Published on January 6, 2026 · Updated on September 2, 2026
Battery storage stopped being the optional half of a solar project somewhere in the last eighteen months. Across most of Europe and North America the binding constraint on new generation is grid access, and a battery is one of the few things a developer can add to a site that makes a connection easier for the operator to say yes to.
The market has moved with it. BloombergNEF counted 112 GW and 307 GWh of new storage worldwide in 2025, and forecasts 158 GW and 459 GWh for 2026, a 41% jump in a year when solar and wind additions slowed. Europe's operational fleet passed 100 GWh for the first time, with 36 GWh added during 2025 and utility-scale taking more than half of new installations. In the United States, storage accounts for 749 GW of the 2,061 GW sitting in interconnection queues at the end of 2025.
Those figures describe demand. They say nothing about how much harder it has become to place a battery well. Queue rules were rewritten in Germany and Great Britain. Planning law for standalone BESS tightened, then moved again. Noise turned into a permitting blocker. And connection offers now arrive with operating restrictions attached, which drags the revenue question forward into site selection, where most teams have no tooling for it.
The short version
A Battery Energy Storage System stores electrical energy and releases it later. At utility scale, developers use one to shift generation into higher-priced hours, to hold a grid connection that curtailment would otherwise waste, and to make an application competitive in a queue that now scores projects on quality rather than on the order they arrived.
Whether a given parcel can carry a battery comes down to five things: available capacity at a nearby substation, the position of that node in the network, planning and fire rules on the land, flat and accessible terrain, and noise distance to the nearest receiver.
This guide is your complete introduction to utility-scale BESS development. Whether you're exploring battery add-ons for existing PV sites or targeting standalone opportunities near substations, it covers what a BESS is and how developers use one, what changed in the market during 2026, how to judge whether land can carry a battery, when to run that check, how to size a hybrid, and where Glint Solar's BESS workspace fits into the workflow. There's a frequently asked questions section at the end.
A Battery Energy Storage System (BESS) is a system that captures electrical energy, either from the grid or from an on-site generation source like a solar array, and stores it for later use. Simple enough as a description. The effect on utility-scale development is not.
At scale, a BESS separates the moment energy is produced from the moment it's delivered. That matters most where solar generation is abundant during the day but demand and pricing peak in the evening. Spain shows how sharp the gap has become: 41% of the country's April 2026 solar production landed in hours with negative prices, and its solar capture factor fell to around 18% in February 2026 against 71% in the same month a year earlier. A battery on the same connection turns some of that wasted output into revenue.
Most utility-scale battery systems are built from modular, containerized units that store energy chemically and convert it back to electrical energy on demand. A standard 20-foot container holds roughly 5 MWh. The components inside are:
Once commissioned, a BESS can charge from solar or from the grid when power is cheap or plentiful, hold that energy for hours, and discharge when prices rise, demand spikes, or curtailment would otherwise throw the generation away. That time-shifting is why batteries are now designed in from the start rather than bolted on at the end.
Lithium iron phosphate is still the default, but chemistry and configuration have become genuine decisions rather than a formality, particularly for anyone looking at long-duration storage.
Lithium iron phosphate (LFP) dominates grid-scale work. It combines workable energy density, thermal stability, long cycle life and a mature supply chain, and its behavior is well understood. For most 1 to 4 hour assets it remains the obvious choice.
Sodium-ion moved faster than expected during 2026. CATL signed a 60 GWh sodium supply agreement with HyperStrong in April, the largest sodium order placed to date, followed by a memorandum with Alfen covering 5 GWh of deployments in Europe. Commercial deliveries in China begin in September 2026, with international rollout scheduled for mid-2027. Grid storage doesn't need the energy density that makes lithium indispensable in cars, which is why cycle life, safety and cost are enough to make sodium viable here first. Treat it as a live procurement option on space-tolerant sites rather than a settled one.
Flow batteries, vanadium redox and zinc-bromine among them, separate power from energy capacity, which supports 6 to 12 hour durations with low degradation. Footprint and capital cost are both higher, so they suit sites where a long-duration need is clearly monetizable.
Four forces are doing most of the work, and all four have shifted since the start of the year.
Europe has more than 1.7 TW of renewable and hybrid projects waiting in connection queues across 16 EU member states and Great Britain. Germany's grid connection applications ran from around 400 GW at the end of 2024 to more than 720 GW during 2025, against 78 GW with confirmed connection commitments. Italy's queue reached 316.6 GW across 5,749 requests by 30 June 2026, of which only 4% holds a firm connection contract, and 74% of it sits in three southern and island zones.
