Executive Summary

The economics of public EV charging have inverted. After a decade of footprint-first expansion, the market now has more than a thousand players across the charging value chain, according to PwC Strategy&’s 2025 EV charging market outlook. Most public networks are still lossmaking, prioritizing network growth over near-term profit, and even at maturity Arthur D. Little projects operating margins of just 1 to 8 percent, with real profitability unlikely before the end of the decade. When pricing converges across a commoditized market, the physical network stops being a moat. The defining question of the EV charging business model is no longer how many chargers a company can plant, but how much each one can earn.

This paper argues that survival in a saturated market depends on a strategic shift: operating a charging network not as an energy retailer but as a flexibility asset that earns from how it is used and from the grid services it can provide. Three software levers drive that shift. Utilization turns reliability into demand, because where price is matched, drivers reward the network that simply works. Flexibility opens a second revenue line, as smart charging, demand response, and vehicle-to-grid (V2G) move from pilot to bookable income — Octopus Energy’s Power Pack tariff already pays UK drivers around £620 per vehicle per year against a standard variable bill. And an interoperability substrate of open protocols is what makes both levers possible.

Regulation is compressing the timeline. The UK already enforces a 99 percent reliability floor on its rapid network, and the European Union’s incoming Network Code on Demand Response will turn flexibility participation from optional into expected. For charge point operators (CPOs), and for the utilities now entering their market, the decisive choice is whether to build, buy, or partner for the software that earns more from each charger they already own.

The Math Changed: Why Footprint Stopped Being a Moat

For most of the last decade, the public charging strategy was simple: secure sites, plant hardware, and assume that footprint would compound into pricing power. That assumption has expired. With more than a thousand players competing across the value chain (PwC Strategy&), public charging has commoditized faster than it has matured. The Rabobank 2025 infrastructure review counts 1.22 million public charge points across 32 European markets, up 19 percent year on year — supply is no longer the constraint.

Two structural facts now define the market. First, profitability is thin even in the optimistic case: Arthur D. Little’s analysis of US public charging projects mid-term operating margins of just 1 to 8 percent, and most networks are not there yet. Second, the bottleneck has moved from the parking lot to the grid connection. gridX’s 2025 European market analysis finds that operators now cite local grid capacity, not site acquisition, as their primary constraint, with new connections taking 18 to 24 months in many markets.

When every operator can match a competitor’s price per kilowatt-hour, and when adding hardware no longer guarantees a return, the moat has to come from somewhere other than the asphalt. Are EV charging stations profitable on the strength of footprint alone? The market’s own consolidation — a steady cadence of exits and mergers among smaller operators — answers that. The networks that endure will be the ones that earn more from each charger, not the ones that simply own the most.

What changed The footprint playbook The software playbook
Source of advantage Site count: own the most locations and hardware Software capability: earn the most from each charger
Primary revenue Energy margin: sell kWh at a price the market can match Stacked revenue: charging session plus flexibility and grid services
How profit is reached Scale of footprint: assumes more chargers compound into pricing power Utilization and flexibility: reliability drives demand, software opens new income
Scaling economics Capital-heavy: each site adds fixed cost regardless of use Asset-light at the margin: new revenue lines run on assets already deployed
Regulatory posture Compliance as cost: reliability and standards treated as overhead Compliance as moat: the 99% floor and flexibility rules become advantages

The Marginal kWh and the Marginal Software Capability

Here is the shift in one line: in a saturated market, the marginal kilowatt-hour no longer differentiates a charging network, but the marginal software capability does. Electrons are fungible and priced by a market that can match any number. Software capability is not. The ability to keep a site running at 99 percent uptime, to shift load away from a demand peak, to bid a fleet of connected vehicles into a flexibility market — those are differentiated, defensible, and increasingly the difference between a network that earns and one that bleeds.

This reframes ev charging revenue from a single line — energy sold at a margin the whole market can undercut — into a stack. The same physical asset can serve a charging session, defer a grid upgrade, and sell flexibility back to the system. None of that is visible on the hardware spec sheet. All of it lives in the software layer. The rest of this paper examines the three levers that turn that capability into money: utilization, flexibility, and the interoperability substrate beneath both.

