As an Amazon Associate we earn from qualifying purchases.
A gaming PC with a 1000 W power supply does not draw 1000 W. It draws what its parts demand at that moment, and that number moves constantly: low at the desktop, high in a demanding game, and briefly far higher still during microsecond spikes. This article is about what a gaming PC actually draws, measured at the wall, tier by tier. It is not a sizing guide. If you are still choosing a power supply, see our PSU sizing guide for the headroom math.
The Three Numbers Behind One Wattage Figure
Three different numbers get called the wattage of a gaming PC, and confusing them is the source of most arguments about power draw. The first is the number on the power supply box, the maximum the unit can deliver, not what it delivers. The second is the direct current load inside the case: graphics card, CPU, motherboard, memory, drives and fans combined. The third is the alternating current draw at the wall, which is what a plug-in meter shows. The wall figure is always higher than the internal figure, because the power supply loses some energy as heat while converting wall power to the direct current the parts use.
Board power is the most useful single number for a graphics card. NVIDIA publishes total graphics power for its reference designs: 115 W for an RTX 4060, 200 W for an RTX 4070, 220 W for an RTX 4070 Super, 285 W for an RTX 4070 Ti, 320 W for an RTX 4080 and 450 W for an RTX 4090. These are specifications, not measurements: a card that is not the bottleneck in a given game draws less. The whole machine adds a CPU, a motherboard, memory, storage and fans on top of the card, and then power supply efficiency divides the total again. That chain is why a 200 W graphics card sits in a machine that draws roughly 356 W from the wall while gaming, a figure compiled by appliancecostcalculator.net.
Idle, Sleep and Light Use
A gaming PC spends most of its day doing very little, and that is where its draw is smallest. At the desktop, modern processors and graphics cards drop into deep power-saving states, and the whole tower typically draws between 45 and 75 W, per pcbottleneckcalculators.com. The same site puts the spread by tier at 50 to 80 W for a budget build, 70 to 100 W for a mid-range build and 90 to 120 W for a high-end build, as published by howmanywatts.org. That is an 80 to 85 percent drop from gaming load.
The graphics card is a small part of the idle figure: published idle draw for current NVIDIA cards runs from about 10 W on an RTX 4060 to about 25 W on an RTX 4090, per techsearchers.com. Sleep is lower still, at roughly 4 W, according to appliancecostcalculator.net, so a machine left on overnight at the desktop draws far more than one put to sleep.
Gaming Load by Build Tier
In gaming the graphics card takes over, accounting for roughly 40 to 60 percent of total system draw, according to howmanywatts.org. The table below gives whole-PC wall figures compiled from published board power and wall-meter data by appliancecostcalculator.net. Treat these as typical figures for the named combinations.
| Build tier | Typical combination | Whole PC at the wall while gaming |
|---|---|---|
| Budget 1080p | RTX 4060 with a Ryzen 5 class CPU | Around 261 W |
| Mid-range 1440p | RTX 4070 with a mid-range CPU | Around 356 W |
| Upper mid-range | RTX 4070 Ti with a mid-range CPU | Around 450 W |
| High-end 4K | RTX 4080 with a high-end CPU | Around 489 W |
| Flagship 4K | RTX 4090 with a flagship CPU | 600 to 800 W |
The pattern is clear: moving up a graphics card class roughly doubles the wall figure from budget to flagship, while the CPU and the rest of the system add a fairly steady 120 to 150 W on top. Both ends sit far below the wattage of the power supply usually installed, which is deliberate: a 750 W unit in a machine that draws 356 W while gaming runs at under half load, close to its efficiency sweet spot. Resolution and frame rate move the gaming figure within a tier: a card that is not the bottleneck draws less, and a frame-rate cap can cut draw sharply, because the card stops working the moment it has produced the frames the monitor needs.
Where the Watts Go
Inside the tower, the split is lopsided. The graphics card takes 40 to 60 percent of the total, the CPU takes 15 to 25 percent, and the motherboard, memory, storage, fans and connected devices share the rest, according to howmanywatts.org. The CPU share is smaller than many builders expect, because games rarely load every core at once. A Ryzen 5 7600 typically draws between 55 and 70 W while gaming, according to measurements reported by Gamers Nexus and TechPowerUp and cited in this site’s own CPU comparison. An Intel Core i9-14900K can draw up to 253 W under full load, per its published maximum turbo power, though gaming rarely reaches that ceiling.
The rest of the system is a steady background load of roughly 50 to 100 W for the motherboard, memory, drives and fans, with higher-end boards, more drives and more fans at the top of that range. A monitor is not part of the tower figure: a typical monitor draws 30 to 50 W of its own from a separate wall plug. None of these change much between idle and gaming, so the gap between the two states is almost entirely the graphics card.
Transient Spikes, the Hidden Part of the Draw
Average draw is not the whole story. Modern graphics cards do not draw power at a steady rate. When the GPU scheduler loads a new shader workload or the driver submits a heavy batch of draw calls, the card can demand far more power for a few microseconds than its board power suggests. These spikes are invisible to software monitors, which typically sample every 100 milliseconds, and to a plug-in wall meter, which averages over time. They are the reason a PC can shut down even when its average draw is far below the power supply size.
