Guarantees of Origin.

GEO 59/2025, in force November 2025, extends the GoO regime to energy storage. Order no. 27/2025 places standalone BESS operators on the same level as electricity producers for green certificate purposes. By 1 January 2027, Romania targets full AIB (Association of Issuing Bodies) membership and a fully operational cross-border tradable GoO market.

What this means for BESS project finance:

A standalone BESS that buys cheap off-peak power (coal, nuclear, overnight) and discharges during peak hours does not produce renewable energy. Its GoO eligibility is limited to energy discharged that can be tracked to a renewable source.

But a BESS that sources its charging energy from a contracted renewable PPA — even a synthetic PPA with a wind or solar generator — can potentially issue GoOs against the renewable energy it has taken in and discharged. The market is not yet standardized on methodology. ANRE must set the framework by June 2026.

If GoO revenue becomes bankable — even at €2–5/MWh, where European GoO markets currently clear — it adds €1–3M annually to a 150MW/300MWh project’s revenue stack. For a lender, contracted or quasi-contracted GoO revenue is exactly the kind of stable, predictable income stream that helps support DSCR. Watch this space closely.

The Financing Conversation That Needs to Happen Now

The call hasn’t launched yet. The consultation window — where the grant conditions can still be shaped — closes around June 5th.

That timing is not coincidental. The developers and lenders who engage with the consultation now, who push for clarity on share pledge treatment and in-natura security structure, who ask for a precise definition of “change in the nature of ownership” — those are the ones who will have a bankable project structure ready when the call opens in Q2 2026.

Everyone else will be improvising at credit committee.

The capital stack outlined here is available today. EBRD is lending. Romanian banks have InvestEU guarantees. Infrastructure equity understands the asset class. The grant is the accelerant — but only for developers who have structured everything else correctly first.

The technical parameters that Romanian BESS lenders require — DSCR 1.35× stress-tested across the full tenor, revenue stack of EUR 14–20M annually for a 150MW/300MWh configuration — don’t change with the grant. The grant changes the equity story. The debt story is built on the same fundamentals as before.

Start with the fundamentals. Then add the grant.

Working through a blended finance structure for a Romanian BESS project — or trying to understand how the grant conditions interact with your existing lender relationships? DM me to start the conversation!

Sources:

1. Ghidul Solicitantului — Instalații de stocare autonomă FM (May 21, 2026); Ministerul Energiei; energie.gov.ro

2. Anexa nr. 5 — Model Contract de Finanțare (Ghid stocare autonomă FM)

3. Anexa nr. 11 — Model Scrisoare Bancară (Ipotecă)

4. EBRD — “EBRD supports battery storage expansion across central Europe” (May 2026)

5. EBRD — “EBRD launches InvestEU programme in Romania’s financial sector” (May 2026)

6. ESS News — “EBRD lends €70 million to Slovenian BESS developer across four countries” (May 22, 2026)

7. ESS News — “Enery secures $298 million project financing for CEE, begins construction of 1 GWh BESS in Romania” (January 27, 2026)

8. Aukera Energy — “Aukera secures €60 million debt facility to finance 250 MW / 500 MWh battery storage project in Romania”

9. ESS News — “European Commission approves Romania’s battery storage subsidy scheme” (March 9, 2026)

10. PV Magazine — “AccelerateEU backs 200 GW storage goal but omits dedicated mechanism” (April 24, 2026)

11. Energy Industry Review — “From Vision to Reality: Romania Takes Decisive Steps Towards Creating a Market for Guarantees of Origin”

12. energynomics.ro — “Romania readies the European recognised RES Guarantees of Origin in 2027”

13. GEO 59/2025 — Ordonanța de urgență privind garanțiile de origine (November 2025)

14. Order 27/2025 — ANRE (assimilates BESS operators to producers for GoO purposes)

15. EC Decision C(2026)1588 final — State Aid Scheme SA.108847 (March 6, 2026)

16. fi-compass.eu — “Combination of financial instruments and grants” (EU double-funding rules)

In 2022, Dynamic Containment in the UK was the single most valuable thing a battery could do.

Average revenues: over £20 per MW per hour. Maximum — briefly — approaching £80/MWh. In that year, DCo alone accounted for 63% of all BESS revenue in Great Britain, worth approximately £100,000 per MW to an average system. The developers who built projects around that revenue stack weren’t speculating. They were being perfectly rational.

By February 2024, Dynamic Containment Low had dropped to below £0.50/MWh.

Not a correction. Not a rebalancing. A 7-fold collapse in prices in under two years, driven by a tripling of installed BESS capacity while ancillary service demand expanded by only 50%. The frequency services that once represented 80% of BESS revenues in the UK fell to 20% of the stack by 2024. Energy arbitrage — which had been 8% of revenue in 2022 — became the primary revenue source.

The batteries kept operating. The revenue moved. That part is important — BESS is resilient, and in the UK, the market adapted. But the project finance world adapted more slowly.

The lenders who had sized debt against 2022 DCo projections were now sitting on models that were structurally wrong. Not by a small margin. Wrong in a way that made refinancing conversations uncomfortable and new deals difficult. What the market needed was a revenue stream that didn’t depend on a single ancillary product holding its price.

That’s where BESS tolling came from. Not from policy. From a credit committee.

🇬🇧 How the UK Built It — and Where It’s Going

The first major UK BESS tolling deal closed in June 2024.

Gresham House signed a two-year tolling agreement with Octopus Energy covering 568 MW and 920 MWh of battery capacity — the largest tolling deal in the GB market at the time. The fee: £57,000 per MW per year, fixed, regardless of market conditions. For context: that figure exceeded the average BESS revenue in 2024. Two parties with different risk appetites, one contract.

That was twenty-three months ago. The market has moved considerably faster than anyone expected.

Standalone tolling agreements in the UK went from 3 in 2024 to 15 in 2025 — with a further 4 project finance deals involving tolling counterparties, bringing the total to 19. The UK also closed Europe’s largest single BESS project finance deal: Thorpe Marsh, at 1.4 GW.

The contracted revenue framework that made that possible is tolling. Shell Energy Europe completed a 7-year tolling deal for BW ESS’s 100 MW / 331 MWh Bramley battery. Engie executed the GB’s first tolling agreement on a co-located BESS asset. And Modo Energy obtained FCA authorisation for a GB BESS benchmark index — making financial swaps and derivatives against BESS revenue a near-term reality.

But the deal that best illustrates where the UK market has arrived is from February 2026.

Drax signed a 15-year tolling agreement with Zenobē for a 200 MW / 800 MWh project at Coalburn in Scotland. No inflation indexation. Flat fixed capacity fee for fifteen years. Drax gets full operational control and dispatch rights, keeping all revenues except capacity market and certain ancillary income. No upfront capital cost to Drax; construction, maintenance, and availability risk stays with Zenobē throughout.

Fifteen years. No CPI. That’s not a risk mitigation strategy. That’s a position.