Lever 1 — Utilization: Turning Reliability Into Revenue

In a market where price is matched, the network that simply works wins the next session. Reliability is not a compliance checkbox; it is the demand engine of the EV charging business model. A charger that fails a driver once loses not one transaction but a relationship, and in a commoditized market that driver has a dozen alternatives within range.

That makes reliability a margin lever rather than a cost center. Utilization, not price, is what amortizes a charger’s largely fixed capital cost: a site running at high uptime spreads that cost across more sessions, while a site that fails intermittently strands it.

Regulation has already made this floor explicit. The UK’s Public Charge Point Regulations 2023 enforce a 99 percent average annual reliability standard across the rapid (50kW and above) network, backed by financial penalties and a 24/7 helpline requirement. Reliability is now a legal minimum in a major European market, not an aspiration, and the operators that turn it into a utilization advantage do so on the strength of their management software. That is what a CSMS and CPMS development practice delivers, with vendor-agnostic monitoring turning a heterogeneous hardware estate into a single, observable network. How reliability is engineered, measured, and defended across a multi-vendor fleet is the subject of a companion analysis.

Lever 2 — Flexibility: When the Network Earns From the Grid

The second lever changes the business model outright. A charging network sits on top of a large, controllable, increasingly bidirectional electrical load. Managed correctly, that load is a grid asset, and grid assets get paid. Smart charging shifts demand to cheaper, cleaner hours; demand response sells the ability to curtail or shift load when the system is stressed; V2G turns parked vehicles into distributed storage that can discharge back to the grid at a premium.

This is no longer speculative. Bidirectional charging is crossing from pilot to commercial standard, and the same logic already proven in the renewable-energy market, where demand response aggregators build real businesses on flexibility revenue, now applies directly to charging networks. What each tier of the stack actually pays in 2026 is quantified in a companion analysis.

The proof that it works at the asset level is already running in production. Codibly built an EV fleet aggregator for a global technology solutions provider that performs automated demand response with load shedding across a multi-vendor estate, bridging MODBUS, the ChargePoint API, and OCPP in a single multi-tenant platform. The Flux project validated bidirectional vehicle-to-everything (V2X) energy exchange with major automotive OEMs. These are the mechanics of a second revenue line, running today.

Revenue stream What it is Enabling protocol Who pays
Charging session Energy sold to the driver: the baseline transaction OCPP, OCPI: charger control and roaming The driver
Smart / managed charging Load shifted to cheaper, cleaner hours: lower energy cost, better margins OCPP, ISO 15118: session-level control The operator (cost saved), the driver (lower price)
Demand response Curtailing or shifting load on a grid signal: paid availability and dispatch OpenADR, IEEE 2030.5: grid signaling The grid operator or aggregator
Vehicle-to-grid (V2G) Discharging stored energy back to the grid: bidirectional revenue ISO 15118-20, IEEE 2030.5: bidirectional sessions The grid, energy markets, the EV owner (shared)

Lever 3 — The Interoperability Substrate

Neither utilization nor flexibility is possible without an interoperability layer beneath them. A charging network speaks to its chargers through OCPP, exchanges roaming sessions with other networks through OCPI, signals demand response through OpenADR, communicates with the grid through IEEE 2030.5, and authenticates bidirectional sessions through ISO 15118. Each protocol is a precondition for a specific revenue stream. Together they are the substrate that lets a charger be more than a vending machine for electrons.

This is where the EV charging business model quietly becomes a software architecture decision. A network whose stack was built protocol-first can add a flexibility revenue line by configuration. A network whose stack treats interoperability as a later integration project pays for it twice — once in delayed revenue and again in the engineering cost of retrofitting protocols into firmware that was never designed to carry them. Interoperability is not plumbing to be deferred; it is the asset that determines which revenue streams a network can ever access. The protocols a network speaks today define the markets it can sell into tomorrow — which is why OCPI and OICP roaming and dynamic load management sit at the center of any charging stack built for the next decade.

