A regular and a stranger pull into the same stall
Two cars, one charger. The first driver has the network’s app open before the handbrake is up. Home account, saved card, loyalty tier, all of it riding into your charger over a roaming connection. The second has never heard of the network, doesn’t want an app, and is standing at the unit hunting for a slot to tap a card.
Both are paying customers. Only one is reachable by roaming. That gap is why ad-hoc and roaming aren’t competing access methods — not pricing, not power, not uptime. They serve different people, and turning one on doesn’t subtract from the other.
This piece is about why that’s true, and where the money actually sits in each path.
The split that makes roaming possible
The CPO/eMSP divide is the whole reason roaming exists. The CPO (Charge Point Operator) owns the hardware: the charger, its link to a CSMS over OCPP, the electricity contract, the spot on the map. The eMSP (e-Mobility Service Provider) owns the driver: the app, the account, the RFID card, the contract that says “charge anywhere and I’ll bill you.”
A driver signs with one eMSP. That eMSP roams across many CPOs over OCPI, so its drivers can light up chargers it doesn’t own. The CPO gets a session it didn’t have to acquire; the eMSP bills its own customer and settles with the CPO behind the scenes. Plug & Charge is the same model with the handshake moved into the cable — the car carries the eMSP contract instead of the phone.
Notice the assumption baked into all of it: the driver already has a relationship. Roaming is a settlement between two businesses that both know who the driver is. No contract, no roaming.
Where the money sits in roaming
In roaming, the payment is invisible at the charger. The driver authenticates (app, RFID, or Plug & Charge) and simply charges. The CDR flows over OCPI to the eMSP, the eMSP charges the card on file, and money moves between eMSP and CPO on whatever roaming terms they agreed.
It’s a clean experience precisely because the financial event already happened — at signup, before the driver arrived. The charger is only proving identity. The actual card transaction lives in the eMSP’s books, on the eMSP’s rails, under the eMSP’s terms. The CPO sees a reconciliation line, not a card-present sale.
That’s great for drivers who have an eMSP. It does nothing for the ones who don’t, and there is no roaming move that conjures a contract into existence at the moment someone needs to charge.
Where the money sits in ad-hoc
Ad-hoc is the mirror image. No prior relationship, no settlement to lean on. The driver walks up, keys in an amount, taps a card, and the money flow has to run right there, at the unit, in real time. And it has to be your money flow, settling to your acquirer, because there’s no eMSP in the middle to do it for you.
This is where the real engine lives, and where the work is. AFIR is explicit: public DC at or above 50 kW needs a contactless card reader, and ad-hoc must work with no app and no account. But the regulation mandates the outcome, not the machinery. Behind the tap, you still have to:
- Pre-authorize and pick a strategy — one fixed hold, incremental re-auth, or a driver-declared target. Each swings decline rates and cash flow differently.
- Capture only what was delivered and refund the unused hold when the session lands under the amount.
- Survive the non-happy paths — a session dropping mid-charge, a failed finalization, the charger’s CDR and the acquirer disagreeing on what to settle.
- Produce a fiscal receipt the tax authority accepts — a country-specific round trip where the fiscal signature travels from the invoicing provider out to an unattended terminal and back, and which OCPI carries nothing for.
And every cent of it settles card-present to the operator’s own acquirer: EMV contactless at the unit, rather than the card-not-present rails that tend to weigh more heavily on small AC sessions. There is even a standard for exactly this — OCPI’s DirectPayment module defines the Terminal object and Financial Advice Confirmation so ad-hoc terminal payments ride the same protocol the roaming path already speaks.
None of that lives in the roaming books. It’s a self-contained money flow that the CPO either rebuilds itself, per terminal and per country, or runs through a neutral layer once. That layer is what Bolt is: it bridges payments over OCPI so the same flow works across chargers, CSMS releases, terminals, and acquirers, and the operator stays the merchant. Bolt never holds the funds.
Why ad-hoc doesn’t cannibalize roaming
Here’s the fear worth killing outright: “if I put a card reader on the charger, won’t my roaming drivers just tap instead, and won’t I lose the roaming relationship?”
No, and for a structural reason, not a hopeful one. A roaming driver taps a phone or a car because it’s less friction than digging out a card and keying an amount. The eMSP app remembers them, bills them automatically, often hands them a better rate or points. The card reader doesn’t compete with that. It sits beneath it, for everyone the app was never going to reach.
Roaming captures drivers who pre-committed to a relationship. Ad-hoc captures the ones who structurally can’t be in one at the moment of charging.
Who are those drivers?
- Tourists and cross-border drivers whose eMSP doesn’t roam onto your network. A contract exists, just not one that reaches your charger.
- Rental and car-share drivers in a vehicle that isn’t theirs: no provisioning, no contract, often no app they’d install for one stop.
- First-timers — the EV-curious, the plug-in hybrid that charges twice a month, the driver who simply will not download an app to buy one charge.
- Fleets without your contract — a van that needs power now and will expense the receipt, not negotiate a roaming agreement.
None of them have an eMSP relationship that reaches your charger, and roaming has no mechanism to bill them. Without a card reader they don’t quietly convert into roaming customers; they drive past, or they sit at a working unit wrestling with a QR form that fails. Ad-hoc isn’t pulling demand off the roaming path. It’s collecting revenue that was already leaving the lot.
The two paths are one product
The mistake is treating this as a fork in the road: pick roaming or ad-hoc, optimize one. They’re two intake lanes into the same charger, and a serious CPO wants both wide open. Roaming maximizes utilization from the installed base of eMSP customers across the region. Ad-hoc closes the door on walk-up revenue and keeps you AFIR-aligned on every DC unit you deploy, which matters most on exactly the high-power, high-value, reliability-critical sites where a stranded stranger is the most expensive empty stall you own.
What makes the two feel like rivals is usually that they’re built by different teams on different rails: roaming over OCPI through the CSMS; ad-hoc as a bespoke terminal welded to one charger model, with its own acquirer integration and its own per-country fiscalization work. That’s two systems to keep correct while acquirer APIs, OCPI versions, and CSMS releases all drift independently.
Bridge ad-hoc over OCPI instead — the same protocol roaming already runs on — and it stops being a second system bolted to the side. It becomes the same neutral layer moving money for the drivers roaming can’t see: terminating into your acquirer, emitting a fiscal receipt in each country’s own format, on whatever terminal you’ve deployed. You keep your CSMS. You keep your bank. You answer the regular and the stranger from the same stall.
Roaming or ad-hoc is the wrong question. The drivers already answered it for you, and half of them are holding a card.