A PC that reboots itself mid-game with no blue screen is one of the more alarming faults — and one of the most methodical to diagnose. A sudden restart (versus a BSOD) usually points to power, heat, or a hardware protection cutting off instantly. This guide walks through the causes from most to least common so you can isolate it without swapping parts blindly.
Restart vs blue screen
A blue screen means Windows caught an error and chose to stop; a hard, instant restart means something cut power before Windows could react — typically an overheating shutdown, a PSU that can’t hold load spikes, or an unstable overclock. First, disable “automatically restart” under System Failure settings so any hidden BSOD stays on screen and gives you a code to work with.
Cause checklist
| Cause | Clue | Test |
|---|---|---|
| Overheating | Restarts under load | Monitor CPU/GPU temps |
| Weak/failing PSU | Restarts on GPU spikes | Try known-good PSU |
| Unstable overclock/XMP | Started after tuning | Reset to stock, test |
| Loose power cable | Random, movement-related | Reseat 24-pin, EPS, PCIe |
Check temperatures first
Run HWiNFO and play until it restarts, watching CPU and GPU temps. A modern CPU throttles in the 90s °C and can shut down near 100-105 °C; GPUs cut off around 100-110 °C hotspot. If temps are climbing to those numbers before the reboot, it is thermal — reseat the cooler, repaste, improve case airflow, or clean dust filters. Thermal shutdowns are protective, not the disease itself.
Suspect the power supply
If temps look fine, the PSU is the prime suspect, especially an older or low-wattage unit under a modern GPU. Big GPUs draw brief power spikes far above their rated average; a marginal PSU trips its protection and cuts power instantly. Make sure the PSU comfortably exceeds your system’s needs (a quality unit with real headroom, e.g. 750W+ for a mid-high build) and that the 12VHPWR/PCIe and EPS cables are fully seated.
Roll back overclocks and test
Reset any CPU/GPU overclock and disable XMP/EXPO to run RAM at stock, then test. Unstable memory or an aggressive overclock frequently causes instant reboots under specific loads. If stock is stable, reintroduce one tweak at a time. Also test RAM with MemTest86 and run the PC on a single stick to rule out a bad module.
Work the process of elimination
Random reboots reward patience. Change one variable, then test long enough to be confident — many faults only appear after the system has been under load and hot for a while. Start by exposing hidden blue screens (disable automatic restart), then rule out heat with a temperature log, then power by checking the PSU and reseating every power cable, then instability by resetting overclocks and memory profiles to stock.
Do not overlook the simple stuff
Before assuming a dead component, check the boring causes: a wall outlet or surge strip that cannot hold load, a loose 8-pin EPS or PCIe connector, dust choking the cooler, or a BIOS power setting gone wrong after tuning. A cheap outlet tester and a firm reseat of every cable resolve a meaningful share of “random restart” cases. If you have another known-good PSU, swapping it in is the single most decisive test for load-related reboots.
Keep notes on what you have already ruled out. Intermittent faults are frustrating precisely because they hide, and a written elimination log stops you from re-testing the same thing twice and helps anyone you ask for support zero in faster.
FAQ
It only restarts in demanding games — what does that mean?
Load-dependent restarts point strongly at power delivery or cooling, since those are the moments of peak current and heat. Check temps first, then the PSU and its cables before anything else.
Could a Windows or driver issue cause a hard restart?
Usually those cause a blue screen, not an instant power cut. Disable automatic restart so a BSOD stays visible; if you get a code, it is software or a driver. A truly instant reboot with no code is almost always hardware, power, or heat.
Bottom line: A randomly restarting gaming PC is a process of elimination: disable auto-restart to expose hidden BSODs, check temps with HWiNFO, verify a healthy PSU with fully seated cables, and reset overclocks and XMP. Work load-related causes (heat and power) first — they explain the large majority of no-warning reboots.
