A common criticism of off-peak electricity tariffs is simple:
“If everybody buys an electric car and a heat pump, everybody will use electricity overnight and overnight electricity will no longer be cheap.”
It sounds intuitive. But it treats the electricity system as if demand, generation and consumer behaviour were all fixed.
They are not.
Electrifying transport and heating will undoubtedly increase electricity demand. The important question is not whether demand rises, but how much energy is actually required, when it needs to be consumed, and how intelligently we use the enormous flexibility provided by electric vehicles, heating systems, batteries and other controllable loads.
The numbers show why this distinction matters.
Around 33 million cars and 30 million homes
At the end of 2025 there were approximately 33.4 million licensed cars in Great Britain. Great Britain also has approximately 30 million dwellings, comprising around 25.8 million in England, 1.49 million in Wales and 2.73 million in Scotland. (RAC Foundation)
Imagine the extreme end state:
- every car becomes electric
- every suitable home uses a heat pump
- substantially more electricity is therefore required
That sounds enormous.
But electricity-system planning depends on energy, measured in kWh or TWh, and power, measured in kW or GW.
They are not the same thing.
A 7 kW home EV charger does not mean every EV continuously requires 7 kW.
Most cars spend the overwhelming majority of their lives parked.
The battery only needs enough energy added over time to replace the miles actually driven.
How much electricity would an all-EV car fleet actually require?
Take a deliberately simple assumption:
6,000 miles per car per year
and an EV efficiency of:
4 miles per kWh, or approximately 0.25 kWh per mile.
A typical car would therefore require around:
1,500 kWh per year
for charging.
Across 33.4 million cars that becomes approximately:
50 TWh per year.
That sounds large, but spread across the year it represents an average electrical load of only around:
5.7 GW.
Another way of looking at it is daily energy.
The entire electric car fleet would require roughly:
137 GWh per day.
If every single kWh were forced into a seven-hour overnight window, the average charging load during that window would be approximately:
19.6 GW.
That would clearly be a substantial load.
But this is deliberately an unrealistic worst-case assumption.
It assumes every car:
- charges every night
- charges only overnight
- cannot charge while parked during the day
- cannot respond to renewable generation
- cannot have its charging delayed by a few hours
- cannot respond to network conditions
- cannot participate in vehicle-to-grid services
Real electric cars are almost the perfect flexible electricity load.
A vehicle that arrives home at 18:00 with 100 miles of remaining range and is not required again until 07:00 does not particularly care whether it charges at 19:00, 01:00, 04:00 or partly at lunchtime the following day.
It merely needs to be ready when its owner needs it.
That gives the electricity system hours of flexibility.
Smart charging changes the calculation
NESO’s Future Energy Scenarios 2025 estimates that smart charging could shift as much as 83% of EV peak electricity demand. (National Energy System Operator (NESO))
This is fundamentally different from trying to persuade millions of households to manually turn appliances on and off.
The customer could simply tell the vehicle:
“I need 80% charge by 07:00.”
The charging system, supplier or system operator can then determine when the energy is delivered within those boundaries.
Government policy already recognises this. Smart home chargepoints have been required to incorporate off-peak default charging periods and randomised delays, specifically to avoid millions of chargers responding simultaneously. (GOV.UK)
So the realistic future is not:
33 million cars start charging at midnight.
It is:
millions of batteries become controllable flexible loads spread across many hours.
Most EVs Need Surprisingly Little Energy Each Day
Another important fact is often missed when discussing the impact of millions of electric cars on the electricity grid:
Most cars simply do not travel very far each day.
Department for Transport data estimates that the average car in England travelled around 7,100 miles during 2024.
Spread across the year, that is only:
19.5 miles per day.
An efficient electric car travelling around 4 miles per kWh would therefore need approximately:
4.9 kWh per average day
to replace the energy used for driving.
Even using an ordinary 7 kW home charger, that represents only around:
42 minutes of charging.
The car might be connected to the house for ten or twelve hours overnight, but that does not mean it needs to charge for ten or twelve hours.
This creates enormous flexibility.
A car plugged in at 18:00 and needed again at 07:00 might have a thirteen-hour window in which the electricity system only needs to find around 40 minutes of charging time.
And many cars will not need charging every day at all.
A vehicle with a 60 kWh battery might accumulate several days of normal driving before its owner chooses to recharge it.
This is why looking at the maximum rating of millions of EV chargers dramatically exaggerates their likely impact on the electricity grid.
Thirty million 7 kW chargers do not represent 210 GW of unavoidable demand.
