For most fleets, depot charging wins. If a vehicle sits unused longer than it takes to recharge it, a charger and a service connection cost less than a spare pack per vehicle, and the battery inside that vehicle is already bought and paid for. Battery swapping earns its capital only where three things hold at once: the dwell window is genuinely shorter than the charge time, one pack standard already covers the whole fleet, and every vehicle comes back to the same site.
Key takeaways
The deciding number here is not swap time. It is the dwell window: departure time minus return time, minus every task that is not charging. Yard moves, washdown, loading and unloading, driver handover, mandated rest, inspection. Whatever survives that subtraction is the time a charger actually has. If that slice is longer than a full charge takes, a plug and a service connection beat a spare pack per vehicle, and no amount of station speed changes the answer.
Fleet averages hide the vehicles that set your bill. A depot where most vans cover a modest distance still has a tail running two or three times the average: the relief vehicle, the long route, the one that carries the awkward customer at the end of the day. That tail fixes the charger power you need and the pack count you need, because those vehicles arrive late and leave early. Size the site on the worst three vehicles, then check how many others share that pattern.
Sort the fleet before you compare hardware. Overnight vehicles are parked longer than a full charge takes. Partial-dwell vehicles get hours at the yard, enough for a top-up but not a full pack. Opportunistic vehicles are back at base for under an hour with high daily energy. Only the third class is a real swap candidate, and in most depots it is a minority of the fleet.
| Dwell class | Net window | Plug power per vehicle | Does a top-up cover the next shift? | Can spare packs pay for themselves? |
|---|---|---|---|---|
| Overnight | Longer than a full charge | Low, often AC | Yes, full pack available | No |
| Partial | A few hours | Medium DC | Sometimes, depends on shift energy | Only if the shortfall recurs every day |
| Opportunistic | Under an hour | Very high, or physically impossible | No | Yes, if one pack standard holds |
Peak season adds trips and moves loading later. Winter slows everything that happens outdoors. A vehicle returning forty minutes late cascades into a later washdown and a later plug-in, which is exactly the kind of chronic slippage that turns a comfortable overnight window into a marginal one. Measure the window per vehicle class across four consecutive weeks, worst week included, before you size anything with it.

Here is the calculation, worked once on a hypothetical fleet. Swap in your own numbers, because the method is the deliverable and the example is not. Take 40 electric vans on a single daily shift, returning at 19:00 and departing at 06:00, with a two-hour loading and washdown block before they can be plugged in.
Swapping usually wins on stop time and loses on cost per delivered kWh, and buyers who conflate the two end up owning the wrong asset. Build both stacks on the same energy throughput and the comparison becomes ordinary engineering economics rather than a technology argument.
| Cost line | Depot charging path | Battery swapping path |
|---|---|---|
| Battery capital | Already inside the vehicle | Pack price divided by lifetime kWh throughput |
| Fixed equipment | Charger capex plus service upgrade | Station capex, automation and pack store, amortised |
| Energy | Metered energy at the tariff rate | Same energy, plus handling losses and store cooling |
| Labour | Plug in, walk away | Staffed exchange or maintenance of automation |
| Space | Existing bays and a service corridor | Building, queue lane and pack store |
| Demand charges | Peak around shift change | Flatter profile, higher total kWh |
Pack capital does not disappear because you bought the vehicles. Battery cost sits in the swap stack as a number that has to be recovered over a finite cycle budget: pack price divided by cycles multiplied by usable kWh, adjusted for round-trip efficiency. Every kWh you deliver through that pack carries a share of it.
If the operator owns the spare packs, swapping converts a charging problem into a fleet capital problem, with packs depreciating on the balance sheet. If a third party owns them, the operator pays a margin on every kWh and has handed away control of its energy price. Neither is wrong. Both should be a deliberate choice, signed, before the first pack arrives.
Faster turnaround, bought with more capital sitting in packs. A lower site peak, paid for with more total energy through the site because of double handling and pack-store HVAC. Less parking consumed, at the price of a building that must exist. Those are the same three points that appear in every pros and cons list, stated in the form a finance team can actually price.
One contract question outranks the rest. Packs on exchange duty see more full cycles, more high-rate charging and more physical handling than a vehicle-mounted pack ever would, so warranty terms drafted for a pack that stays in one chassis may not survive exchange duty. Settle that with the supplier before ordering, not after the first warranty claim. On land, depot charging spreads across bays the site already holds, while a swap station needs a dedicated building close to the route, and on an urban site that building may not exist at any price. For a fuller breakdown of the depot-side capital lines, see commercial EV charging station costs explained.

Swapping fails for structural reasons rather than because the idea is unsound. Three constraints decide it, and they are not the ones usually listed.
Swapping only works if every vehicle on site takes the same pack, mechanically and electrically. A fleet bought over seven years across model changes will not hold that line. One new vehicle with a different pack height, a different mounting pattern or a different BMS protocol turns a station into a monument. The station does not degrade gracefully. It stops serving a share of the fleet the day that vehicle arrives.