Adding storage improves how a project uses a constrained node, supports a hybrid application, and gives the operator a reason to prefer it. Our own analysis of Italy's queue and capture rates works through what that looks like in one market.
Midday cannibalization has gone from a forecast to an operating reality. Spain's solar capture rate fell from about 70% in 2023 to 30% in 2025, and by March 2026 the capture price gap had opened to €124.72/MWh in Italy-North against €13.51 in Spain. Italy has held up better, at 82% in 2025 against roughly 87% in 2023 and 2024, with Sardinia already down at 77%. Germany's Solarspitzengesetz made the trend financial in February 2025 by stripping support payments for every hour the wholesale price goes negative, and there were 457 such hours in 2024. A co-located battery is the most direct answer available.
Germany replaced first-come-first-served with a round-based maturity procedure. The first application window opened on 1 April 2026 with a non-refundable fee of €50,000, and projects are scored across four categories: land control, technical concept, applicant capability, and grid and system benefit, the first three weighted at 30% each. Great Britain rebuilt its queue around readiness and strategic alignment criteria, and 283 GW of generation and storage cleared the first round. The European Commission's Grids Package points the rest of the continent the same way, with instructions to member states on connecting grid-friendly projects such as hybrid solar plus storage ahead of the queue.
Two developers competing for the same substation are no longer separated by who filed first. They're separated by evidence, submitted at the same moment.
Flexible Connection Agreements entered German law on 31 January 2025 and are now the default. They attach operating restrictions to the connection offer itself, in four forms: caps on import and export, ramp-rate limits, restrictions on ancillary market participation, and uncompensated re-dispatch. Ancillary service caps of 10 to 90% of installed capacity are the single harshest of them for project economics. Trading some midday production for a connection years earlier is often still a good deal, and the point is that it's now a deal you have to model before you sign.
Shared infrastructure moved in the same direction. Germany legalized cable pooling in January 2025, letting wind and solar operators share one connection and split the capacity-dependent costs. The Dutch regulator allows up to four parties to share a connection from 2 MVA upwards, batteries included. For hybrid and standalone storage, that's the cheapest grid capacity on the market. We've written up five ways a project can serve the grid, which is the frame operators increasingly apply.
Germany's new rules for utility-scale BESS
The full TSO maturity phase timeline, the four scoring categories and what evidence each one wants, grid fee scenarios for the post-2029 tariff, and the FCA stacking economics that drive project IRR. Co-developed with Philipp Schulz of Recurrent Energy, who manages around 150 MWp of PV and 300 MW of BESS development across Germany.
Read the guideWhere BESS development is moving, and what shapes the business case in each market.
| Market | Grid constraint | Revenue mechanisms | Policy and process | Opportunity type |
|---|---|---|---|---|
| United States | 749 GW of storage in queues at end-2025, of a 2,061 GW total | Arbitrage, capacity, ancillary services | 30% ITC for standalone storage, with location-based adders | Standalone, hybrid, retrofit |
| Great Britain | Queue rebuilt on readiness; 283 GW cleared the first round | Capacity Market, Dynamic Containment, frequency response | Gated reform, 15-year capacity contracts | Grid-following BESS, hybrid PPAs |
| Germany | 720 GW of applications against 78 GW of commitments | Spot arbitrage, intraday, reserves, FCA-restricted | Maturity procedure from 1 April 2026; §35 BauGB still moving; §118 EnWG grid fee exemption under review | Hybrid PV plus storage, substation-adjacent standalone |
| Spain | Severe curtailment; capture factor near 18% in February 2026 | Curtailment capture, energy shifting | PERTE grants, hybrid PPAs, grid-forming requirements tightening after the April 2025 blackout | Co-located solar plus storage, retrofit |
| Italy | 316.6 GW queued, 4% with firm contracts, 74% in the south and islands | MACSE capacity auctions, arbitrage, balancing market | Terna publishes connection requests per municipality, refreshed monthly | Co-located and standalone, zone-selective |
| France | Constrained HTA zones published by Enedis; RTE flags prospective limits | Arbitrage, capacity mechanism, ancillary services | PPE3 targets, SimuRacco connection costs and lead times published | Hybrid, standalone near constrained nodes |
| Australia | High in QLD and SA | FCAS, arbitrage, capacity deferral | State-level support including VRET and LTESA | Long-duration and hybrid |
Battery storage is less fussy about land than solar, which is not the same as every parcel working. To get from concept to commissioning a site needs the right combination of grid access, land conditions, permitting environment and technical feasibility. Five factors carry most of the weight during early-stage screening.
