After Building My Gaming PC It Won’t Restart: How to Fix It

If after building my gaming PC it won’t let me restart it and I have to shut it down and turn it on manually, the most likely causes are an incorrect BIOS power setting, unstable memory training, a front-panel wiring mistake, or a power-delivery problem.

A PC that starts normally after a complete shutdown but fails when you select Restart is not necessarily dead. A restart keeps standby power active and asks the motherboard to reset the CPU, memory, and devices without fully removing power. That makes restart failures especially useful clues: the system may pass a cold boot but fail during reset or memory reinitialization.

1. Confirm exactly what “restart failure” means

Before changing hardware, identify the behavior:

• Windows shuts down but the screen stays black: suspect memory training, graphics initialization, BIOS settings, or a graphics-driver problem.
• The PC instantly powers off and does not come back on: check the power supply, motherboard power cables, and short circuits.
• The fans spin but there is no display: check RAM seating, GPU initialization, monitor input, and motherboard debug LEDs.
• The case power button is the only way to recover: avoid holding the button repeatedly; first determine whether the board is hanging during POST.
• Restart works from BIOS but not Windows: focus on Windows drivers, Fast Startup, chipset software, and third-party utilities.

Try one controlled test: enter the BIOS, save no changes, and choose its restart option. If BIOS restart also fails, the problem is below Windows. If BIOS restarts correctly but Windows does not, skip ahead to the driver and operating-system checks.

2. Return BIOS to a known-good baseline

A newly built PC often fails restart because memory overclocking is enabled before the platform is stable. XMP, EXPO, manual memory timings, undervolting, and CPU boost adjustments can all affect warm resets.

Use these BIOS changes first

• Load Optimized Defaults or Load UEFI Defaults.
• Disable XMP, EXPO, memory overclocking, CPU undervolting, and manual power-limit changes.
• Set boot mode to UEFI, not Legacy or CSM, unless an older operating-system installation specifically requires it.
• Temporarily disable Fast Boot in the motherboard firmware.
• Enable Memory Context Restore only after the system is stable at default memory speed. If it is already enabled, disable it for testing.
• Leave PCIe link speed on Auto. If a graphics card repeatedly fails to initialize after restart, test the primary slot at Gen 3 or Gen 4, depending on the board and card.
• Save, shut down completely, wait 30 seconds, then power on and test Restart several times.

Memory Context Restore can shorten later boots by reusing training data, but an unstable memory configuration may fail specifically on warm reboot. Disabling it forces more complete training and is a useful diagnostic, not necessarily a permanent performance setting.

Update the motherboard firmware carefully

Check the support page for the exact motherboard revision and install a stable BIOS version using the board’s built-in flash utility. A newer BIOS can improve DDR5 compatibility, USB initialization, CPU microcode, and restart behavior. Do not interrupt power during flashing, and do not use a BIOS file intended for a different model or revision.

If the board has a failed-boot recovery button or debug display, note the code shown immediately after a failed restart. Codes associated with DRAM, VGA, or CPU are more useful than guessing from fan noise.

3. Check the front-panel wiring

Front-panel connectors do not usually cause a failed software restart directly, but incorrect wiring can make shutdown and power-button behavior confusing. With the PC switched off and unplugged, compare each connector with the motherboard manual.

• Connect the case’s Power SW to the motherboard’s power-switch pins.
• Connect Reset SW only to the reset pins. If the case has no reset button, leave those pins empty.
• LED connectors are polarity-sensitive; the positive wire must match the positive pin.
• Do not connect a two-pin switch plug across unrelated front-panel pins.
• For testing, disconnect the case reset-switch connector entirely. A damaged or stuck reset switch can trigger repeated resets.

You can briefly touch the two power-switch pins with a screwdriver to start the system, but do this only momentarily and only on the correct pins identified in the manual. Never bridge a random pair of header pins.

4. Reseat memory and test one module

RAM is a leading cause of restart-only failures, especially on new DDR5 builds. A module can appear installed while one latch is not fully closed, or two modules may be stable at cold boot but fail during rapid retraining.