Drax is making a long-term directional bet on their ability to extract more value from that asset than the fixed fee they’re paying — for a decade and a half. Which means they believe — not hope, but have committed capital behind — that GB BESS arbitrage and ancillary revenue will remain deep enough to justify that bet across two full electoral cycles and whatever market conditions emerge. The developer gets certainty. The toller gets upside.

Two weeks before the Zenobē deal, Drax had signed a 10-year CPI-indexed tolling agreement with Fidra Energy for 250 MW / 500 MWh at West Burton C. Notably different terms for a different asset — the CPI protection on the Fidra deal reflects the developer’s need for inflation-proofing on a longer-dated commitment. Both structures are bankable. The market is developing term and pricing differentiation, not convergence. That’s what maturity looks like.

For the lender: the capacity fee is contracted revenue. Not modelled. Contracted. That distinction changes everything about how debt is sized.

The toller still has to be bankable. Octopus, Shell, Drax — investment-grade entities or entities able to provide LC backing. The credit committees are defining the quality floor case by case, deal by deal, faster than any regulator could.

🇩🇪 Germany — From First Deal to Pipeline in 14 Months

Here is something that gets missed in most BESS tolling discussions: Germany’s first tolled BESS secured project financing in March 2025.

Not 2020. Not 2022. March 2025. Who took the charge?

Read more

Romania sold electricity in April 2026 at 50 EUR/MWh.

Then bought it back at 250.

Not across different weeks. Not during a geopolitical crisis. On the same day — midday surplus exported, evening peak imported — four hours apart, five times the price. The result: a net monthly outflow of approximately 28 million EUR. Not because Romania lacks generation capacity. Because the capacity it has generates when the sun shines, and the demand peaks when it doesn’t.

The nuclear outage that began on May 4 didn’t create that problem. It widened the gap until even people who weren’t watching Romanian energy markets noticed.

This piece is the full account.

⚛️ What Happened at Cernavodă — and Why the Timing Matters

The insulator that started it

At 19:35 on May 4, 2026, an insulator on a power evacuation transformer at Cernavodă Unit 2 degraded under operating conditions. The plant’s protective systems detected the anomaly and triggered automatic disconnection. Standard sequence. Nuclearelectrica confirmed the cause, announced transformer replacement, and set a restart target of June 1, 2026.

700 MW of constant baseload generation: offline.

Unit 1’s scheduled maintenance window was approved independently, well in advance. It disconnected on May 10 at 11:00 AM, per plan.

The timing overlap was unlucky. Neither outage was mismanaged. Both followed their own engineering logic. The problem is that Romania’s grid had never absorbed both simultaneously — because in thirty years of operation, it hadn’t happened. System planning had not priced that scenario into its architecture.

What followed was the first simultaneous shutdown of both CANDU reactors in Romanian nuclear history.

The reactors themselves — why this matters beyond the headlines

Cernavodă is not just Romania’s only nuclear plant. It is the only CANDU-design reactor in Southeast Europe — a technology renowned for its operational flexibility and exceptionally high capacity factors. Together, the two units contribute approximately 18–20% of Romania’s annual electricity generation. Not as an average across weather conditions, but as a constant. Available on demand. 24 hours a day. Every day.

That constancy is what the rest of the grid is built around. Romania can aggressively expand solar capacity precisely because 1,400 MW of reliable baseload is always there to balance it. Remove that anchor, and every percentage point of renewable penetration becomes harder to manage.

The structural picture behind the three-week crisis

Unit 2 returns around June 1. The immediate pressure eases.

Unit 1 does not. In December 2026, it enters a planned refurbishment programme worth €1.85 billion — one of the largest energy infrastructure investments in Romania’s post-communist history. The refurbishment extends the unit’s operational life by decades. It is the right investment. It also means Romania operates on a single reactor — 700 MW of baseload — from December 2026 through December 2028.

Two years. What happened in May 2026 was a preview. Starting in December 2026, it becomes the structural baseline.

📊 The Grid Mathematics of Losing 23% of Dispatchable Capacity

Installed capacity is not the number that matters

Romania’s total installed generation capacity sits at approximately 20,000 MW. That figure is almost meaningless for understanding what happened on May 10.

The number that matters is dispatchable capacity — generation that can be called when demand peaks, regardless of weather. Wind produces when the wind blows. Solar produces when the sun shines. Hydro depends on reservoir levels. Only nuclear and thermal generation are fully controllable. Of Romania’s dispatchable fleet, Cernavodă represents approximately 23% of peak demand.

That 23% isn’t just capacity removed from the stack. It’s the reliable slice that everything else is planned around. Lose it, and the grid’s weather-dependency increases sharply — and immediately.

What 2,200 MW of imports actually means

On the evening of May 10 at 21:35, Romania was importing approximately 2,200 MW from neighbouring grids.

Two things are worth dwelling on.

First: that figure exceeds Romania’s total hydroelectric output at that hour. Romania’s rivers contribute significant generation capacity — and on May 10, the country was leaning harder on its interconnections than on its own water. Second: when Romania imports from the European interconnected grid, it doesn’t pay its own marginal cost of generation. It pays the marginal clearing price of the grid it’s drawing from.

If Romania’s domestic cost is 80 EUR/MWh and the interconnected grid is clearing at 140 EUR/MWh, Romania pays 140 EUR/MWh. That’s not a failure. That’s how interconnected markets work. But at 2,200 MW scale, on a night when domestic shortfall is structural rather than marginal, the arithmetic becomes expensive very quickly.

💸 How the OPCOM Day-Ahead Market Broke — and Why It Was Predictable

The clearing mechanism

OPCOM’s day-ahead market (DAM) is where Romanian electricity is priced and traded — approximately 75–80% of all traded volume flows through it. Sellers submit generation offers. Buyers submit demand bids. The market clears hourly, setting a unique price for each hour of the following day.

The critical mechanic: the last marginal unit to clear in any hour sets the price for that entire hour. When domestic supply drops and imports must fill the gap, import costs set the marginal clearing price. Every Romanian generator gets paid that import price too — a windfall for domestic producers, a bill for Romanian consumers and grid operators.

On May 11, 2026, Romania’s average DAM price reached approximately 730 lei/MWh — roughly 140 EUR/MWh. Highest in Europe that day.

April 2026 weighted average: 514.65 lei/MWh — approximately 99 EUR/MWh.

42% jump. 48 hours.

Why 140 EUR/MWh is not the story — the 200 EUR/MWh spread is

The 140 EUR/MWh headline number has attracted most of the commentary. It shouldn’t. It’s a symptom.

The structural story is in the spread.

The spread on May 10 was not an anomaly. It was the same arbitrage gap that had been running throughout April — the nuclear outage simply made the DAM average visible in a way that generated coverage.

What the table shows: Romania had a 200 EUR/MWh intraday arbitrage spread throughout April 2026. Normal baseline spreads run 30–40 EUR/MWh. This was five to seven times the baseline. And it cost approximately 28 million EUR in net monthly outflow — exporting at 50, importing at 250, all month long.

