Plug in a Level 2 charger, price it at 20 cents over your utility rate, and wait for the passive income to roll in — that's the pitch a lot of EV charging operators hear before they open their books and discover their single biggest expense isn't the electricity they resell. It's the electricity they don't resell: the peak demand charge a handful of simultaneous fast-charging sessions can trigger, sometimes running higher than the entire month's energy bill.
EV charging is one of the few small-business categories where the cost of goods sold moves by the hour, the revenue model splits into three incompatible pricing structures, and a federal tax credit that shaped a lot of buy decisions quietly expired mid-2026. None of that shows up in a generic small-business chart of accounts. Here's how to build bookkeeping that actually tracks whether a charge point operator is profitable.
Why "sell electricity, keep the markup" isn't the real business model
A busy Level 2 port typically brings in $1,200 to $1,800 a month; a well-sited DC fast charger running at 50–60% utilization can produce $3,000 to $5,000 a month before demand charges and other operating costs are subtracted. Those numbers look attractive until you see what sits below the revenue line. Reported all-in operating margins for charging businesses land in the 15–35% range, with operating expenses — electricity, maintenance, repairs, and networking fees — commonly eating 20–30% of revenue on their own, before payroll, site lease, and software fees are counted.
The gap between "revenue per kWh" and "profit per kWh" is where most new operators get their books wrong. Three things drive that gap, and each one needs its own line in your ledger.
1. Demand charges are a separate cost center, not part of COGS
Commercial electricity bills usually have two components: a per-kWh energy charge, and a demand charge based on the single highest 15-to-30-minute spike of power draw during the billing period. For most commercial customers, demand charges make up 30–70% of the total bill. EV charging is uniquely bad for this metric — a few vehicles fast-charging simultaneously can spike a site's peak draw for only a few minutes and still set (and inflate) the demand charge for the entire month.
That means two charging stations with identical monthly kWh throughput can post wildly different utility bills depending on whether their sessions cluster together or spread out. If your books lump "electricity" into one COGS account, you'll never see this. Split it:
- Energy COGS — per-kWh cost, tied directly to metered throughput
- Demand charge expense — the peak-draw penalty, tracked separately so you can see whether load-shifting or session-limiting software is actually working
Submetering each port (rather than relying on the building's whole-site meter) is what makes this split possible in the first place, and it's also the data trail utilities and incentive programs want to see.
2. Revenue model determines what "a session" even means in your books
Charging revenue isn't one number — it's some combination of four streams, and most real operators mix at least two of them:
- Energy sales — billed per kWh, the closest thing to a "unit price" model
- Time-based pricing — billed per minute of session or per minute of port occupancy after charging completes (an idle fee)
- Session/connection fees — a flat charge per plug-in event
- Subscription revenue — a recurring monthly fee, often bundled with a lower or waived per-kWh rate
Pay-per-use pricing for public Level 2 charging typically runs $0.10–$0.30 per kWh, averaging around $0.20. A subscription model instead books recurring revenue independent of usage that month, which means your revenue-recognition treatment has to differ by stream: energy and session fees are recognized as sessions occur, but subscription fees should be recognized ratably over the covered period, the same way you'd treat any other prepaid service contract. If you're booking a full month of subscription cash as revenue the day it's collected, your monthly P&L will overstate margin in months with light usage and understate it in months with heavy usage — the opposite of what the subscription was supposed to smooth out.
3. The federal tax credit that shaped a lot of purchase decisions just ended
If you bought or are planning to buy charging hardware, know that the Section 30C Alternative Fuel Vehicle Refueling Property Credit — 6% of depreciable project cost for most commercial installs, or 30% if the project met prevailing-wage and apprenticeship requirements, capped at $100,000 per charging port — stopped applying to property placed in service after June 30, 2026 under the One Big Beautiful Bill Act's accelerated termination date. Property placed in service before that date can still claim it on the return for that tax year via Form 8911; anything installed after cannot. If you have a charger sitting in inventory or mid-install right now, the placed-in-service date is the number that determines whether this credit applies at all — worth confirming with your CPA before you assume it's baked into your ROI math.
Beyond the now-closed federal credit, most of the remaining tax treatment is straightforward: charging hardware and the associated electrical work (wiring, wall mounts, panel upgrades) are depreciable capital assets, not expensable equipment, and should sit on your balance sheet and run through your normal depreciation schedule — not get buried in a single "equipment" expense the month you paid for it.
A chart of accounts that actually separates the moving parts
For a small charge point operator (a handful of sites, mixed Level 2 and DC fast), a workable expense breakdown looks like:
- Energy COGS (per-kWh utility charge, metered per port where possible)
- Demand charge expense (peak-draw penalty, tracked by site)
- Networking/software fees (the per-port monthly fee charging-network operators charge for payment processing, uptime monitoring, and app connectivity — easy to forget because it's a recurring subscription charge on your own books, not a variable cost)
- Maintenance & repairs (connector replacement, calibration, vandalism repair)
- Site lease/host revenue share (if you're paying a property owner a cut of session revenue or a flat site fee)
- Depreciation (hardware + installation, on its own schedule per site or per port)
Because energy costs move with utilization and demand charges move with simultaneity of use rather than total volume, per-port and per-site profitability can only be judged once these are separated. A site that looks profitable in aggregate might have one over-clustered fast-charging bank quietly wiping out the margin from three well-utilized Level 2 ports elsewhere on the property.
Keeping records this granular — one ledger with clear, git-versioned history of every session-level revenue posting and every utility bill split into energy vs. demand — is exactly the kind of structured, auditable bookkeeping that plain-text accounting is built for. Beancount.io gives charge point operators plain-text, version-controlled accounting: every account split above lives as a readable, diffable ledger entry instead of a black-box QuickBooks category, so when a demand-charge spike or a subscription-revenue quarter looks off, you can trace exactly which session or which bill caused it. Get started for free and see why finance-minded operators are moving off opaque spreadsheets and onto a ledger they can actually read.