Laptop Sleep vs Shutdown Energy Cost Calculator

Introduction to the sleep-versus-shutdown energy question

Every time you step away from a laptop you make a small energy decision: leave it asleep so it wakes instantly, or shut it down so it draws almost nothing. The folklore on both sides is confident and mostly wrong. One camp insists that sleep is "basically free"; the other insists that booting burns so much power that you should never shut down. Both claims are testable, and the numbers involved are small enough that intuition is a poor guide. This calculator settles the question the only way it can be settled — by pricing the two policies against each other over a defined away period, including the energy of the boot itself.

The power figures come from a standardised vocabulary. The US EPA's ENERGY STAR Program Requirements, Product Specification for Computers, Eligibility Criteria Version 8.0 defines Off Mode as "the lowest power mode which cannot be switched off (influenced) by the user", correlating to ACPI system level S5, and Sleep Mode as a low power mode the machine enters automatically, most commonly correlating to ACPI system level S3, suspend-to-RAM. The definitions matter because they make explicit something users routinely get wrong: Off Mode is a mode, not the absence of one. A shut-down laptop with its adapter plugged in is still energised, and the specification treats its consumption as a measurable quantity with its own symbol, P_OFF.

That distinction is what makes the comparison non-trivial. If shutdown were genuinely zero watts, any away period longer than the boot would favour shutting down and there would be nothing to calculate. Because off-mode power is small but real, the saving you get from shutting down is only the difference between the two modes — and that difference has to pay back the energy your machine spends booting when you return.

How to use this comparison: entering power, away time and rate

Work through the form from top to bottom. Every field affects the answer, and two of them are commonly mis-entered.

1. Sleep power draw (W). The wall power your laptop and its adapter draw while asleep. Measure it with a plug-in meter if you can; a single reading taken a minute after the lid closes is enough.
2. Off-mode power draw (W). The wall power with the machine shut down but the adapter still plugged in. This is almost never zero. If you genuinely unplug the adapter every time, enter 0.
3. Length of one away period (hours). This is the single stretch you are away for — an overnight, a lunch break, a meeting — not the yearly total. Getting this right is what makes the break-even meaningful.
4. Away periods per day. How many such stretches happen in a typical day. One overnight is the common case; a machine that is closed for lunch and again overnight is two.
5. Boot power draw (W) and boot duration (s). The average wall power while the machine starts up, and how long that lasts. Together these give the boot-burst energy the shutdown policy has to repay.
6. Electricity rate ($/kWh). Your all-in marginal rate, delivery charges included.

Press Calculate (or the Enter key from any field) to get the annual cost of each policy, the break-even away time, and a chart showing where the two curves cross. Press Reset to return to the reference values described below.

Formula derivation: from watts to a genuine break-even

Start with the definition of energy. Power P in watts sustained for t hours is P·t watt-hours; dividing by 1000 converts to kilowatt-hours, and multiplying by the rate r in dollars per kilowatt-hour gives cost. For a state held for L hours per away period, n away periods per day, across a 365-day year:

The unit chain here is the first place calculators go wrong. P is in watts, so P·L·n is watt-hours per day; multiplying by 365 gives watt-hours per year; only then does the division by 1000 produce kilowatt-hours. Dividing before multiplying, or treating L as a yearly total while n is a daily count, produces answers wrong by three orders of magnitude.

The second error is subtler and more common: charging both states for the same hours. Sleeping and shutting down are mutually exclusive policies, not simultaneous loads. In any given away period the machine is either asleep or off, never both, so the two annual figures below are alternatives to be compared — never added, and never summed into a "total standby cost".

Policy A — always sleep. The machine holds P_s for the whole away period and resumes in a second or two, which is negligible:

Policy B — always shut down. The machine holds the lower P_o for the away period but must also pay one boot burst each time you come back. That burst is a genuine energy cost, and omitting it is what lets naive calculators overstate the saving:

The boot burst itself is just power times time again, with the time expressed in seconds and so divided by 3600 to reach hours:

Setting the two policy costs equal and solving for L gives the quantity users actually want — the away time at which shutting down starts to win. The rate r, the period count n and the 365 all cancel, which is why the break-even is a property of the hardware alone and not of your electricity bill:

The denominator is the whole story. If sleep and off-mode power are close, the gap is tiny, the break-even is long, and sleeping wins for anything short of a holiday. If your machine sleeps badly — the usual outcome on a Modern Standby laptop — the gap is wide, the break-even collapses to minutes, and shutting down wins almost immediately. When P_o ≥ P_s the denominator is zero or negative, no break-even exists, and the calculator says so rather than dividing by zero.

