The problem: two limits at almost every site

A DC fast charging site runs into two limits almost as soon as it is busy.

The first is the service. Every charger draws power the utility has to deliver. When a site needs more than the existing service can carry, the next step is often a service upgrade, and in many territories that means a long wait and a large bill. The chargers you planned sit idle, or never get ordered.

The second is the demand charge. Most commercial tariffs bill for the highest power the site draws in a billing period, not only for the energy it uses. A charging site is built to create exactly that kind of peak. Two or three vehicles plugging in at once can set a demand charge that lasts all month.

A battery addresses both. It charges slowly from the service you already have, then discharges when vehicles plug in. To the utility meter, the site looks smaller and smoother than it is. That lets you add chargers before an upgrade, or instead of one, and it shaves the peaks that drive the bill.

That is the simple version. The harder question is which sites actually make money, and for how long.

Why it matters: the lasting case and the incentive case

There are two reasons to put a battery at a charging site. They behave very differently over time, and treating them as one is the most common mistake we see.

The lasting case is grid capacity and demand charges. Both exist wherever you build. Neither depends on a program staying open. If a battery lets you open a site months sooner, or skip an upgrade, or cut a demand charge every month for ten years, that value holds whatever happens in the state legislature.

The incentive case is everything paid by a program: state performance incentives, utility and ISO demand response, export tariffs, and the federal investment tax credit. These streams can be large. In some territories they are larger than the lasting case. But they open, fill and reset on their own schedules. Some programs run out of money as capacity fills, while others keep enrolling but reset their rates at the end of each program period. A program can close to new applicants, cut its rates or change its rules between the day you screen a site and the day you file.

Incentives set the pace. They decide which territories go first, and they can turn a marginal site into a strong one. But a site that pencils only while a program is open is a site with an end date. Our team screens every territory twice: once with incentives on, once with them off. A territory qualifies only if its sites hold up with incentives switched off, or if the incentive window is long enough to cover the build.

One caveat on the lasting case. Several utilities now offer EV-specific rates or demand charge holidays for charging sites. These are good for operators. They also shrink the demand charge a battery can save. A site can screen well on a standard commercial tariff and fail once its real rate is applied. Rate design has to be checked at every site, before anyone sizes a battery.

The value stack, in plain language

1. Grid capacity and faster energization. The utility service upgrade you avoid or defer, plus the charging margin you earn by opening months sooner. This is the stream most specific to EV charging. It is also the one standard storage tools do not value, because it depends on site facts like service size, upgrade cost and upgrade timeline, not on the tariff.

2. Demand charge reduction. Shaving the peaks that charging sessions create. This includes time-of-use demand charges and ratchets, net of any EV rate or demand charge holiday.

3. Energy shifting. Charging the battery when power is cheap and using it on site when it is expensive, which cuts the energy charges on the bill. At most charging sites this is a smaller stream.

4. Capacity charge reduction. In PJM, NYISO and ISO-NE, part of a large customer’s bill is set by its load during a few system peak hours each year. Cutting load in those hours lowers capacity charges for the following year.

5. Demand response and grid programs. Utility programs, state performance programs and ISO programs, often through an aggregator, pay for battery output during called events.

6. State incentives and rebates. Upfront or performance payments. Some programs pay per kWh in blocks that step down as capacity fills. Others pay per kW of performance over several years.

7. The federal investment tax credit. A credit on eligible project cost. The rate depends on project size, labor compliance, equipment sourcing and site adders. More on this below.

8. Export compensation. Some states pay for energy a battery sends to the grid. New York’s Value Stack is an example: it pays market-based energy and capacity values, and some components are locked for ten years (NYSERDA, Value Stack Reference Guide for Storage Developers). At a busy charging site the battery mostly serves the site’s own load, so exports are usually small. Each kWh the battery discharges counts either as a bill saving or as export compensation, never both.

Not every stream stacks. Some programs bar enrolling the same equipment in two programs. And demand shaving competes with program events for the same stored energy: a battery that just emptied itself shaving a charging peak at 4 PM has nothing left for a program event at 5 PM. Every territory needs a stacking rule (which programs can be combined) and a dispatch priority (what the battery does first when two uses collide). Those rules belong in the site controller’s settings, not just in the financial model.

