When fleet managers first look at their electricity bills after transitioning to EVs, the energy cost per kWh is usually the first number they focus on. It is the most legible line item. But it is rarely the largest problem. The demand charge, which most commercial electricity tariffs attach to the highest 15-minute or 30-minute power draw in the billing period, frequently accounts for 30 to 50 percent of the total bill. And it spikes hard when a dozen vans all plug in at the same time after returning from afternoon routes.
We built Fleetvolts specifically to address this. What follows is a practical breakdown of where the actual savings come from in a commercial EV fleet, and what does not move the needle as much as it looks like it should.
The bill structure most operators see too late
A typical commercial electricity tariff has two distinct cost components: the energy charge and the demand charge. Energy charges are simple. You use 500 kWh, you pay the per-kWh rate multiplied by 500. Time-of-use tariffs add rate variation by time of day, but the math is still straightforward.
Demand charges work differently. Your utility records the highest sustained power draw during the billing period, measured across 15-minute or 30-minute intervals. That peak value, expressed in kilowatts, gets multiplied by a demand charge rate that typically runs between $10 and $22 per kW per month in commercial tariff structures common to the US Midwest and Mid-Atlantic. If your fleet draws 180 kW at 5:00 PM on one Tuesday because 14 vans all started charging simultaneously, you pay for that peak all month, even if every other charging session was perfectly spread out.
A mid-size delivery fleet running 20 electric vans out of a single depot can see demand charges between $2,000 and $4,500 per month depending on how charging is timed. That number is not fixed. It is entirely a function of scheduling.
Where the savings actually concentrate
The biggest savings in EV fleet charging come from two places: flattening the demand peak and shifting sessions away from peak-rate windows. These are related but distinct problems, and they interact in ways that catch operators off guard.
Flattening the demand peak requires knowing how many vehicles will return to the depot at roughly the same time, what their state of charge (SOC) will be when they arrive, and how long each vehicle needs before its next dispatch. If a vehicle is not leaving until 7:00 AM the next day, it does not need to start charging at 4:30 PM. That flexibility is the scheduling margin Fleetvolts works with.
We track each vehicle's route, its real-world energy consumption per mile given current temperature and payload estimates, and its expected return SOC. That data lets us answer the question "which vehicles need to charge first, and which can wait?" at the vehicle level, not the fleet average level. When you can delay the five vans that came back at 80 percent SOC and prioritize the three that returned at 22 percent, you cut the simultaneous draw significantly without any vehicle failing to be ready for tomorrow.
Time-of-use rate shifting: valuable but secondary
Shifting charging sessions to off-peak rate windows does save money, particularly on tariffs with wide peak-to-off-peak spreads. A tariff with a peak rate of $0.18/kWh and an off-peak rate of $0.07/kWh offers real savings for a fleet consuming 800 to 1,500 kWh per night. At that consumption range, the difference between all-peak and all-off-peak charging runs roughly $88 to $165 per billing period.
That is meaningful, but it is smaller than most operators expect when they first learn about TOU rates. And it competes with the demand charge math. If you delay all your charging to start at 11:00 PM when rates drop, and 18 vehicles all plug in at once, you may have just created your monthly demand peak at 11:01 PM. You saved on the energy rate and paid it back with interest on the demand side.
The correct approach is to interleave TOU optimization with demand peak management, treating them as a joint constraint rather than applying one then the other. The scheduler needs to know both the rate schedule and the depot's power capacity ceiling simultaneously.
A scenario worth walking through
Consider a Midwest-based produce distribution fleet operating 18 electric vans with a shift structure that returns all vehicles to the depot between 3:30 PM and 5:00 PM. The depot has a 200 kW service capacity and a demand charge of $16 per kW per month. The vehicles return with SOC values ranging from 15 percent to 72 percent depending on route length and ambient temperature.
Without any scheduling intervention, if 14 of those 18 vans connect and begin charging simultaneously at the maximum 7.2 kW rate, the depot draws approximately 101 kW just from charging, on top of whatever HVAC and lighting load exists. On a warm afternoon with refrigeration running, it is easy to see 140 to 160 kW of total demand. At $16 per kW, a 155 kW monthly peak means $2,480 in demand charges alone.
With vehicle-level SOC data and a per-vehicle next-dispatch time, a scheduler can sequence charging so no more than 8 to 10 vehicles charge at full rate simultaneously, staggering start times across the 3:30 to 10:00 PM window. The vehicles with lowest SOC and earliest next dispatch go first at full rate. Others get delayed or rate-limited. The depot peak comes down to the 90 to 110 kW range. At 100 kW monthly peak, the demand charge drops to $1,600. That is $880 per month from scheduling alone, with every vehicle fully charged and ready.
What does not help as much as it should
We are not saying that hardware investments do not matter. Upgrading service capacity, adding on-site battery storage, or installing smart chargers with load management firmware all contribute to cost reduction. But they require capital expenditure, permitting, and often months of lead time. Scheduling optimization can start producing results with the chargers and service connection you already have.
We are also not saying that every fleet will see the same magnitude of savings. Fleets running very short routes with high return-SOC averages have less scheduling flexibility because every vehicle needs a full charge before the next day regardless. The savings potential scales with the variance in return SOC and the variance in next-dispatch times across your fleet. More variance means more scheduling room to work with.
The role of accurate energy forecasting
None of the scheduling math works well if you do not know what SOC a vehicle will return with. A vehicle that was supposed to return at 45 percent but actually returns at 18 percent because of an unexpected route extension changes the calculus entirely. The scheduler needs to know that in advance, not when the van pulls in.
This is where real-time telemetry integration and route-aware energy forecasting become load-bearing parts of the cost reduction system. Fleetvolts ingests GPS and telematics data from vehicles on active routes, updates expected return SOC continuously as routes deviate, and adjusts the charging schedule before the vehicles arrive. A depot manager does not need to manually intervene when a driver gets stuck in traffic for 45 minutes; the system requeues the affected vehicle automatically.
The billing result shows up three to four weeks later when the electricity statement arrives. The demand charge line item is either a continued surprise or a number you recognize because you watched it get managed. Getting to the second outcome requires treating charging not as an operational afterthought but as a schedulable resource with its own constraints and cost structure.