What does each kWh a home battery delivers really cost?

Analysis · Published by SEO Dons Ltd · Updated 7 October 2026

The question behind every price table

Every solar battery cost table divides a price by a capacity: £890 per kWh, £630 per kWh. It is the natural number, and it is the wrong one for deciding whether to buy, because capacity is only the size of the tank. What the house gets is the electricity that comes out of it. A 13.5 kWh battery in a home that can only use 3,000 kWh a year has far more tank than it will ever empty.

So this post asks a different question: what does each kWh that a home battery delivers cost, once you spread its price over a decade of use?

How the calculation works

1. Run the site’s savings model for a London home using 3,000 kWh a year, with 4 kWp of solar on Octopus Flux, or with no solar on Octopus Go or Agile. The model gives the saving in pounds a year and the kWh the battery puts back into the house, or exports in the peak, in that year.
2. Take the price: Octopus’s published battery-only prices (April 2026) of £3,947 for about 5 kWh, £5,336 for about 10 kWh and £7,999 for a Tesla Powerwall 3, and the official benchmarks from multiplying DESNZ’s 2025/26 medians by the size (£6,500, £8,900 and £8,505).
3. Divide the price by the delivered kWh over ten years, scaled by the model’s 90% average usable capacity over a decade of fade. That is the battery cost per delivered kWh.
4. Compare it with the model’s saving per delivered kWh: the annual saving divided by the annual kWh delivered. That figure already allows for the cost of the electricity used to charge the battery and for the 10% lost on the round trip.

Rates are Octopus’s 26 September 2026 rates; the model has not yet been re-run on the 7 October ones, which carry no VAT. It assumes one full battery cycle a day at most.

With solar on Octopus Flux

For a London home with 4 kWp of solar on Octopus Flux, the model delivers the same kWh a year from batteries of up to 13.5 kWh whatever the usage between 2,000 and 6,000 kWh, because on Flux it can export what the home cannot use in the 4pm to 7pm peak. That makes the throughput simply a function of battery size:

Three readings. The smallest battery costs the most per kWh delivered, because the installation cost is spread over the least electricity. The 10 kWh battery is the cheapest per kWh delivered at Octopus’s prices, 18.0p, and the Powerwall 3 costs a little more at 20.0p despite delivering 35% more, because it costs 50% more. And no row repays within ten years: the earning per delivered kWh (10.6p to 12.8p) is below the cost (18.0p at best), which is the same result as the paybacks in the last column.

Without solar, size is capped by the house

With no solar the battery can only move cheap night-rate electricity into the day, and a home can only use so much of that. On Octopus Go for the same 3,000 kWh home:

The model delivers 3,000 kWh a year from a 13.5 kWh battery in this home, which is simply all the electricity the house uses: it has hit the ceiling. The Powerwall 3 therefore costs 29.6p per delivered kWh, against 20.2p for the 10 kWh pack, for an extra £14 a year of saving. On Octopus Agile the pattern repeats, with the 10 kWh battery at 20.2p, the Powerwall 3 at 29.6p and the savings per delivered kWh at about 11.4p. A home with a larger appetite changes the answer: at 4,500 kWh a year on Go the model delivers 4,206 kWh from a 13.5 kWh battery and the extra capacity earns £197 a year more than 10 kWh does. See 13.5kWh batteries for who that suits.

A 12-month back-test on real prices shows the same ceiling: see the no-solar battery page for what a 5 kWh and a 10 kWh battery earned on Agile.

The price that makes ten years work

Turn the question round. What would a battery have to cost to repay itself in ten years, at the model’s 90% average capacity? That is simply the annual saving times ten times 0.9:

Every published price is above its ten-year break-even: by 43% for the 10 kWh battery with solar on Flux, by 21% for the 10 kWh battery without solar on Go, and by between 56% and 152% for the other four cases. It does not follow that a battery is a bad buy: manufacturers’ warranties run 10 to 15 years, the model holds today’s tariffs for the whole decade, and it puts no value on backup in a power cut, on a battery that also carries the solar inverter, or on the chance that tariffs get more volatile. It does mean that the honest range for a simple payback on the bill saving alone, at Octopus’s published prices, is 11 to 23 years depending on size and tariff, and 13 to 16 years for a Powerwall 3 across the 14 regions in the Powerwall payback by region post.

What to do with this

• Compare quotes by delivered kWh, not capacity. Ask which tariff the saving assumes, what the installer thinks the battery will deliver in a year, and divide the price by that over ten years. A quote that cannot say how many kWh the battery will deliver is quoting capacity only.
• Size the battery to what the home can use. Without solar, the throughput stops rising once the battery can cover the home’s usable evening and overnight use. Past that point extra capacity is cost without earning.
• Fit the battery with the panels, or choose the tariff first. The cost side falls when a battery is fitted with new solar (E.ON Next lists batteries from £1,900 with its panels against £3,995 alone, and the supplier comparison sets out the other suppliers’ prices), and the earning side rises on a tariff such as Octopus Flux.
• Check the price you are quoted against the cost guide. The battery cost guide sets out the published prices and the official medians by size.