Charging power is easy to sell and hard to deliver. A charger rated at 350 kW will hold that figure for a small part of a session on most vehicles, and the reason lies inside the cell rather than in the equipment.
The limit is the anode
On discharge, lithium leaves the graphite structure. On charge, it has to enter it, and intercalation into graphite has a finite rate. When ions arrive at the particle surface faster than they can be accommodated, the anode potential falls below zero volts against lithium metal, and lithium deposits on the surface instead of entering the structure.
That plated lithium is a compound problem. Part of it re-intercalates on rest, part reacts with electrolyte to form new interphase and is lost, part becomes electrically isolated, and the deposits can grow into structures that eventually threaten the separator. So the charging limit is not a comfort setting. It is the boundary of a mechanism that costs capacity and, at the extreme, creates a safety hazard.
Three variables move that boundary. Temperature, because intercalation kinetics slow sharply in the cold. State of charge, because a fuller anode has less room and a lower potential margin. Age, because a cell whose anode porosity has been reduced by earlier degradation has higher local current density at the same pack-level current.
Why CC-CV exists
The standard profile has two phases.
During constant current, the charger holds a fixed current and the cell voltage rises. Most of the charge is delivered here.
The cell then reaches its upper voltage limit while still short of full. Holding the current would push it above that limit, so the charger switches to constant voltage and the current tapers as the cell approaches equilibrium. Charging ends when current falls to a defined cut-off, often around C/20 or C/50.
Two consequences follow. The tapering phase is unavoidable within a simple CC-CV scheme, and it is why the last portion of charge always takes disproportionately long. And the measured internal resistance affects when the transition happens, so an aged cell with higher resistance hits its voltage limit earlier and spends more of the session in the slow phase, even before considering any degradation-related current derating.
What real charging profiles look like
Production fast charging is not one constant current step. It is a sequence of decreasing current steps, computed by the BMS against a plating-avoidance limit that depends on temperature, state of charge and estimated cell condition. Power is highest in the first minutes, when the pack is relatively empty and the current limit is high, then steps down repeatedly.
This is why the honest metric is time from 10 to 80 per cent rather than peak kilowatts. A vehicle peaking at 350 kW for ninety seconds and then settling to 120 kW can lose to a vehicle that holds 200 kW steadily.
Worked example. A 77 kWh usable pack charging from 10 to 80 per cent moves about 54 kWh. At a session average of 150 kW the transfer takes roughly 22 minutes; at a session average of 90 kW it takes about 36 minutes. Peak power appears nowhere in that calculation.
Two current reference points show the spread. The current Porsche Taycan peaks at 320 kW on its 800 volt pack, sustains more than 300 kW for up to five minutes, and covers 10 to 80 per cent in 18 minutes. BYD's Super e-Platform, launched in 2025 in the Han L and Tang L, pairs a 1,000 volt pack with cells rated for 10C charging and draws up to 1,000 kW on matching chargers, adding roughly 400 km of CLTC range in five minutes under ideal conditions. Cells rated for 4C to 6C charging are now common in premium Chinese-market vehicles, and the constraint is shifting from the cell towards grid connections and cooling hardware.
Some systems now use plating-detection feedback rather than fixed lookup tables. The voltage relaxation signature of re-intercalating lithium can be observed after a charge step, and adaptive strategies use it to push closer to the actual limit rather than to a conservative margin. Expect more of this as onboard diagnostics improve.
Temperature, and the case for preconditioning
Charge acceptance falls steeply as temperature drops. Below roughly 0 degrees Celsius, plating begins at quite modest currents, which is why every capable EV restricts or blocks fast charging on a cold pack.
Preconditioning warms the pack, typically using waste heat, a heat pump or resistive heating, so that it arrives at the charger inside a suitable window rather than warming up during the session. Where a vehicle knows the destination is a charger, it can start early. Where it does not, the driver pays the difference. This is why the same car can post very different charging curves in the same weather depending on whether navigation was routed to the charging stop.
There is an upper bound too. Charging hot accelerates film growth and other thermally driven mechanisms, so the objective is a window rather than maximum temperature. Heat removal during the session matters as much as heat addition before it, and the heat generated scales with the square of current.
Cell and pack design for fast charge
Thinner electrodes shorten the diffusion path and lower the local current density per unit area, which improves rate capability and reduces specific energy. This is the central compromise: a fast-charging cell stores less energy per kilogram than an energy-optimised cell of the same chemistry.
Electrode engineering can shift the compromise. Lower tortuosity through structured or laser-patterned electrodes, graphite particle size and surface treatment, and electrolyte formulations with higher ionic conductivity all improve charge acceptance without simply thinning the coating.
At pack level, cooling capacity and its uniformity set the sustained rate. A pack that can move heat out of the middle of a large-format cell as fast as the cell generates it holds power longer than one that cannot, regardless of cell specification.
Voltage architecture and the infrastructure side
Power is current multiplied by voltage, and resistive losses scale with the square of current. Raising the pack voltage from roughly 400 to roughly 800 volts halves the current for the same power and cuts resistive heating in the cables, connectors and busbars by around three quarters. That is the engineering case for 800 volt systems, and it also allows thinner, lighter, more flexible charging cables. The dependency runs both ways: the same Taycan that peaks at 320 kW on an 800 volt charger is limited to 150 kW at a 400 volt post, because the power then passes through an onboard voltage converter.
Above a certain power, cables need active cooling regardless. For heavy vehicles, the Megawatt Charging System (standardised as SAE J3271, rated up to 1,250 volts and 3,000 amps, about 3.75 MW) has been developed because passenger car connectors cannot deliver the power a truck needs in a driver rest break.
Communication matters more than it looks. ISO 15118 defines the vehicle to charger interface that carries the current and voltage limits the BMS calculates, along with authentication and, in later versions, bidirectional power transfer. A charging session is a negotiation between the vehicle's limits and the charger's capability, and the vehicle is normally the one setting the ceiling.
What to tell someone who asks if fast charging is bad
The published fleet evidence points the same way. Occasional DC fast charging on a thermally managed pack in temperate conditions has a small effect relative to temperature and calendar ageing. Frequent fast charging on a pack without active thermal management, or in cold conditions without preconditioning, has a considerably larger one.
The useful advice is therefore about conditions rather than abstinence: precondition in the cold, avoid charging a pack that is already hot from hard driving, and avoid holding the pack at high state of charge after a fast session. Those three habits matter more than the number on the charger.
Informational and educational content only. Not professional, financial, legal, or engineering advice.