AMD Video Driver Keeps Crashing | Fixes That Stick

Guide / By

AMD video driver crashes often come from a messy install, unstable clocks, or overlays; a clean reinstall plus a few checks stops it.

A crash can feel random until you track the pattern. One minute a game is fine, the next you’re staring at a black screen, a frozen frame, or a desktop with no warning. Windows might bounce back with a driver reset message, or it might reboot.

This walkthrough keeps the order sane. You’ll triage the symptom, remove common conflicts, do a clean driver install, then rule out heat and power problems. You’ll end with a stable setup you can trust.

Symptoms And Fast Triage

Match what you see to a likely trigger. Then run the quick checks below before you reinstall anything.

Crash Pattern Likely Trigger First Move
Game closes to desktop Overlay hook, shader cache mess, unstable OC Disable overlays, reset GPU tuning
Black screen, audio keeps playing Driver reset, cable/port handshake, power dip Swap cable/port, lower refresh rate
Full reboot under load PSU limit, hotspot spike, undervolt too tight Remove tuning, watch temps
Crashes right after updates Driver mismatch, corrupted install Clean reinstall the driver
Crashes in browser or Discord Hardware acceleration conflict Turn off hardware acceleration

If you see artifacts on the desktop (sparkles, blocks, odd colors), stop heavy testing until you’ve reset tuning and checked heat. Artifacts are a stability red flag.

  • Reboot and set a baseline — Test one game for 10–15 minutes so you can tell if each change helps.
  • Reset GPU tuning — In AMD Software, set tuning back to default and disable undervolt/overclock profiles.
  • Disable overlays — Turn off Xbox Game Bar, Discord overlay, Steam overlay, and any FPS counter.
  • Check the display chain — Re-seat DisplayPort/HDMI, try another port, and avoid adapters while testing.
  • Lower refresh rate temporarily — Drop to a lower refresh rate to rule out a shaky high-refresh link.

Keep notes so you can undo changes after testing easily.

Why AMD Video Driver Keeps Crashing On Windows 10 And 11

Driver crashes tend to land in a few repeat buckets. The safest way through is to test one bucket at a time, with one clean change at a time. That keeps your results clear.

Leftovers from old installs

Stacked upgrades can leave behind broken profiles, old shader data, or mismatched driver pieces. A PC can look fine in light use, then crash as soon as a game leans on a newer graphics path.

Unstable clocks or voltage

Even “small” tuning can be stable in one title and fail in another. GPUs jump between clock states fast, and a tight undervolt can fall over when a scene spikes power draw or VRAM use.

Overlays and capture hooks

Streaming tools, RGB suites, and monitoring apps can all hook into rendering. Two hooks may coexist, then the next update adds another and the driver starts resetting.

Windows display behavior changes

Windows updates can shift fullscreen handling, power settings, or scheduling. If the driver install was already shaky, a Windows change can make the crash show up more often.

Heat, power, or a loose connection

A hotspot spike or a power dip can trigger a driver reset that looks like “the driver crashed.” In that case, the driver is the symptom, not the root cause.

Clean Reinstall The AMD Driver The Right Way

This is the best fix when amd video driver keeps crashing after updates, after swapping driver versions, or after a crash streak that started out of nowhere. The goal is simple: remove the driver and leftovers, then install a fresh copy with default settings.

Prep before you uninstall

  • Download the driver package — Get the correct AMD driver for your exact GPU so it’s ready to run offline.
  • Disconnect the internet — Unplug Ethernet or turn off Wi-Fi so Windows doesn’t install a driver mid-cleanup.
  • Create a restore point — It’s quick insurance if another component acts up.

Remove the current driver cleanly

  • Run a cleanup tool — Use AMD Cleanup Utility or Display Driver Uninstaller to remove AMD display drivers and profiles.
  • Restart when prompted — Let the tool finish the reboot cycle so Windows loads a basic display driver.

Install fresh and keep it plain

  • Install with default options — Avoid extra presets and keep settings stock for the first test.
  • Restart once more — A second restart clears pending driver services and components.
  • Reconnect the internet — Only after the driver is installed and the PC is stable.

Test the same game again. If it’s stable now, you likely cleared a corrupted install or a broken profile. If it still crashes, move on to conflicts.

If Windows swaps the driver on its own, your clean install won’t stay clean. Staying offline during uninstall and install is often enough. If a driver change lands right after you reconnect, pause updates while you test.

Stop Conflicts From Overlays, Browsers, And Games

Some crashes show up only in a few apps because those apps lean on overlays, hardware acceleration, or a specific graphics path. Treat this like a conflict sweep. Change one item, test, then keep or revert.

Turn off hardware acceleration where it matters

Browsers, Discord, and launchers can use the GPU for UI rendering and video decode. When a hook clashes, you can get driver resets even when you’re not gaming.

