Since 2026, Austrian grid operators charge less for the grid between 10:00 and 16:00 from April to September, the “Sommer-Nieder-Arbeitspreis”, or SNAP. The idea is simple: in summer the grid is full of solar power at noon, so power used at that time should be cheaper, and people should move their consumption there.
My water boiler already moves, it heats in the cheapest 4 hours of the day (how that works). So I added SNAP to the price the boiler looks at and expected a nice saving. It saved 5 cents in a whole summer.
What SNAP is worth
At Wiener Netze the grid fee for a household is 6.98 ct/kWh, and 5.58 ct/kWh in the SNAP hours, both net. With the Vienna usage levy of 7 % and 20 % VAT, the difference is 1.80 ct/kWh.
In my Flux task that builds the price I really pay, SNAP is one row of the fee table and one function:
networkFee = (t, f) => {
m = date.month(t: t)
h = date.hour(t: t)
return if f.hasSnap and m >= f.snapMonthFrom and m <= f.snapMonthTo and h >= f.snapHourFrom and h < f.snapHourTo then f.snap else f.network
}
The task writes this price as a forecast, and the boiler window is calculated from it. So the boiler sees SNAP before it decides.
What it changed
I compared the boiler window with and without SNAP for every day from April to September, in 2026 with the real SNAP, and in 2025 as if it had been there already:
| April to September | 2025 | 2026 |
|---|---|---|
| Cheapest 4 h without SNAP within 10:00 to 16:00 | 167 of 183 days | 174 of 183 days |
| Days on which SNAP moved the boiler window | 15 | 6 |
| What the moved windows saved | 0.24 EUR | 0.05 EUR |
On 95 % of the days in 2026, the cheapest 4 hours were inside the SNAP hours anyway, without SNAP. On the few days where SNAP moved the window, the cheaper grid fee just tipped a close race.
In the post about heating water I wrote 7 days and 9 cents. That was a first calculation over a whole year with the energy price only. These numbers use the full price I pay, for the SNAP months only. The result is the same.
Why
Because the exchange price already does what SNAP is meant to do. The same solar power that SNAP is about pushes the day-ahead price down at noon. From April to September 2026, the evening between 16:00 and 22:00 was on average 13.7 ct/kWh more expensive than the SNAP hours, even without SNAP. SNAP adds 1.8 ct to a gap of almost 14 ct. A window that already sits at noon does not move for that.
So it does nothing?
Not quite. Everything I draw from the grid in the SNAP hours is cheaper, whether I moved it there or not. From April to September 2026 I drew 1414 kWh from the grid, 874 kWh of it in the SNAP hours, mostly while the boiler was heating. Only those 874 kWh got cheaper, by 1.80 ct each:
| April to September | From the grid, all hours | With SNAP | Without SNAP | Difference |
|---|---|---|---|---|
| 2025 | 1155 kWh | 234.12 EUR | 246.50 EUR | 12.38 EUR |
| 2026 | 1414 kWh | 306.86 EUR | 322.57 EUR | 15.71 EUR |
About 5 % less for the summer half of the year, with 2025 calculated as if SNAP had been there.
But that is a cut in the grid fee, not a reward for moving anything. The boiler was at noon because of the exchange price, and it would be there without SNAP too.
With a fixed price per kWh it is different. Then SNAP is the only reason to run something at noon, and moving the boiler from the night to the SNAP hours saves those 1.8 ct/kWh. With a dynamic tariff, the exchange price got there first.