The important number is the energy the cars actually require.
For the average car, that requirement is surprisingly small.
Intelligent charging then allows those few kWh to be distributed across the quietest periods of the day or night.
Rather than asking:
“What happens if 30 million cars all charge at 7 kW at midnight?”
the more relevant question is:
“How do we distribute roughly five kWh per vehicle across the many hours each car spends parked?”
That is a much more manageable electricity-system problem.
Heat pumps are different
Heat pumps require more careful treatment because heating cannot be shifted as freely as vehicle charging.
A house needs to remain warm.
But a heat pump is also not normally a huge appliance drawing its rated maximum continuously.
Unlike a gas boiler that can deliver a large burst of heat and then stop, a well-designed heat pump generally operates for longer periods at lower output.
Modern units modulate their compressors rather than simply running at maximum output whenever heating is required. Government-backed research specifically notes that modern heat pumps are most efficient when running for extended periods at lower output. (GOV.UK)
Assume, illustratively, an average heat-pump electricity consumption of around:
3,000 kWh per home per year.
Across 30 million homes this would represent around:
90 TWh per year
or an average system demand of approximately:
10.3 GW.
But heating is extremely seasonal.
A January evening is very different from a June night.
That is why the challenge created by heat pumps is primarily a winter peak-capacity problem, rather than simply an annual energy problem.
NESO explicitly recognises this. Its future scenarios expect electrified heating to increase winter demand, but also describe heat-pump demand as having a steadier daily pattern than today’s gas boilers and having some scope for flexibility. (National Energy System Operator (NESO))
NESO estimates flexibility could shift approximately 36% of heat-pump peak demand. (National Energy System Operator (NESO))
That flexibility can come from several places.
Homes themselves store heat.
Hot-water cylinders store heat.
Some systems can include dedicated thermal storage.
A house can sometimes be heated slightly more before a constrained period and slightly less during it without the occupants experiencing a noticeable temperature change.
The important distinction is that this should be automated rather than requiring households to constantly manage their heating.
This does not mean today’s grid could support total electrification tomorrow
This point matters.
It would be misleading to claim that today’s unused overnight capacity alone could accommodate every car and every heating system becoming electric immediately.
It could not.
Electricity generation, transmission and distribution capacity will need to expand as the economy electrifies.
NESO already models exactly that transition.
An earlier NESO scenario for 2035, for example, modelled approximately 27 million battery electric vehicles and around 15 million homes with heat pumps, with total annual electricity demand rising from roughly 300 TWh to around 450 TWh. (National Energy System Operator (NESO))
The electricity system therefore grows alongside electrification.
We should not compare the electricity demand of a largely electrified 2040s economy with the generation infrastructure of 2026 and assume nothing else changes.
Britain’s electricity demand already varies enormously
In 2025 Britain’s electricity demand reached an annual high of approximately 45.9 GW.
Its lowest recorded demand was only 12.9 GW. (National Energy System Operator (NESO))
Those figures do not represent a permanent 33 GW block of spare capacity — generation conditions, network constraints and seasonal requirements matter — but they demonstrate just how variable electricity demand already is.
The system is designed around peaks.
If flexible electrification fills parts of today’s valleys without creating larger unmanaged peaks, the result can actually be better utilisation of infrastructure that already exists.
A network that carries 45 GW for a few hours but only 20–25 GW for long periods is economically inefficient.
Adding controllable demand during quieter periods can spread the fixed cost of that network across more electricity.
That should help make electricity cheaper, not more expensive.
Overnight does not have to be the only off-peak period
This is perhaps the biggest flaw in the argument that electrification inevitably destroys cheap overnight electricity.
Our energy system is changing.
Historically, Britain’s lowest demand generally occurred overnight because households and businesses were asleep while large conventional generators continued operating.
A renewables-heavy system behaves differently.
There will increasingly be periods when electricity is abundant because the wind is blowing strongly overnight.
But there will also be summer afternoons when solar generation is extremely high.
Instead of defining cheap electricity permanently as:
23:00 to 06:00
the system can combine a predictable core off-peak period with dynamic flexibility rewards when additional demand would help the grid.
Reward people for consuming electricity when Britain has too much of it
Imagine a sunny summer Sunday.
Solar generation is extremely high.
Wind generation is healthy.
Industrial demand is relatively low.
Rather than paying renewable generators to curtail production, the system could issue a national or regional flexibility signal:
Excess Energy Event: 12:00–15:00
Customers with flexible appliances could then receive additional credit for increasing consumption.
An electric car could automatically charge.
A home battery could fill.
