Code 43 on a GPU in Device Manager: The Fixes That Actually Work

Code 43 is the most misread error in Device Manager. People see “Windows has stopped this device because it has reported problems” and immediately assume the card is dead. Most of the time it isn’t. In my years of cleaning up driver messes, I’ve pulled maybe twenty cards back from code 43 with nothing more than a proper uninstall and a reseat. A few were genuinely dying. The trick is knowing which is which before you waste an afternoon.

Windows 10 and 11 use code 43 as a generic “the device talked to me wrong” flag. It fires when the OS can’t communicate correctly with the GPU driver or the hardware itself. NVIDIA, AMD, and Intel cards all get it. Integrated GPUs get it. Even USB controllers and Bluetooth adapters can throw the same code. So the message tells you less than you’d hope. You have to work through the causes in order of likelihood, and the order matters.

Start with the free stuff before you touch a driver

A plain restart clears temporary conflicts between Windows and the GPU driver. Windows reinitializes the hardware and reloads drivers on boot, and I’ve seen that alone resolve code 43 more than once. Sounds lazy. It works.

If the restart doesn’t stick, power the machine fully off and open the case. Pull the card, reseat it in the PCIe slot until it clicks, and check every power connector, 6-pin, 8-pin, or both, fully seated. Dust and small vibrations over years cause poor contact between the card and the motherboard, and the symptom looks exactly like a dead GPU. Also check the display cable at both ends, whether that’s HDMI, DisplayPort, or DVI. On a laptop, plug in an external monitor to see whether the fault lies in the GPU or the panel.

Those two steps sound too basic to matter. In practice, a large share of code 43 reports trace back to exactly this territory.

The driver fix, done properly

Outdated or corrupted drivers are the number one cause, and there’s a right way to reinstall them. Microsoft’s own steps are a decent baseline: right-click Start, open Device Manager, find the device with the yellow exclamation mark under Display adapters, right-click it, choose Uninstall, then right-click your PC name at the top of the list and select Scan for hardware changes. Windows reinstalls the driver. Check the General tab under Device status afterward. “This device is working properly” means you’re done.

That’s the quick pass, and it’s fine for a first attempt. If code 43 comes back within a day or two, the quick pass wasn’t enough. That’s when I stop using Device Manager and reach for Display Driver Uninstaller.

Why DDU beats the Windows uninstall

The ordinary uninstall leaves registry entries and leftover files behind. When you install the new driver on top of that residue, the conflict just rebuilds itself. DDU strips the vendor’s driver out at a much deeper level.

  • Run it in Safe Mode for the most stable result. Normal mode works, but the tool’s own guidance recommends Safe Mode.
  • If you insist on Normal mode, use its recommended cycle: Clean, reboot, clean again, reboot. Tedious, but it catches leftovers.
  • Set a system restore point first. It’s normally safe, but “normally” isn’t “always.”
  • Exclude the DDU folder from your antivirus. Security software can interfere with the cleanup mid-run.

After the cleanup, install the current driver directly from NVIDIA, AMD, or Intel. Skip the OEM version unless you have a specific reason, because OEM builds lag badly and occasionally ship broken. Reboot before you judge the result.

When a recent update is the culprit

Sometimes a driver update that landed after a Windows update breaks compatibility with an otherwise healthy card. If code 43 appeared right after an update, roll the driver back to the version that worked. Newer is not automatically better, and vendor release notes don’t always catch conflicts with a specific Windows build.

Also check Windows Update itself. Open Settings, go to Windows Update, and run Check for Updates. Windows updates often carry driver compatibility fixes and kernel changes that affect how the OS talks to hardware.

Firmware and platform-level problems

If the driver path is clean and the error persists, the fault may sit below Windows. An outdated BIOS can cause a modern GPU to miscommunicate with the motherboard. Go to your motherboard manufacturer’s site, find your exact model, and follow their official update instructions. Do not interrupt the flash. Power loss during a BIOS update can leave the system unbootable, and that’s a much worse day than code 43.

There’s a subtler possibility too: another PCIe device interfering with the card. Open System Information (msinfo32), expand Hardware Resources, and look at Conflicts/Sharing for overlapping components. If something overlaps, move the GPU to a different PCIe slot or temporarily disable the unnecessary device. Testing with only the GPU installed isolates this fast.

Deciding whether the card is actually dying

This is the fork in the road, and here’s where most guides get vague. My rule: if a known-good GPU runs fine in the same machine, and your card throws code 43 in a second system too, you’re looking at hardware. VRAM instability, overheating, or a failing connection are the usual suspects.

Watch for the supporting signs. Flickering or artifacts during gameplay, random driver crashes, fan failures. Those push the diagnosis toward failing hardware rather than software. A clean Windows install is the last software card to play. Back up your files, build a USB with the Media Creation Tool, and install fresh. If code 43 shows up on that clean system before you’ve added any other software, treat the card as bad.

One more thing worth knowing: a fix that works on one machine often doesn’t transfer. Code 43 is a symptom with many causes, so the person on a forum insisting it’s always the driver and the person insisting it’s always the VRAM are both half right. Work from cheap and reversible to expensive and permanent, and you’ll get the right answer faster than either of them.