Heat pump payback at the Q4 2026 Ofgem cap
At Ofgem's rates for 1 October to 31 December 2026, a gas boiler and a heat pump deliver a kWh of heat for almost exactly the same money. That single fact decides the payback question.
The math, plainly
Four steps, in this order. Take the gas standing charge off the bill first, because it buys no gas: at 29.68p a day that is £108.33 a year. Divide what is left by 7.97p to get the gas kWh bought. Multiply those kWh by the boiler's seasonal efficiency to get the heat the house actually received, because a boiler is not a perfect converter. Divide that heat by the heat pump's delivered SPF and price the result at 26.32p.
The third step is the one that decides the answer, and it is the one most comparisons skip. Dropping it is the same as assuming a 100% efficient boiler, which quietly hands the gas side an advantage it does not have.
What that gives, per kWh of heat
| Gas boiler at 92% (a good modern condensing unit) | 8.66p |
| Gas boiler at 85% (this file's mid-band assumption) | 9.38p |
| Gas boiler at 76% (bottom of the published assumption band) | 10.49p |
| Heat pump at SPF 2.55 (bottom quartile of the national trial) | 10.32p |
| Heat pump at SPF 2.78 (trial median, 428 homes) | 9.47p |
| Heat pump at SPF 3.05 (top quartile of the national trial) | 8.63p |
Read down that table and the conclusion is unavoidable: at this cap the two technologies cross over somewhere around the median. A median heat pump replacing a median boiler changes the fuel bill by almost nothing. What does change is the standing charge, because a house that has given up gas stops paying the £108.33 a year that comes with the meter, and that is where most of the modelled saving on this site comes from.
So what is the payback
Against the three published post-grant prices (Octopus £4,064, E.ON Next from £4,946, Aira £6,729), running-cost savings alone repay the heat pump in about sixteen years at the most favourable combination in the published evidence, forty to sixty-seven years at the trial median, and never where the delivered SPF lands in the bottom quartile against an efficient boiler. Those are not typographical errors and they are not an argument against heat pumps. They are an argument against buying one on a fuel-bill promise.
The case that does survive the arithmetic is the £7,500 Boiler Upgrade Scheme grant, which arrives at installation rather than over two decades, the carbon delta, and a heat-pump tariff. Octopus publishes the structure of Cosy Octopus rather than a rate: nine hours a day at 51% below the regional day rate, three hours at 50% above. Whether that beats the cap depends entirely on how much of your heat you can move into those nine hours, which is a function of your cylinder, your fabric and your habits, so this file does not put a number on it.
Why there is no deep-retrofit payback figure here
Fabric, glazing and solar all cut demand, and this file does not model that. A payback number computed by dividing the whole £35,000 to £70,000 deep-retrofit capex by a saving that only values the gas-to-electricity swap would be arithmetic without meaning, so it is not published. The calculator reports capex, the grant stack and the net cost for the deep case, and quotes payback on the heat-pump leg alone.
Where payback collapses further
Delivered SPF under 2.6, which the national trial shows is a quarter of real installations, not an edge case. Fabric not addressed first, so the flow temperature has to be raised and the SPF falls with it. Oversized units cycling. The lever is design quality: flow temperature of 40 to 45°C, radiators sized with margin, a weather compensation curve that is actually commissioned, and a delivered-SPF commitment with a twelve-month review.