Three numbers at three different levels
Italy’s national day-ahead price (PUN), the exchange price for electricity, averaged 115.9 €/MWh in 2025 — 11.6 cents per kilowatt-hour, up 7 per cent on 2024. It is the figure that appears in supplier offers and in scenario analyses.
The price actually borne is a different one. In 2025 a connection consuming between 20 and 499 MWh a year paid an average of 24.8 cents per kilowatt-hour net of VAT, which a company recovers. That is the band a typical retail site falls into: in the same year the average Italian non-household connection consumed 27,337 kWh.
| Item | c€/kWh | Share |
|---|---|---|
| Energy and supply | 14.72 | 59% |
| Network costs | 4.12 | 17% |
| Taxes and levies, excluding VAT | 5.98 | 24% |
| Total net of VAT | 24.82 | 100% |
Within the 14.72 cents of the first line, the exchange price accounts for roughly four fifths. The remainder is the supplier’s spread: margin, risk cover, imbalance and dispatching costs, and the network-loss uplift the regulator recognises on low-voltage supply. In total, in the order of 3.1 cents per kilowatt-hour.
What a renegotiation can achieve
That composition sets the boundaries of a supply tender: the genuinely contestable part is the spread, not the invoice as a whole.
| Lever | What it acts on | Effect on annual spend |
|---|---|---|
| 20% reduction in the spread | 3.1 c€/kWh | about −2.5% |
| 10% reduction in the whole energy component | 14.7 c€/kWh | about −5.9% |
| 10% reduction in kilowatt-hours consumed | every volume-based item | about −9% |
Only the last line also acts on network charges, levies and excise duties, which are applied predominantly per kilowatt-hour withdrawn. A kilowatt-hour not consumed is worth the full price; a kilowatt-hour negotiated well is worth a discount on the only share of the price the supplier controls.
In terms of effect on spend, then, a successful renegotiation is worth roughly a third of a 10 per cent reduction in consumption. Both are worth pursuing, but telling them apart determines where procurement effort and technical effort should go.
The European comparison: where the gap forms
Same consumption band, same year, same source.
| Item | Italy | EU-27 average |
|---|---|---|
| Energy and supply | 14.72 | 12.03 |
| Network costs | 4.12 | 6.44 |
| Taxes and levies, excluding VAT | 5.98 | 3.87 |
| Total net of VAT | 24.82 | 22.34 |
An Italian company pays 11 per cent more than the European average, but the gap is not distributed as one might expect: on network costs Italy sits below the average. The difference concentrates in the energy component and in taxation, the two items an individual company does not control.
One operational consequence follows: in Italy avoided consumption carries a higher unit value than the European average, because it is the only part of the price determined inside the company.
Shifting consumption towards lower-priced hours
Shifting load towards cheaper hours was for years the first recommendation of any analysis. The 2025 hourly profile scales it back: the 1am-8am band was worth 95 per cent of the daily average and the 9am-7pm band 94 per cent. Solar generation pushed daytime prices down almost to night-time levels. The only window still materially more expensive is the evening, 8pm to midnight, at 112 per cent of the average.
Two practical consequences:
- shifting load from day to night now produces a marginal benefit. On an hourly-indexed contract the gain is in the order of one per cent; on fixed time bands it depends on the tariff applied rather than on market movements, so it has to be verified in the contract;
- the evening band still offers room to act. Those are the most expensive hours of the year and, in retail and food service, also the fullest.
The resulting order of priority is: switch off what is not needed, reduce what stays on, shift what remains. The reverse order delivers materially smaller benefits for the same effort.
How to verify your own price on an invoice
- Add up the amount due net of VAT and divide it by the kilowatt-hours for the period. The result is the effective unit price, to be compared with the national average of 24.8 cents. A figure above it deserves analysis, and the market is not necessarily the cause.
- Separate the fixed charges and the capacity charge from the rest. These are monthly amounts and amounts per kilowatt of contracted power: they do not fall with lower consumption. What remains is the part efficiency acts upon, and the figure to use in any business case.
- Compare the energy component with the exchange price for the month. The difference is the spread being applied. A value materially above 3 cents justifies a request for clarification from the supplier; one materially below justifies a careful reading of the indexation clauses and penalties.
- Isolate the kilowatt-hours consumed during closing hours. This is consumption nobody decided to buy: it is the first available margin for reduction and it is measurable without changing how work is organised.
A line-by-line breakdown of a store invoice is in Anatomy of a store’s energy bill; the effect of lower energy spend on margin is covered in How much energy weighs on retail EBIT.
Sources and disclaimer
ARERA, Annual Report 2026, Volume 1, chapter 2, based on GME data: 2025 average day-ahead price (PUN), prices by time band, hourly profile and average consumption of non-household users · Eurostat, table nrg_pc_205_c, electricity price components for non-household consumers, annual data 2025 in euro per kWh, consumption band IB (20 to 499 MWh per year) · Eurostat, table nrg_pc_205, half-yearly 2025 prices, used as a cross-check. Band figures are national averages: an individual invoice may differ, and the method to verify it is in the final section. Informational document.