Romanian energy analysts at AEI described this as “rezultatul unei proaste planificări” — the result of poor planning. I’d frame it differently. It’s the result of a grid designed for one generation mix operating in a fundamentally different one. The solar capacity was correctly built. The storage capacity to monetise it was not.

🔋 The BESS Revenue Case — Built From the Actual Numbers

What May 10 would have paid

Let me build this precisely.

A 150 MW / 300 MWh BESS asset connected to Romania’s transmission grid on May 10 would have run a standard DAM arbitrage cycle.

Charge during midday solar surplus at approximately 50 EUR/MWh. Discharge during the evening peak at approximately 250 EUR/MWh. Spread: 200 EUR/MWh. Volume: 300 MWh.

DAM arbitrage revenue: 60,000 EUR. One asset. One day.

That’s not the full picture.

The frequency services layer

When 1,400 MW of synchronous baseload generation disappears, grid frequency deviates from 50 Hz. Transelectrica must restore it using FCR (Frequency Containment Reserve) and aFRR (automatic Frequency Restoration Reserve) assets — both compensated at ancillary market clearing prices. Under normal grid conditions: 60–70 EUR/MWh equivalent. Under grid stress, they spike above that range.

A BESS pre-qualified for FCR and aFRR would have been actively dispatched during the frequency excursions caused by the dual nuclear outage. Ancillary revenue on top of the DAM arbitrage cycle. This is the part of the BESS commercial case that rarely gets modelled properly by developers coming to the asset class for the first time: the revenue stack is not just correlated with grid stress events. It is positively driven by them. The asset earns most in the exact conditions where the grid needs it most.

The balancing market — the third layer

Transelectrica’s real-time balancing mechanism corrects imbalances between scheduled and actual generation. Prices are the most volatile in the stack — typically 80–120 EUR/MWh under normal conditions, with significant upside during structural stress. On May 10, Romania was managing imbalance throughout the day. A BESS with balancing dispatch capability captured a third revenue stream on top of DAM and ancillary.

What the full annual stack looks like under normal conditions

On days like May 10, the stack compresses into a few hours. Under normal operating conditions across a full year — DAM arbitrage at 30–40 EUR/MWh spread, FCR/aFRR at 60–70 EUR/MWh, balancing market participation at 80–120 EUR/MWh — a 150 MW/300 MWh BESS in Romania generates EUR 14–20 million annually.

May 10 was not the base case. It was the stress case becoming visible.

January 2027: The fourth layer activates

Under GEO 59/2025, Guarantees of Origin (GoO) cross-border certificate trading becomes available to Romanian BESS operators from January 2027. This adds a fourth commercial mechanism — green certificate income sitting on top of the three-layer energy market revenue stack.

Projects that commission in 2026 and are operational in January 2027 capture the full stack from their first operating year. Projects that commission later join a more mature GoO market with more participants and more competition.

The timing is not trivial.

📅 The 2026–2028 Window — Why This Moment Doesn’t Come Back

Unit 2 returns around June 1. Pressure eases. This isn’t the beginning of a permanent crisis.

It’s the preview.

In December 2026, Unit 1 enters refurbishment. Romania runs on 700 MW of nuclear baseload — not 1,400 MW — for two years. During that window, four structural conditions converge:

Condition 1: Lower dispatchable backup. The midday surplus / evening peak differential that produced the 200 EUR/MWh spread in April doesn’t moderate. It widens. One reactor means less dispatchable cushion on days when solar is low and demand is high.

Condition 2: Higher renewable penetration. Solar capacity additions accelerate under AccelerateEU targets. More solar means more midday surplus, more frequent export at low prices, and a deeper evening deficit. The 28 million EUR monthly outflow number doesn’t improve — it grows.

Condition 3: PICASSO aFRR integration matures. Romania’s participation in the European aFRR balancing platform deepens through 2027–2028. Cross-border ancillary market access supports clearing prices in the FCR/aFRR stack and creates new dispatch opportunities for BESS assets pre-qualified under Transelectrica’s ancillary service framework.

Condition 4: GoO trading activates. January 2027.

These four conditions overlap for the duration of Unit 1’s refurbishment — December 2026 through December 2028. Projects that commission during this window operate at the peak convergence of structural grid need and commercial revenue opportunity.

Post-2028, Unit 1 returns. The dispatchable balance improves. The exceptional spread conditions moderate. Projects commissioning after that date operate in a more normalised market.

The entry window is specific. It doesn’t repeat.

🏗️ What Bankable BESS Architecture Looks Like in This Context

How independent engineers are re-reading Romanian revenue projections

When an IE assesses a Romanian BESS project, they model expected DAM spreads under conservative grid conditions — historically 30–40 EUR/MWh. They model FCR/aFRR clearing prices against observed market data. They apply LFP degradation curves (0.7% annually under warranty, approximately 1.0% post-warranty), model augmentation CAPEX at year 12–15, and stress-test DSCR at 1.35× across the full debt tenor — not just year one.

May 10 is now a data point in that picture. An IE reviewing a Romanian BESS project in late 2026 has observed 200 EUR/MWh DAM spreads during a structural grid event — one that will recur systematically during Unit 1’s refurbishment, not as a random shock.

IEs won’t adopt May 10 as the base case. Conservative methodology doesn’t work that way. But it shifts the framing of what qualifies as a stress scenario, what the p90 downside looks like, and what range of spread conditions produces debt service coverage. The 30–40 EUR/MWh base assumption now has a documented outlier at 200 EUR/MWh that occurred because of a structural condition lasting two years.

That changes the credit committee conversation. Not dramatically — but at the margin, and the margin is where deals close or don’t.

The four things lenders actually need to see

1. A multi-mechanism revenue stack. DAM arbitrage, FCR/aFRR, balancing market, GoO from January 2027. Single-mechanism projects — pure arbitrage plays, pure ancillary — face revenue concentration risk questions from lenders. The stack diversifies that risk and demonstrates real grid service capability across operating conditions.

2. A technically credible OT architecture. Network Code on Cybersecurity requirements apply to BESS assets connected to Romania’s transmission network. NIS2 extends those obligations. The EMS must demonstrate operational technology isolation from corporate IT, no persistent remote access channels through Chinese OEM systems, and a cybersecurity posture that survives independent technical review. Independent engineers are asking about this now. Lenders are asking about it now. Preparing these answers at credit committee stage is structurally late.

3. A financial model built around conservative spreads, not May 10. The stress test that matters is not whether the project is viable at 200 EUR/MWh spreads — it obviously is. The test is whether it covers at 1.35× DSCR under 35 EUR/MWh normal conditions, with May 10–type events modelled as periodic upside rather than baseline assumption. Projects that need May 10 to cover their debt service don’t survive IE review.

