Working note, 30 October 2026.
1. Roadmap
This note works through the socio-economic analysis that led Gentofte Municipality to cancel district heating expansion in eight remaining areas in 2026, and explains why the legal advice treated the result as binding. The analysis is the project proposal prepared by NIRAS for the municipality [2]. Its headline is that individual air-to-water heat pumps are about 277 million kr. cheaper for society than district heating over twenty years, about 19 per cent. Section 2 states the legal rule that turns this number into an obligation. Section 3 rebuilds the headline as a simple calculation per MWh of heat, which shows that the comparison is between a capital-heavy and an energy-heavy technology. Section 4 stress-tests the result in three directions and shows that the most important driver is one that receives little attention in the public debate: how fast the households in the heat pump scenario are assumed to convert.
2. The legal rule
The legal rule is that a municipality may only approve the project that is socio-economically most advantageous. The Executive Order on project approval (projektbekendtgørelsen, BEK nr. 1091 of 8 September 2025) requires in § 6 that the municipal council ensures that a project “ud fra en konkret vurdering er det samfundsøkonomisk mest fordelagtige projekt”, and § 22, stk. 2, bars approval of a project with poorer socio-economics than a relevant scenario. For an expansion of a district heating network, individual heating with heat pumps is a relevant scenario.
The legal advisers drew the consequence directly. The opinion from Poul Schmith/Kammeradvokaten concludes that continued expansion in the areas concerned cannot be carried out without acting “i strid med bestemmelserne i varmeforsyningsloven og projektbekendtgørelsen” [3]. The municipality’s own legal memo concludes that the council must approve the proposal to change the heat supply status and that “Udbygningen skal derfor sættes i bero” [1]. The word that carries the argument in every step is samfundsøkonomisk: the expansion is not judged too expensive for the utility or for households, but worse for society than the alternative.
3. The headline as a calculation per MWh
The analysis compares two scenarios for the same buildings. The project is individual air-to-water heat pumps in 2,648 buildings currently heated with gas or oil; the alternative is district heating supplied by a new 35 MW seawater heat pump. Annual heat demand is \(H=85{,}605\) MWh. The calculation period is 20 years (2030–2049) and the discount rate is \(r=3.5\) per cent. All costs are in market prices, that is, factor prices multiplied by a net tax factor of 1.28.
Table 1 reproduces the present values. Investments enter as annuities, and only the part of each annuity that falls inside the 20-year window is counted, which handles assets that outlive the window (the network has a lifetime of 45 years) without an explicit scrap value.
| Present value, million kr. | Heat pumps | District heating | Difference |
|---|---|---|---|
| Investment | 306 | 1,074 | −768 |
| Fuel and electricity | 686 | 280 | +406 |
| Operation and maintenance | 165 | 64 | +101 |
| CO2e | 31 | 45 | −14 |
| Other environmental | 2 | 5 | −3 |
| Total | 1,190 | 1,467 | −277 |
Table 1. Socio-economic present values, 2030–2049 [2].
The present values are easier to read as a cost per MWh of heat. Let \(V_k\) denote the present value of cost category \(k\), and let \(A(r,T)=\left(1-(1+r)^{-T}\right)/r\) denote the present value of one krone per year for \(T\) years; with \(r=0.035\) and \(T=20\), \(A=14.21\). If heat demand were \(H\) in every year, the constant annual cost per MWh equivalent to \(V_k\) would be
$$v_k=\frac{V_k}{A(r,T)\,H}. \tag{1}$$
This is a simplification, because in the project scenario buildings convert gradually, so heat from heat pumps is not \(H\) in every year. The conversion reads the present values on a common scale; it is not a reconstruction of the annual cash flows.
| kr. per MWh of heat, market prices | Heat pumps | District heating |
|---|---|---|
| Investment | 252 | 883 |
| Fuel and electricity | 564 | 230 |
| Operation and maintenance | 136 | 53 |
| CO2e and environment | 27 | 41 |
| Total | 978 | 1,207 |
Table 2. Present values from Table 1 converted with equation (1).
Table 2 shows the structure of the comparison. District heating is a capital-heavy technology: 883 kr. of its 1,207 kr. per MWh is investment in network, production and building units. The heat pump is energy-heavy: more than half its cost is electricity (and, before conversion, gas and oil). District heating saves about 420 kr. per MWh on energy and maintenance, because a large seawater heat pump is more efficient and pays a lower electricity tariff than many small ones, but spends about 630 kr. per MWh more on capital. The net difference is about 230 kr. per MWh, or 19 per cent.