Tesla demonstrated a pack swap in 2013 and never brought it to market as a product. The instructive part is the arithmetic, not the demo. A swap network needs duplicate packs and station capital in place before the first customer arrives, spread across every route you intend to serve. A fast-charging site reuses one asset across every vehicle that turns up, including vehicles you have never seen. One model carries inventory risk on a schedule the operator does not control. The other carries it on the grid connection, which is at least a known quantity once the utility signs off.
Modern packs are part of the vehicle floor and structure. Structural integration buys rigidity, crash performance and mass, and all three push against removability. Vehicle engineering is not converging on a universal cassette, so a buyer betting on exchange is betting against the direction of the products they will be offered in five years.
Every pack still has to be charged, and the pack store adds handling losses and cooling load on top of the same energy bill. Then peak congestion returns: after a shift change, the station is recharging a stack of depleted packs at once. Unless the site is deliberately oversized on packs, the swap station pulls the same grid capacity as the depot it replaced, just later in the evening.
Battery swapping is the right answer inside a narrow band, and inside that band it is clearly right. The band is defined by operating geometry, not by technology preference.
Forklift and AGV fleets have exchanged batteries for decades, with a battery room, a bank of chargers and a trained operator on shift. The playbook already exists, including the parts people rediscover the hard way: matching charger voltage to the fleet, keeping the battery room ventilated, and training whoever performs the exchange. What changes with road vehicles is pack mass, structural integration and public-road regulation, not the underlying concept.
The industrial product landscape shows why material handling standardized early. evaisun’s own catalogue lists industrial chargers at 24V, 36V, 48V, 72V and 80V with CE and UL by TUV, and lithium packs in 80V, 76.8V, 51.2V and 48V versions, so a warehouse can standardize a voltage across a fleet and keep it for years. A mixed road fleet spanning three OEMs, three pack architectures and two BMS dialects cannot do the same thing. If your operation runs forklifts or AGVs alongside road vehicles, the material handling side is where exchange economics are already proven; optimizing warehouse efficiency with advanced forklift battery chargers covers the operating discipline that goes with it.
Transit and shuttle operators, taxi and ride-hailing depots, port drayage running fixed loops. What they share is one procurement authority, one route network and one place vehicles come home to. That is the topology swapping needs. Even there, the case hinges on dwell: a taxi depot whose cars idle six hours overnight is a charging site with a queue problem, not a swap site. The trade-offs on that duty cycle are set out in taxi and ride-hailing fleet charging: balancing speed, cost, and vehicle uptime.
You are signing a pack standard for the life of the asset. Write four things into the procurement document: the pack envelope and mounting interface, the BMS communication protocol, the swap mechanism interface, and an exit clause that covers the day the next vehicle model does not fit. Without the exit clause, the station owns your replacement cycle.

Interconnection lead time outranks equipment lead time on almost every depot project, and it is the schedule risk buyers forget to put on the critical path. A site needing a substantial service upgrade waits on a utility study, transformer availability and queue position, and none of those respond to a purchase order. Confirm the local process with your utility early and design around the answer, because the number varies by region, by voltage level and by how busy your feeder is.
Demand charges are billed on peak kW, and many commercial tariffs include a ratchet clause that lets one bad evening set the billing demand for months afterwards. Depot charging concentrates its peak into the hour or two around shift change, which is exactly where the ratchet bites. A site that charges forty vans at once at 18:30 pays for that simultaneity all year.
Depot charging is a short, tall peak. A swap station is a flatter profile spread across twenty-four hours, scaled up by the number of packs in the store. That flattening is swapping’s genuine utility advantage: same energy, less peak. It is not free, because the station still needs a connection sized for the stack it recharges, but the capacity can often be bought at a lower tier and managed downwards.
A battery-buffered charger flattens the depot peak with one battery instead of a fleet’s worth of spare packs. This is the comparison a buyer should actually run: cost per kW of peak avoided by spare packs versus cost per kW avoided by a buffer battery. On most sites the buffer wins on capital by a wide margin, because it is sized for the peak, not for the vehicles. The control side of that design is covered in how load management EV charging improves reliability and cuts costs.
Tariff structure decides whether any of this works. Time-of-use windows and flexible connection agreements can make a swap station’s controllable charging genuinely cheap, but only if the site’s metering and contract permit that charging pattern. Read the tariff before the design is frozen, not after.
The month that decides this comparison is the coldest one, not the annual average. A cold pack accepts limited charge current until its cells warm, so a depot charger delivers well below nameplate for the opening stretch of every session. The nine-hour window you calculated in June quietly shrinks in February, and the vehicles most affected are the ones that arrived latest.