4. An augmentation plan. LFP chemistry degrades. 0.7% annually under warranty conditions, approximately 1.0% post-warranty. Over a 20-year project life, the effective capacity of the asset shrinks meaningfully without augmentation. Modelling that CAPEX event at year 12–15 is not optional — it’s one of the first things a lender’s technical advisor checks before FID. Projects that omit it get restructured. Sometimes they get restructured after commitment, which is more expensive than doing it correctly in the model.

🎯 What Developers Should Be Doing Before the Window Opens

This section is not strategic. It’s operational.

ATR applications: Transelectrica’s grid connection queue runs two to four years. If you’re targeting a 2027 commissioning date and your ATR process hasn’t started, your commissioning date is aspirational. Start the CTE study now — the ATR agreement is what validates the grid connection, and you need it before lenders will take the technical package seriously.

Revenue model calibration: Review your DAM spread assumptions against the post-May 2026 dataset. Not to build 200 EUR/MWh into the base case. To understand whether your existing p50 and p90 assumptions hold under the Unit 1 refurbishment condition, and whether your DSCR coverage at normalised spreads remains where you think it is.

NIS2 / Network Code technical review: If you haven’t had an OT architecture review against Network Code on Cybersecurity and NIS2 implementation obligations, do it before submitting to lenders. The alternative is receiving IE comments during technical due diligence — at the moment when project timelines are most sensitive to delay.

OEM contract review: The supply chain decisions you make in 2026 — specifically on EMS architecture and the remote access provisions buried in OEM service contracts — determine your cybersecurity posture for the project’s operating life. This is not a procurement question. It is a bankability question. Review the contracts before you sign them, not after.

GoO eligibility: Ensure your development structure and registration process qualifies for GoO cross-border certificate trading under GEO 59/2025. If the corporate structure or project registration doesn’t qualify, January 2027 arrives without the fourth revenue layer. Fixing it retroactively is possible. It is also slower and more expensive than structuring correctly upfront.

The grid stress-tested itself on May 10, 2026.

The results were public, legible, and precise: 140 EUR/MWh spot price, 2,200 MW of imports in a single evening, a 200 EUR/MWh DAM spread that nobody captured because there wasn’t enough storage on the Romanian grid to capture it. Behind that single event: a 28 million EUR monthly outflow that had been running since April, driven not by crisis but by a grid architecture selling cheap and buying expensive every single day.

From December 2026, the conditions that made May 10 visible become the structural baseline.

Battery storage is not the only answer to that problem. It is the fastest deployable answer — 18 to 24 months from FID to commissioning, which puts projects squarely within the Unit 1 refurbishment window if development decisions are made now. The projects that commission in 2026–2027 operate through the peak convergence of grid need and commercial opportunity. The ones that commission in 2029 operate in a different market.

The arithmetic isn’t complicated. The decisions are.

— Tudor, RENEWD

In December 2025, Romanian Waters had its systems compromised. Two weeks later, Oltenia Energy Complex — one of Romania’s largest power producers — reported a significant cyber incident affecting operational systems.

Neither attack was sophisticated. Both were preventable. Both are now on the public record as evidence that Romanian energy infrastructure is an active target, not a theoretical one.

They’re also the reason NIS2 in Romania isn’t an academic discussion about European regulatory alignment. It’s a compliance obligation with teeth, a penalty regime with €10 million ceilings, and a board accountability framework that makes “we delegated this to IT” a legally inadequate defence.

The Romanian NIS2 framework has been in force since August 2025. The T1 registration deadline passed in September. The compliance clock is running for every energy operator in scope — and the scope is wider than most assume.

⚖️ The Romanian NIS2 Framework — What’s Actually in Force

What legislation governs NIS2 compliance in Romania?

Romania transposed the EU NIS2 Directive (2022/2555, adopted 14 December 2022) through GEO 155/2024 — the Government Emergency Ordinance published on 31 December 2024 in Monitorul Oficial nr. 1332. Parliamentary confirmation followed through Law 124/2025, in force from 10 July 2025.

The operational layer sits in two implementing orders from DNSC (Directoratul Național de Securitate Cibernetică), both effective 20 August 2025:

• Order DNSC nr. 1/2025 — entity registration rules, notification procedures, the NIS2@RO submission platform

• Order DNSC nr. 2/2025 — risk management measures, the impact assessment methodology, the maturity self-assessment framework

DNSC is the competent authority. CERT-RO, integrated within DNSC, is the national CSIRT handling incident notifications. For the energy sector specifically, ANRE (Autoritatea Națională de Reglementare în Domeniul Energiei) functions as the sectoral authority and may interface with DNSC on energy-specific cybersecurity requirements.

This is not proposed legislation. It is not pending secondary regulation. It is in force. Entities in scope have obligations now.

🏭 Who Is In Scope — The Energy Sector in Annex I

Which energy operators fall under NIS2 in Romania?

NIS2 covers the energy sector through Annex I, which Romania’s GEO 155/2024 implements across six sub-sectors. Five mirror the EU Directive. The sixth is a Romanian-specific addition.

(a) Electricity — the broadest sub-sector. Covers:

• Electricity undertakings (umbrella category)

• Producers — PV, wind, hydro, thermal, cogeneration, nuclear

• Distribution system operators (DSOs): E-Distribuție, Delgaz Grid, DEER

• Transmission system operator (TSO): Transelectrica

• NEMO: OPCOM

• Market participants providing aggregation, demand response, or energy storage services — this explicitly captures BESS standalone operators, BESS co-located with generation, balancing service providers, and aggregators. Battery storage is not a grey area in Romanian NIS2 scope. It’s named.

• EV charging point operators

(b) District heating and cooling — ELCEN, RADET successors, Termo Iași and equivalent municipal operators

(c) Oil — transmission pipeline operators, refining and treatment facilities, storage entities: Conpet, Petrom, Rompetrol, ADR

(d) Gas — supply undertakings, DSOs, TSO (Transgaz), storage operators (Depogaz), LNG, refining

(e) Hydrogen — producers, storage operators, transport operators (still emerging in Romania)

(f) Modernization Fund beneficiaries — Romania’s own addition, not present in the EU Directive. Entities receiving non-reimbursable EU funds under OUG 60/2022 enter scope regardless of company size. A 5MW BESS SPV with three employees is in scope if it received Modernization Fund financing. Size exemptions don’t apply.

If your operation falls under any of these categories — producer, storage operator, grid service provider, balancing participant — you are in Annex I. The question from there is not whether you have obligations, but which tier of obligation applies.

📊 Essential vs Important — The Classification That Determines Your Penalty Exposure

What is the difference between Essential and Important entities under Romanian NIS2?

The classification determines your penalty ceiling and the intensity of DNSC’s supervisory attention.

