A prebuilt PC that powers on, boots to Windows, and then shuts off abruptly after two to ten minutes is one of the most common failure patterns in desktop gaming systems. The symptom usually looks like this: the machine starts normally, games or benchmarks run fine at first, and then the screen goes black instantly — not a crash to desktop, not a blue screen, just a total power-off, as if someone yanked the cord. Sometimes the fans spin back up on their own and the system reboots into a “Windows was not shut down correctly” prompt. In other cases, the PC stays dead until you flip the power supply switch off and on again, or you have to unplug it for thirty seconds before it responds at all.
It is urgent, but do not panic. If your PC shuts off under load and reboots repeatedly, stop gaming on it and diagnose before you keep pushing it — repeated thermal or power cutoffs can shorten component lifespan, and in rare cases a failing power supply can damage other parts. The good news is that roughly eight in ten cases come down to thermal paste, dust, or a tripped protection circuit, all of which are fixable at home in under an hour. Work through the fixes below in order, since they run from a five-minute software check to a sixty-minute hardware teardown.
Likely Causes
| Cause | How to confirm | Fix difficulty |
|---|---|---|
| CPU or GPU overheating (dried paste, dust, dead fan) | Watch temperatures in HWiNFO64; shutdowns when CPU package hits 95–105°C or GPU hotspot exceeds ~105°C | Easy to Moderate |
| Power supply failing or underpowered (OCP/OPP tripping) | Shutdowns only under heavy load; PSU makes clicking sounds; replacements of another PSU stop the problem | Moderate |
| Unstable memory (XMP/EXPO profile) | Shutdowns happen even at idle or light use; errors in MemTest86 or Windows Memory Diagnostic | Easy |
| Loose or degraded power cables (24-pin, EPS 8-pin, PCIe) | Shutdowns when the case is bumped or moved; connectors feel slack when gently wiggled | Easy |
| Overheating VRMs or blocked case airflow | Case exhaust air is scorching; shutdowns cluster in warm rooms; motherboard VRM area is too hot to touch | Easy to Moderate |
| Faulty motherboard, front-panel short, or BIOS bug | Problem persists after PSU swap and repaste; shutdown happens even in BIOS with no OS loaded | Hard |
Fix 1: Disable XMP/EXPO Memory Overclocking
What causes it
Prebuilt builders frequently enable XMP (Intel) or EXPO (AMD) memory profiles to hit advertised RAM speeds. Some boards ship with marginally stable profiles, and instability can hard-freeze or power-cycle the system minutes into a session, especially after a BIOS update changes memory training.
How to check
Restart and enter BIOS (usually Delete or F2 during POST). Look for the memory speed on the main page — anything above JEDEC defaults (DDR5-4800/5600 or DDR4-2133/2666) means a profile is active. Also check Reliability Monitor (search “Reliability Monitor” in the Start menu); “Hardware error – WHEA” entries timed to the shutdowns point at memory or CPU instability.
What to do
In BIOS, find the overclocking or “Ai Tweaker”/”Extreme Tweaker”/”OC” menu, locate XMP or EXPO, and set it to Disabled or “Auto.” On AMD boards also set “Memory Context Restore” to Disabled while testing. Save with F10 and reboot. Use the PC for a full day of gaming; if shutdowns stop, you can later re-enable the profile and update the BIOS to the latest version before trying again.
How to undo it
Re-enter BIOS, reselect the XMP/EXPO profile, and save. Memory settings reset safely at any time; nothing is degraded by toggling them.
Fix 2: Clean Dust From Intakes, Heatsinks, and Fans
What causes it
Prebuilt systems often sit on carpet and run for years without service. Dust mats into the radiator fins, GPU heatsink, and intake filters, choking airflow so heat soaks up until a sensor trips the emergency shutdown.
How to check
Shut down, unplug the power cord, and hold the power button for ten seconds to drain residual charge. Pull the side panel and inspect the CPU cooler fins, GPU heatsink, and radiator. If you cannot see daylight through the fins, or fans are fuzzy with dust, this is your problem. Confirm with HWiNFO64: idle CPU package temps above ~60°C or a CPU cooler fan reporting 0 RPM are dead giveaways.
What to do
Take the case outside or to a garage. Use compressed air in short bursts, holding fans stationary with a finger so they do not overspin. Clean intake filters with warm water and dry them fully. Blow through the GPU and CPU heatsinks from the opposite side of the airflow until dust stops coming out. Reassemble and monitor temps in HWiNFO64 — idle temps should drop noticeably.