The Convergence: When Utilities and CPOs Become Each Other

The saturated market is also a converging one. Utilities are entering charging because a charging network is both a retail margin and a grid asset they have a structural interest in controlling. The pattern is most visible in Europe, where utility-origin operators — EnBW, Fortum, ESB, EQUANS, and others — now sit among the most active CPOs. Across European CPO and utility engagement in the field, the same signal holds: many of the most engaged operators are utility-backed.

The convergence runs both ways. As CPOs reach toward flexibility revenue, they need grid-grade software — the kind of standards-compliant, dispatchable control that utilities have always required. As utilities reach toward retail charging, they need CPO-grade driver experience and roaming. Each is acquiring the other’s software competencies, and in a thin-margin market neither can afford to build that competency slowly. This is the same interoperability-as-grid-revenue logic that turns regulatory mandates into grid revenue assets, now playing out across the utility-CPO boundary. The strategic implication for both sides is the make-or-buy decision examined next.

Build, Buy, or Partner: The Software Economics of the EV Charging Business Model

Once an operator accepts that software capability is the moat, the question becomes how to acquire it. There are three paths, and in a thin-margin market the choice is consequential.

Building in-house gives full control but costs 12 to 18 months and several senior protocol engineers — talent that is scarce and expensive — before the first session is billed. Buying a SaaS platform is fast but introduces per-charger and per-transaction pricing that compounds against a 1 to 8 percent margin, and it locks differentiation inside a vendor’s roadmap. The third path is partnering on a pre-built accelerator with custom services on top: certification-ready protocol codebases the operator owns outright, deployed in weeks, with no per-charger scaling penalty and full freedom to modify.

Dimension Build in-house Buy SaaS Partner (accelerator + custom)
Time to first revenue 12–18 months: protocol engineering from scratch Fast: live in weeks Weeks: pre-built, certification-ready codebase
Cost model High fixed cost: several scarce senior engineers Per-charger / per-transaction fees: scales against thin margins One-time license + implementation: predictable, flat at scale
Source-code ownership Full: but at full cost and risk None: differentiation locked in a vendor roadmap Full: client owns and may modify the code
Protocol breadth Limited by in-house expertise: hard to staff Fixed to the vendor’s stack: add-ons cost extra Broad: OCPP, OCPI, OpenADR, IEEE 2030.5, ISO 15118
Scaling penalty None on licensing: but ongoing maintenance burden High: cost rises with every charger added None: no per-charger fee on owned code
Best fit Operators with deep protocol teams and time Small pilots, short horizons Operators needing speed, ownership, and flexibility revenue

The ownership point is decisive at thin margins. A SaaS fee that scales with charger count is a tax on growth precisely when growth is how a network reaches the utilization that makes it profitable. A perpetual license that the operator owns and hosts carries no such tax. The path is fast as well as ownable: for the retail energy provider APG&E, Codibly built a custom DERMS platform in 12 weeks, standards-ready for OpenADR and IEEE 2030.5 and positioned for multi-market flexibility participation. For operators that want to run the software in-house eventually, a build-operate-transfer model hands over both the platform and the team’s capability. The same revenue-stacking discipline visible in merchant battery storage economics applies here: the asset earns more when the software lets it sell into more than one market.

From Energy Retailer to Flexibility Asset

The saturated charging market will not be won by the operator with the most chargers. It will be won by the operator that earns the most from each one — through reliability that drives utilization, through flexibility that opens a second revenue line, and through an interoperability substrate that makes both durable. The marginal kilowatt-hour is a commodity the whole market can match. The marginal software capability is the moat that remains.

For CPOs and the utilities entering their market, the strategic work of 2026 is to stop treating software as a cost of running chargers and start treating it as the mechanism that makes chargers pay. Regulation has already set the clock: a 99 percent reliability floor is enforced in the UK today, and the EU’s Network Code on Demand Response is coming. The operators that re-architect now, around the levers that turn a network into a flexibility asset, are the ones that will still be operating when the consolidation settles.