Build-versus-buy models for charging software usually compare a development quote against a monthly fee and pick whichever number is smaller over three years. That comparison omits most of what building actually costs.
The honest version compares two ongoing obligations rather than a project against a subscription.
What buying actually costs
- The recurring fee, on whatever basis it scales.
- Integration work into your billing, support and reporting systems.
- Constraints on your commercial model where the platform cannot express it.
- Exposure to price changes at renewal.
- The cost of leaving, if you ever do.
What building actually costs
The recurring items are what decide this, and they are the ones most often left out of the model entirely.
- Initial development, which is the number everyone has.
- A permanent engineering function, because the protocol evolves and chargers misbehave.
- Security response capability, including out of hours.
- Driver-facing support for an application you now own.
- Infrastructure, monitoring and the operational discipline to run a service.
- The opportunity cost of engineering attention that is not on your core business.
Where the crossover actually sits
It depends on the pricing model you are comparing against. Per-charger pricing scales with fleet size, so a large network reaches crossover on volume alone. Revenue share scales with success, so a high-utilisation network reaches it that way.
Licence or self-hosted pricing may never cross over, because the recurring exposure is bounded rather than proportional.
Non-financial reasons that legitimately override the math
A commercial model no platform supports. A requirement to hold data in a specific way for regulatory reasons. A product where charging is embedded in something else and cannot be a separate experience.
Where one of these applies, the arithmetic is secondary. Where none applies, the arithmetic should decide, and it usually favours buying for longer than operators expect.
The middle options the binary hides
White label gives brand ownership without engineering obligation. A routing layer preserves the freedom to change platform without building one. Deep configuration of a bought platform covers more requirements than teams assume before trying.
Most operators who conclude they must build have not evaluated these, and at least one of them usually resolves the pressure at a fraction of the cost.
The question that settles it
Ask what happens on the day the protocol version you rely on is superseded, a security issue is disclosed in a library you depend on, and a driver cannot start a session at midnight. If you build, all three are yours, permanently.
Organisations that can answer that comfortably should build. Organisations that hesitate have their answer.