Windows 11 Keeps Restarting? Find the Cause

Right-click Start, System, scroll to About, Advanced system settings, click Settings under Startup and Recovery, uncheck “Automatically restart” under System failure. Do this first, before anything else. Windows 11 hides blue screens by default — when a BSOD happens, it flashes the error for half a second and restarts so fast you never see the code. You think “my computer randomly restarted” when what actually happened is a crash with a specific, identifiable cause. With that box unchecked, the next crash stays on screen until you manually restart, and the error code on that blue screen is worth more than every generic troubleshooting step combined.

A software engineer’s desktop started rebooting mid-compile two or three times a day — no warning, no blue screen, no error message. He’d already replaced his RAM, reseated his GPU, bought a new power supply. Three hundred dollars in parts. We pulled up Event Viewer on a remote session. Kernel-Power Event ID 41, BugcheckCode 0x00000124 — WHEA_UNCORRECTABLE_ERROR, a hardware error reported by the CPU itself. Turned out his Ryzen 7 5800X had a manual overclock set a year ago and forgotten. The chip had degraded just enough that it couldn’t sustain that voltage anymore. Reset BIOS to defaults, problem gone. He’d spent $300 because he never checked Event Viewer.

About 40% of random restart cases we diagnose remotely are GPU driver crashes. Another 25% are Windows Update related. Maybe 15% are genuine hardware — overheating, failing RAM, PSU that can’t hold load. The remaining 20% scatters across corrupted system files, third-party software conflicts, and power plan issues.

Event Viewer

Win+R, type eventvwr.msc, Enter. Expand Windows Logs, click System. Look for the most recent Kernel-Power Event ID 41 marked Critical. Double-click it. BugcheckCode is the field that matters — if it’s 0, the power just cut physically (PSU failed, power strip tripped, overheating thermal cutoff). Anything other than 0 means Windows crashed and you can Google the hex code for the specific BSOD.

The event right before Kernel-Power 41 is often the real clue. A BugCheck event gives you the actual stop code and faulting driver. A WHEA-Logger event means the hardware itself reported a fault. And if there’s nothing — normal operation logs then abrupt silence — that pattern almost always means the power got yanked. PSU dying, thermal shutdown, or the outlet lost power.

If the blue screen names a .sys file — nvlddmkm.sys is NVIDIA’s display driver, atikmdag.sys or atikmpag.sys is AMD, igdkmd64.sys is Intel integrated, rtkvhd64.sys is Realtek audio — that tells you exactly what to roll back. Our BSOD guide covers the full crash dump analysis with BlueScreenView. If the screen goes black and stays black instead of restarting — fans running but no display — that’s a black screen problem rather than a restart loop.

GPU Drivers

GPU driver crashes are the number one software cause of random restarts, and it’s not close. NVIDIA’s been the worst offender lately — driver 555.85 caused widespread restart issues on RTX 40-series, and the 560.xx branch introduced TDR failures on some RTX 3060 and 3070 models. The dead giveaway beyond Event Viewer is if restarts happen during gaming, video playback, or anything that loads the GPU.

Check your driver version in Device Manager, Display adapters, right-click GPU, Properties, Driver tab. Go to nvidia.com/drivers and download one version back from what you have. During installation, check “Perform a clean installation.” For a deeper clean, download DDU (Display Driver Uninstaller), boot into Safe Mode, let DDU strip everything, then install fresh.

AMD had a nastier problem because it didn’t always look like a GPU issue. The Adrenalin 24.x drivers on RDNA 3 cards crashed specifically during hardware-accelerated video in Chrome and Edge — you’d be watching YouTube, screen goes black, system reboots, and you’d blame your internet. Quick workaround was disabling hardware acceleration in Chrome settings. Real fix was rolling back to Adrenalin 23.12.1 or updating past 24.7.1.

If restarts happen during web browsing, document editing, or sitting idle — not GPU-heavy tasks — the GPU probably isn’t the issue.

Windows Update

Windows Update causes restart loops two ways. Either the update itself is buggy and crashes the system on every boot — infinite loop, never reach the desktop — or it installed a broken driver that lets you boot but crashes during normal use.

For the infinite loop: force-shutdown three times in a row (hold the power button until the PC turns off, turn it back on, repeat). On the third failed boot, Windows enters Recovery. Troubleshoot, Advanced options, Uninstall updates, “Uninstall latest quality update.” This removes the last patch without needing to log in.

For restarts during normal use after an update, check what installed recently — Settings, Windows Update, Update history. Correlate dates. Specific bad KBs: KB5039212 caused boot loops on systems with Docker or WSL2 with Hyper-V. KB5040442 triggered BitLocker recovery prompts and restart loops on TPM systems. KB5034765 broke camera and audio drivers on several Lenovo and Dell models. Uninstall the suspect KB from the same screen, then grab Microsoft’s “Show or Hide Updates” tool (wushowhide.diagcab) to block it from reinstalling. Our Windows Update guide covers the full rollback procedure.

One thing that isn’t really “random” at all: Windows Update forcibly restarts to finish installing patches, and picks a time it thinks you’re away. If restarts happen at roughly consistent times — 3 AM, or whenever you step away — that’s not a crash. Settings, Windows Update, Advanced options, set your actual Active hours.

Power and Hardware

Fast Startup is under Control Panel, Power Options, “Choose what the power buttons do,” uncheck “Turn on fast startup.” It hibernates the kernel instead of doing a full shutdown, which means drivers that failed on one boot carry their broken state into the next boot. If restarts happen within the first 10-20 minutes after booting, this is worth disabling. Our shutdown guide covers the full diagnostic. If the machine restarts specifically when coming out of sleep, that’s a sleep mode issue — hybrid sleep or a GPU driver failing during resume.

If software fixes don’t stick — restarts happen even in Safe Mode, or the pattern is completely random with no correlation to what you’re doing — something physical is failing. CPU overheating causes immediate thermal shutdowns with no BSOD at all, and if your machine is more than three years old the thermal paste has probably dried out. PSU failures show up as restarts under heavy GPU load with temperatures looking fine. RAM issues usually come with blue screens — IRQL_NOT_LESS_OR_EQUAL, MEMORY_MANAGEMENT — run Windows Memory Diagnostic first, then MemTest86 from a USB drive for at least four passes if the built-in test comes back clean.

For system file corruption, run DISM /Online /Cleanup-Image /RestoreHealth then sfc /scannow from an admin Command Prompt. If your computer won’t turn on at all — no lights, no fans — that’s a different problem, and if you’ve been through everything here and still restart randomly, we can read Event Viewer crash dumps remotely and identify the exact faulting driver or hardware fault in about fifteen minutes.