Essential Entities (Entități Esențiale):

• Penalty cap: €10 million or 2% of global annual turnover, whichever is higher

• Triggered by: Annex I sector + large enterprise (≥250 employees, OR turnover >€50M, OR balance sheet >€43M)

• Also triggered by: impact assessment scoring HIGH on any of the six Order 2/2025 dimensions, regardless of size

• Includes by default: Transelectrica, OPCOM, major DSOs, large producers

Important Entities (Entități Importante):

• Penalty cap: €7 million or 1.4% of global annual turnover, whichever is higher

• Triggered by: Annex I or Annex II sector at medium-enterprise size (50–249 employees, OR turnover €10M–€50M)

Out of scope: Micro and small enterprises below medium-size thresholds — unless they trigger a “regardless of size” provision. Modernization Fund beneficiaries (letter f) are the most common such trigger in the energy sector.

There is also an impact methodology trap that catches small operators with disproportionate grid criticality. Order 2/2025 establishes six scoring dimensions — fundamental rights, national economy, health and human life, financial impact, defense and public order, cross-sectoral impact — each scored High (10), Medium (5), or Low (0). A 10MW BESS that is the sole fast-frequency response asset in a specific grid segment can score HIGH on national economy and defense dimensions. That score triggers Essential classification regardless of headcount.

The self-assessment of this methodology is not straightforward. It requires OT/energy domain knowledge to apply correctly — which is precisely why most SPVs shouldn’t be doing it without sector-specific support.

🔒 What You Actually Have to Do — The 10 Mandatory Measures

What cybersecurity measures does Article 21 NIS2 require?

Every entity in scope must implement ten specific cybersecurity risk management measures under Article 21 of the Directive, transposed into GEO 155/2024:

1. Risk analysis and information system security policies — documented, approved by the management body

2. Incident handling — detection, classification, response, escalation procedures

3. Business continuity — backup management, disaster recovery, crisis management protocols

4. Supply chain security — assessment of direct suppliers and service providers, including EPC contractors, O&M providers, SCADA integrators

5. Security in network and information system acquisition, development, and maintenance — vulnerability handling, disclosure procedures

6. Effectiveness assessment — policies and procedures for testing whether your cybersecurity measures actually work

7. Cyber hygiene and training — basic practices embedded, regular staff training, management body training

8. Cryptography and encryption — policies for data in transit and at rest

9. Human resources security, access control, asset management — who has access to what, and how that access is controlled

10. Multi-factor authentication and continuous authentication — MFA is not optional; secured voice, video, and text communications

Measure 4 — supply chain security — is where most energy operators underestimate their exposure. Article 21(d) obligations don’t stay within your perimeter. If your EPC contractor has access to your SCADA system, or your O&M provider manages your BMS remotely, their cybersecurity posture is now part of your compliance obligation. The obligation cascades downstream whether or not your contractors are aware of it.

📅 The Compliance Timeline — Where You Are Now

What are the NIS2 deadlines for Romanian energy entities in 2026?

The T1 registration deadline of 19 September 2025 has passed. For entities that registered, the compliance sequence runs from their confirmed DNSC registration date:

Milestone Timing Obligation T1 19 Sept 2025 (passed) DNSC registration via NIS2@RO platform or alternative channel T2 T1 + 60 days (essential) / T1 + 150 days (important) DNSC confirms registration T3 T2 + 30 days Management body formally designates cybersecurity responsible persons T4 T2 + 60 days Risk level assessment submitted to DNSC per Order 2/2025 methodology T5 T4 + 60 days Maturity self-assessment of cybersecurity risk management submitted T6 T5 + 30 days Remediation plan (plan de măsuri) submitted for identified gaps

For Essential entities, the full cycle from T1 runs approximately 8–9 months. For operators who haven’t registered yet: the T1 deadline has passed, which creates sanctionable exposure, but registration now followed by a structured compliance programme is substantially better than continued inaction. DNSC’s enforcement posture has been calibrated toward entities demonstrating no compliance effort. Document everything.

One practical note on the platform: the NIS2@RO system has had functionality gaps since launch. Order 1/2025 explicitly provides for alternative submission channels — email, physical delivery, courier — precisely because the platform isn’t fully operational. Alternative submission is legally valid and creates a documented submission record. Platform limitations are not a legal defence for non-registration.

🔋 What This Means Specifically for BESS and Renewable Energy Operators

The energy storage sector has a specific attack surface that generic cybersecurity frameworks don’t address. BESS assets connect BMS (battery management systems), EMS (energy management systems), SCADA, and in most operational deployments, vendor remote access portals — multiple OT/IT convergence points that each represent an entry vector.

The December 2025 attacks on Oltenia Energy Complex weren’t targeting the finance function. They targeted operational systems. The asymmetry in Romanian energy cybersecurity is that OT environments — the systems that actually control physical energy assets — have historically been managed with far less security rigour than IT environments. NIS2 closes that gap legislatively. The 10 mandatory measures apply to operational technology as much as to office networks.

For BESS operators specifically: your BMS vendor’s remote access credentials, your EMS connectivity to OPCOM for balancing market participation, your SCADA interface with Transelectrica’s grid monitoring system — all of these are now within the scope of your Article 21 obligations. The supply chain security requirement (Measure 4) means your Chinese BMS manufacturer’s firmware update process is a documented risk you are required to assess.

This is also increasingly a bankability issue. Lenders are beginning to integrate NIS2 readiness into financing covenants, particularly for projects seeking institutional capital through IDB or Catalyst Europe instruments. A project that reaches credit committee without a documented NIS2 compliance programme faces questions that delay drawdown. The full picture of how institutional lenders are assessing BESS technical risk is covered in the debt committee series.

Board Accountability — The Governance Layer That Changes Everything

GEO 155/2024 is explicit on management body liability. The board must approve cybersecurity risk management measures, oversee implementation, receive regular training, and be informed of significant incidents. This is not a delegation to the IT director. It is a governance function with personal liability attached.

January 2026 draft secondary legislation formalises the role of Cyber Security Governance Manager — a designated position for organisational cybersecurity posture accountability, separate from IT management.

“We have an IT team handling this” is not a governance defence under GEO 155/2024. The board approved the risk management framework, or it didn’t. If an incident occurs and the board had no oversight mechanism, the liability picture is clear.

The ISO 27001 defence doesn’t hold either. Certification covers approximately 60% of NIS2 requirements. The gaps — OT-specific controls, DNSC submission formats, the 24-hour incident early warning timeline, supply chain obligations — are material. ISO is a useful starting point. It is not a compliance finish line under Romanian law.

Incident Reporting — The 24-Hour Clock

When a significant incident occurs, the reporting timeline is non-negotiable:

• 24 hours — early warning to CERT-RO: nature of incident, suspected cause, any cross-border impact

• 72 hours — full incident notification: updated assessment, severity classification, indicators of compromise, initial response actions

• On request — intermediate report to DNSC

• Maximum 1 month — final report with root cause analysis, remediation steps, lessons learned

The 24-hour window is the one that creates operational pressure. Most energy operators don’t have a documented incident classification framework that can determine within 24 hours whether a cybersecurity event meets the NIS2 notification threshold. Building that framework is part of Measure 2 (incident handling). It needs to exist before the incident, not during it.

