PC keeps restarting randomly - causes and fixes

No blue screen. No error. The screen goes black and the PC boots itself back up like nothing happened. Random restarts are frustrating precisely because they leave so little evidence - but the pattern of when they happen tells you a lot about why.

First: turn off automatic restart

Windows hides crash information by default. It's configured to restart instantly when it hits a critical error, which means what looks like a random restart may actually be a blue screen you're never seeing.

Go to System > Advanced system settings > Startup and Recovery > Settings and untick Automatically restart. Now, if Windows crashes, you'll see the blue screen and its stop code instead of an instant reboot. That stop code turns a mystery into a searchable fault.

If you make this change and the restarts continue with no blue screen at all, that's significant: it points away from Windows and towards hardware - most likely power or temperature.

Check the pattern

Restarts under load (gaming, rendering, heavy multitasking) point to power delivery or temperature. The system pulls maximum power and generates maximum heat exactly when these restarts happen.

Restarts at idle or random times are more likely RAM, drivers, or a failing motherboard component. Pure idle restarts are also a classic sign of marginal RAM that fails on background tasks.

Restarts at cold boot or first use of the day suggest PSU capacitor degradation - aged capacitors perform worst when cold.

Restarts that started after a hardware change - new GPU, new RAM, moved house, new room - point at whatever changed. A new GPU on an old PSU is the most common version of this story.

Check temperatures

Overheating protection triggers hard shutdowns and restarts with no warning. Run HWiNFO64 and watch CPU and GPU temperatures under the workload that normally causes a restart. A CPU spiking past 95°C or a GPU past its limit will cut out to protect itself.

If temperatures are the problem, the fix is usually cleaning, fan curves, or thermal paste - see our temperature guide for the full thresholds and what to do about them.

Test the RAM

Faulty RAM causes restarts that follow no pattern at all. Run MemTest86 from a USB stick overnight - any errors mean a stick needs replacing. With two sticks, test each individually to find the culprit. Also worth trying: disable XMP/EXPO in the BIOS temporarily. If the restarts stop with RAM at default speeds, the profile is unstable on your board - common with budget motherboards and four-stick configurations.

Check Event Viewer

Open Event Viewer > Windows Logs > System and look for events around the times of the restarts. Event ID 41 (Kernel-Power) confirms the system lost power without shutting down cleanly - that's the hardware-restart signature. A WHEA-Logger error just before it often identifies the failing component: CPU, cache, or PCIe device.

No events at all before the restart, just the Kernel-Power entry? That's another arrow pointing at the PSU - the system lost power too quickly to log anything.

The PSU question

If you've ruled out temperature and RAM, and the restarts happen under load or with no blue screen ever appearing, the power supply is the leading suspect. A degraded PSU causes exactly this fault and is nearly impossible to test conclusively at home - it'll pass idle checks and fail only at the load levels that trigger the restarts. Our PSU failure symptoms guide covers this in detail.

The substitution test - a known-good PSU of adequate wattage - is the definitive home check if you have access to one.

Less common causes worth knowing

• Failing VRM on the motherboard - the power delivery between PSU and CPU. Restarts under CPU load specifically, sometimes with VRM temperatures spiking in HWiNFO
• Front panel power button fault - a sticking or shorting power switch can trigger resets. Disconnect the case's reset switch header as a test; they fail more often than people expect
• GPU power connector issues - a partially seated or damaged PCIe power cable causes restarts the moment the card draws full power
• Surge-damaged components - if the restarts began after a storm or power cut, the PSU or motherboard may have taken damage that only shows under stress

When to stop and send it in

If you've worked through temperatures, RAM, Event Viewer, and the software checks, what remains - PSU, VRM, motherboard - needs bench equipment to separate. We can load-test the PSU properly, monitor power delivery at the component level, and catch the fault in the act rather than guessing from symptoms. Random restarts are almost always diagnosable; the difficult part is doing it without the right tools.