Interconnection is the first gate. Standalone BESS projects are sited as close to a substation as the land allows, because every extra meter of line extension adds cost, permitting complexity and voltage drop. In Germany the distance question has a legal edge to it: the §35 BauGB privilege for standalone storage turns on a 200-meter radius from the substation, and authorities and developers still disagree on whether the whole project area has to sit inside it or only the boundary.
Proximity on its own is the naive version of the question. The sharper one is what your project does to the network at that point. A battery in a zone with surplus generation behaves differently from the same battery near concentrated load, and operators are starting to price that difference. Germany's grid fee reform is working toward a dynamic tariff, potentially from 2029, where the charge depends on capacity and location.
Permitting rules for BESS vary more than solar rules and are frequently less mature. Batteries are often classified as industrial infrastructure, which can rule out agricultural or residentially zoned land. What to check:
Batteries don't need to follow the sun, but they do need compacted, level ground and heavy equipment access for installation and long-term maintenance. Workable parcels tend to share four traits:
Hybrid structures make better use of a connection and spread the infrastructure cost. The three common shapes are greenfield solar plus storage under a single interconnection application, a battery retrofitted onto an existing PV site where curtailment is eating yield, and a shared substation build across a wind, solar and battery portfolio.
Some of the fastest-moving storage projects don't start with clean land at all. Former peaker plants and substations with decommissioned capacity, utility-owned grid access points with surplus land, and previously shelved solar parcels all come with the one thing that's scarce. For the checks that tend to eliminate a parcel late, there's a companion piece on what makes a site ideal for battery storage.
Plenty of teams still test BESS feasibility after the solar layout is fixed or the interconnection request is already filed. By then the useful decisions have been made. Substation headroom goes fast, landowners move on, queue positions fill, and a site that's been internally deprioritized costs real time to revive.
Round-based connection procedures punish that sequence directly. Under a scored round, a project that hasn't resolved land control and technical concept isn't late, it's unscoreable. So experienced teams now test battery viability at the same moment they screen for solar, or before it.
| Development stage | Traditional approach | BESS-inclusive approach |
|---|---|---|
| Site lead intake | Land team shares a parcel with GIS | Parcel is reviewed for solar and battery viability together |
| Feasibility screening | GIS checks irradiance, slope and basic zoning | Grid access, BESS setbacks, fire codes and noise distance are screened in the same pass |
| Internal discussion | Solar-only layout and economics | PV and BESS reviewed as one concept, with a modeled ratio behind it |
| Interconnection planning | Submit a solar MW request | Submit a hybrid or BESS-first application, with evidence of efficient grid use |
| Queue round or scoring | File early and wait for position | Present land control, technical concept and system benefit as scored evidence |
By building BESS into early-stage workflows, teams create optionality: they can move forward as solar-only, storage-only or hybrid, depending on how permitting, offtake and grid timelines actually land.
Hybrid is already the default shape for new utility-scale solar. The live question is the ratio between PV output, battery power and battery capacity, and with connection capacity scarce, that ratio decides what the project earns for its whole life.
It's a hard question to answer early because it needs a price series, a dispatch strategy and a grid limit, none of which sit in a yield model. Most teams either send it to a consultant or approximate it in a spreadsheet, and both routes arrive too late to change the site decision.
Glint Solar's PV-BESS Hybrid Analysis models combined performance across the plant lifetime using historical day-ahead spot prices in 18 markets, capturing the revenue effects that standalone PV yield modeling misses. Set a grid limit below PV inverter power to switch curtailment on, then compare four scenarios: full hybrid optimization with arbitrage, green energy storage with prices, green energy storage without prices, and arbitrage only. Outputs run to energy timeseries, state of charge, grid utilization by hour and revenue comparisons, all exportable. It's AC-coupled today, and the grid limit is static across a simulation, so treat it as an early-development answer rather than a financing model.
Battery storage carries more early-stage complexity than solar: interconnection, noise limits, fire setbacks, and a revenue case that depends on hours rather than annual totals. Glint Solar exists to bring those checks forward into the weeks when a decision is still cheap. Seven parts of that matter for BESS work.
Interconnection is the bottleneck, so grid data belongs in the screening step rather than in a separate study. Glint Solar carries grid line and substation layers by voltage level as part of built-in coverage across 35 countries, with 350+ layers in Germany, 200+ in Italy and 150+ each in France and the UK.