The ATX 3.0 specification, released by Intel in March 2022, exists largely because of these spikes. It requires a power supply to handle 200 percent of its continuous rated power for 100 microseconds, and 300 percent at the graphics card connector for the same duration, according to graphicscardhub.com. The standard also sets longer excursion limits, per sohoopc.com. Older ATX 2.x units were only required to support about 120 percent for brief bursts.
Measured spikes are large. Tom’s Hardware measured transient spikes up to 550 W on RTX 4090 systems, as cited by ofzenandcomputing.com. Evetech documents RTX 4090 transients of 600 to 650 W and RTX 5090 transients of 700 to 750 W, against board power figures of 450 W and 575 W. A spike adds almost nothing to total energy use, but it can trip the over-current protection of a power supply not built to absorb it, which is why an ATX 3.0 or 3.1 unit with a native 16-pin cable is the cleanest match for a flagship card.
Efficiency Curves, or Why the Wall Figure Is Higher
The gap between the direct current load inside the case and the alternating current draw at the wall is the power supply efficiency, and efficiency is not a single number. It changes with load, following a curve that peaks near 50 percent load and falls away at both ends, especially at very low load. The 80 PLUS program publishes minimum efficiency at set load points for 115 V input. Bronze must reach 82 percent at 20 percent load, 85 percent at 50 percent load and 82 percent at full load. Gold must reach 87, 90 and 87 percent at the same points. Platinum must reach 90, 92 and 89 percent. Titanium adds a 90 percent minimum at 10 percent load.
The arithmetic is simple: wall draw equals internal load divided by efficiency. A machine whose parts demand 350 W of direct current draws about 389 W from the wall through a Gold unit running at 90 percent efficiency, a figure pcbottleneckcalculators.com gives as approximately 388 W. The same 350 W load through a Bronze unit at 85 percent efficiency draws about 412 W. The difference is heat lost inside the power supply. Load position matters too: at the desktop the same unit runs at 6 percent load, far below the published efficiency points.
How to Measure Your Own PC
The only way to know your own machine is to measure it. A plug-in wall meter is the direct tool: it sits between the wall outlet and the power cord and shows alternating current draw in real time. Measure three states: the desktop at idle after the machine has settled, a demanding game running uncapped for at least ten minutes, and a synthetic stress test if you want the ceiling, keeping in mind that a stress test loads CPU and graphics card together in a way games rarely do.
Software monitoring answers a different question. Tools that read component sensors show the direct current draw of the graphics card and CPU, not the wall figure. To convert, add the rest of the system, roughly 50 to 100 W, and divide by the power supply efficiency at that load. Software also samples slowly, so it misses the microsecond spikes. Use both: software for the split between parts, and the wall meter for the true total.
What the Draw Means Day to Day
Watts times hours gives energy use. A mid-range machine drawing 356 W while gaming uses about 1.42 kilowatt-hours in four hours; idling at 100 W for four hours uses 0.4 kilowatt-hours, and sleeping at 4 W for sixteen hours uses about 0.06. The daily total is roughly 1.9 kilowatt-hours, with gaming at 73 percent, idle at 24 percent and sleep at 3 percent, per appliancecostcalculator.net.
Usage patterns matter more than the number on the power supply box. Cutting idle time, putting the machine to sleep instead of leaving it at the desktop, and capping frame rates where the extra frames add nothing visible all move the daily figure more than swapping a component within the same tier.
FAQ
How much wattage does a gaming PC use at idle?
A typical gaming PC draws 50 to 120 W at the desktop, depending on tier: budget builds sit at 50 to 80 W, mid-range builds at 70 to 100 W and high-end builds at 90 to 120 W, according to howmanywatts.org. Sleep drops the figure to roughly 4 W, and most of the idle draw is the motherboard, memory, drives, fans and connected devices.
Does a bigger power supply make a PC draw more power?
No. A power supply delivers only what the parts demand, so a 1000 W unit in a machine that needs 350 W draws about the same as a 550 W unit. The difference is efficiency at that load: a larger unit runs further below its efficiency peak at idle, so it can waste slightly more energy at the desktop, but the gaming figure is essentially unchanged.
Why does my PC shut down when average draw is far below the power supply size?
Transient spikes are the usual cause. A graphics card can demand two to three times its board power for microseconds, and if the power supply cannot absorb that spike, its over-current protection trips and the machine shuts down. The ATX 3.0 specification requires units to handle 200 percent of rated power for 100 microseconds precisely to cover these spikes. An older unit can trip even when the average draw is far below its size.
How can I measure my own PC power draw accurately?
Use a plug-in wall meter between the outlet and the power cord, and record the figure at idle, during at least ten minutes of uncapped gaming, and under a stress test if you want the ceiling. Software monitoring shows component draw, not the wall figure, and it samples too slowly to catch transient spikes. Combine both: software for the split between parts, and the wall meter for the true total.
Conclusion
A gaming PC draws far less than its power supply size most of the time: roughly 50 to 120 W at idle, 260 to 490 W while gaming for budget to high-end builds, and 600 to 800 W for flagship builds, with microsecond spikes above those averages. The number on the box is a ceiling, board power is a specification, and the wall meter is the truth.