Grounding the inputs in the ENERGY STAR duty cycle

The EPA does not merely define the modes; it also publishes how long a typical notebook is assumed to spend in each. Equation 1 of the Version 8.0 specification computes typical energy consumption as a weighted average of the four mode powers over the 8760 hours in a year:

Table 5 of the specification gives the conventional notebook weightings as 25% off, 35% sleep, 10% long idle and 30% short idle. In other words the EPA's own model of a typical laptop year has the machine shut down for roughly a quarter of it and asleep for a third — so the choice this calculator addresses covers about 60% of a notebook's calendar. The specification's worked appendix example measures a dual-core notebook at 0.5 W off, 1.0 W sleep, 6.0 W long idle and 10.0 W short idle, giving 8.76 × (0.5×0.25 + 1.0×0.35 + 6.0×0.10 + 10.0×0.30) = 35.7 kWh per year. Those two low figures, 0.5 W and 1.0 W, are the defaults this page loads, and they represent a well-behaved machine using classic S3 sleep.

Many current laptops are not that machine. Windows systems increasingly use Modern Standby (the S0 low-power idle state) rather than S3, which the specification classifies as an Alternative Low Power Mode. Section 3.5.2 permits ALPM power to substitute for both sleep and long-idle power in the TEC equation only when the measured ALPM power is "less than or equal to 10 watts". A machine drawing 6 W asleep is therefore still certifiable — six times the sample notebook's figure — which is exactly why measured values beat assumed ones here.

Reading the output and deciding what to do

The result panel reports five things, and they are meant to be read in this order:

• The verdict. For the away period you entered, which policy costs less and by how much per year. This is the headline; everything else supports it.
• Break-even away time. The L_* derived above, expressed in hours and minutes. Away periods longer than this favour shutting down; shorter ones favour sleep. If no break-even exists — because off-mode power is not lower than sleep power — the panel says so explicitly rather than printing a meaningless number.
• Annual cost of each policy. Two alternatives, not two line items. They are never added together.
• Boot-burst energy. The watt-hours one start-up costs, and the annual total across every boot. Seeing this number is usually what ends the "booting uses more energy than sleeping all night" argument, because it is measured in fractions of a watt-hour.
• Annual energy in kWh. Useful if you would rather attach your own grid emissions factor than a price.

Two cautions about magnitude. First, per-laptop amounts are small — often well under a dollar a year — and a result in the tens of dollars usually means the sleep wattage entered belongs to an idle machine with the screen off rather than a genuinely sleeping one. Second, small per-device numbers still matter at scale: multiply by the number of machines before deciding the answer is negligible. A 300-seat organisation setting a sleep policy is making a three-hundred-fold version of this decision, and the same arithmetic drives the emissions figure as drives the invoice.

Worked example: an ENERGY STAR notebook left overnight

Take the sample notebook measured in the Version 8.0 specification appendix — 0.5 W in Off Mode and 1.0 W in Sleep Mode — left overnight for one 14-hour away period per day. Assume a 30 W average draw over a 40-second boot, and the EIA national residential average of $0.1844 per kWh.

1. Boot-burst energy.
   E_boot = 30 W × 40 s / 3600 s per h = 0.3333 Wh per boot.
2. Break-even away time.
   L* = 0.3333 Wh / (1.0 W − 0.5 W) = 0.667 h, that is 40 minutes. A 14-hour overnight is twenty-one times longer, so shutting down sits well past the crossover.
3. Annual energy, sleep policy.
   (1.0 W × 14 h × 1 × 365) / 1000 = 5.110 kWh per year.
4. Annual energy, shutdown policy.
   ((0.5 W × 14 h) + 0.3333 Wh) × 1 × 365 / 1000 = (7 + 0.3333) × 365 / 1000 = 2.677 kWh per year.
5. Annual cost of each policy.
   5.110 × 0.1844 = $0.94 for sleep; 2.677 × 0.1844 = $0.49 for shutdown.
6. Annual saving from shutting down.
   $0.94 − $0.49 = $0.45 per year. The 365 boots consume 0.122 kWh between them, roughly 2.4% of what the sleep policy would have spent.