Land Screening sweeps an area the size of a US state or a medium-sized European country against a constraint profile you already use, and returns opportunity areas you can sort and filter, each with its buildable area and the parcels behind it.
The value is in what it removes. A realistic utility-scale profile eliminates the overwhelming majority of any region, and what survives is the short list worth a phone call. The parcel breakdown behind an opportunity area also tells you whether it means one landowner or fifty, which is often the difference between a viable project and a two-year assembly exercise. Per-parcel buildable area is close to what the Buildable Area tool returns for the same parcel without matching it exactly, so treat the output as approximate. Land Screening isn't enabled out of the box and is live in selected regions, US states first, with France and Germany close after.
A site that looks clean from the road can fall apart on terrain, zoning or fire code. Area Insights scans a selected area against every constraint layer and reports a status per layer at three levels: Cleared where nothing interacts, Caution where at least one geometry does, and Danger where the area is fully covered. Slope and parcel sections go deeper, including buildable area and parcel inclusion.
The underlying mechanic is the constraint profile, the same object that drives Buildable Area, which turns those rules into a buildable footprint you can measure.
Layout work still sits inside engineering departments at most developers, which puts a bottleneck in front of every feasibility check. For repeatable systems like 2-hour or 4-hour containerized BESS, that wait is avoidable. A BESS profile bundles layout, batteries and MV stations, and a set is the building block replicated to fill the area, either from a predefined arrangement or built in a drag-and-drop editor. No AutoCAD, no desktop install, browser-only.
Noise compliance has become one of the most sensitive permitting hurdles for standalone and hybrid BESS, particularly near housing. Glint Solar's noise modeling produces a site noise map during design, using the European CNOSSOS method, so acoustic risk surfaces before permitting turns into the blocker.
Within about 1 km, deviation from industry reports is typically below 3 dBA in rural areas. Results are preliminary estimates and need regulatory confirmation before formal submission.
Noise emissions from utility-scale BESS
The TA Lärm limits your site will be held to, where the noise actually comes from, which mitigations move the number and by how much, and the order to run them in so you're not paying for an acoustic study on a site you were always going to drop.
Read the guideHybrid Analysis sits in the same workspace as the land screen, which is the point. Revenue attaches to a parcel while the parcel is still a decision, rather than to a project you've already committed to. The hybrid workspace covers the dispatch strategies, price markets and grid-utilization outputs described above.
Containerized infrastructure is hard to explain to landowners, planners and local officials who have no reference point for it. Flat PDFs rarely convey safety buffers, access routes or scale. 3D visualization, a 3D object library for fences, equipment and planting, and Beautify's photorealistic renders exist for that conversation. Beautify runs on a token allowance and uses a third-party AI provider, so image data isn't guaranteed to be processed inside the EU.
Map export produces a page-formatted snapshot with legends for permit applications, available in seven languages. And because land, GIS, commercial and permitting teams all touch the same projects, Site Control keeps parcels, contacts, comments and files against the project as a single source of truth, with Charts for pipeline views and the Project API for syncing to a CRM.
Aurinkokarhu, a Finnish developer, moved into battery and hybrid projects and ran into the two bottlenecks this guide keeps returning to. Noise analysis meant commissioning external consultants, which cost several thousand euros and added weeks of delay at the earliest stage of development. Layout work meant manual AutoCAD processes measured in hours.
"Work has become significantly more efficient. An initial layout can now be generated in just a few minutes."
Aurinkokarhu, on accelerating their pivot to BESS and hybrid
A simple analysis takes about two minutes on their account. The wider point isn't the speed, it's which questions become cheap enough to ask before land is secured.
See your own market screened for BESS
Send us a region or a shortlist you're already working on. We'll run it against a utility-scale constraint profile, show you what gets ruled out and why, put a layout with fire setbacks on a parcel that survives, and model the hybrid ratio against day-ahead prices for that market. Half an hour, on your sites, not a slide deck.
Book a demoThe questions that come up in development meetings, rather than the ones a search box asks.
Price the restrictions rather than the headline. A German Flexible Connection Agreement can carry four: caps on import and export, ramp-rate limits, caps on ancillary market participation, and uncompensated re-dispatch. TSO guidance on ramp rates sits at 6 to 20% per minute, and some DSO proposals go as low as 1% per minute, which is close to unworkable commercially. Ancillary caps run from 10 to 90% of installed capacity and are usually the harshest single term. Model each against the dispatch strategy you're actually underwriting rather than against nameplate, then compare restricted revenue starting in year two with firm revenue starting in year seven, discounted. In congested zones the restricted offer normally wins. The exception is a project whose case rests on frequency services.
