Building your own charging platform looks like independence. Count the five-year cost, the eight subsystems and the energy upside, and the honest answer is usually buy, or hybrid.
If you run a growing charging network, sooner or later someone asks the question out loud: should we just build our own software? You already own the hardware, the sites and the customer relationships.
The charge point management system (CPMS), the platform that controls the chargers, prices the sessions, bills the drivers and keeps the network compliant, can feel like the last piece you do not fully control. Owning it looks like independence.
The honest answer has three parts, not two. For a small number of operators, building makes sense. For most, buying an open platform is the faster path to profitable scale. For a growing number, the real answer is a hybrid: buy the hard parts, build only what differentiates you.
What decides it is rarely the licence fee. A CPMS is a stack of at least eight subsystems rather than a single application, and the five-year cost of owning them runs in three layers, of which the visible one is the smallest. One industry analysis puts a basic build for 10,000 charge points at around €2.85M upfront, with in-house maintenance running up to 50% of the build cost every year.
Two things rarely make it into the business case at all. A platform with energy capabilities earns money back, turning the network into a grid asset rather than a cost line. And uptime, support and compliance are permanent staffed obligations rather than features you ship once. This guide works through each in turn.
A CPMS is a stack, not a single app
Before you compare building with buying, be precise about the job. A CPMS is not a single application. It is a stack of subsystems that each has to work, scale and stay compliant.
When teams underestimate building, it is because they are picturing one of these and forgetting the other seven:
- The charger connection: persistent Open Charge Point Protocol (OCPP) connections (versions 1.6J and 2.0.1) to every charger, session authorisation, remote diagnostics, error handling and over-the-air firmware updates. This is the part most people mean by "CPMS", and it is roughly an eighth of the work.
- Pricing and tariffs: dynamic and scheduled tariffs, spot-price-linked pricing, per-segment rates, and the logic to show a compliant price before a session starts.
- Payments: integration with a payment service provider (PSP), authorisation and capture, settlement and payouts, refunds, chargebacks and fraud prevention.
- Roaming: Open Charge Point Interface (OCPI) connections to roaming hubs, bilateral agreements with e-mobility service providers (EMSPs), charge-data-record (CDR) reconciliation and failed-session handling.
- Energy and load management: dynamic load balancing across a site, smart charging, and grid-flexibility participation, the difference between adding chargers and paying for a grid upgrade.
- Back office: invoicing and VAT handling across every country you operate in, reporting, quoting, support ticketing and the customer and asset data model everything else depends on.
- Deployment tooling: bulk onboarding, configuration and testing of chargers, trivial at 20 chargers, a product in its own right at 2,000.
- Hosting, security and the integration layer: cloud infrastructure that scales with your network, security hardening and certification, plus the application programming interfaces (APIs) and webhooks that connect all of the above to your other systems.
Each of these is a subsystem with its own roadmap, failure modes and compliance surface. "Can we build this?" is the wrong question, because a good team can. The right question is whether maintaining eight subsystems is the best use of the team you have.
The real cost is the five-year total
The build-versus-buy decision is almost never won or lost on the upfront price. It is won on total cost of ownership, the five-year cost to run, support and evolve the stack above. TCO has three layers, and the visible one is the smallest.
Layer 1, direct software cost. Licence or build cost, development and integration, hosting, and the transaction fees on payments and roaming. This is the number most procurement exercises stop at. For scale, one industry analysis estimates a basic build for 10,000 charge points at around €2.85M upfront, before anything below.
Layer 2, operational cost. The work that never ends: people operations (commissioning, onboarding, roaming agreements, support), financial operations (reconciliation, fraud, working capital), supporting assets (RFID cards, per-charger connectivity) and electricity.
The recurring bill here dwarfs the build. Industry estimates put in-house maintenance at up to 50% of the original build cost per year, and an agency-built platform at roughly 20% per year. On a €2.85M build, that is a seven-figure annual line either way. These are vendor-published industry estimates, so treat them as order-of-magnitude and model your own.
Layer 3, missed revenue. The cost that never shows up on an invoice: slow time-to-market while you build, churn from a poor driver experience, revenue left on the table because your pricing is inflexible, and the risk of a compliance gap. For a growing network, this layer is often the largest of the three, and it is the one a "we saved on licence fees" business case ignores.
Add the three layers and the conventional wisdom, build for scale and buy for speed, mostly inverts. Building concentrates cost and risk in your own team. Buying an open platform spreads it across the vendor's whole base, and the unit economics improve as the vendor scales.
For a wider view of what to check in a platform, see our guide on how to choose an EV charging platform in Europe.
Energy can turn the cost around
Here is the part most build-versus-buy comparisons miss entirely. A charging network is also a grid asset. The same chargers that earn from drivers can earn from the energy system, and that revenue offsets the platform cost rather than adding to it.
With energy and load management built into the platform, a network can do three things that pay. It can shift and balance load to avoid expensive grid upgrades. It can sell flexibility into balancing and reserve markets: Virta operators in the Nordics already earn from the Frequency Containment Reserve for Disturbances (FCR-D), where regulator auction prices set real, recurring revenue. And it can turn on-site renewable surplus into charging revenue and CO2 cashback.
For an operator weighing TCO, this is the decisive point. A platform that only costs money is a liability, while a platform that turns the network into an earning grid asset can significantly reduce, and in favourable cases offset, its own total cost of ownership. Build your own, and you forgo that upside unless you also build the energy-market integrations, which are a stack of their own.
Reliability is a revenue number
Every failed session is a lost transaction and a driver who may not come back. Before you compare platforms, work out what one hour of network downtime costs you in lost sessions, lost roaming income and inbound support calls, then ask how many of those hours a small in-house team can realistically absorb in a year.
