The gap between citing AFIR and quoting it
AFIR gets invoked constantly in EV-charging payment conversations — as a reason to buy a thing, a reason to panic about a date, a reason a vendor’s roadmap suddenly grew a deadline. Far fewer people have read the regulation and can tell you what it obligates, on which chargers, by when.
So here is the regulation in plain terms, with the marketing scraped off. AFIR — the EU’s Alternative Fuels Infrastructure Regulation — is the rule that turned “drivers should be able to just pay” from a nice idea into a legal obligation at public charge points. It is a regulation, not a directive, and that distinction matters: it applies directly across member states, with no national transposition step in between. The obligations land on the operator of the public charge point — not the carmaker, not the driver.
The core idea is small and stubborn. At a public charger, a driver must be able to pay ad-hoc — no app, no account, no prior registration, no membership in anyone’s network. Show up cold, pay, charge, leave. Everything else in this article is detail hanging off that one sentence.
Ad-hoc is the whole point
“Ad-hoc” is the load-bearing term, so it’s worth being precise. It does not mean “an app that doesn’t require signup.” It does not mean “a QR code that opens a web form where you type your card number.” It means a one-off payment a first-time visitor can complete on the spot, without joining anything.
The regulation is explicitly trying to kill the failure mode every EV driver has lived through: you pull up to a charger run by a network you’ve never heard of, and the only way to start it is to download that network’s app, create an account, add a card, and wait for a verification email — in the rain, at 20% battery. AFIR’s answer is that roaming-by-app cannot be the only door. There has to be a way in that assumes nothing about you.
That is the spirit. The letter then splits by power.
The split that actually matters: power thresholds
AFIR does not treat all chargers the same, and this is where most summaries get sloppy. The dividing line is power — not the type of current at the connector.
For high-power public chargers — 50 kW and above — the obligation is specific and hardware-shaped: the charge point must accept payment through a card reader at the unit. The operator can satisfy this with a payment card reader, or — the option everyone actually ships — a device with contactless functionality that reads payment cards. What is not allowed at this power level is satisfying the ad-hoc duty with only a QR code or a payment link. A stranger has to be able to pay card-present, at the wall, with the card in their hand.
In practice that means contactless. A contactless reader accepts tap-to-pay bank cards, which is also why a wallet tap — Apple Pay, Google Pay — works at the same reader with nobody building anything extra. EMV does the heavy lifting; the regulation just insists the reader is physically there.
For lower-power public chargers — below 50 kW (typically AC, though the threshold is power, not current type) — AFIR gives the operator an extra option, not a softer obligation. The ad-hoc duty is identical: a stranger must still be able to pay without an account. But below the threshold, a QR code or payment link that takes the driver to a secure payment page is added to the list of acceptable mechanisms, alongside readers — so an operator can satisfy the rule without bolting a certified terminal onto every post. The logic is economic. Lower-power points tend to be cheap, numerous, and slow, and mandating a certified payment terminal on each of millions of them would be disproportionate. A fast charger is a different animal: high value per session, far more expensive hardware, and a driver standing there for minutes, not hours.
So the mental model is simple:
- 50 kW and above: a card reader is required.
- Below 50 kW: a QR code or payment link may stand in for the reader.
- Either way: a stranger must be able to pay without an account.
That’s the substance.
The deadlines, including the one nobody likes
Two clocks, and people routinely conflate them.
The first clock covers new high-power stations. AFIR’s date of application is 13 April 2024; high-power units deployed from that date onward must already meet the card-reader requirement. If you’re commissioning new fast chargers now, there’s no grace period to argue about — the obligation is live on the equipment going into the ground today.
The second clock is the retrofit clock, and it’s the uncomfortable one. High-power points already operating before 13 April 2024 must be brought into compliance by 1 January 2027. Two things to be precise about. First, this is a retrofit obligation, not a “replace on natural end-of-life” obligation — the deadline applies whether or not the unit was otherwise due for replacement. Second, it is scoped, not universal: the retrofit clock attaches to pre-existing 50 kW+ points on the core trans-European (TEN-T) network or in secure parking areas, not literally every legacy fast charger on every back-lot in Europe.
Even so, that date is the source of most of the AFIR urgency in the market right now — and it’s the part that gets quietly dropped from pitches that only want to talk about new builds.
What the regulation does not do (and why that’s the trap)
Here’s the part worth slowing down on, because it’s where operators get hurt.
AFIR tells you the outcome: a stranger can pay, at the wall, at your fast charger, by these dates. It says almost nothing about how to run that payment correctly once the reader is mounted. It does not specify your pre-authorization strategy — whether you place one fixed hold, re-authorize incrementally as energy flows, or let the driver declare a target amount. Each choice swings decline rates, driver experience, and your cash flow, and the regulation is silent on all of it.
It says nothing about the non-happy paths: the session that drops mid-charge, the hold that needs partial capture and a refund of the unused remainder, the finalization that fails, the day your charger’s record and your acquirer’s record disagree and someone has to reconcile them. It says nothing about fiscalization — producing a receipt the tax authority will actually accept, which is a jurisdiction-specific round trip between an invoicing provider and the unattended terminal, not a PDF you email. It says nothing about routing across multiple acquirers, jurisdictions, and currencies, and nothing about staying correct as acquirer APIs, charger-to-backend protocol versions, and charger-management software all move underneath you.
This is the trap. “AFIR compliance” is sold as a box: bolt a certified reader on the charger and you’re done. But the reader is the easy part — a solved problem you can buy off a shelf.
What turns a compliant reader into a working payment is the engine behind it: holds, captures, refunds, reconciliation, and a fiscally valid receipt — correctly, every time, in every country you operate, and still correct after the next protocol release.
It’s worth being fair about where that line falls. The companies that build chargers, terminals, and back-office platforms are excellent at exactly that. A charger maker should pour its effort into charging; a terminal supplier into a reader that taps cleanly in the cold. The payment engine is simply a different discipline — and welding it permanently into hardware that ships on a five-to-ten-year cycle is what leaves operators exposed when the money rules move faster than the metal.
Reading it without fear
None of this is cause for panic, and the regulation is not trying to trap you. AFIR is, on balance, good for the industry: it forces the one experience drivers have always wanted — pay the way you’d pay for anything else — and it does so in a technology-neutral way. It mandates a card reader on fast chargers; it does not mandate whose reader, whose bank, or whose charger-management platform. You keep your acquirer, with card-present economics rather than card-not-present treatment — a difference that matters most on small sessions. You keep your charging network. You keep the freedom to swap a terminal or change countries later.
What AFIR gives you is a clear floor and a hard date. What it leaves entirely to you is the money flow standing on that floor — the part that is genuinely hard to build, easy to get subtly wrong, and expensive to maintain across a continent of differing tax rules and moving APIs.
Bolt’s view is simply that this engine should be built once for the whole market, sitting neutrally behind any reader, any acquirer, any charger-management system. Read the regulation. Then make sure what’s behind the reader is as compliant as the reader itself.