Result. In the analysis, district heating costs about 1,207 kr. per MWh of heat and individual heat pumps about 978 kr. per MWh (market prices, 2030–2049, \(r=3.5\%\)). District heating wins on running costs by about 420 kr./MWh and loses on capital by about 630 kr./MWh. Under § 22, stk. 2, the sign of the total decides the matter: a project with the higher total cannot be approved.
4. Stress-testing the result
The result can be stress-tested by asking how much a single input would have to change to close the 277 million kr. gap. Three directions are informative.
Investment. The district heating investment would have to be \(277/1{,}074\approx26\) per cent lower for the two scenarios to tie. The analysis reports that the ranking survives a ±20 per cent change in investment, which is consistent with this break-even.
Energy prices. Energy costs are higher in the heat pump scenario, so higher energy prices narrow the gap. If all fuel and electricity costs scale by a common factor \(1+\delta\), the gap closes when \(\delta\,(686-280)=277\), that is, at \(\delta\approx0.68\). Energy prices would have to be almost 70 per cent higher across the board before district heating came out ahead. The reported ±20 per cent sensitivity is far inside this range.
Conversion speed. The most important assumption is the speed at which buildings convert to heat pumps in the project scenario. District heating is assumed to reach 90 per cent of buildings in 2030, whereas heat pumps are assumed to spread linearly from 4.8 per cent in 2030 to 95 per cent in 2049. Until a building converts, it keeps its gas or oil boiler, and that fuel is counted in the project’s fuel costs. The analysis also reports a case in which heat pumps are installed as fast as district heating would have been rolled out. In that case the project costs 1,358 million kr. instead of 1,190, and its advantage falls from 277 to 109 million kr., or 7 per cent.
| Present value, million kr. | Gradual conversion | Fast conversion | Change |
|---|---|---|---|
| Investment | 306 | 642 | +336 |
| Fuel and electricity | 686 | 458 | −228 |
| Operation and maintenance | 165 | 257 | +92 |
| CO2e and environment | 33 | 0.5 | −32 |
| Total | 1,190 | 1,358 | +168 |
Table 3. The heat pump scenario under gradual and fast conversion [2].
Table 3 decomposes the effect. Converting faster pulls heat pump investments and maintenance forward, where they are discounted less, and saves fuel and emissions from the gas and oil boilers that would otherwise keep running. The first effect dominates: deferring the heat pump investments saves more in present value than the extra years of gas and oil cost. About \(168/277\approx61\) per cent of the headline advantage therefore comes from the assumption that households convert slowly. Heat pumps still win under fast conversion, so the ranking holds, but the size of the advantage depends heavily on this assumption.
Result. The ranking survives all three stress tests: investment would need to fall by about 26 per cent, or energy prices rise by about 68 per cent, before district heating came out ahead. The size of the advantage is fragile. With fast conversion it falls from 277 to 109 million kr., so about 60 per cent of the headline comes from the assumed timing of conversion.
5. Utility and household economics
The socio-economic result is consistent with the utility’s own economics. In the district heating scenario, the present value of the utility’s revenue is 1,052 million kr. and its costs are 1,143 million kr. (capital 763, fuel 348, operation 32), a deficit of 91 million kr., about 9 per cent of revenue. Under the non-profit rule the deficit would have to be recovered through higher prices for all of the utility’s customers.
The household comparison is closer. For the 2,385 small customers, the average annual cost including VAT is 30,508 kr. with a heat pump and 33,003 kr. with district heating, so district heating is about 8 per cent more expensive. For the 131 large customers, district heating is about 9 per cent cheaper (160,483 against 176,886 kr.). The consumer comparison therefore does not point in one direction for all buildings, but the legal test is the socio-economic one.
6. Where the analysis is less clean than the headline
Two points deserve attention. First, the comparison is between two scenarios that differ in timing as well as technology, and Section 4 shows that the timing difference accounts for most of the headline gap. A reader who takes 277 million kr. as the pure technology difference overstates it; the fast-conversion figure of 109 million kr. is closer to that. Second, the per-MWh figures in Table 2 assume a constant heat volume, which is not true in the project scenario. They are useful for seeing the cost structure, not for pricing a single building.
References
- Gentofte Kommune (2026). Notat om pligt til at godkende projektforslag for ændring af varmeforsyningsstatus. 12 May 2026. gentofte.dk
- NIRAS (2026). Projektforslag: Aflysning af otte udbygningsområder for fjernvarme, version 3. Prepared for Gentofte Kommune, 4 March 2026. gentofte.dk
- Poul Schmith/Kammeradvokaten (2026). Notat til Gentofte Kommune om fjernvarmeudbygningen. 12 May 2026. gentofte.dk