  • Disable browser acceleration — Turn off hardware acceleration in Chrome, Edge, or Firefox and restart the browser.
  • Disable Discord acceleration — Turn off hardware acceleration and the in-game overlay, then restart Discord.
  • Pause capture tools — Close recorders, FPS counters, and any overlay you don’t need for the test.

Reset game-side extras that trigger crashes

Start by removing the extras, confirm stability, then add them back. This keeps you from blaming the driver for a single setting that a game patch broke.

  • Clear shader caches — Clear the shader cache from AMD Software and clear the game cache if the launcher offers it.
  • Disable ray tracing for a test — Run without RT to see if the crash is tied to a specific path.
  • Cap frame rate — Set an FPS cap so the GPU isn’t pinned at 100% in menus and light scenes.
  • Switch display mode — Try true fullscreen and borderless to see which is stable on your setup.

Trim background apps that hook graphics

RGB suites and motherboard utilities can collide with graphics hooks. You don’t need to uninstall everything. Just stop them from running for a test.

  • Clean boot Windows — Disable non-Microsoft startup items, reboot, then test the crash-prone app.
  • Re-enable items slowly — Turn startup items back on one at a time until you find the trigger.

Try turning off a few driver-side features for a test. Enhanced Sync, Radeon Anti-Lag, Radeon Boost, or sharpening can crash a picky game. Switch them off, test, then add them back one by one.

Stabilize Temps, Power, And Clocks

If crashes land during heavy load, treat it as a stability check. A driver reset can be caused by heat, power delivery, or clocks that are right on the edge.

Return the GPU and RAM to safe settings

  • Reset AMD tuning to stock — Use default clocks, default voltage, and default power limits.
  • Disable third-party OC tools — Close tools that apply profiles at startup and remove auto-start tuning.
  • Test RAM without XMP/EXPO — Run memory at default settings for a short test to rule out RAM instability.

Power issues can be sneaky. If your card uses more than one power plug, avoid a daisy-chain cable. Use separate cables when possible, then retest.

Check thermals with a repeatable load

Pick one demanding game or one stress test and run it for 10 minutes. Watch GPU temp, hotspot temp, fan behavior, and whether clocks crash hard right before the reset.

  • Clear dust and restore airflow — Clean filters, check case fans, and make sure the GPU fans spin freely.
  • Re-seat the GPU — Remove the card, check the slot, then re-seat it firmly with the latch clicked.
  • Check PCIe power cables — Re-seat connectors and use separate cables if your GPU draws high power.

Rule out monitor and cable quirks

Sometimes the “crash” is the display link dropping. A weak cable or a picky mode can make it look like a driver reset.

  • Swap the cable — Use a known-good cable rated for your resolution and refresh rate.
  • Change ports on both ends — Try another GPU output and another monitor input.
  • Test at a lower refresh rate — If stability returns, stick with that mode until you replace the cable.

Crashes Still Happen After These Fixes

If you’ve done the clean reinstall, removed tuning, tested with overlays off, and checked thermals and cables, you’re down to deeper causes. Now the job is isolation: Windows damage, a single game bug, or hardware that can’t hold load.

Use a known-stable driver version

  • Roll back one release — Install a prior AMD driver that you know worked well for your games, then retest.
  • Switch versions cleanly — Use the same cleanup process when changing versions so leftovers don’t pollute the result.

Repair Windows system files

System file damage can break graphics components in odd ways. Built-in checks can repair common problems without reinstalling Windows.

  • Run SFC — Use sfc /scannow in a Command Prompt opened as Administrator, then reboot.
  • Run DISM — Use DISM /Online /Cleanup-Image /RestoreHealth, reboot, then test again.

Separate game bugs from driver crashes

If only one title crashes, treat it like a game issue first. Verify files, remove mods, and reset configs before you blame the driver.

  • Verify the install — Use Steam/Epic/Battle.net verify tools and remove reshade or mods while testing.
  • Test a second heavy title — If another demanding game also crashes, it points back to system stability.

Know when hardware is the likely cause

Repeat signs across multiple games, reboots under load, or desktop artifacts can point to a failing GPU, weak PSU, or unstable RAM. If amd video driver keeps crashing across clean installs and clean boots, hardware testing is the last step that gives a clear answer.

  • Test the GPU in another PC — If the same crashes follow the card, the card is the prime suspect.
  • Try another PSU — A borderline PSU can look fine until the GPU pulls hard spikes.
  • Check Windows Event Viewer — Look for display driver reset events near the crash time to confirm the pattern.

Handy references can save time when you’re swapping drivers or checking what changed on your PC. Keep these bookmarked so you’re not hunting mid-crash:

  • AMD driver download page — Get the correct package for your exact GPU and Windows version.
  • AMD cleanup tool page — Use the official cleanup tool when you want a fresh driver base.
  • DDU site — Use Display Driver Uninstaller for a deeper clean after many driver swaps.
  • Windows update history page — Check what installed on the day the crashes began.