A hot-water cylinder could heat.
A heat pump could slightly preheat the property.
Commercial refrigeration could shift some demand.
Industrial users could increase suitable processes.
The customer would not need to watch the electricity market.
Their devices could respond automatically within limits they had previously approved.
A simple reward-credit model
Under a national domestic tariff, normal off-peak electricity might remain inexpensive and predictable.
But when the electricity system has excess generation, an additional flexibility credit could be offered.
For example:
Normal off-peak rate: 7.5p/kWh
Then during a renewable surplus event:
7.5p electricity – 5p flexibility credit = effective 2.5p/kWh
During more extreme surplus conditions the reward could theoretically approach the entire electricity cost.
This changes consumer behaviour without creating complicated tariffs.
People would still know their normal rate.
The flexibility payment would simply be a bonus for helping the electricity system when additional demand has value.
The objective should be to flatten the curve
The ideal electricity system does not have everyone using as little electricity as possible.
It has generation and demand matched efficiently.
Consider a simplified winter day.
An unmanaged system might look like:
Morning: demand rises sharply
Midday: demand falls
Evening: enormous peak
Overnight: demand collapses
Now introduce electrification intelligently.
EV charging moves away from the evening.
Some water heating moves overnight.
Home batteries charge during low-demand periods.
Heat pumps operate steadily rather than producing a huge recovery peak.
Industrial flexible loads respond to system conditions.
The valleys become higher while the peaks remain controlled.
That is beneficial.
The expensive part of an electricity network is not necessarily supplying another kWh at 03:00.
It is building enough generation, substations, cables and transformers to supply everybody simultaneously during the highest-demand half hour of the year.
Local networks still matter
National generation capacity is only part of the issue.
A country might theoretically have sufficient generating capacity while a particular street transformer or local distribution circuit becomes overloaded.
That is why intelligent charging should also respond to local network capacity, not just wholesale electricity prices.
If twenty EVs are connected to one residential transformer, the system does not need to charge all twenty simultaneously.
A managed charging system could distribute available capacity between them.
One vehicle might charge from 23:30–01:00.
Another from 01:00–03:00.
Another from 03:00–05:00.
The customer only needs the requested charge level by the requested departure time.
This avoids expensive local reinforcement where it is unnecessary while identifying places where genuine reinforcement is required.
Electric vehicles could eventually become part of electricity supply
The opportunity becomes even larger with vehicle-to-grid technology.
A future fleet of tens of millions of electric vehicles represents an extraordinary quantity of battery storage.
Most of those vehicles will be stationary at any particular moment.
Even if only a fraction of their battery capacity were available to the electricity system, the theoretical flexible resource would be enormous.
NESO has already trialled vehicle-to-grid systems where charging and discharging were altered in response to system conditions while protecting the driver’s requirements. (National Energy System Operator (NESO))
The electric car therefore should not simply be regarded as another appliance creating demand.
It can potentially become both a flexible consumer and a distributed energy store.
So will overnight electricity always remain cheap?
Not necessarily every night.
There will be cold, still winter nights when electricity is relatively scarce.
There will be windy nights when Britain has enormous quantities of inexpensive generation.
There will increasingly be summer afternoons when solar makes electricity abundant.
The objective therefore should not be to promise that electricity between two arbitrary clock times will always be the cheapest electricity produced.
It should be to maintain a predictable low-cost off-peak tariff for consumers while using automated flexibility to move discretionary demand towards the periods when electricity is genuinely abundant.
That distinction matters.
Electrification does not mean 30 million kettles switching on at once
The worst way to electrify Britain would be to install millions of EV chargers, heat pumps and batteries and allow every device to operate independently without considering the electricity system.
Fortunately, these are precisely the technologies that give us greater control over when electricity is consumed.
NESO’s own modelling concludes that household and business flexibility could reduce peak demand substantially, with smart EV charging capable of shifting 83% of EV peak demand and heat-pump flexibility shifting 36%. (National Energy System Operator (NESO))
So the question should not be:
“What happens if everybody charges their car at the same time?”
The question should be:
“Why would we design an electricity system that allowed everybody to charge their car at the same time?”
Cheap electricity should reward behaviour that makes the entire system cheaper.
Off-peak charging is part of that.
But the next generation of electricity pricing should go further.
It should make flexible demand almost invisible to the consumer:
Tell us when the car needs to be ready.
Tell us what temperature the house should be.
Let the electricity system work out the cheapest and most efficient time to supply the energy.
That is how electrification can increase electricity consumption without simply creating ever-larger peaks.