Every charging project starts the same way: a hardware shortlist. Spec sheets are comparable, prices are quotable, and a charger is a satisfying thing to have opinions about. The platform gets chosen afterwards — often defaulting to whatever the hardware vendor bundles.
We'd argue that's exactly backwards. The charger you buy determines what a bay can deliver; the software you run determines what your network can ever become — how reliably it operates, how it bills, how it scales across brands, and whether you can leave. Choose the platform first, and the hardware decision gets easier, because the platform tells you what to demand from it.
Why software-first is the right order
Hardware is plural; software is singular. Over a network's life you will own multiple charger brands — through procurement rounds, acquisitions, model discontinuations and replacements. Nobody ends up single-vendor for long. That makes the one constant — the platform operating them all — the decision with compounding consequences. A platform that treats mixed fleets as normal makes every future hardware choice free; a bundled app that only manages its maker's chargers makes every future choice a migration.
The failure modes live in software. Drivers rarely meet a charger whose power electronics failed; they constantly meet chargers that are online-but-unusable, stuck, or mispriced — operational failures the platform detects, prevents or prolongs. Your reliability number (now a regulatory number) is mostly a software achievement.
Lock-in is a software phenomenon. A charger is yours; the question is who it answers to. As covered in our vendor-questions guide, the exit story — repointing hardware, exporting data — is decided by the platform relationship, not the hardware one.
The hardware questions that actually matter
Once a platform is chosen, the hardware shortlist stops being about brochure kW and becomes about operability. The questions we'd put to any charger vendor:
- Which OCPP versions does the shipping firmware support — properly? Not "planned", not "compliant on paper". Ask the platform vendor which charger models they've proven in production; battle-tested pairings are worth real money in avoided integration pain.
- How does it behave offline? Local authorisation lists, offline session recording, safe load-management fallbacks — connectivity will blip, and the charger's autonomous behaviour is a spec-sheet footnote that becomes an operational headline.
- How is firmware updated, and how often does the vendor ship it? A charger is a computer with a decade of duty ahead; its update story is its security story. A vendor with no OTA discipline is selling future site visits.
- What can be diagnosed and fixed remotely? Every fault that needs a van costs more than the part. Rich error codes, remote reset granularity, per-connector control — these determine your MTTR more than build quality does.
- What does commissioning look like? Can it be pre-provisioned so site commissioning is verification, not data entry? Serial-number scanning, config-by-profile, sensible defaults?
- What's the security posture? Certificate support and OCPP security profiles, signed firmware, lockable local interfaces.
- Parts, warranty and the ten-year question. Who repairs it in year seven? A bargain charger from a vendor who exits the market is the most expensive unit on the site.
Notice what's absent: maximum kW barely features. Power ratings are commodity facts; the list above is where total cost of ownership actually hides.
The bundle trap, briefly
Hardware-bundled software isn't automatically bad — it's automatically unexamined. It arrives as a default, skipping the due diligence a platform decision deserves, and its incentives are structural: it exists to sell that vendor's chargers, so multi-brand support, exit tooling and honest hardware comparisons cut against its purpose.
The test is simple: evaluate the bundled platform against the same twelve questions you'd ask an independent one. If it passes, use it with confidence. It usually doesn't — the API is partial, the white-label is a logo, the exit answer is uncomfortable — and it's far better to learn that before the fleet standardises on it.
The takeaway
Chargers are the visible investment; the platform is the operating system of the business. Choose the operating system deliberately, let it inform the hardware spec, and every subsequent decision — brands, models, replacement cycles — becomes reversible. Do it the other way round and the free software that came with the chargers quietly becomes the most expensive line in the project.
AmpNexus operates mixed-brand fleets across OCPP 1.6–2.1 with full API access and a clean exit story — see the platform or book a demo.