You sell kilowatt-hours, but your grid operator bills kilowatts. That one letter costs a charging site real money every year, and it sits on an invoice line most operators never open. The month you were busy once is the month your grid cost goes up, even if you sold no more energy than usual.
What does your grid invoice actually measure?
It measures power, not volume. In Flanders the digital meter reads your draw every fifteen minutes and keeps the highest quarter-hour value of each month. The capacity tariff then bills the average of those twelve monthly peaks, with a floor of 2.5 kW. For 2026 the Flemish regulator VREG puts the average at around 53.39 euro per kilowatt per year, excluding VAT. The Dutch form differs but the principle matches. On a large-consumer connection you pay monthly for contracted transport capacity and, on top of that, for kWmax: the highest quarter-hour load that actually occurred in that month.
A capacity tariff is the part of your grid bill charged on the highest power your connection drew during a period, not on the number of kilowatt-hours that passed through it. Two sites that sell exactly the same energy will get different grid bills as soon as their sessions land differently across the day.
Here is how one Dutch grid operator splits it, in the band a mid-size charging site often falls into:
| Component | What it measures | 2026 rate, excl. VAT |
|---|---|---|
| Fixed transport charge | per month | 36.75 euro |
| Contracted transport capacity | per kW per month | 2.0228 euro |
| kWmax, highest quarter-hour peak | per kW per month | 3.0966 euro |
| Energy transported | per kWh | 0.0198 euro |
Put those two variable lines side by side and the picture flips. One extra kilowatt of peak costs roughly what transporting 155 kilowatt-hours costs in that same month. A charger sitting at full power five minutes longer than it needed to therefore hurts more than one additional session.
How does a capacity tariff affect your charging price?
A capacity tariff raises your cost per kilowatt-hour at exactly the moments when several things happen at once on your connection. Because you are billed on the highest quarter-hour peak rather than on total draw, a single busy quarter is spread across every kilowatt-hour you sold that month. A site that halves its peak without selling one kilowatt-hour less lowers its cost per session by doing so. That is why the peak belongs inside your tariff model: when your price moves with demand, sessions spread across more quarter-hours and you buy less grid for the same sales. Price your energy cost alone and you leave that lever untouched.
Why a spot-following tariff can raise your grid cost
A tariff that only tracks the energy price sends everyone to the same hour. That is by definition the cheap hour, and it is usually the hour the rest of your connection picks too. Your energy cost falls, your peak rises, and the capacity part of your bill is not inside that spot price. Spot plus a fixed margin is therefore a poor base tariff and a worse roaming tariff. It does work as a discount layer for members and subscribers, sitting on top of a base tariff that weighs demand and your own peak. How to bound such a base tariff is covered in our post on setting a price floor and ceiling on charge points.
What does one careless quarter-hour cost?
Enough to pay back an investment in control. Two worked examples with round numbers, based on the published rates above.
- DC example, Netherlands. A site with four 150 kW points. Without control they run together at full power for one quarter-hour: 600 kW. With control the peak stays at 300 kW. That saves roughly 930 euro of kWmax in that month. If contracted capacity can then drop by 300 kW, add roughly 7,300 euro per year.
- AC example, Flanders. Two 11 kW points on one digital meter. If they overlap for a quarter-hour every month, your monthly peak reads 22 kW. Split across two moments it reads 11 kW. At 53.39 euro per kilowatt per year that is around 590 euro per year, on identical sales.
These are illustrations, not measurements from a specific site. The order of magnitude is the point: grid costs run into hundreds of euro per year for every kilowatt avoided, while margin per kilowatt-hour is counted in cents.
Why load balancing on its own is not enough
Load balancing throttles after the cars have arrived. It protects your connection and you need it as a safety net, but the driver pays for it in charging time. Someone pulling in at 18:00 to a site splitting 100 kW across four cars gets a slow session and remembers it. Price acts earlier in the chain. It shapes the decision to come now or later, before the connector goes into the car. The two work together: price flattens demand, limiting catches whatever still coincides.
There is a trap in it. Announce one cheap window and you build a new peak inside it. A model that varies price per charger and per quarter-hour spreads load across a wider band instead of pushing everything into the same hour. That is also why a demand-driven pricing model does something different from a simple day and night rate.
What is changing on the contract side?
Grid operators are building contract forms that reward this. In July 2024 the Dutch regulator ACM approved two alternative transport rights: a time-block variant granting full transport rights inside agreed windows, and a time-duration variant covering a set number of hours per year. Both trade flexibility for a lower transport tariff. Netbeheer Nederland states they become available once grid operators have adapted their business processes, and that no concrete date is set for that yet.
For an operator that is worth planning around. Price your site today so demand moves toward the cheap windows, and you can switch such a contract on later without rebuilding your operation. Skip it, and you sign for a peak capacity you pay for twelve months a year and use for five quarter-hours.
What would you do differently on Monday?
Start by pulling your twelve monthly peaks from your grid operator and finding the quarter-hour each one landed in. In most cases those are a handful of moments, not a structural pattern. Then work out what one kilowatt of peak costs you per year and set it against your margin per kilowatt-hour. If that ratio shifts your view, it is time to stop building your tariff on the energy price alone. Our free potential analysis works through what demand-driven pricing could do on your points, and the per-charge-point subscription lets you start without switching CPO platform.
Which quarter-hour from last year is actually sitting on your invoice?
Frequently asked questions
Does a capacity tariff also apply to a site with a large connection?
The Flemish capacity tariff applies to connections with a digital meter on the distribution grid. Larger charging sites on a transformer connection fall under a different tariff structure, but the underlying idea is the same: there is a price on your highest quarter-hour power. In the Netherlands that line is called kWmax, and a monthly charge on contracted transport capacity sits on top of it.
Can I lower my monthly peak without losing revenue?
Usually yes, because the peak lands in a handful of quarter-hours per year rather than across your whole utilisation. If your price is higher at those moments, part of the sessions shift to neighbouring hours where capacity is still free. You sell the same kilowatt-hours across a wider window. The share that genuinely cannot wait pays the higher price and helps cover the grid cost of that quarter.
What is the difference between a capacity tariff and an energy tariff?
The energy tariff is what you pay your supplier per kilowatt-hour, and it moves with the market. The capacity tariff is what you pay your grid operator per kilowatt, and it moves with your own behaviour. That distinction matters: you have little influence over the energy price but real influence over your peak. A pricing model that only watches the energy price is steering the line you cannot control.
Is load balancing not cheaper than dynamic pricing?
Load balancing is necessary and you need it on your connection as a safety net regardless. It only solves the wrong problem when it is your single instrument, because it throttles the session after the driver has arrived. That costs charging speed and therefore satisfaction. Price acts earlier and shapes whether someone comes now or later. In practice you switch on both.
How do I find out which quarter-hour caused my monthly peak?
Your grid operator supplies the quarter-hour values for your connection, and many meters also expose them through a local port or a portal. Line those up against your charging sessions from the same day and you can usually see within one quarter which combination of chargers coincided. That is the starting point: one or two recurring moments often explain most of your annual capacity cost.
Does this still work if my charge points are in roaming?
Yes, but then it matters whether your tariff change actually reached every eMSP. A price you send but that is never applied steers no driver at all. So check per partner that the tariff settled in the CDR matches what you configured. Without that feedback loop you are managing a peak with a price nobody sees.
Sources
- Capaciteitstarief VREG, Vlaamse Regulator van de Elektriciteits- en Gasmarkt
- Aansluiting en transport voor grootverbruikers, elektriciteit tarieven 2026 Stedin
- Zo werkt het: alternatieve transportrechten Netbeheer Nederland