Smart Charge Leader · 27 October 2026
Connectivity is free; the money is in the clearing
Jakob Kleihues puts the industry's connectivity problem at 250,000 integrations, and argues that anyone still charging for the pipe has already lost the argument.
By RiDERgy editorial
Jakob Kleihues tells friends, family and investors that charging an electric car at a public station still often does not work, and he says about 90 percent of the people he tells agree with him. He is no critic outside the industry: he co-founded ENAPI, which brokers transactions between charge point operators, the firms that own the hardware, and mobility service providers, the firms whose app a driver taps.
Connectivity between those two sides should be given away, because the fee that pays for it is the very thing that pushes the largest customers to build their own link instead.
The arithmetic is his. Five hundred operators multiplied by five hundred providers produces 250,000 bilateral integrations, each of which needs a handshake, a set-up, testing and then permanent maintenance. Kleihues points at what the roaming hubs did before his company existed: they carried about 80 percent of the integrations but only 20 percent of the transactions, because the biggest operator and the biggest provider had already connected to each other directly. A fee completes the argument. At a million sessions a month, even five cents a session is a standing invitation to route around the broker.
He knows the cost because he lived on the other side of it. At Bonnet, an operator he worked with, broken connections surfaced as customer service cost rather than as an outage on a dashboard, because every token that failed to sync and every protocol read differently by the other side became a support ticket. That is slow, expensive, and invisible in a market that counts charge points instead.
The commercial pressure runs the other way from what a hub would want. ENAPI reports 1,050,000 charge points connected in production and more than 450 partners onboarded, at roughly two a week handled by one team member. Scale like that is what allows a company to give the pipe away: the connection is the loss leader and the reconciliation is the product.
Which brings the argument to where the money actually moved. Brokering the transaction lets the broker compare the charge detail records, the itemised billing lines an operator issues for each session, against the sessions it authorised, and find the three failures that matter: a record for a session that never happened, a session with no record at all, and one session billed twice under different identifiers. He can also apply a discount at record level rather than as a flat invoice reduction, by re-rating the tariff for a single partner. Once prices vary by the hour, that distinction stops being cosmetic.
The gap between that model and a working market is more mundane. When ENAPI started, many of its partners had not implemented the tariff module of the protocol at all, which meant two parties could connect but could not automatically tell each other what a kilowatt-hour costs. Connectivity without tariffs is a handshake about nothing. That is why the first move toward dynamic pricing is not an algorithm, in his account, but two prices a day: a night rate and a day rate, carried cleanly through the tariff module to every provider that needs them.
Public pricing stays near-static, set by charging speed rather than by the electricity cost behind it: a DC session simply costs more than an AC one, as the host framed it. He is not starry-eyed about what comes next: ask about flexibility and he reaches for the old joke about the internet being a fad, then offers the number that explains why the sector keeps returning to it. By his recollection, something like 75 percent of battery capacity until 2030 will sit in electric cars that are driving around. He sees charge point operators moving from optimising utilisation to trading flexibility, and is candid that nobody has settled how that will work.
The vehicle side of the same problem is where his caution earns its keep. Version 2.3 of the open charge point interface, the common language between chargers and backends, adds a booking module that reserves a specific connector in advance. Kleihues sees it working behind a fence, for logistics and depot charging, where the slot belongs to a known vehicle. In public charging he reports that many of the people he speaks to do not believe it will work at scale, because the reserved bay can simply be occupied by someone else. Heavy trucks are the opposite case: every significant player is connected on both sides, drawn by volumes a van fleet never generates and by vehicles almost always moving, and therefore almost always needing to charge.
Reliability is the part he is most willing to be judged on: uptime is published continuously for Europe and the United States, and although more than 94 percent of production code is written with AI, the components that must not fail are handled differently.
The strongest objection is that an intermediary is one more thing to break, and that the open protocol it mediates is free by design. A provider could connect to operators directly at no licence cost, and a broker that gives connectivity away is cross-subsidising it from clearing, which only a broker with enough volume can afford. Kleihues answers the first half with measurement: the providers he works with report materially better success rates through a harmonising layer than through bilateral connections, because someone is cleaning the partner data before it breaks. The second half he accepts as a consequence rather than a flaw, and the comparison he would draw is a card network, which does not charge shops for the terminal.
Jaap Burger made the adjacent point in S2 Ep. 9: mobility providers have to become stacked service providers rather than invoicers. Kleihues shows what the stack looks like once its bottom layer has been priced at zero.
What follows is a change in what buyers ask for. Operators and providers should treat connectivity as a commodity bought on uptime and a published status page, and write the uptime obligation into the contract as seriously as the price. Regulators should watch conformance, meaning that implementations of a common standard behave alike, rather than approving intermediaries, because the twenty-fifth variant of a standard is the actual problem. Fleets should ask which broker sits behind the app, and what happens to their session record when a link fails.
The pipe was never the business. The confidence that the invoice is right, at scale, is.
RiDERgy Commentary
We would put the point more strongly than he does, because we are the customer for what he is describing. Connectivity is not only a billing matter for the parties exchanging invoices. It is the foundation that companies like ours optimise against. Energy management at a depot runs on two things it does not own, the price of a quarter-hour and the limit of a connection, and both arrive over exactly the interfaces Kleihues wants to commoditise. We depend on that layer as much as any provider does.
Which is why energy requirements belong inside the connectivity standards rather than downstream of them. Concretely: price intervals at fifteen minutes, layered so that wholesale price, network charge and supplier margin arrive distinguishable instead of pre-blended; the same intervals bidirectional, so energy flowing back out of a vehicle is priced as carefully as the energy going in; and capability for bidirectional charging early, not once every other harmonisation is finished. A vehicle that cannot export is a battery we cannot use, and a tariff that arrives as one number per day cannot be optimised against at all. Bake those into the enabler and the optimisation happens. Leave them out and the optimisation layer is capped by the pipe beneath it, however good it is.
6 min read