Two conclusions fall out of this. The boot-energy objection is real but quantitatively trivial at overnight timescales — 0.33 Wh is about what a 1200 W hair dryer uses in one second. And the saving from shutting down an efficient laptop is also trivial: 45 cents a year. The honest answer for a well-behaved S3 machine is that the energy stakes are low in both directions and you should choose on convenience. The stakes only become interesting when sleep power is high, which is exactly what the next table covers.

Comparison table: how the break-even moves with sleep power

The break-even depends only on the boot burst and on the gap between sleep power and off power — the electricity rate cancels out of it entirely. Holding the boot burst at 0.3333 Wh (30 W for 40 s) and off-mode power at 0.5 W, and pricing a single 14-hour overnight period per day at $0.1844 per kWh, the picture changes sharply as sleep power climbs:

The shutdown column never moves, because the shutdown policy does not care how badly the machine sleeps. The sleep column scales linearly with sleep power, and the break-even falls as its reciprocal. That is the most useful thing on this page: the decision is driven almost entirely by one number you can measure in two minutes with a plug meter, and hardly at all by your electricity rate. If your laptop sleeps at 0.6 W, stop thinking about it. If it sleeps at 6 W, shut it down whenever you leave the desk for more than a coffee.

Measuring your own sleep, off-mode and boot power

Every number in the table above is only as good as the wattages behind it, and the honest position is that you should measure rather than assume. The measurement is not difficult. IEC 62301, Household electrical appliances — Measurement of standby power, is the international standard that defines how off-mode and standby power are measured: a stabilised supply, a settled reading, and averaging over a defined period rather than snapping a single instantaneous value. You do not need laboratory equipment to apply the same discipline at home.

• Plug meter, settled reading. Put a plug-in energy meter between the wall socket and the laptop's adapter. Sleep the machine, wait at least two minutes for background tasks to quiesce, then read the average watts — not the peak. Repeat with the machine shut down. Cheap meters resolve to 0.1 W at best and many read 0.0 W below about 3 W, so for off-mode power prefer a meter that accumulates kWh over several hours and divide.
• Accumulate for the small numbers. Off-mode draw of 0.3 W is below the resolution of most consumer meters. Leave the machine off overnight and read the accumulated watt-hours instead: 0.3 W for 12 hours registers as 3.6 Wh, which almost any meter can show.
• Boot power needs the same treatment. Watch the meter through a full cold boot to the desktop, note the rough average and the elapsed seconds. If your meter accumulates, zero it, boot, and read the watt-hours directly — that number goes straight into the boot fields without needing the power and duration split at all.
• Include everything on the same circuit. A dock, an external monitor in standby, a powered hub and an always-lit charger all keep drawing while the laptop sleeps. If they stay connected under both policies, measure the whole setup at the wall; the calculator treats whatever you enter as the total for that state.
• Beware the battery. A laptop that is charging draws far more than one that is asleep. Take readings only when the battery is already full, otherwise you will record charging current and conclude that sleep is catastrophically expensive.

Reference wattages and where the defaults come from

The page loads with the values from the appendix sample calculation in the ENERGY STAR Computers Version 8.0 specification — Off Mode 0.5 W, Sleep Mode 1.0 W — because those are published figures from a primary source rather than folklore. The remaining defaults are illustrative and are flagged as such:

• Sleep power, 1.0 W (sourced). The EPA sample notebook. Real classic-S3 machines cluster between roughly 0.5 W and 2 W. Modern Standby machines routinely sit far higher; ENERGY STAR will still certify one at up to 10 W in Alternative Low Power Mode.
• Off-mode power, 0.5 W (sourced). The EPA sample notebook again. It is not zero, and it cannot be, because Off Mode is defined as a mode the user cannot switch off. Unplugging the adapter is the only genuine zero.
• Boot power 30 W and boot duration 40 s (not sourced — illustrative). No standards body publishes a boot-energy allowance for notebooks, so this page does not pretend one exists. The 30 W figure is chosen to sit meaningfully above the specification's 10.0 W short-idle measurement for the same sample notebook, since a boot loads the CPU and storage far harder than an idle desktop. Substitute your own measurement; the break-even scales in direct proportion to it.
• Electricity rate $0.1844/kWh (sourced). The EIA national average residential price for May 2026. Your own marginal rate is better, and state averages span more than a factor of four.