Where to Start

If you’re an energy operator in Romania — BESS, PV, wind, gas, a balancing service provider, an EPC contractor with direct beneficiary clients — the first question is scope. Are you in Annex I? Under which sub-sector letter? Essential or Important classification?

From there, the compliance sequence is the T1-T6 calendar above. Registration, designation, risk assessment, maturity self-assessment, remediation plan. Each step has a defined output. None of them are optional.

The framework is clear. The deadlines are running. The December 2025 incidents established that Romanian energy infrastructure is a live target — not a future scenario.

If you’re uncertain about your NIS2 scope status or where you sit on the T1-T6 timeline, the 5-minute scope assessment is the right starting point. Or schedule a 30-minute discovery call if you already know you need to move.

— Tudor

Want to read more of my stuff and gain access to exlcusive pieces on Substack? Consider becoming a paid subscriber to support my work 👇

Subscribe now

Nobody warns you about the access problem!

The EU announces AccelerateEU. The IDB launches with a €100 billion mobilisation target. Catalyst Europe comes online. The EIB creates an equity fund. A PPA counter-guarantee pilot follows. The capital press releases stack up like delivery confirmations on a package that never arrives.

Here’s the uncomfortable reality of AccelerateEU BESS financing in 2026: these instruments are genuinely funded and operational, and most BESS developers in Romania and CEE are still not accessing them. Not because the projects are unviable. Because the access architecture is more complex than the policy announcements suggest — and the developers who get in first are the ones who mapped it before they needed it.

This article maps five AccelerateEU financing instruments for BESS projects: what each one is, what type of project it serves, how you actually access it, and what the common mistakes look like at each entry point. The sequencing section at the end is where most of the value sits — because using these instruments correctly is as much a timing problem as a structure problem.

🏦 What Is AccelerateEU and Why Does It Matter for BESS?

AccelerateEU is the EU’s emergency capital architecture for strategic clean energy infrastructure, formalised in April 2026 following the reclassification of BESS as primary strategic infrastructure — the same category as power transmission lines and gas interconnectors.

The reclassification came with a capital mandate: €660 billion per year in clean energy investment through 2030, three times the annual average that funded the first decade of wind and solar.

The instruments doing the mobilising are primarily debt, not grants. The Industrial Decarbonisation Bank (IDB), Catalyst Europe, InvestEU, the Strategic Infrastructure Investment (SII) Fund, and the PPA Counter-Guarantee Pilot represent the five main access points for BESS project finance through the AccelerateEU architecture. A sixth instrument — Romania’s €150 million standalone scheme — operates alongside these as a national-level grant layer.

For a 150MW/300MWh BESS project in Romania generating EUR 14–20M annually across DAM arbitrage, FCR/aFRR, and ID, the difference between accessing and not accessing these instruments is typically 300–500 basis points on the cost of debt — and in some configurations, the difference between a project that reaches financial close and one that doesn’t.

1. 🏛️ The Industrial Decarbonisation Bank — First-Loss Guarantees for Commercial Debt

What is the IDB and how does it work for BESS financing?

The Industrial Decarbonisation Bank is the flagship AccelerateEU instrument. It operates as a first-loss guarantee provider — it absorbs the first tranche of loss on qualifying BESS projects, which allows commercial lenders and institutional investors to participate at senior debt rates they wouldn’t otherwise accept for merchant battery storage exposure.

The IDB doesn’t lend directly to projects. That’s the detail most developers miss. It provides guarantees to participating financial institutions — typically national promotional banks and commercial lenders with clean energy mandates — who then underwrite the project debt with the IDB guarantee backstopping their credit exposure. Getting IDB support means working through an IDB-participating institution, not applying to the IDB directly.

For a Romanian BESS project, the practical question is: which banks operating in the Romanian market are IDB-participating institutions, what are their current credit criteria for storage projects, and how is the IDB guarantee sized relative to total project debt? These questions have answers, but they’re not in the public documentation.

What this instrument is best for: Projects above 30MW with a clear taxonomy compliance case — meaning documented grid services contribution (FCR, aFRR, balancing market participation), not merely clean technology classification. Merchant-only arbitrage projects face taxonomy questions at credit committee level.

Where developers stall: Arriving at the IDB gateway without an established relationship with a participating institution. The IDB’s institutional process doesn’t wait while developers build bank relationships. The developers accessing this instrument in 2026 started the relationship in 2024.

DSCR requirement: IDB-backed project finance structures require a minimum 1.35× DSCR stress-tested across the full loan tenor — not year one only. Augmentation CAPEX at year 12–15 must be modelled as a reserve, not an assumption.

2. 🔋 Catalyst Europe — Direct EIB Loans for Utility-Scale Projects

What does Catalyst Europe offer BESS developers?

Catalyst Europe provides up to €150 billion in loans backed by the EU budget for strategic technologies, including utility-scale BESS. Unlike the IDB, Catalyst Europe lends directly — the EIB Group is the lender of record, not a guarantor behind a commercial institution.

This changes the due diligence process significantly. EIB direct lending involves the EIB’s own technical advisors, environmental assessment, and procurement review — all completed before disbursement. The timeline is longer than commercial bank lending. The standards are higher. That’s by design: the EIB is deploying public capital and applies corresponding scrutiny.

For projects in the 10–50MW range, Catalyst Europe is typically not the primary instrument. It functions best as a co-financing layer — covering the portion of the debt that exceeds a single commercial lender’s appetite.

Understanding how to size the Catalyst Europe tranche relative to commercial debt, and how to sequence the two credit processes so they close simultaneously, is one of the more technically demanding aspects of structuring a BESS financing.

What this instrument is best for: Projects above ~50MW where the debt quantum exceeds the appetite of any single commercial lender. Catalyst Europe steps in where the commercial market has a capacity ceiling, not where it has a risk problem.

Where developers stall: Underestimating the EIB technical review process. Projects that arrive at Catalyst Europe without completed independent technical advisor reports, without SoH modelling that a technical advisor has already reviewed, and without a procurement specification that addresses the EIB’s procurement standards — these projects don’t get declined, they get deferred. Indefinitely, in practice.

Key procurement standard: UL 9540A safety certification, active DC protection to aerospace-grade specifications, SoH algorithm integration from day one of operations. Pass/fail requirements, not scoring criteria.

3. 💼 InvestEU — The Mid-Market Access Point

How does InvestEU support BESS project finance for mid-market developers?

InvestEU’s €26.2 billion budget guarantee operates through national implementing partners — commercial banks, development banks, leasing companies — each with their own sector focus, minimum deal size, and credit criteria. The Sustainability Guarantee under the EIF is particularly relevant for BESS developers who don’t yet have the balance sheet to qualify for direct IDB or Catalyst Europe support.