Depot mitigations exist and they work. Vehicle-side pack heaters and thermal management bring the pack into its acceptance band. Scheduled preconditioning warms the pack before the vehicle returns, so the window is not spent heating. Operators can also accept a longer window or a lower state of charge at departure, which is a schedule decision rather than an equipment one.
Do not forget the charger has an operating envelope of its own. AiPower’s installation guide specifies an installation environment of -20 degrees C to 45 degrees C, which is the kind of figure that belongs on the layout drawing next to the bay it serves. A cabinet in a spot that collects afternoon sun on a July wall can fall outside the upper end just as surely as a January morning falls below the lower end.
Swapping handles cold differently. Packs charge indoors, in a conditioned store, at whatever rate the operator chooses, and the vehicle leaves with a warm pack. That is a real winter throughput advantage, and it is paid for by heating and ventilating a building all winter, plus the electrical load of doing it.
Winter also raises energy per kilometre through cabin heat and traction losses, so charge time grows exactly when charging is slowest. Rerun the arithmetic from Section 2 on the coldest month and the longest route. If the window still closes there, it closes all year. If it does not, that is the number that should drive the decision, not the summer case.

These are not two equipment choices. They are two construction projects with different approvals, different footprints and different labour models, and comparing them on hardware cost alone is how projects get to the planning stage and stop.
Run the decision in order, and stop at the first no. Does the dwell window exceed the full charge time for most vehicles? Is one pack standard fixed for the asset’s life, and does every vehicle return to this site? Can the site obtain its peak power inside the project schedule? Does the coldest month still fit inside the window? Any no sends you back to depot charging, or to a buffered depot that flattens the peak instead. Only when all four are yes is it worth running the cost per delivered kWh on both paths with your own numbers.
| Criterion | Depot charging | Battery swapping |
|---|---|---|
| Capital per vehicle | Charger share plus service upgrade | One or more spare packs per vehicle |
| Best dwell window | Longer than a full charge | Shorter than a full charge |
| Site peak demand | Tall peak around shift change | Flatter curve across 24 hours |
| Pack standardization | Not required | Mandatory for the asset’s life |
| Shift-change throughput | Scales with each plug added | Scales in discrete station increments |
| Cold weather | Slow until the pack warms | Warm pack on hand, heated store needed |
| Footprint | Existing bays plus clearance | Dedicated building, queue lane, pack store |
| Best-fit fleet | Overnight depot fleets | Captive short-dwell single-pack fleets |
The pilot that earns the signature is boring and cheap. Instrument twenty to thirty vehicles for four weeks and log return times, departure times and energy per shift, charger-side and vehicle-side. Rebuild the model from that data before signing anything, because a documented pilot costs a fraction of an order of packs that turns out to be the wrong size or the wrong voltage.
Two situations override the arithmetic. If the fleet is growing or procurement is multi-OEM, the standardization bet is the wrong bet regardless of the numbers. If the site is leased short-term, you are building someone else’s asset, and a depot charger that can be relocated is worth more than a station that cannot.
On the depot side, evaisun’s product lines cover DC fast chargers, AC and portable EV chargers and industrial charging hardware, with power modules alongside them. The SDM Series and AVR Series are among those module products, and one DC EV charger variant is offered with a 30kW Power Module. On the pack and material handling side it builds industrial lithium batteries and 24V to 80V industrial chargers, with solutions stated to carry UL and CE certification and a team of 60+ engineers behind the specification work. If you are holding a dwell dataset and a half-finished charger schedule, settle the sizing before the capital request goes forward rather than after. The broader design context sits in fleet depot charging for 50+ vehicles.
No, not in most cases. Swapping adds packs, station capital and pack-store running costs on top of the same energy bill, so cost per delivered kWh usually rises even where stop time falls. It becomes competitive only when a short dwell would otherwise force expensive peak power or lost vehicle hours worth more than the pack capital.
It depends on turnaround time and swap frequency. Divide pack turnaround, which covers queueing, exchange, handling and recharge, by the interval between that vehicle’s swaps, then multiply vehicle count by one plus that ratio. Fleets running multiple shifts per day need a visibly higher ratio than single-shift fleets, because there are fewer idle hours to absorb recharging.
Rarely, and never cheaply. Retrofitting means changing the pack envelope, the mounting interface and the BMS communication protocol on vehicles already in service, then adding a pack store with its own approvals. The economics only close for a fleet buying new vehicles on one platform under a single procurement authority.
No. Swapping concentrates cycling rather than reducing it. Total energy through the fleet stays the same, but it now passes through fewer packs, which see more full cycles, more high-rate charging and physical handling. Ask the pack supplier to confirm in writing that the warranty covers exchange duty before you order.
US
English中国大陆
简体中文German
Deutsch
Russian
Русский
French
FrançaisSpanish
español
Portuguese
PortuguêsJapanese
日本語
Italian
Italia
Korean
한국인Subscribe the newsletter to get updated to news and promotions