Connecticut as a worked example

Connecticut shows how much the rules of one program shape the design of every site. The figures below come from published program documents. Program terms change, so confirm each figure against the current Program Manual before relying on it.

How the program pays

Connecticut’s Energy Storage Solutions (ESS) program is administered by the Connecticut Green Bank with Eversource and United Illuminating. In its Year Five decision, PURA moved the program to a performance-based design called Construct 5, effective April 1, 2026. Under Construct 5, non-priority commercial customers receive no upfront or enrollment incentive. They are paid a performance incentive over ten years, based on how their battery performs in utility-called events (ESS Program Manual, rev. Feb. 11, 2026, Section 2).

The published commercial performance rates are front-loaded (Connecticut Green Bank, Feb. 2, 2026):

The class is set per host customer by annual peak demand: the highest billed peak in the 12 months before application. Small is under 200 kW, medium is 200 to 500 kW, and large is over 500 kW (Program Manual, Section 6.1). A busy DC fast charging site will often fall in the large commercial class.

To put the rates in scale: a 999 kW battery that delivered its full rating in every event would earn about $275,000 a year in years one to five at the large commercial rate, about $175,000 a year in years six to ten, and about $2.25 million over ten years. That is an upper bound. The program pays on measured average AC output during events, taken from inverter data, not on nameplate (Program Manual, Sections 6.5 and 6.6). Commercial applicants also pay a $350 non-refundable application fee, and the rate is locked when the reservation is issued, not at completion (Program Manual, Sections 3.1 and 4.1).

When events happen

The Construct 5 event design, per Table 4 of the manual (Program Manual, Section 2):

Utilities usually give day-ahead notice, but the manual also allows real-time dispatch (Program Manual, Section 5). Events can fall on weekends and holidays.

Why the battery’s controls decide the economics

The performance incentive pays on the average kW a battery delivers across all of the season’s events. A missed or opted-out event counts as zero kW and pulls the average down (Program Manual, Sections 6.5 and 6.6). Construct 5 has no clawback for declining an event, but the zero still costs money.

At a charging site, that creates a daily conflict. The dispatch window runs from noon to 9 PM. That is also when many charging sites see their busiest sessions and set their demand peaks. If the battery spends its charge shaving a 2 PM peak, it may be empty for a 5 PM event. A few missed events in a 40-event summer pull down the whole season’s payment.

In our modeling of a 999 kW, 3.3 MWh Connecticut site, event performance is the single largest operating risk. At 75% average performance, the indicative return turned negative. So the site controller has to put program events ahead of peak shaving, hold enough charge on event days, and coordinate with the chargers’ own load management. That is a design requirement, and it belongs in the OEM technical RFI before equipment is chosen.

Rules that shape the design

A few other Construct 5 requirements change what gets built (Program Manual, Sections 3.1 and 4.2.1):

Export. Systems with an interconnection approval dated January 1, 2026 or later must be approved to export to the grid. A C&I customer whose base load is higher than the battery’s maximum dispatch can request an exemption, at the administrators’ discretion. Downtown sites on secondary network service, where utilities often restrict export, need this checked first.

Islanding. The battery must be able to provide backup power to the customer during an outage. Depending on the site, that can mean a transfer switch or microgrid controller and a critical-load panel, which adds hardware and design cost.

Data. The battery must connect to the utility’s DERMS platform and report data at 15-minute intervals or better.

Interconnection first. An active interconnection application is required before a reservation letter is issued. Approval is not required at that stage, but withdrawing the application puts the reservation at risk.

Size. The often-quoted limit of 150% of peak demand or 2 MW, whichever is greater, applied to the old upfront incentive. Our team found no equivalent size rule for Construct 5 in the manual (Program Manual, Sections 6.3 and 6.4). Confirm the sizing rule with the administrator before you size up.

Stacking. The same equipment cannot be enrolled in both ESS and ConnectedSolutions, and for new commercial applicants the manual offers no path to ISO-NE capacity market participation (Program Manual, Section 2). Demand charge savings at the host site are a separate stream the manual does not restrict.

None of these are obstacles on their own. Each one is a question to answer at the screening stage, before money goes into studies.

