Code 43 is Windows’ way of saying “the driver reported a problem and I stopped the device.” On a GPU, the symptoms are unmistakable: the discrete graphics card vanishes from games and GPU-accelerated apps, Device Manager shows a yellow warning on the display adapter, and the status reads “Windows has stopped this device because it has reported problems. (Code 43).” Your laptop keeps running on integrated graphics, so everything looks normal until you need real GPU power.
The error is frustratingly non-specific by design — it’s a messenger, not a diagnosis. The fixes below work the problem the way a technician does: driver state first (most common), then power and thermal causes, then the hardware verdict. If you’re dealing with driver issues more broadly, our clean NVIDIA install guide is a useful companion.
Why this happens
- Corrupted GPU driver files. Partial updates, crashed installers, or disk errors leave the driver in a state that fails initialization.
- Incompatible driver version. A too-new (or too-old) driver for your specific GPU and Windows build fails the driver’s own self-check.
- Windows Update installed the wrong driver. The automatically delivered version conflicts with the vendor’s driver stack.
- Insufficient or unstable power. On desktops, a weak PSU rail; on laptops, a failing adapter or power-management fault can trigger Code 43.
- GPU overheating. Thermal damage or clogged cooling causes initialization failures that the driver reports as Code 43.
- Hardware failure. Dead VRAM, a failed GPU core, or broken solder joints — the last resort diagnosis.
- Conflicting software. Overclocking tools, mining software remnants, or multiple GPU utilities fighting over the card.
How to fix it
Work through these in order — diagnose first, then act.
Fix 1: Restart cleanly and check for transient causes
Code 43 can be a one-time initialization failure. Rule that out in two minutes.
- Shut down the laptop completely (not sleep, not hibernate) and wait 30 seconds.
- If the battery is removable, remove it; otherwise hold the power button 30 seconds with the charger disconnected to drain residual power.
- Reconnect power and boot normally, then check Device Manager.
- If the error is gone, it was a transient power-state failure — monitor for recurrence before doing more.
This “full power drain” resets the GPU’s power state machine, which a normal reboot doesn’t always do. It’s the cheapest fix in this guide and resolves a real (if small) percentage of cases.
Fix 2: Roll back the driver
If Code 43 appeared after a driver or Windows update, revert immediately.
- Open Device Manager > Display adapters, right-click the GPU, and choose Properties > Driver.
- Click Roll Back Driver, select a reason, and confirm.
- Restart and check whether the error clears.
- If rollback is unavailable, note the current driver version — you’ll need it for the clean reinstall in Fix 3.
Don’t skip this just because “the driver was working yesterday.” Driver updates are the single most common Code 43 trigger, and rollback is a two-minute test. If it works, disable automatic driver updates so the bad version doesn’t return.
Fix 3: Clean-reinstall the GPU driver
The heavy hitter: purge the corrupted driver state completely and install fresh.
- Download the latest driver for your exact GPU from NVIDIA or AMD .
- Run the installer and choose Custom > Clean install (NVIDIA) or Factory Reset (AMD) to wipe all previous driver state.
- Restart and verify in Device Manager that the yellow warning is gone.
- Test with a GPU workload (a game or benchmark) — idle-desktop success isn’t proof the driver initializes under load.
For NVIDIA cards specifically, our detailed clean install walkthrough covers the full procedure including the pre-install preparation that makes the difference between a clean install and a merely fresh-looking one.
Fix 4: Remove conflicting GPU software
Overclocking utilities and abandoned GPU tools can wedge the driver into Code 43.
- Uninstall MSI Afterburner, EVGA Precision, ASUS GPU Tweak, or similar tools via Settings > Apps.
- Check Task Manager > Startup for lingering GPU utility processes and disable them.
- After uninstalling, repeat the clean driver install from Fix 3 — these tools leave registry and driver hooks behind.
- Retest before reinstalling any tuning utility; add them back one at a time if you need them.
An unstable overclock is a classic Code 43 source: the driver fails its self-test at the overclocked settings and reports the device as faulty. If you were overclocking, reset to stock clocks first — the “hardware failure” may just be an unstable tune.
Fix 5: Check power delivery
GPUs are power-hungry, and marginal power produces exactly this error.
- On a laptop, verify you’re using the original manufacturer charger at the correct wattage — underpowered third-party chargers cause GPU initialization failures.
- Set Settings > System > Power & battery > Power mode to Best performance and retest.
- In the manufacturer’s power utility (or BIOS), check that the discrete GPU isn’t power-limited or disabled.
- On a desktop, verify all PCIe power connectors are seated and the PSU meets the GPU’s requirements.
Laptop dGPUs can draw 80W+ on their own; a 65W charger on a laptop that shipped with 200W will absolutely cause GPU errors. This is also why Code 43 sometimes appears only on battery — the system can’t feed the GPU.
Fix 6: Address thermals
Heat-damaged or heat-stressed GPUs fail initialization and report Code 43.
- Monitor GPU temperatures with HWiNFO or the vendor’s tool — sustained idle temps above 70°C on a laptop suggest cooling failure.
- Clean dust from vents and fans with compressed air; replace dried thermal paste if the machine is old and you’re comfortable opening it.
- Elevate the laptop for airflow and retest — if Code 43 clears when cool but returns under load, thermals are confirmed.
- Check that fans actually spin under GPU load; a dead fan on a dGPU is a common finding.
Thermal Code 43 has a signature: the error appears after gaming or heavy use, clears after the machine cools, then returns. If you see that pattern, cooling service — not driver reinstalls — is the fix.
Fix 7: Test the hardware verdict
When every software fix fails, it’s time to determine whether the GPU itself is dead.
- Boot a Linux live USB and check whether the GPU is detected and functional there — if yes, it’s a Windows/driver problem you haven’t fully solved.
- On a desktop, move the card to another PCIe slot or another machine; on a laptop, this isn’t possible — instead, run the manufacturer’s built-in hardware diagnostics (Dell, HP, and Lenovo all ship GPU tests).
- Check for artifacting (visual glitches, colored squares) which indicate VRAM failure rather than driver issues.
- If diagnostics confirm GPU failure on a laptop, the repair is a motherboard replacement — get a quote before deciding, as it may exceed the machine’s value.
A Code 43 that survives a clean driver install, correct power, good thermals, and Linux testing is a hardware fault. At that point, further software attempts waste time — the decision becomes repair versus replace.
Closing
Code 43 is a process of elimination with a clear gradient: transient power state, bad driver version, corrupted driver, conflicting software, power delivery, thermals, hardware. The overwhelming majority resolve in the first four fixes — which is why driver hygiene (clean installs, blocking bad auto-updates) prevents most recurrences. If you reach Fix 7 with the error intact, you’ve earned the hardware verdict honestly, and you can act on it without second-guessing.