1. Shut down, switch off the PSU, unplug the power cable, and press the case power button once.
2. Remove both memory modules and inspect the slots for dust or obstruction.
3. Install one module in the motherboard’s recommended single-stick slot, commonly A2, which is the second slot from the CPU socket.
4. Boot at default BIOS memory settings and test Restart three times.
5. Repeat with the other module, then test both modules in the recommended paired slots, commonly A2 and B2.
6. Only after stable testing should you re-enable XMP or EXPO.

Memory training can take several minutes on the first boot after a reset or BIOS change. Do not interrupt it immediately. However, if the board repeatedly loops or hangs indefinitely, return to defaults and test modules individually.

5. Verify power delivery and cable installation

A restart briefly changes the power state of the CPU and graphics card. A marginal connection may survive a cold start but fail when voltage rails transition quickly.

• Push the motherboard’s 24-pin ATX connector fully into place.
• Connect the 8-pin CPU EPS cable near the processor; do not substitute a PCIe/GPU cable.
• Use every CPU power connector the motherboard manual recommends for a high-power processor.
• Connect the graphics card with the correct PCIe power cable. For a high-power graphics card using a compact high-current connector, insert the plug completely and avoid a sharp bend immediately at the connector.
• If the PSU is modular, use only cables supplied for that exact PSU model and series.
• Remove unnecessary USB devices, RGB hubs, storage drives, and expansion cards while troubleshooting.

Estimate whether the PSU has enough headroom

Add the CPU’s realistic package limit, the GPU’s board power, and roughly 75 watts for the motherboard, fans, memory, and drives. Then allow at least 30% headroom for transient loads:

(CPU 180 W + GPU 320 W + system 75 W) × 1.3 = 747.5 W.

That example points toward a quality 850-watt power supply rather than a heavily loaded 750-watt unit. The calculation is not a replacement for the graphics card or CPU manufacturer’s recommendation, but it explains why a system that appears fine in games can still fail during power-state changes.

6. Eliminate graphics and driver causes

If the PC restarts but the monitor remains black, remove the graphics card and test the processor’s integrated graphics if your CPU provides it. Connect the monitor to the motherboard video output. If restart becomes reliable, inspect the GPU seating, PCIe power connection, display cable, and PCIe slot.

Inside Windows, install the latest stable chipset driver for your platform and a current graphics driver. If the issue began after a graphics-driver update, use the installer’s clean-install option or remove the existing driver in Safe Mode before reinstalling. Also install motherboard LAN, Wi-Fi, and storage-controller drivers when applicable.

Temporarily disable overlays and hardware-control utilities, including RGB, fan-control, GPU-tuning, and motherboard management software. These programs can interfere with shutdown and restart through ACPI power commands.

7. Check Windows power settings

• Open Control Panel > Power Options > Choose what the power buttons do.
• Temporarily disable Turn on fast startup, then perform a full shutdown and test Restart.
• Run Windows Update, including optional hardware-driver updates only when they are appropriate for your board.
• Open an elevated Terminal and run sfc /scannow, followed by DISM /Online /Cleanup-Image /RestoreHealth.
• Check Event Viewer under Windows Logs and System for Kernel-Power, WHEA-Logger, display-driver, or ACPI errors around the failed restart.

Fast Startup primarily affects shutdown rather than Restart, so disabling it is a diagnostic step, not a guaranteed fix. WHEA errors, memory-related crashes, or display-driver resets are stronger evidence of an underlying stability problem.

8. Test the motherboard outside the case

If the problem remains, remove the motherboard from the case and place it on its box. Connect only the CPU and cooler, one RAM module, the PSU, and a graphics device if required. Disconnect front USB, RGB, storage, and expansion cards. Start it using the power-switch pins.

This test can reveal an extra brass standoff beneath the motherboard, a trapped screw, a pinched cable, or a case short. Add components back one at a time and test Restart after each addition.

When to replace or return a part

After defaults, a BIOS update, single-stick memory testing, correct cabling, and an outside-the-case test, a restart failure that follows the motherboard is strong evidence of a board or CPU-platform fault. If a different known-good PSU fixes it, replace the PSU. If one RAM stick consistently fails while the other passes, replace the memory kit rather than continuing to increase voltage or timings.

Do not repeatedly force the system off during failed boots, and do not raise memory voltage to hide instability before the basic wiring and firmware checks are complete. A stable default configuration is the reference point that makes every later performance setting safe to evaluate.