A commercial transit depot with 80 battery-electric buses doesn't look like a power plant. But plug in all 80 during the afternoon peak window and you've added somewhere between 4 and 6 MW of load to a distribution feeder that wasn't designed to handle it. If three similar depots exist on the same substation, you have a reliability problem that no amount of peak shaving on the utility's generation side will fix.
This is the reality distribution operators are navigating right now. Fleet electrification is moving faster than most distribution infrastructure plans assumed. The question is no longer whether EV fleets will stress local grids; it's whether those fleets can be operated in a way that turns a potential grid problem into a grid asset.
The Load Profile Problem
Unmanaged commercial EV charging has a brutal load profile. Drivers return vehicles at the end of a shift, the depot connects chargers, and the grid sees a near-simultaneous ramp of high-draw Level 2 or DC fast chargers. At a transit depot, that ramp can be steep: 50 kW per bus, times 80 buses, comes to 4 MW arriving in a 30-minute window around 5 or 6 pm. That overlaps almost perfectly with the residential evening peak.
The impact on a distribution feeder is significant. Depending on feeder topology and the upstream substation's loading margin, that depot arrival event can push a feeder close to or over its thermal limit. Transformers operating near rated capacity experience accelerated aging. Voltage at the far end of the feeder can drop below acceptable bounds. And because the depot is a single customer with a large block of load, there's no statistical averaging to help; when buses come back, they all come back.
This is a qualitatively different problem from residential EV adoption, where millions of small chargers create a more diffuse load that's harder to aggregate but easier to smooth through time-of-use pricing signals. A 100-vehicle commercial depot is a discrete, predictable event. It has a fleet manager who knows the operational schedule. It has a charging management system that knows the state of charge of each vehicle. That means it's also a controllable event, if the right interfaces exist between the fleet operator and the distribution operator.
From Load to Dispatchable Resource
The shift from "large load" to "dispatchable DER" requires two things: technical controllability and an operational framework that creates an incentive to participate.
On the technical side, most commercial fleet charging management systems expose APIs or OCPP-compliant endpoints that allow external dispatch signals to modify charging rates per session or per station. A depot with a fleet management system that has this capability can, in principle, receive a dispatch signal from a distribution operator or aggregator and reduce its aggregate draw from 4 MW to 2 MW for 90 minutes, then resume full charging rate overnight.
The state-of-charge constraint is the key variable. A fleet manager's non-negotiable requirement is that every vehicle is at or above its minimum departure state of charge by the time the first shift rolls out the next morning. Everything else, including when overnight energy is delivered, is flexible. That flexibility window is often 8 to 12 hours at a typical transit depot, which is a substantial dispatchable window for a grid resource.
We're not saying every fleet is equally controllable. Some depots operate around the clock with continuous vehicle turnover, and the charging flexibility window is narrow. School bus fleets are almost perfectly predictable, with vehicles returning at roughly 3:30 pm and departing at 6:30 am. Transit depots vary by route structure. Delivery fleet operators may have two return waves per day. The dispatch model has to reflect the specific operational pattern, not a generic fleet profile.
What Distribution Operators Need to Track
For a distribution operator adding fleet depots to their DER coordination model, the key variables are different from those they track for batteries or solar.
Battery storage has a state of charge, ramp rate, and round-trip efficiency. Solar has a generation forecast tied to irradiance. A fleet depot has a state of charge spread across N vehicles, a minimum departure SoC requirement per vehicle, a departure time schedule, and a maximum simultaneous draw from the charger infrastructure. The dispatch model has to respect all of these constraints or the fleet operator won't agree to participate.
Take a concrete scenario: a regional parcel delivery depot in the Mountain West with 60 medium-duty electric delivery vans. Vehicles return in two waves: roughly 40 vehicles between 5 and 6 pm, and 20 more between 7 and 8 pm. First departure is at 6 am the next day. Maximum charger capacity is 3.6 MW (60 vehicles at 60 kW each). The average vehicle returns with 25% state of charge and needs to depart at 90%. With roughly 65 kWh usable capacity per vehicle and a 60 kW charger, each vehicle needs just over an hour of full-rate charging to reach departure SoC.
That means the 40-vehicle evening wave has nearly 12 hours of flexibility before the 6 am constraint binds. For the distribution operator, that 40-vehicle block represents up to 2.4 MW of shiftable load, with a 10-hour window in which to deliver it. That's a meaningful flexibility resource, comparable to a small distributed battery installation, with the difference that no capital expenditure is required from the grid side.
Integration with Load Forecasting
The challenge is incorporating fleet load into a distribution-level forecast in a way that captures its shape and controllability. A static "the depot adds 4 MW at 5 pm" entry in a load model will systematically overestimate peak demand if any managed charging is in place, and will fail to capture the overnight load that replaces it.
What we've found works better is treating the managed depot as a flexible load block: a quantity of energy (total kWh the fleet needs between return and departure) with a maximum instantaneous draw limit and a time window constraint. The optimizer can then schedule that energy block across the available window in a way that minimizes peak load, avoids coincident peaks with other DERs, and respects real-time feeder loading signals.
This is different from dispatch optimization for a battery. A battery can discharge back to the grid; a vehicle charger cannot (unless bidirectional V2G hardware is in place, which is still uncommon at the fleet scale). The fleet depot is a load-shifting resource, not a storage resource in the traditional sense. The optimization objective is about when to draw, not whether to draw. That distinction matters for how the resource enters the dispatch problem.
The Aggregator Angle
DER aggregators with fleet charging in their portfolio face an additional layer of complexity: the fleet operator is their customer, not their asset. The agreement structure matters. A fleet manager who signed up for a demand-response program expecting to reduce their bill has different expectations from one who signed a formal DER aggregation contract with specific availability requirements and performance penalties.
Most fleet operators in early managed charging programs are willing to accept event-based load curtailment of 1 to 4 hours, 10 to 20 times per year, in exchange for demand charge reductions or incentive payments. That's a reasonable starting point. It's a much narrower availability window than a dedicated battery storage asset would offer, but the zero incremental hardware cost changes the economics.
For aggregators building a portfolio that includes fleet depots alongside dedicated storage and solar, the dispatch intelligence has to understand which resource to call first. Fleet depots with narrow flexibility windows are best used for predictable peak events where you can commit to delivery in advance. Batteries are better suited for real-time frequency response and ancillary market participation where availability requirements are stricter and response times are shorter.
What Voltsynth Tracks for Fleet DERs
When we built the fleet charging module into our dispatch optimizer, the hardest part wasn't the optimization math. It was getting the departure schedule and SoC data updated in near-real-time rather than relying on a fixed daily template. Vehicles break down. Routes change. A dispatch signal based on stale departure schedules will fail when the actual operational picture diverges from the model.
The interface that matters most is between the fleet charging management system and the dispatch platform. Our approach is to poll available OCPP endpoints or fleet management APIs every 15 minutes to update per-vehicle SoC, expected departure time, and charger availability. Those updates feed directly into the dispatch scheduler, so the flexibility window and energy requirement estimates stay current. A fleet depot where 12 vehicles are in maintenance and offline reduces the available flexibility by 20%; the optimizer needs to know that before issuing a curtailment signal.
Fleet electrification will continue to accelerate. Distribution operators who develop the operational frameworks to treat commercial depots as dispatchable resources now will have a significant advantage as fleet sizes grow. The operators who treat them as uncontrolled loads will keep rolling transformer upgrades to stay ahead of peak demand events that, with better coordination, didn't need to happen at all.