Tax credit and equipment sourcing

Two thresholds matter for batteries at charging sites. First, the One Megawatt Exception: a project with maximum net output under 1 MW AC can claim the 30% base credit without meeting prevailing wage and apprenticeship requirements. Exactly 1.0 MW does not qualify, and units with common ownership behind the same interconnection are added together (26 CFR 1.48E-3). Second, the foreign entity of concern cost tables split at 1 MWh. Systems above 1 MWh use the grid-scale table, where battery cells carry about 52% of the equipment cost, so Chinese cells cannot reach the 2026 threshold (Carina FEOC Compliance Guide). A typical 1 MW, 3 MWh site sits on both sides of those lines. FEOC rules are interim, and Treasury guidance may change them.

In our indicative Connecticut modeling, hardware price and sourcing, more than development cost, decided whether a site made money. Development was roughly 5% of a first site’s capital, so cutting it sharply moved the return by about a point. The tax credit was worth roughly $100 per kWh of hardware, which means a compliant system can cost about that much more than a non-compliant one and still come out ahead. Whether waiting for compliant supply is worth it turns on that price gap. These are indicative results, not tax advice. Your tax advisor makes the final call on eligibility, adders and sourcing.

How Carina approaches it: territory by territory

A charging portfolio is not one project repeated fifty times. Each utility territory has its own tariffs, programs, interconnection rules and permitting path. So the unit of work is the territory, not the site. Our team works in five steps.

1. Confirm where the money is. Before any site work, we research the territory’s programs: what is open, how much capacity is left, the size and per-company caps, and when funds get reserved. That means talking to program administrators, not only reading the manual.

2. Rank territories and sites. We screen each territory with incentives on and off, then rank territories and the sites within them. Weak territories drop out early, while it is cheap.

3. Build one design package per utility territory. A standard one-line diagram, layout, equipment list, utility application and permit template, built once and reused at every site in the territory. Where drawings must be stamped, they are sealed by a Professional Engineer licensed in your state under that engineer’s own license and insurance.

4. Screen sites in batches. We run desktop screens on 10 to 20 sites at a time. Each screen checks the tariff and any EV rate, service capacity, program fit, export and islanding constraints, and permitting path. Only the sites that pass go forward.

5. Develop the survivors in parallel. Sites that pass move through applications, permits and notice to proceed on the shared package. The more sites in a territory, the less each one costs to develop.

Permitting varies far more than the program rules. A statewide siting path in one state and a local special use permit in the next can mean very different schedules, and some jurisdictions have paused storage approvals altogether. Our BESS moratorium tracker follows those pauses. We do not promise permit outcomes or timelines; we find out early which sites face the hardest path.

We work alongside your team, not in place of it. Some operators are new to batteries and want a full development department. Others already have a storage lead and need execution capacity in hard jurisdictions. Either way, you keep your process, your vendors and your decisions, and you keep control of the portfolio. More on how we engage is on our How Carina Works page.

What you should do

Screen every territory with incentives on and off. If a territory only works while a program is open, size your commitment to the time the program has left.

Check the tariff and any EV rate before sizing. Pull 12 months of interval data or bills. Model the real rate, including any EV rate or demand charge holiday, before choosing a battery size.

Confirm program capacity with the administrator before spending on studies. Published capacity and real availability can differ. A short call can save a survey, a site plan and a utility application on a site the program cannot fund.

Design the controls around the program. Where a program pays on event performance, the site controller must put events ahead of peak shaving. Put that requirement in the equipment RFI, not in a commissioning punch list.

Plan sites in batches per utility territory. One design package per territory, batched screens and batched surveys bring the cost per site down as the portfolio grows.

If you are weighing batteries across a charging network, send us your charging locations. Our team will tell you which territories have money for batteries, which sites can use it, and what to do first.

Sources

• Connecticut Energy Storage Solutions Program Manual, rev. Feb. 11, 2026
  
• Connecticut Green Bank, ESS program updates, Feb. 2, 2026
  
• 26 CFR 1.48E-3, prevailing wage and apprenticeship, One Megawatt Exception
  
• Carina FEOC Compliance Guide for BESS
  
• NYSERDA, Value Stack Reference Guide for Storage Developers