Turning the break-even into a habit

Once you know your break-even away time, the policy writes itself. Compare it to the natural gaps in your day:

• Break-even under about 15 minutes. Your machine sleeps badly. Shut down whenever you leave the desk for a meeting, and certainly overnight. This is the common case on Modern Standby hardware with an active network.
• Break-even between roughly 15 minutes and 2 hours. Sleep through short breaks, shut down overnight and at weekends. This is the typical S3 laptop and the case where the calculator earns its keep.
• Break-even beyond a few hours. Your machine sleeps extremely well. Sleeping overnight costs pennies a year; shut down only for multi-day absences, and unplug the adapter then so off-mode draw goes to a genuine zero.
• No break-even at all. If measured off-mode power is not lower than sleep power — some machines keep charging circuitry or wake-on-LAN alive identically in both — shutting down cannot save energy, and the calculator reports that rather than inventing a crossover.

Modelling limitations you should know about

This is a two-state comparison with a boot correction, and it is deliberately simple. The following assumptions are baked in, and each of them can move the answer:

• The year is uniform. Every day is assumed to contain the same number of identical away periods, 365 times over. Weekends, holidays and travel all break that pattern. For a machine that is off for three weeks a year, run the calculator twice and weight the results.
• Power is constant within a state. Real sleep power is not flat. Modern Standby machines wake periodically to service mail and updates, so measured sleep power is an average over a duty cycle rather than a steady draw. Off-mode power varies with battery state of charge and with whether USB ports stay powered.
• Resume from sleep is treated as free. Waking from S3 takes a second or two, so its energy is far below the resolution of this model. If you are modelling hibernate — where resume involves a full read of the memory image from disk — put that burst into the boot fields instead of ignoring it.
• Boot energy is a single lump. Real start-up power ramps and falls, and post-login background work continues well past the desktop appearing. Entering an average power and a duration approximates the integral; entering accumulated watt-hours from a meter is more accurate.
• Everything is measured at the wall. The figures include adapter conversion losses, which is correct for a billing calculation, but they do not describe battery drain. A laptop asleep on battery is a different measurement with different numbers.
• Only electricity is priced. Component wear from power cycling, the minutes you spend waiting for a boot, and the value you place on picking up exactly where you left off are all real costs that this model does not attempt to quantify.
• One flat rate. Tiered tariffs, time-of-use pricing and demand charges are not modelled. Since almost all of the energy in question is consumed overnight, a time-of-use customer should enter their off-peak rate rather than a blended average.

Within those bounds the arithmetic is exact — power times time, divided into kilowatt-hours, priced at your rate — and the break-even is an algebraic identity rather than a heuristic. The uncertainty lives entirely in the inputs, which is why the measurement section above is longer than the formula section.

Frequently asked questions about laptop sleep and shutdown power

Sources. Mode definitions, the Table 5 notebook mode weightings (25% off, 35% sleep, 10% long idle, 30% short idle), the Equation 1 typical-energy-consumption formula, the 10 W Alternative Low Power Mode substitution limit and the sample notebook measurements of 0.5 W off and 1.0 W sleep are taken from the U.S. EPA, ENERGY STAR Program Requirements — Product Specification for Computers, Eligibility Criteria Version 8.0 (Rev. April-2020). The default electricity price of $0.1844 per kWh, and the Idaho and Hawaii extremes, are from the U.S. Energy Information Administration, Electric Power Monthly, Table 5.6.A — Average Price of Electricity to Ultimate Customers by End-Use Sector, May 2026 data released 23 July 2026. Standby and off-mode measurement practice follows IEC 62301, Household electrical appliances — Measurement of standby power. The boot power and boot duration defaults are illustrative only: no standards body publishes a notebook boot-energy allowance, so measure your own rather than relying on the placeholder.