Duration follows the revenue stack and the connection terms, not the site. Where arbitrage carries the case, the spread between the cheapest and dearest hours sets how many hours you can profitably fill, and energy capacity past that point earns nothing. Where an ancillary cap bites, capacity you aren't allowed to dispatch earns nothing either. Two-hour German projects delivered around €259,000 per MW in 2025 against return expectations of 15 to 18%. The useful output of the exercise isn't the optimum, it's how flat the revenue curve is across two or three ratios. A flat curve means the ratio isn't your risk and the argument can stop.
It turns on an interpretation nobody has settled. The §35 BauGB privilege for standalone storage is tied to a 200-meter radius from the substation, and authorities and developers disagree on whether the entire project area has to fall inside that radius or only the project boundary. Screen at both readings and know which one your site depends on. The framework is also still moving: a draft amendment circulated in March 2026 proposes replacing federal privileging under §35(1) with a new §35(1a) tied to municipal land use plans, and abolishing overriding public interest status under §35(2) for these projects. A site that only works under the current privilege carries regulatory duration risk, and that belongs in the decision now rather than at planning submission.
Ask what the regime rewards. Where efficient use of an existing connection is scored, and that's the direction of the EU Grids Package, Great Britain's strategic alignment test and Germany's grid and system benefit category, a single hybrid application presents better than two competing ones. Where you're chasing two revenue stacks on different timelines, splitting preserves optionality and doubles your exposure to the same congestion. Worth remembering that submitting one project at several nodes is the behavior that broke first-come-first-served, and the reforms were designed to penalize it.
In Germany, land control, technical concept and applicant capability carry 30% each, and grid and system benefit 10%. Three of those four are things you can evidence during screening: land optioned or signed, a layout with real setbacks that survives the constraint set, and a technical concept consistent with the node you're applying to. In an oversubscribed round the difference between 14 and 11 out of 18 is a binding connection offer against waiting for the next cycle. The first window opened on 1 April 2026 with a non-refundable €50,000 fee, which puts a price on applying with a thin file.
Preliminary modeling on the CNOSSOS method will tell you whether you're in the right order of magnitude at a given receiver, typically within 3 dBA of industry reports in rural areas at up to about 1 km. That's enough to choose between container positions, to size a 3 to 5 m barrier, and to decide whether to walk away. It isn't enough to submit. The sequence that saves money is model first, reposition or drop the site, then commission the formal study only on what you've already decided to pursue. Doing it the other way round is what Aurinkokarhu was paying for: several thousand euros and weeks of delay per site, on sites nobody had qualified yet.
Assume the exemption might not hold. Storage commissioned before the end of August 2029 currently gets a 20-year exemption from grid fees under §118(6) EnWG, and BNetzA's AgNes process, with its January 2026 orientation paper, signals that a full exemption is incompatible with European law. Options under discussion include capacity-based charges, dynamic tariffs linked to congestion, and fees on net withdrawal, with different treatment for standalone, co-located and multi-use assets. New rules are expected to enter force on 1 January 2029. The detail that changes an underwriting model is that BNetzA has indicated Vertrauensschutz can no longer be assumed, so operating assets could be exposed retroactively. Run at least one scenario with the exemption gone.
Four things, roughly in the order they bite. Space inside the site boundary once fire setbacks are applied, which is usually tighter than it looks on the as-built. Whether the connection agreement permits storage at all. Headroom on the existing connection and inverter. And noise distance to the nearest receiver, which on a retrofit is often shorter than you'd accept on a greenfield site, because the PV plant was permitted against limits that didn't contemplate ventilation fans. The number that justifies the project is curtailed volume on the existing asset, not the nameplate you could physically fit.
In some markets, yes, and it's the cheapest grid capacity available. Germany legalized cable pooling in January 2025, letting wind and solar operators share one connection and split the capacity-dependent costs. The Dutch regulator allows up to four parties to share a connection from 2 MVA upwards, batteries and electrolysers included. Software keeps the combined export inside the limit. The screening question becomes which connections near your pipeline have an owner willing to share, and that's a commercial conversation you can start well before a connection application.
Germany's queue is scored, Britain's is gated, and Italy's is 96% unconfirmed. The projects that get built out of that will be the ones whose developers could show, early, that the site works on grid position, planning, noise and revenue at once. That evidence is now assembled during screening, weeks before an application, and by the land and GIS teams rather than by engineering.
If storage is part of your 2026 pipeline, the practical next step is to run one real market through a constraint profile you trust and see what survives.