The mechanism: InvestEU provides first-loss protection to local banks and funds lending to mid-market clean energy projects. The same guarantee can be accessed through a Romanian commercial bank or through a pan-European infrastructure fund, at very different terms and with very different requirements.

This is where the implementation partner selection matters enormously. Mapping which implementing partners are active in Romanian BESS, what their current deployment appetite looks like, and how their criteria interact with your specific project structure is the first operational question — and it’s not answered on the InvestEU website.

What this instrument is best for: Mid-market projects (10–50MW) where the developer has a strong project structure but not the institutional relationships required for direct IDB or Catalyst Europe access. InvestEU effectively democratises access to EU guarantee backing through the local banking network.

Where developers stall: Treating InvestEU as a single instrument with a single access point. It’s a network. The right implementing partner for a 20MW BESS project in Dobrogea may be a Romanian commercial bank you’ve already worked with — or it may be a pan-European infrastructure fund with a specific CEE mandate. Identifying the right entry point requires market knowledge that isn’t publicly available.

4. 🏗️ The SII Fund — Equity Co-Investment for Build-and-Refi Structures

What is the SII Fund and when should BESS developers use it?

The Strategic Infrastructure Investment (SII) Fund is a €500 million EIB-anchored equity fund that co-invests directly in grid infrastructure alongside private infrastructure managers. It takes a minority stake in the project company rather than providing debt — making it an equity instrument, not a guarantee or loan.

For developers structuring a build-and-refi — where equity holds through construction and stabilisation, then refinances at a lower cost once operational data is established — the SII Fund represents a specific co-investment opportunity that can bridge the equity gap between developer capital and institutional infrastructure fund minimums.

The critical requirement: SII Fund co-investment is structured through partnerships with private infrastructure managers, not direct project applications. A developer seeking SII Fund equity needs either an existing relationship with one of the fund’s private co-investment partners, or an introduction pathway through the EIB’s project team. This is more relationship-dependent than any other instrument in the AccelerateEU architecture.

What this instrument is best for: Projects where the equity capitalisation table has a gap — where developer equity plus any grant contribution falls short of the 20–30% equity requirement for non-recourse debt, and where the developer doesn’t have a pre-existing institutional equity partner.

Where developers stall: Trying to approach the SII Fund without an existing private infrastructure manager relationship. The fund doesn’t take direct approaches from project developers. The access path runs through private partners — which means the developer needs to be having conversations with infrastructure managers before the project reaches the equity closing stage.

5. ⚡ The PPA Counter-Guarantee Pilot — Unlocking Non-Recourse Debt for Merchant Projects

How does the PPA counter-guarantee pilot change BESS bankability?

The €500 million EIB pilot provides credit enhancement for long-term Power Purchase Agreements between BESS developers and SME offtakers. It solves a specific bankability problem: non-recourse project finance for BESS typically requires demonstrated revenue certainty, which merchant projects don’t have by definition.

Here’s the mechanism. A long-term PPA with an SME counterparty creates contracted revenue — but SME creditworthiness is typically too weak to anchor a project finance structure. The counter-guarantee wraps the SME’s credit risk, allowing the PPA to function as contracted revenue for debt underwriting purposes. A project that was previously financed on a partial-recourse or corporate-guarantee basis can now access non-recourse project finance.

This matters enormously for the Romanian market. The ability to convert what was a purely merchant project into a partially contracted revenue structure widens the commercial debt market significantly — and changes which projects can reach financial close without developer balance sheet support.

What this instrument is best for: Developers who have identified a long-term SME offtaker but whose counterparty creditworthiness isn’t sufficient to anchor project finance. Also valuable as a structural tool for projects that need a contracted revenue floor to pass a bank’s revenue stress test.

Where developers stall: This instrument is in pilot phase. Uptake is currently concentrated in Iberian markets where hybrid solar-plus-storage PPA models are more mature. Romanian developers who engage with this instrument now — while it’s still in pilot — have an advantage in the implementation partner conversations that will define how it deploys in CEE markets.

#How to Sequence AccelerateEU BESS Financing Instruments

The choice between these five instruments isn’t really a choice — it’s a sequence. And the sequencing has a specific logic.

The sequencing constraint is real. The national grant application has to be filed before the debt structure is finalised, because the grant terms affect permissible revenue stack assumptions in the project finance model.

The Catalyst Europe or IDB tranche has to be sized before the commercial bank approaches, because its presence changes what commercial lenders will underwrite. The SII Fund equity decision has to come before the equity capitalisation table is closed.

None of this is obvious from the instrument descriptions. All of it is sequential in ways that, if you get wrong, require you to restart upstream. The developers who access the 2026–2028 window fastest are those who mapped the instrument stack before they needed it.

The Romania-Specific Picture

Romania sits in an unusually strong position within the AccelerateEU architecture for two reasons.

First, the €150 million standalone BESS scheme — one of the most concrete grant opportunities in CEE — mandates a 2:1 capacity-to-power ratio at a €69,000/MWh cap through competitive bidding. This aligns exactly with the duration configuration that current IRR modelling supports for the Romanian market. The grant rewards the most bankable projects, not the earliest applications.

Second, GoO cross-border certificate trading comes online from January 2027 under GEO 59/2025, adding a revenue layer to the stack that most existing financial models don’t yet include. For projects commissioning in late 2026 or early 2027, this is upside available from year one — but only if the project has been registered with the national issuing body as part of the grid connection process. It’s not automatic.

The PICASSO aFRR integration matures in parallel. Romanian projects prequalified for aFRR delivery are accessing a cross-border balancing market where clearing prices sit at 60–70 EUR/MWh — above what the domestic markets alone delivers.

The window for capturing peak ancillary pricing — 2027 to 2033 — compresses after 2028 as market saturation drives frequency service prices down. Projects that commission in this window and build a revenue stack across FCR, aFRR, balancing, DAM arbitrage, and GoO certificates are structurally better positioned for the post-2033 arbitrage-dominant market than any single-mechanism project.

The Preparation Gap

The instruments are funded. The access architecture exists. And yet the most common outcome for BESS developers engaging with AccelerateEU in 2026 is not a funded project — it’s a deferred one.

The gap isn’t a financing gap. It’s a preparation gap. Projects that reach the IDB gateway without a bankable EPC, or without a DSCR model stress-tested across the full tenor at 1.35× minimum, or without a taxonomy compliance document, or without a relationship with an IDB-participating institution — these projects don’t get declined. They get deferred, which in institutional capital deployment terms means indefinitely.

If you’re currently working through any of the access questions raised here, the conversation starts with a message.

The window is open. The preparation required to walk through it is demanding. Both of those things are equally true.

— Tudor Grigore, RENEWD

There is a version of the energy storage opportunity that looks obvious from the outside.

The EU has committed to 200GW of storage by 2030. AccelerateEU reclassified BESS as strategic infrastructure in April 2026. Six separate institutional mechanisms — IDB, Catalyst Europe, InvestEU, the SII Fund, the PPA counter-guarantee pilot, the Investment Booster — are now funded and operational. The capital exists. The mandate is clear. The only remaining question appears to be which projects get built.