The bar is explicit in some markets. The United States federally funded National Electric Vehicle Infrastructure (NEVI) programme requires each charging port to maintain an average annual uptime above 97%, with uptime defined as hardware and software both online and able to deliver power (NEVI minimum standards). Europe has no single mandate yet, but your roaming partners and fleet customers judge you on it anyway.
The gap between the standard and the experience is real. The J.D. Power 2025 U.S. EVX Public Charging Study found 14% of EV owners had visited a charger and left without charging, and that 60% of those failures were an out-of-service charger.
This is where a bought platform is hardest to match, because the number is a product of scale. The Virta platform runs at 99.99% uptime, and automatic fault recovery resolves up to 20–30% of charger errors on its own, rebooting the unit remotely when no session is running and returning it to service without anyone being dispatched.
Uptime at that level is not a feature, it is a subsystem with a rota attached. Monitoring, alerting, diagnostics, firmware rollback and people who answer at three in the morning all have to exist before the number does, and building them is a large part of the true cost of building.
Support is a staffed 24/7 operation
Support is the line that build business cases forget, and the one your customers will ask about first. Chargers fail in front of drivers who are standing in the rain at an unfamiliar site, and those drivers expect a phone number rather than a web form.
Done properly, that means staffed phone lines around the clock in every language your network serves, a second line of technical support for your own operations team, and someone accountable when a roaming partner disputes a batch of sessions. Each of those is a hiring plan, a training programme and a shift pattern that has to hold over Christmas.
For comparison, Virta runs 24/7 phone support for EV drivers in five languages (English, Finnish, Swedish, German and French), technical support for business customers, and a named Customer Success Manager for each CPO and EMSP.
Build, and every one of those shifts is yours to fund whether the phone rings or not. Buy, and the same capability is spread across the vendor's entire customer base. Of everything in this comparison, support is the clearest case where a shared cost beats an owned one.
Standards and compliance keep moving
Whichever path you take, the compliance bar is the same, and it moves. If you operate public charging in the European Union, the Alternative Fuels Infrastructure Regulation (AFIR), Regulation (EU) 2023/1804, sets the baseline.
From 13 April 2024, every new public charge point of 50 kW or more must accept ad-hoc contactless card payment, with the price shown per kWh before the session starts. From 1 January 2027, that ad-hoc obligation extends to all public charge points of 50 kW or more, not only new ones.
Payment is only the visible half. AFIR also obliges operators to make static and dynamic data about their public charge points available through their member state's national access point, which means a data pipeline pointing at a different endpoint in every country you operate in, kept accurate as stations come and go.
Underneath the regulation sits the protocol layer, and that moves too. OCPP versions advance, OCPI versions advance alongside them, roaming hubs revise their certification requirements, and each change has to be implemented, tested against real hardware and rolled out across a live network without dropping sessions. Managed roaming exists as a service precisely because that work never finishes.
If you buy, meeting all of this is the vendor's problem, spread across their whole base. If you build, it is yours alone, every time the rules change, which is the Layer 3 compliance risk made concrete.
Open standards keep it reversible
The single best protection against lock-in, by a hardware vendor if you build or a software vendor if you buy, is open standards. OCPP, maintained by the Open Charge Alliance, is the common language between chargers and management software.
When your hardware and platform both speak OCPP, you can change one without replacing the other, which is what makes a network migration to a different platform possible at all. The NEVI standards make exactly this a requirement: chargers must be able to switch network providers without hardware changes.
Open standards are what turn buying into a reversible decision, and what make the hybrid path possible in the first place.
Hybrid: buy the core, build the edge
The sharpest operators have stopped treating this as binary. The hybrid approach buys the proven platform for the parts that are hard, regulated and undifferentiated, and builds in-house only the thin layer that is genuinely yours. The split is the whole point.
- Buy the core: OCPP connection management, payments, roaming, uptime and monitoring, energy management and regulatory compliance. None of this wins you a customer; all of it will sink you if it breaks.
- Build the edge: your pricing and packaging logic, your driver or fleet-manager app, a data product, a loyalty model, the part that differentiates your network.
What makes the hybrid real rather than a slogan is the platform's API surface. A platform with genuine APIs, webhooks and developer tooling lets you build on top without forking the core, so you get enterprise-grade reliability from day one and still ship the experience that sets you apart.
A platform without that surface forces you back into the binary. So when you evaluate buying, the test is not only what it can do today, but what you can build on it tomorrow.
Which path fits which operator
Build if software genuinely is your product, you can permanently fund a team to own all eight subsystems as a roadmap rather than a project, and you operate at a scale where that fixed cost spreads across enough chargers to be rational.
Buy an open platform if your product is a reliable, compliant, profitable network rather than a software platform, and you would rather put capital into sites, hardware and customers than into a permanent development team.
Go hybrid if you are mostly in the buy camp but have one or two places you truly differentiate: buy the core, build the edge on the platform's API.
For most growing charge point operators (CPOs) in the 50 to 500 charger range, the honest answer is buy or hybrid. The decision, in one line: build if you want to own the software, buy if you want to own the network, go hybrid if you want to own just the part that makes you different.
The bottom line
Building your own CPMS buys control, but it also buys a permanent obligation: eight subsystems, three layers of cost, a 24/7 support operation and a compliance bar that keeps moving, all carried alone, while forgoing the energy revenue a good platform earns back.
For most growing operators, an open platform delivers the same independence without that recurring cost, and turns the network into a grid asset that helps pay for itself. For the ones with something genuinely different to offer, a hybrid buys the core and builds just the edge.
Either way, put your capital where the return is: more sites, more chargers, more loyal drivers.
To see what running your network on an open, API-first platform looks like in practice, explore Virta Hub, our CPMS.