How to undo it
Not applicable — cleaning is strictly beneficial. The only “undo” is reattaching any panel or filter you removed.
Fix 3: Reseat All Power Cables
What causes it
Shipping and handling loosen the 24-pin ATX, the 8-pin EPS CPU cable, and PCIe GPU cables. A partially seated connector increases resistance, heats up, and trips the power supply’s protection under load — often minutes into a game.
How to check
Unplug the PC, hold the power button for ten seconds, then open the case. Firmly press the 24-pin motherboard connector, the 8-pin EPS at the top-left of the board, and both ends of every GPU power cable. Wiggle gently: any connector that shifts visibly was loose. Also check that the PSU’s own rocker switch is fully on and the voltage selector (if present) matches your region.
What to do
Unplug each cable and reseat it until it clicks fully home. If your GPU uses a daisy-chain pigtail and the card has two or more 8-pin inputs, run a separate cable to each input rather than chaining. Route cables away from sharp case edges. Reassemble and test with a demanding game for thirty minutes.
How to undo it
Return cables to their previous routing if you changed it, but reseating itself needs no reversal.
Fix 4: Fix Case Airflow and Cooling Curves
What causes it
Many prebuilts ship with one exhaust fan, intake fans set to quiet curves, or fans mounted to pull air the wrong direction. Heat accumulates in the case, and VRMs or the GPU eventually hit a protection threshold and cut power.
How to check
Run a demanding game for ten minutes with the side panel off and a house fan blowing in. If shutdowns stop with the panel off, airflow is your bottleneck. Check fan RPM in HWiNFO64 — case fans idling under 400 RPM or pinned at a fixed low speed suggest flat BIOS fan curves.
What to do
Enter BIOS and open the fan control section (often Q-Fan Control, Smart Fan, or Fan Control depending on the brand). Set CPU and case fans to a standard curve reaching 100% duty by about 75°C. Confirm intake fans are at the front/bottom blowing in and exhaust fans at the rear/top blowing out. Add an intake or exhaust fan if a mounting position is empty. Keep the PC off carpet and give the rear exhaust at least 4 inches of clearance.
How to undo it
In the BIOS fan menu, select “Restore Defaults” or set the curve back to its previous shape; fan changes are reversible at will.
Fix 5: Repaste the CPU
What causes it
Factory thermal paste on prebuilts is often a low-grade compound applied thinly, and it dries out within a couple of years. Once it cracks, the CPU spikes past its thermal trip point (commonly 95–105°C depending on the platform) and the board cuts power instantly.
How to check
Log temperatures with HWiNFO64 during a game session. If CPU package temperature jumps from 60°C to 98°C+ within seconds of load starting, or the system shuts off while “CPU (Tctl/Tdie)” or “CPU Package” reads near its max, the cooler-to-CPU interface has failed. Also confirm the cooler pump/fan spins at all.
What to do
You need isopropyl alcohol (90%+), lint-free cloths or coffee filters, and a quality paste. Unplug everything, remove the side panel, and disconnect the cooler. If it is an air tower, remove it from the board; if it is an all-in-one liquid cooler, you can usually repaste with the radiator still mounted. Clean the old paste completely from both the CPU heat spreader and the cooler cold plate. Apply a pea-sized dot (about 4–5 mm) to the center of the CPU, remount the cooler, and tighten screws in a diagonal pattern until snug — not cranked. Reconnect the cooler fan/pump to CPU_FAN/PUMP. If the cooler was an AIO, verify its pump header reads a constant RPM in HWiNFO64 afterward.
How to undo it
Repasting is reversible only in the sense that you can redo it — there is nothing to restore. If temps get worse, the cooler was not mounted evenly; clean and remount it.
Fix 6: Update BIOS and Chipset Drivers
What causes it
Prebuilt vendors sometimes ship boards on launch BIOS versions with known power-delivery or memory-training bugs fixed in later updates. A buggy AGESA or microcode can cause sporadic power cycles under specific load patterns.
How to check
Press Win+R, type msinfo32, and note the “BIOS Version/Date” line. Compare it against the latest BIOS listed for your exact motherboard model on the manufacturer’s support page (you can identify the board model in msinfo32 under “BaseBoard Product”). Also open Device Manager or your vendor’s utility to check the chipset driver version.
