Screen freezes, flickers black, then returns with a “display driver stopped responding” notification — one of the most disruptive Windows 11 faults. With NVIDIA’s driver the pattern is unmistakable: games crash to desktop, video playback blacks out, or Event Viewer fills with Event ID 4101 entries naming nvlddmkm.

These crashes are rarely random. Windows 11 uses Timeout Detection and Recovery (TDR): if the GPU does not respond within seconds, Windows restarts the graphics driver instead of blue-screening. That restart is the flicker you see. Repeated flickers mean something keeps hanging the GPU — a corrupt install, a driver conflict, unstable clocks, heat, or marginal power.

The fixes below run from least to most disruptive. Test stability after each step — a few minutes of gaming is enough. Do not skip the clean-install step; it resolves more cases than everything else combined.

Why this happens

  • Corrupted driver remnants: partial updates, interrupted installs, or leftover files can leave the NVIDIA stack broken in ways ordinary reinstalls do not fix.
  • Windows Update driver conflicts: Windows 11 sometimes installs its own older NVIDIA driver alongside yours, and the two fight for control of the GPU.
  • Unstable GPU clocks: factory or user overclocks and aggressive boost behavior can push the GPU past stability as it ages or heats up.
  • Overheating: dust-clogged vents or dried thermal paste let the GPU hit its thermal limit, causing hangs that TDR reads as driver failure.
  • Insufficient or unstable power: on desktops, an aging or undersized PSU — or a loose PCIe cable — starves the GPU under load.
  • Conflicting software: overlays, screen recorders, GPU utilities, and some antivirus hooks can destabilize the driver.
  • Aging hardware: VRAM or solder-joint degradation produces the same TDR signature as a software fault, worsening over months.

How to fix it

Work through these in order — diagnose first, then act.

Fix 1: Stop Windows Update from replacing your NVIDIA driver

Before touching the driver, close the tap that keeps re-breaking it. Windows 11’s automatic driver updates are a leading cause of repeat crashes: you install a stable driver, and days later Windows silently downgrades it.

  1. Open Settings > Windows Update > Advanced options > Optional updates and decline any queued display driver.
  2. To block driver updates entirely, open the Group Policy editor (gpedit.msc) or set the registry value ExcludeWUDriversInQualityUpdate to 1.
  3. Alternatively, in System > About > Advanced system settings > Hardware > Device Installation Settings, choose “No” for automatic driver downloads.
  4. Restart, then confirm in Device Manager that your currently installed NVIDIA driver version has not changed.
  5. From here on, install NVIDIA drivers only from NVIDIA’s official driver download page — never let Windows Update manage the GPU.

Fix 2: Perform a true clean install with DDU

Display Driver Uninstaller (DDU) wipes the driver, registry entries, and services completely — fixing the majority of recurring cases.

  1. Download the latest stable NVIDIA driver for your GPU (prefer a release a few weeks old over a brand-new one) but do not run it yet.
  2. Download DDU, disconnect from the internet (to stop Windows Update interfering), and boot into Safe Mode.
  3. Run DDU, select GPU > NVIDIA, and choose “Clean and restart.”
  4. After the restart (still offline), run the NVIDIA installer, choose Custom installation, and tick Perform a clean installation.
  5. Reconnect to the internet, restart once more, and test with a game or video workload.

Fix 3: Switch driver branches (Game Ready vs. Studio)

Sometimes the bug is in the driver release itself. If Game Ready crashes, the longer-validated Studio branch may be rock solid on the same hardware, and vice versa.

  1. Note your current driver version and branch in the NVIDIA Control Panel (Help > System Information).
  2. Use DDU as in Fix 2 to remove the current driver completely.
  3. Install the latest Studio driver instead of Game Ready (or Game Ready if you were on Studio).
  4. Test for at least 30 minutes under your normal workload before judging stability.
  5. If the alternate branch is stable, stay on it until the next branch release and re-evaluate then — NVIDIA’s driver download portal lists both branches for every supported GPU.

Fix 4: Reset clocks and check thermals

An overclocked or overheating GPU crashes any driver, however clean the install. Rule out clocks and heat first — especially on thin gaming laptops.

  1. Reset any overclock: uninstall or reset MSI Afterburner or similar tools, and restore default clocks in the NVIDIA Control Panel.
  2. Watch the GPU temperature in Task Manager (Performance > GPU) during a game or benchmark — sustained readings above 87°C on a laptop warrant attention.
  3. Clean the vents with compressed air and elevate the rear of the laptop for better airflow.
  4. If crashes cluster under heavy load but never at idle, heat or power delivery is the likely trigger, not the driver version.
  5. As a diagnostic, cap the frame rate or enable V-Sync to reduce sustained GPU load and see whether crashes stop.

