Electric Vehicles: Where the Energy Actually Goes

The Efficiency Gap

A petrol engine is a heat engine, so it is bound by the Carnot relationship between its combustion and exhaust temperatures. In practice a modern engine converts 20 to 30 percent of the fuel's energy into work at the crankshaft under favourable conditions, and considerably less in city driving where it idles and runs far from its efficient operating point. The rest leaves as exhaust heat and radiator heat.

An electric motor is not a heat engine. It converts electrical energy to rotation through magnetic fields, and modern permanent-magnet and induction motors reach 90 to 95 percent efficiency across most of their speed range. Adding inverter losses, battery charge and discharge losses and transmission losses, roughly 77 percent of the energy drawn from the socket reaches the wheels.

Two further differences compound in city driving. An electric motor produces full torque from standstill and consumes nothing while stationary, so there is no idling loss. And regenerative braking runs the motor backwards as a generator during deceleration, recovering 60 to 70 percent of kinetic energy that a friction brake would have turned into heat. A combustion car discards all of it.

The honest counter-argument is the upstream one. If the electricity comes from a coal plant at 37 percent efficiency, the well-to-wheel figure falls considerably. Even then electric vehicles come out ahead in most analyses, because the coal plant's 37 percent still beats the engine's 20, and the advantage grows every year the grid decarbonises - which is the structural point: an electric vehicle gets cleaner while parked, and a combustion vehicle does not.

What Is Actually in the Battery

A lithium-ion cell stores energy by moving lithium ions between a graphite anode and a metal-oxide cathode. The cathode chemistry is what the various names refer to, and it sets the trade-offs.

Nickel-manganese-cobalt and nickel-cobalt-aluminium chemistries offer the highest energy density, 250 to 300 watt-hours per kilogram at cell level, which buys range. They also require cobalt, much of it from the Democratic Republic of Congo, where artisanal mining has documented child labour and safety problems. Manufacturers have progressively reduced cobalt content for this reason as much as for cost.

Lithium iron phosphate has taken a large share of production since around 2020. It carries 150 to 200 watt-hours per kilogram, so it gives less range for the same mass, but it uses no cobalt or nickel, survives two to four times as many charge cycles, is markedly more resistant to thermal runaway, and costs less. For a vehicle whose typical journey is short, the density penalty rarely binds.

Sodium-ion cells entered commercial production in the mid-2020s, trading further density for materials that are abundant everywhere and low-temperature performance that lithium chemistries struggle with. Solid-state cells, which replace the liquid electrolyte with a solid one, promise higher density and better safety and have been close to production for longer than their advocates would like.

Degradation is better than early fears suggested. Fleet data shows typical capacity retention around 90 percent after 200,000 kilometres, and battery warranties of eight years or 160,000 kilometres have become standard. Recycling now recovers more than 95 percent of nickel, cobalt and copper in commercial plants, though volumes remain small because few vehicles have reached end of life.

The Carbon Debt

Building a battery emits carbon dioxide - between 60 and 100 kilograms per kilowatt-hour of capacity depending on the electricity used in the factory. A 75-kilowatt-hour pack therefore starts life with a debt of roughly 5 to 7 tonnes, which is why an electric vehicle leaves the showroom with a larger footprint than an equivalent petrol car.

It then pays that debt back through use. The break-even distance depends almost entirely on the grid: roughly 15,000 kilometres in a country running on hydro and nuclear, 30,000 to 40,000 on a European average grid, and considerably more where generation is coal-heavy. Against a typical vehicle life of 200,000 kilometres or more, the debt is repaid early in every case examined by mainstream life-cycle studies.

Two factors are improving the arithmetic on their own. Battery factories are increasingly sited next to cheap clean electricity, which cuts manufacturing emissions directly. And grids decarbonise over a vehicle's life, so the car charging in 2035 is cleaner than the same car charging in 2025 without anything being done to it.

The mining question is real and separate from the carbon question. Lithium extraction uses large volumes of water in arid regions of the Atacama, nickel mining has caused deforestation in Indonesia, and cobalt carries the labour issues noted above. These are legitimate concerns about how materials are obtained, and they apply to a one-time stock of material that is then recycled, rather than to a fuel that must be extracted continuously.

What a Fleet Does to the Grid

Electrifying a national car fleet raises electricity demand by roughly 20 to 30 percent. That sounds alarming and is manageable, because the timing is almost entirely flexible. A car is parked around 95 percent of the time and typically needs only a few hours of charging per week.

That flexibility is worth more than the demand costs. Charging that shifts to windy nights or sunny midday absorbs generation that would otherwise be curtailed, and managed charging schemes are already among the largest sources of demand-side flexibility available. Uncontrolled charging does the opposite: if everyone plugs in at 18:00 on arriving home, the fleet lands squarely on the existing evening peak.

Vehicle-to-grid goes further and lets cars discharge back. A fleet of a million vehicles with 60-kilowatt-hour packs holds 60 gigawatt-hours, which is comparable to all the grid-scale batteries in several countries combined. Commercial deployment remains limited by standards, warranty terms and the question of who pays for the extra cycling, but the physical resource is undeniably there.

The local network is where the real constraint sits. National generation can cope; a residential distribution transformer sized in 1975 for lighting and a kettle may not cope with six cars charging at 11 kilowatts. This is the same reinforcement problem that distributed solar creates from the other direction, and it is why electrification of transport and heat is in practice a distribution network question rather than a generation one.