What to do
Download the latest non-beta BIOS from the board manufacturer, put the file on a FAT32 USB stick, and use the BIOS’s built-in flasher (EZ Flash on ASUS, Q-Flash on Gigabyte, M-Flash on MSI, Instant Flash on ASRock). Do not power off during the flash. Afterwards, re-enter BIOS and select “Load Optimized Defaults,” then re-enable XMP if Fix 1 showed it was safe. Install the current chipset driver from AMD or Intel’s support pages.
How to undo it
Most boards let you flash an older BIOS version using the same USB method, though downgrading is sometimes blocked. Defaults can be reloaded from the BIOS at any time.
Fix 7: Test With a Different Power Supply
What causes it
Budget prebuilt PSUs are the single most common genuine hardware culprit. Capacitors age, protection circuits become hair-trigger, and the unit cuts out the moment transient spikes exceed its real (not rated) capacity — typically a few minutes into a heavy game.
How to check
Symptoms that point at the PSU: shutdowns only under GPU-heavy load, clicking or buzzing from the PSU, a burning smell, or a system that will not power on again without being unplugged for a minute. There is no safe home test of PSU health beyond swapping it — software readings are unreliable.
What to do
Borrow or buy a quality replacement sized at roughly 100–150 W above your system’s realistic draw. Swap the unit: connect the 24-pin, 8-pin EPS, and GPU cables from the new PSU, leaving the old unit’s cables entirely disconnected (do not mix modular cables between units). If shutdowns stop, the old PSU was the fault and should be retired, not reused.
How to undo it
Swap the original unit back in if you need to return the loaner — five cable connections, fully reversible.
Fix 8: Rule Out the Motherboard and Front-Panel Wiring
What causes it
A shorted front-panel connector, a standoff misaligned under the board, or a failing motherboard voltage regulator can mimic every other cause on this list. This is the diagnosis of last resort.
How to check
Enter BIOS and let the PC sit on that screen for twenty minutes without touching anything. If it still shuts off with no OS and no load, the problem is hardware-level, not Windows. Inspect the bottom of the case for an unused brass standoff touching the board, and check that the front-panel header wires (power SW, reset SW) are on the correct pins per the board’s manual and not frayed.
What to do
Remove any stray standoff, reseat or replace the front-panel connectors, and try powering the board using only the bare minimum: one RAM stick in the second slot from the CPU, GPU installed, no extra drives or USB devices. If the system is stable in that configuration, reconnect parts one at a time until the fault reappears. If it shuts down even minimal, the board itself is likely failing — contact the prebuilt vendor’s warranty support, since replacing a motherboard on a proprietary prebuilt can void coverage.
How to undo it
Reinstall any components you removed and reconnect front-panel wiring as documented in the motherboard manual.
If Nothing Works
If the machine still cuts out after every fix above, stop swapping parts and use the warranty. Most prebuilt vendors include one to three years of coverage, and by this point you have documented evidence — HWiNFO64 logs, a PSU swap, a repaste — that the fault is not user error. Open a support ticket with the vendor, describe the exact shutdown pattern (minutes into load, instant power loss, no BSOD), and list the steps you have already taken. If the warranty has expired, a local repair shop can bench-test the board and PSU with proper equipment for far less than the cost of replacing components blindly. Do not keep gaming on a machine that shuts off under load; repeated hard cutoffs risk damaging the drive and motherboard.
FAQ
Why does my PC shut off after exactly the same amount of time?
A consistent timer points strongly at a thermal threshold: heat builds at a predictable rate until a sensor trips. The more repeatable the interval, the more likely the cause is cooling (paste, dust, dead fan) rather than random power faults.
Can overheating really shut a PC off instantly with no warning?
Yes. Emergency thermal protection is deliberately instant — the board cuts power the moment a critical sensor crosses its limit, bypassing Windows entirely. That is why you see a black screen instead of an error message.
Should I disable overclocking permanently?
Not necessarily. Disable XMP/EXPO first to confirm it is the trigger. If shutdowns stop, update the BIOS, re-enable the profile, and test again. Many instability cases disappear after a memory-training update.
How do I know if it is the power supply or the motherboard?
The cheapest discriminator is a PSU swap, since Fix 7 takes twenty minutes versus hours of board troubleshooting. If a known-good PSU fixes it, you are done. If it persists through a PSU swap, thermal service, and a BIOS update, assume the board and pursue warranty replacement.
