Commercial battery storage payback in 2026: a worked example

Why payback only makes sense as a worked example

Ask what the payback is on a commercial battery and the honest answer is: it depends entirely on your tariff, your load shape, and how many value streams you can stack. A figure quoted without your half-hourly data is a guess. So rather than offer a headline number, this guide works a single realistic case from the ground up, shows every assumption, and lands on a simple payback and an internal rate of return (IRR) you can sense-check.

The business below is unnamed and illustrative. Your own numbers will differ. The point is the method, not the answer.

The site: a mid-size manufacturer

Take a single-shift engineering firm on a half-hourly metered supply. Its key figures, taken from twelve months of meter data:

• Annual consumption: roughly 850,000 kWh.
• Peak demand: around 320 kW, concentrated 08:00-18:00 on weekdays.
• Agreed import capacity: 350 kVA, with available-capacity (kVA) charges levied monthly whether used or not.
• A 120 kWp rooftop solar array already installed, currently self-consuming about 55% of what it generates and exporting the rest at a low rate.
• Day-rate unit cost around 28p/kWh, with red-band distribution (DUoS) charges and peak unit rates pushing the effective cost of late-afternoon energy towards 38p/kWh on winter weekdays.

This is a textbook candidate. There is a clear daily peak, an existing solar asset that is being under-used, and capacity charges that bite.

Sizing the system

The aim is to cover the worst of the weekday afternoon peak and soak up surplus solar at midday. From the half-hourly profile, a 250 kW / 500 kWh system, a “2-hour” battery, fits. It can shave the peak by up to 250 kW for two hours and store enough midday solar to redeploy into the late-afternoon shoulder.

The battery uses lithium iron phosphate (LFP) cells for thermal stability and cycle life, runs at 88-92% round-trip efficiency, and carries a warranty in the order of 6,000-10,000 cycles or 10-15 years to roughly 70% retained capacity.

What it costs

At commercial scale in 2026, cabinet-style systems of this size sit around £200-£450 per kWh installed, depending on enclosure, switchgear, and grid-connection works. For a 500 kWh system, that is a project in the region of £175,000 once design, G99 application, protection, and commissioning are included. Prices have fallen hard from the ~£800/kWh of 2020, which is much of why these projects now stack up.

The stacked savings

No single saving justifies the spend. The case is built by stacking several, each modelled from the meter data.

1. Peak-shaving

Discharging across the weekday afternoon peak cuts the most expensive units and trims red-band DUoS charges, which apply only in those peak distribution windows. Shaving around 250 kW for two hours a day across roughly 240 working days, at a blended saving of about 12p/kWh versus charging off-peak, contributes on the order of £28,000-£32,000 a year. See peak-shaving for how the red-band mechanics work.

2. Reduced capacity (kVA) charges

By holding peak grid draw below the previous high-water mark, the site can review its agreed capacity. Trimming a slice of unused kVA saves a few thousand pounds a year in standing charges. Conservatively, £3,000 a year.

3. Energy arbitrage

Charging on cheap overnight power and discharging into the daytime peak captures the spread between off-peak and peak unit rates. On the nights solar cannot fill the battery, arbitrage against a day-night spread of 15-20p/kWh adds roughly £6,000-£8,000 a year. More on this in energy arbitrage.

4. Solar self-consumption uplift

Storing midday surplus rather than exporting it lifts solar self-consumption from about 55% towards 85%+, displacing imported units at the full day rate instead of exporting at a poor rate. On this array that is worth around £6,000 a year.

Taken together, the modelled gross annual benefit is approximately £45,000. We deliberately exclude grid-flexibility revenue from the core case.

Grid revenue: upside, not a foundation

Larger behind-the-meter systems can earn from the Capacity Market, frequency response (Dynamic Containment, Moderation and Regulation), the Balancing Mechanism via a P415-registered aggregator, and the Demand Flexibility Service. On bigger assets this can add £20,000-£100,000+ a year. It is revenue, not a grant, it is site-specific, and it is never guaranteed, so we keep it out of the payback maths and treat anything it delivers as a bonus.

Tax and VAT

Two reliefs change the real cost materially, and both are widely misstated, so be precise.

Annual Investment Allowance (AIA). The £1m AIA lets a business deduct 100% of qualifying plant and machinery against taxable profits in the year of spend. At £175,000, the whole project sits inside the AIA. Batteries are special-rate plant and machinery, so they do not qualify for 100% Full Expensing; where spend exceeds the £1m AIA, a 50% First-Year Allowance applies to the excess. For this project the full cost is relieved via the AIA. At 25% corporation tax, that is around £44,000 off the tax bill.

VAT. Commercial battery storage is standard-rated at 20% VAT. A VAT-registered business recovers that in full, so it is a cash-flow timing item, not a cost. The 0% VAT relief applies only to domestic and charitable buildings (until 31 March 2027, then 5%); it does not apply to a commercial install.

So the £175,000 headline becomes a net-of-tax investment of roughly £131,000 once the AIA relief is taken, with VAT recovered separately.

The numbers that matter

• Net investment after AIA relief: ~£131,000
• Modelled annual benefit (excluding grid revenue): ~£45,000
• Simple payback: 131,000 / 45,000 ≈ 2.9 years on the net-of-tax figure, or about 3.9 years against the gross £175,000

A sub-4-year payback is at the strong end. It is driven by an existing solar array, capacity charges worth attacking, and a clean daily peak. Many sites land in the broader 4-10 year range typical of well-stacked commercial batteries; some do not stack up at all, and we will say so.

IRR

Over a conservative 12-year life, with degradation trimming benefit modestly each year and no grid revenue assumed, £45,000-ish of annual benefit against £131,000 net delivers an IRR comfortably in the high-twenties percent. Add even modest, realistic flexibility revenue and that rises. We model IRR on cautious assumptions so the downside, not the marketing case, is what you sign off.

Honest limits

This worked example flatters the technology because the site suits it. Change the inputs and the answer moves: a flat overnight load with no peak, no solar, and spare capacity may never justify a battery. Payback also depends on tariffs holding their shape, on a workable G99 connection (timescales run 8 weeks to 12 months, so apply early), and on disciplined operation. Read what a commercial battery costs for the full price picture across system sizes.

Get a free desk feasibility from your half-hourly data

The only way to turn this method into your number is to run it against your actual meter data. We model peak-shaving, arbitrage, capacity savings, and solar uplift from your half-hourly readings, then return a fixed-price quote within 7 working days. If storage does not pay back on your site, we will tell you. Request a quote or call +44 7707 970661 to start a no-obligation feasibility.