That version is accurate and almost entirely useless.

The capital exists the way the ocean exists for someone who needs fresh water. The volume is not the constraint. The structure is. Knowing which instruments apply to which project at which stage of development — and in which sequencing, against which eligibility criteria, through which institutional relationships — is a category of knowledge that most developers working in Romanian and CEE markets don’t have. Not because they’re unsophisticated.

Because this architecture was assembled over 24 months of parallel policy processes across six separate EU institutions, and the operational guidance is scattered, contradictory in places, and in several cases not yet published.

This piece is an attempt to map it. Not the summary version — the working version.

1. 🏗️ The Five Instruments — What They Are and How They Actually Work

The Industrial Decarbonisation Bank (IDB)

The IDB is the flagship. €100 billion mobilisation target over 10 years, operating as a project finance engine for clean technology deployment. It is not a grant instrument. It is a first-loss guarantee provider — meaning it absorbs the first tranche of loss on qualifying projects, which allows commercial lenders and institutional investors to participate at senior debt rates they wouldn’t otherwise accept for merchant BESS exposure.

The mechanics matter. The IDB doesn’t lend directly to projects. It provides guarantees to participating financial institutions — typically national promotional banks and commercial lenders with clean energy mandates — who then underwrite the project debt with the IDB guarantee backstopping their credit exposure. To access IDB support, a project needs to go through an IDB-participating institution.

The list of participating institutions is not comprehensive across all EU markets. Identifying the right entry point for a Romanian project — which national promotional banks are IDB-participating, what their credit criteria are, how IDB guarantee coverage is sized against the project — is the first operational question, and it’s not answered in the public documentation.

Eligibility requires demonstrating “significant contribution” to climate objectives under the EU taxonomy — which for BESS means documented grid services contribution, not merely clean technology classification.

A project that delivers FCR and aFRR to a TSO is clearly eligible. A project that primarily does merchant arbitrage without documented ancillary service capacity may face taxonomy questions at credit committee level.

Catalyst Europe

Up to €150 billion in loans backed by the EU budget for strategic technologies. Catalyst Europe operates through the EIB Group and is specifically designed for utility-scale infrastructure where commercial debt is insufficient or unavailable. The practical coverage for BESS: projects above roughly 50MW where the debt quantum exceeds the appetite of any single commercial lender.

The critical distinction from IDB: Catalyst Europe provides the loan directly, not the guarantee. This changes the due diligence process significantly. EIB direct lending involves the EIB’s own technical advisors, environmental assessment, and procurement review — all of which must be completed before disbursement. For a developer accustomed to commercial bank timelines, the EIB direct lending process can feel slow. It is thorough for a reason: the EIB is lending public money and applies corresponding standards.

For projects in the 10–50MW range, Catalyst Europe is typically not the primary instrument — it’s a co-financing layer for the portion of the debt that exceeds commercial bank appetite. Understanding how to size the Catalyst Europe tranche relative to the commercial debt tranche, and how to sequence the two credit processes so they close simultaneously, is one of the more technically demanding aspects of structuring a BESS financing.

InvestEU — Infrastructure and Innovation Windows

InvestEU’s €26.2 billion budget guarantee operates through the European Investment Fund (EIF) for SME-focused applications and through national implementing partners for infrastructure-scale projects. The Sustainability Guarantee under the EIF is particularly relevant for BESS developers who don’t yet have the balance sheet to qualify for direct IDB or Catalyst Europe support — it provides first-loss protection to local banks and funds lending to mid-market clean energy projects. ⚡

The nuance here is implementation partner selection. InvestEU operates through a network of implementing partners across member states — commercial banks, national development banks, leasing companies — each with their own sector focus, minimum deal size, and credit criteria.

The same InvestEU guarantee can be accessed through a Romanian commercial bank or through a pan-European infrastructure fund, at very different terms and with very different requirements. Mapping which implementing partners are active in Romanian BESS, what their current deployment appetite looks like, and how their criteria interact with your project structure is not something you can determine from the InvestEU website.

The Strategic Infrastructure Investment (SII) Fund

€500 million EIB-anchored equity fund, designed to co-invest directly in grid infrastructure alongside private infrastructure managers. This is an equity instrument — it takes a minority stake in the project company rather than providing debt. For developers who need equity capital alongside senior debt, or who are structuring a build-and-refi where equity holds through construction and stabilisation, the SII Fund represents a specific co-investment opportunity.

The critical requirement: SII Fund co-investment is structured through partnerships with private infrastructure managers, not direct project applications. A developer seeking SII Fund equity needs either an existing relationship with one of the fund’s private co-investment partners, or an introduction pathway through the EIB’s project team. This is relationship-dependent in a way that the other instruments are not.

The PPA Counter-Guarantee Pilot

€500 million EIB pilot providing credit enhancement for long-term Power Purchase Agreements between BESS developers and SME offtakers. The bankability challenge it solves: non-recourse debt for BESS typically requires demonstrated revenue certainty, which merchant projects don’t have.

A long-term PPA with an SME counterparty creates contracted revenue — but SME creditworthiness is typically too weak to anchor a project finance structure. The counter-guarantee wraps the SME’s credit risk, allowing the PPA to function as contracted revenue for debt underwriting purposes.

This instrument is in pilot phase. Uptake is concentrated in Iberian markets where hybrid solar-plus-storage PPA models are more mature. For Romanian BESS, the PPA counter-guarantee opens a structuring path that didn’t exist 18 months ago — converting what was a purely merchant project into a partially contracted revenue structure that widens the commercial debt market significantly.

2. 📐 What a Real Capital Stack Looks Like

Here is what the financing architecture actually looks like for a utility-scale Romanian BESS project in 2026, layered correctly:

Equity layer (20–30% of total project cost): Developer equity plus potentially SII Fund co-investment for the balance sheet portion that can’t be raised commercially. Equity holds through construction and the stabilisation period.

National grant layer: Romania’s €150 million standalone scheme, capped at €69,000 per MWh via competitive bidding. For a 30MW/60MWh project, the maximum grant contribution at €69K/MWh caps at approximately €4.14 million. This reduces the equity requirement but must be structured carefully to avoid state aid restrictions on the revenue stack model submitted with the grant application. But what about debt? Let’s see how that works! 👇

Read more

The April 2026 launch of the AccelerateEU framework represents a fundamental pivot in the Union’s energy architecture. No longer framed as a temporary crisis-management tool, AccelerateEU serves as a definitive investment signal, transitioning the EU toward a proactive posture of energy sovereignty. By reclassifying Battery Energy Storage Systems (BESS) from speculative “green-tech” to primary strategic infrastructure, the framework provides the regulatory certainty required for the large-scale deployment of institutional capital.

I’ve created this new podcast series to discuss larger topics in a more condensated manner. If you’re liking it please consider subscribing below

Subscribe now