Fix 5: Disable overlays and conflicting software

Overlays inject themselves into the rendering pipeline, and many “driver crashes” are really overlay crashes. Discord, Steam, GeForce Experience’s own overlay, screen recorders, and GPU monitoring tools are the usual suspects.

  1. Disable the NVIDIA in-game overlay in GeForce Experience (Settings > General > In-Game Overlay off).
  2. Turn off Discord’s overlay (User Settings > Game Overlay) and Steam’s overlay (per-game properties).
  3. Close or uninstall GPU tweaking and monitoring utilities temporarily — they hook the driver directly.
  4. If you use a third-party antivirus with “gaming mode” hooks, test with it disabled for one session.
  5. Re-enable tools one at a time over several days; the crash’s return identifies the culprit, as in our installer troubleshooting guide .

Fix 6: Adjust the TDR timeout

When the GPU is slow rather than stuck — common on older hardware — extending the timeout stops Windows from killing a driver that would have recovered on its own. Apply this after the driver is clean and thermals are verified. For the full TDR walkthrough, see our display-driver recovery guide .

  1. Open Registry Editor and navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers.
  2. Create a DWORD (32-bit) value named TdrDelay and set it to 10 (seconds, up from the default 2).
  3. Optionally create TdrDdiDelay set to 10 as well for the same reason.
  4. Restart the computer for the new timeout to take effect.
  5. If crashes persist unchanged with a 10-second timeout, the GPU is genuinely hanging — revert the values and continue to Fix 7.

Fix 7: Test the hardware and update the system BIOS

When software fixes are exhausted, suspect hardware: VRAM faults, a weakening PSU (desktops), or a BIOS with buggy PCIe power management. Dual-GPU laptops should also check for integrated/discrete conflicts.

  1. Run a GPU stress test (FurMark or a demanding game) and note whether artifacts, checkerboard patterns, or immediate crashes appear — artifacts point to VRAM/hardware.
  2. On desktops, reseat the GPU, check both PCIe power connectors are fully latched, and test with a known-good PSU if one is available.
  3. Update the motherboard or laptop BIOS from the manufacturer’s support site — PCIe and power-management fixes are common in BIOS changelogs.
  4. On laptops, try forcing the discrete GPU in the NVIDIA Control Panel (Manage 3D settings > Preferred graphics processor) to rule out switching glitches.
  5. Check Microsoft’s Windows support resources for any known TDR-related issues with your current Windows 11 build, and consider a repair install of Windows as the last software resort.

Closing

Recurring NVIDIA crashes yield to methodical work: block Windows Update from meddling, wipe the slate with DDU, test the other driver branch, then rule out clocks, heat, and overlays before touching the registry. If artifacts survive every software fix, the GPU needs professional attention; no driver version will change that.

Frequently asked questions

Why does my NVIDIA driver keep crashing on Windows 11?
The most common causes are a corrupted driver installation, a conflict between Windows Update’s driver and NVIDIA’s own driver, an unstable overclock or factory-overclocked card, overheating, or an underpowered power supply. Timeout Detection and Recovery (TDR) restarts the GPU when it stops responding, which is what you experience as a crash.
Will a clean install of the NVIDIA driver stop the crashes?
Often, yes. A standard install leaves behind registry entries and configuration files that can keep reintroducing the fault. Using Display Driver Uninstaller (DDU) in Safe Mode to fully remove the driver, then installing the latest stable release with the ‘clean installation’ option, resolves a large share of recurring crash cases.
What is Event ID 4101 and is it related to NVIDIA crashes?
Event ID 4101, ‘Display driver nvlddmkm stopped responding and has successfully recovered,’ is Windows logging a TDR event for the NVIDIA driver. Occasional recoveries are tolerable, but frequent 4101 events mean the GPU is repeatedly hanging — a sign of driver corruption, instability, or a hardware problem that needs the fixes in this guide.
Should I use Game Ready or Studio drivers to stop crashes?
If you mostly play games, use the latest Game Ready driver; if you do creative work, try the Studio driver, which is validated for stability over a longer cycle. Switching branches sometimes fixes persistent crashes because each branch carries different bug fixes. Either way, prefer a release that is at least a few weeks old over a day-one launch driver.