The tap is the easy part
A driver taps a card, the hold goes through, electrons flow. That moment feels like the transaction. It isn’t — it’s a promise. The actual money, the part that pays for the hardware, the electricity, the staff, the loan on a DC unit that cost an order of magnitude more than its AC sibling, arrives days later. It comes in a lump, from an acquirer, against a batch of sessions the operator now has to recognize one by one.
That gap between the tap and the payout is where charging businesses lose money they can’t see. Not because the rate is wrong, but because they can’t read the money clearly: when it lands, which sessions it covers, and whether the total matches what the chargers reported. Settlement is the unglamorous half of payments — and it’s the half that decides whether a CPO makes payroll.
”You keep your own bank” is a cash-flow sentence
Bolt is non-MoR by design: the operator stays the merchant of record, and the money never routes through us. That usually gets framed as a liability or compliance point, and it is one. But the part operators feel every month is the cash-flow part.
When you keep your own acquirer, settlement runs on a direct line — card network, to your acquirer, to your account, on your acquirer’s clearing schedule. No intermediary balance. No platform wallet. No vendor float parked between the driver’s card and your bank, holding your money a few extra days as a structural feature of that model. The funds are yours from the moment they clear, on a timetable you read in your own acquirer’s terms rather than ours.
Contrast that with how many “all-in-one” payment layers work by design: they’re the merchant, they collect, they hold, and they pay you out on their cadence. That model is genuinely simpler to switch on — one contract, one counterparty — and for some operators that trade is worth it. But the float it creates is invisible until you need the cash and it isn’t there yet. Keeping your own bank deletes that layer.
Keeping your own bank carries a second, distinct effect: it keeps you on card-present economics. A physical tap at the charger is a card-present transaction, which avoids the card-not-present treatment an app-or-web flow invites. On slow, frequent, low-value AC sessions, that difference is the margin. It’s a separate lever from the float — but it pulls the same direction.
So neutrality isn’t only about who’s liable when something goes wrong. It’s about who’s holding your money while nothing goes wrong.
A payout is a lump. A business runs on lines.
The operational problem surfaces on reconciliation day. Your acquirer doesn’t pay you per session — it pays you per batch: one deposit covering a window of captures, net of the deductions your acquirer takes out, landing whenever its clearing cycle says. You get a number. Your accounting needs a hundred lines.
Tying that lump back to individual sessions is the actual work of settlement, and it’s harder than it sounds, because three systems describe the same charge differently:
- The charger and CSMS know the energy session over OCPP — kWh delivered, start and stop, the CDR.
- The payment side knows authorizations and captures — the initial hold, the final settled amount.
- The acquirer knows a batch deposit — money, in your account, by date.
Each is a different clock and a different identifier. The session that started Tuesday night may capture Wednesday and settle Friday, in a batch with two hundred others. Reconciliation means stitching the CDR, the capture, and the deposit line into one story per charge — and flagging the cases where they disagree.
Where the numbers refuse to match
They disagree more than you’d like. The common mismatches:
- A driver pre-authorizes a generous hold, charges less, and the unused hold is released rather than captured. The authorization and the settled amount differ on purpose — but your books need to know that’s a release, not a loss.
- A session drops mid-charge — connectivity, vehicle, a yanked cable. You captured for energy actually delivered, which is neither the hold nor the driver’s intended target.
- The CDR and the capture diverge because the charger metered one figure and the payment finalized against another. Someone has to decide which is canonical and reconcile to it.
- A finalization fails at the acquirer and retries, so the capture lands a cycle late and shows up in the wrong batch.
OCPI hands you clean primitives for this — a Terminal object, a pre-authorization tied to a tariff, start/stop, and a Financial Advice Confirmation carrying the final captured amount. That FAC is the hinge: the authoritative final figure you reconcile the acquirer’s deposit against. But OCPI standardizes the connection, not the bookkeeping. It hands you the confirmation; it doesn’t reconcile Friday’s deposit back to Tuesday’s session, doesn’t model partial captures and released holds, and doesn’t tell you which clock is right when two systems disagree.
That reconciliation engine is the thing every CPO ends up building. Bolt builds it once — across any charger over OCPP, any terminal, any acquirer.
The result is one ledger that already knows a deposit line maps to these sessions, that this difference is a released hold and that one is a dropped session — before anyone opens a spreadsheet.
Timing is a strategy, not a fact
Settlement timing isn’t fixed; it bends to choices made at pre-authorization. A single fixed hold, incremental re-authorization, or a driver-declared target each change when and how much gets captured, which ripples straight into cash flow and decline rates. A hold that’s too high spooks the bank into a decline; too low and you can’t capture what the driver actually used. The pre-authorization strategy you pick upstream is a cash-flow lever downstream — and OCPI deliberately says nothing about which to use.
Then there’s fiscalization, with its own timing and its own round trip. The driver gets a legal receipt by link at the end of the session — not a PDF, but a document the tax authority accepts. That receipt depends on a jurisdiction-specific signature traveling from the invoicing provider to the unattended terminal and back. OCPI carries nothing for it; Bolt bridges it both ways, per market. The point for settlement: the fiscal record and the financial record have to agree, per session, or your reconciliation is fiction. A captured amount with no matching fiscal receipt is a problem you want to catch the same day — not at audit.
The quiet version of “it just works”
For the driver, settlement is invisible: park, key in an amount, tap a card — Apple Pay and Google Pay ride the same contactless tap — charge, get a receipt by link, unplug. They never think about it.
For the operator, “it just works” means something sharper. The money lands in their bank, on a schedule they own, with no one sitting on the float. Every deposit line is already reconciled to the sessions behind it. Releases, partial captures, and dropped sessions are labeled rather than guessed. And the fiscal receipt is matched to the capture, per charge, in every market they run.
Because none of it is welded to one charger model or one acquirer, the operator can swap terminal, acquirer, or market and keep the same ledger — riding their CSMS over OCPI the whole way. The tap is the promise. Settlement is whether the promise is kept: on time, in full, in your account. That’s the part worth getting right, because it’s the part that pays the bills.