Low utilization is the problem that keeps charging station operations managers up at night. The equipment is installed, the meter is running, the lease is being paid — and the cars aren’t coming.
Start with the baseline. Average utilization of DC fast chargers in the United States sits somewhere between 11% and 16%: Paren’s 2025 State of the US DC Fast-Charging Industry report puts the annual average near 16%, while Stable Auto measured 11.0% in July 2024 rising to 12.9% by June 2025. More instructive is the geographic spread — over 32% in Washington, D.C., versus 3.2% in Alaska in the same Paren dataset.
That spread says something uncomfortable: utilization is largely decided before construction. By the time you are optimizing operations, most of the outcome is already locked in by site selection and capacity planning. A serious answer to “how do I raise utilization” has to separate what can be fixed in planning from what can be recovered in operations. This article walks through both, at three levels: macro (supply-demand and seasonality), meso (station sizing, power allocation, user mix), and micro (site selection, reliability, charging speed).
1. Macro: supply and demand, seasonality and volatility
1.1 Supply and demand: utilization is a relationship between cars and chargers, not a measure of effort
Utilization is, at bottom, charging demand divided by charging supply. When a station underperforms, the first question is not “what am I doing wrong” but “how many EVs are within reach of this site, and how many plugs did I install?”
EV population grows in years; charging hardware can be deployed in months. Early movers accept a period of low utilization in exchange for scarce prime locations; latecomers enjoy healthier utilization but find the best sites already taken. Neither strategy is wrong — what is wrong is picking sites on instinct and hoping operations will fix the rest.
To judge whether a site can be fed, look at the EV population and growth rate within a 3–5 km radius, and at real charging demand in the area — including whether competitors’ chargers are queuing. The national average of around 16% tells you nothing about your site; your catchment may sit at 5% or at 30%.
1.2 Seasonality and volatility: averages lie
Winter is often peak season, not a trough. Cold weather shrinks range — Recurrent Auto’s winter study of 30,000 EVs found vehicles retain on average about 78% of their range at 32°F — so drivers charge more often and each session runs longer. Extreme summer heat also slows charging curves and stretches sessions. Holiday travel (Chinese New Year, Thanksgiving, Christmas) spikes demand within days.
This is why “average utilization” is a misleading KPI: a station at 8% off-peak and 25% on holidays averages 12% — and still loses customers to queues on the days that matter. Track two curves: average utilization and peak utilization. Pricing is the most effective lever for flattening both — raise prices at peak, discount off-peak, offer booking discounts — and pull demand off the spike.
2. Meso: station sizing, power allocation, positioning and user mix
2.1 Station size: more plugs does not mean more revenue
Every additional plug dilutes per-port utilization. A 4-port site and a 20-port site are different businesses: small sites fill up, queue, and shed customers; large sites have capacity but may never be fed.
Look at the head of the industry: EVgo reported that in Q2 2024 about 23% of its DC fast charging stalls achieved 30% or greater utilization. Note the number — 23% of stalls. Even at a leading operator, most stalls run underutilized. Utilization is a distribution, not an average. Analyze per-port, not per-site.
The optimal station size follows from your demand curve and the queue time you can tolerate — neither “bigger” nor “cheaper” is automatically right. Before expanding, look at the per-port utilization distribution and find the ports that are always idle and the ones that are always queued; then decide whether to remove, relocate, or add.
2.2 Power allocation: there are two utilizations, and most operators watch only one
Time utilization (how long a port is occupied) and energy utilization (how much electricity actually flows) are different numbers. The worst state is an occupied port delivering 30 kW — the time metric looks fine, the revenue does not.
Total site capacity (transformer capacity) is the real ceiling. With eight ports, a 480 kW site and an 800 kW site deliver very different average power per port. Dynamic load balancing can raise energy utilization meaningfully without adding a single piece of hardware — the highest-ROI operations lever available, provided the hardware and platform support it.
A quieter killer: power-module failure causing derated output. One module down, the charger still “works” — but at 70–80% power, sessions run longer, and energy utilization slips without showing up in any downtime report. Replace failed modules before customers complain.
2.3 Positioning and user mix: your users decide your chargers
Charging scenarios define demand patterns. Highway corridors serve transit users who want speed and reliability, dwell briefly, and care little about price. Urban destination sites (malls, offices, hotels) host long dwell times, low power needs, and price-sensitive users. Fleet depots run fixed routes on schedules — high utilization, thin margins.
Mismatches are common and expensive: destination sites installed with 180 kW hardware serving users who park for two hours; highway sites with 60 kW hardware that drivers route around. Pricing structure must match users too — per-kWh, per-minute, idle fees — each changes charging behavior, and behavior changes utilization.
3. Micro: site selection and traffic, station “age” and reliability, charging speed
3.1 Site selection and traffic: the homework that must happen before groundbreaking
Vehicle traffic (not foot traffic), dwell time from nearby POIs (food, retail, hotels), accessibility, driver visibility, and competitor density within 3–5 km — these are the fundamentals. Site selection error is the number-one cause of low utilization and it cannot be fixed in operations.
If expansion is on the roadmap: traffic surveys, EV population analysis, user research, and competitor mapping must come before any lease negotiation. Answer “how many EVs will be here in five years, who will charge, and how long will they stay” before discussing rent.
3.2 Station “age” and reliability: reliability is a demand-side variable
Chargers age: modules derate, cables wear, screens fail, connectivity drops. One failed charger in a 4-port site cuts capacity by a quarter. Worse, reliability shapes demand through reputation. Drivers do not remember that your station worked 90% of the time; they remember the one time it didn’t. A single failed session can cost you a high-frequency user and everyone they tell.
The operational response is concrete: remote monitoring and alerting, scheduled inspection, spare-parts inventory, and making availability a KPI alongside utilization. Winter deserves special attention precisely because it is peak season — canopies and thermal management directly protect the availability you will be paid for. Industry guidance (e.g., the Electrification Coalition) recommends deploying redundancy based on worst-case winter estimates.
3.3 Charging speed: match, don’t maximize
Vehicles accept power according to their charge curve. A 350 kW charger feeding a car that peaks at 100 kW delivers no more than the car can take; the premium buys future compatibility. But too-slow chargers lengthen sessions and worsen queues, capping time utilization.
Speed selection = user mix × dwell time × vehicle acceptance curve. Truck fleets, taxis, and private owners want entirely different hardware. Do the user research before you specify: what do your users drive, what is their maximum charge rate, how long will they stay?
Conclusion: two checklists
For existing stations (do this now): disaggregate the data — per-port utilization, energy utilization, availability, average vs peak. Fix reliability first, then optimize power allocation, then use pricing to flatten peaks.
For new stations (before breaking ground): user research and traffic surveys first, charger specification second, port count last. Site selection and specification decide about 80% of utilization; operations decide the remaining 20%. The homework is cheaper than the remediation.
Data note:
·The average utilization rate of DC fast chargers in the US ranges from 11% to 16% (Paren’s full-year 2025 report puts the figure at around 16%;
·Stable Auto’s data for July 2024 to June 2025 shows a rise from 11.0% to 12.9%); the inter-state variance ranges from 3.2% to 32.2% (Paren);
·In Q2 2024, around 23% of EVgo’s fast charging stalls had a utilization rate of 30% or higher;
·Winter range retention stands at around 78% (Recurrent Auto’s survey of 30,000 vehicles, conducted under 32°F conditions).
All other figures are industry empirical estimates.
