What a PC Left On All Day Actually Costs

Measure the idle draw and do one multiplication. Watts over a thousand, times hours, times your unit rate. Everything else in this topic is people estimating a number they could measure in ten minutes.

And the honest headline: a desktop idling all day costs less than most people fear, and far less than one hour of anything that makes heat.

A note on this guide: the method is arithmetic and the measurement is yours. Tariffs vary widely and this deliberately gives no currency figures — the site does not publish prices, and yours would be wrong anyway. See how we test.

The problem, as it appears

Somebody says leaving the machine on all night is expensive. Somebody else says shutting it down wears it out. Neither has a number, and the power supply says 750 W on the side, which sounds alarming.

That 750 W is a ceiling the machine almost never approaches, and it is the source of most of the worry here.

Why the wattage on the box disappoints

Because it describes what the supply can deliver, not what the machine draws.

A power supply rated at 750 W is stating a maximum. What the machine actually pulls depends on what it is doing, and at the desktop with nothing running that is a small fraction of the rating — modern processors and graphics cards drop steeply when idle.

So the number people multiply by is the wrong number, usually by a large factor. The right one is measured, and it is the only input to this whole question you do not already have.

What actually matters

The arithmetic, once

Watts ÷ 1000 = kilowatts. × hours = kilowatt-hours. × your unit rate = cost.

That is the whole method. A machine at 60 W for 24 hours is 1.44 kWh per day. Multiply by your rate and you have a daily figure; by 365 and you have a yearly one.

You do not need anyone else's estimate, and estimates are what make this topic confusing — a figure quoted for someone else's machine on someone else's tariff tells you nothing about yours.

Measure idle, not peak

The measurement that matters is the machine sitting at the desktop doing nothing, because that is what "left on" means.

A plug-in meter gives you that directly, and one that records over time is better still — it captures the mixture of idle, light use and load that a real day contains rather than a single instant.

This page contains affiliate links. If you buy through them, Ugly Fixes may earn a commission at no extra cost to you. Products are selected independently.

This is the same measurement used for a completely different purpose in how much power a gaming PC actually uses, where the question is UPS sizing rather than cost.

Sleep is the setting that matters

The interesting comparison is not on versus off. It is on versus sleep.

Sleep drops draw to a few watts — close to negligible against anything else in the house. So a machine that sleeps when unattended has essentially solved this problem, and the remaining gap between sleep and fully off is small enough that convenience reasonably decides it.

If you want one change from this page, it is checking that sleep is actually enabled and actually working, because a machine that never sleeps because something keeps waking it is drawing idle power around the clock without anyone choosing that.

Doing work overnight changes everything

The figures above assume idling. A machine downloading, transcoding, running backups or hosting something is not idling — it is somewhere between idle and loaded, and that is where the numbers stop being trivial.

That is a reason to measure rather than to assume, because the gap between the two states is most of the answer.

The wear argument is out of date

The belief that shutting down and starting up wears a machine comes from an era of mechanical drives and different thermal behaviour.

Solid-state storage is indifferent to power cycles, and thermal cycling on contemporary components is not the concern it once was. Decide this on convenience and cost; the wear consideration no longer carries weight.

Perspective: what actually costs money

Anything that makes heat. A fan heater, a kettle, a tumble dryer, an electric shower — each draws more in an hour than a desktop idling for a day.

If the goal is a lower bill, the machine is not where the money is, and it is worth saying so plainly on a page that could easily sell the opposite.

The mistake to avoid

Multiplying by the number on the power supply. It is the most visible figure and it is a ceiling the machine reaches rarely if ever. Doing the arithmetic with it produces a frightening annual number that bears no relationship to reality, and it is the single reason this question gets asked at all.

The second mistake is optimising this before measuring anything else in the house. A desktop is a modest, steady load, and the large numbers live in heating, hot water and drying. Ten minutes with a meter tells you where you actually stand — and on most desks the honest conclusion is that leaving the machine on is not the problem.