If you’ve noticed your graphics card running hotter than usual, there are a few common culprits likely to blame. Modern cards often feature robust cooling solutions that help it maintain maximum clock speeds and performance, but some problems are simply inevitable with any GPU.
In many cases, overheating isn’t caused by a single issue but by a combination of smaller problems that add up, pushing your card to constantly hit that 90°C+ mark. Here are some of the most likely reasons your graphics card is overheating and how to fix them.
Dust Build-Up
When was the last time you cleaned the inside of your PC? If your answer is “I don’t remember,” there’s a good chance it’s crammed with dust.
Even if you have a tempered glass side panel and it looks relatively clean, there’s still a good chance that dust has accumulated deep inside your graphics card heatsink. You can’t see it, but it can have a major impact on your GPU’s ability to stay cool. Even a thin layer of dust can trap heat and reduce your graphics card’s cooling efficiency, causing temperatures to rise dramatically.
When I disassembled my PC to dust it for the first time in a while, I was shocked by how much dust came out of what had seemed like a relatively clean graphics card.
Besides visible dust on the card, common symptoms of dust build-up include a hot-running GPU under load, louder fan noise, and high idle temperatures.
Fortunately, the solution is straightforward: remove the side panel of your computer, hold the graphics card’s fans in place with one hand, and give the card a good dusting with a can of compressed air. Do this somewhere outside if possible to avoid cleaning up afterward. If a lot of dust comes out, congratulations—you’ve likely found the problem. Make sure to keep a regular dusting schedule going forward.
Poor Case Airflow
If your whole computer tends to run hot (including the CPU), it’s no surprise that the graphics card is overheating as well. GPUs release heat inside the case and rely directly on good case airflow to stay cool. What complicates things is that poor case airflow usually isn’t caused by a single factor, but rather a combination of them.
For starters, if your PC is placed in an obstructed area where it can’t eliminate heat easily, such as under a desk or on a carpet, try moving it to a location with unrestricted airflow, such as the top of your desk. This alone can reduce internal temperatures by a few degrees.
Another factor that can raise case temperatures is poor airflow due to messy cable management. That’s right—managing cables isn’t just for show. Moving cables out of the way so that fans can do their job can have a surprisingly large impact on case temperatures, and in turn, on GPU temps. Taking an hour to open the PC and route cables away from fans is a simple and effective solution.
If that doesn’t help, there’s one more solution for a hot-running PC. Many cases come with low-quality stock fans, which may not move enough air.
Upgrading to decent fans can make a big difference. You don’t even have to spend much—if you don’t care about RGB, a couple of Arctic F12 fans mounted in the right spots can greatly improve your PC’s cooling and lower GPU temperatures. Most importantly, make sure you have at least one exhaust fan mounted at the rear.
Too Aggressive Overclock
If you’ve overclocked your graphics card to get more performance out of it, or if you have one of those factory OC models, you might not realize just how much hotter this can make your graphics card run. This is especially true if you haven’t combined the overclock with an optimized fan curve and perhaps a light undervolt.
The worst part is that if your overclock is causing the graphics card to overheat, it’s probably doing more harm than good to your performance anyway. So, the best solution is to underclock the graphics card on purpose, even if that means going slightly below the speed that was advertised on an OC model’s box.
For instance, if your ASUS TUF Gaming RX 9070 XT OC is running hot at the 3080MHz boost clock, try switching it to the Default Mode that’s set to 3060MHz. And if that doesn’t help, go all the way down to AMD’s advertised boost clock for this card, which is 2970MHz (or possibly go even lower).
Aside from playing with the clock speed, you should consider undervolting the graphics card.
AMD has a handy automatic undervolting solution in its drivers, and NVIDIA users can try a manual undervolt using MSI Afterburner. This can significantly reduce GPU temps, though it may also cause instability, especially if you’re still keeping an overclock on the GPU, so play around with the settings a bit until you find something that works.
Faulty Fans or Fan Curve Misconfiguration
If your graphics card is relatively old or has spent many hours running, there’s a possibility that one of the fans has worn out and is no longer spinning as fast as it should. You can check this yourself by putting the graphics card under load and watching the fans spin.
A slow or non-spinning fan is a clear indicator (assuming zero fan speed isn’t enabled and you’re actually stressing the GPU). You can also try setting the fan speed to maximum manually in the drivers and see if they respond.
Depending on your graphics card vendor, you might be able to perform a fan health test using its proprietary software. For instance, my old Sapphire AMD RX 580 had a fan health test integrated into Sapphire TriXX that I could use to check the fans, so check your vendor’s software to see if it offers something similar.
Assuming your GPU’s fans work, you might be surprised to learn that they could still be the culprit. Most graphics cards use fan curves that are optimized for sound rather than performance.
This means that even under maximum load, the graphics card’s fans might not exceed around 30% of their maximum speed. If you care more about cooling performance than a bit of extra noise or minor fan wear, it’s a good idea to create your own custom fan curve and optimize it as you like. This is a nearly universal solution that will cool down any graphics card, so don’t skip this step.
Dried-Out Thermal Paste and Pads
If you’ve never changed your graphics card’s thermal paste and pads, there’s a very good chance that they’re dried out and crusty. The thermal paste’s main job is to transfer heat from the GPU to the heatsink, so even a small inefficiency in this process can have a major impact on the GPU’s temperatures.
As reported by Tom’s Hardware, most graphics card vendors tend to use low-quality thermal paste that degrades after a few months and causes overheating. Even if they used good paste, a few years of heavy use is enough to dry it out to the point where replacement is necessary to restore cooling efficiency.
Similarly, thermal pads that keep secondary components like the VRAM and VRMs cool also dry out. If you’re already opening up the card to replace the paste, you might as well replace the pads too.
This is actually one of the first maintenance steps I take on any used graphics card I buy.
Both paste and pads are rather inexpensive, so the only real concern is accidentally damaging the card. Only attempt this if you’re willing to take the risk or know what you’re doing; otherwise, leave it to the professionals and have it done at a computer repair shop.
Opening up your graphics card can result in permanent damage.
Misaligned Heatsink
If you’ve already replaced the thermal paste in your graphics card but it’s still overheating, there’s a small possibility that the heatsink wasn’t reinstalled correctly. It might also have shifted slightly during shipping or rough handling, though that’s less likely. Uneven pressure on the GPU and other components can cause poor contact, which in turn can make the graphics card overheat or even crash.
If you’re new to disassembling graphics cards, you might be surprised by just how much pressure is actually needed between the heatsink and GPU. That’s why it’s important to tighten the heatsink screws using a criss-cross pattern to create even pressure. Again, if you’re inexperienced and don’t want to risk damaging the card, it’s best to leave this job to professionals.
An overheating graphics card is one of the most common issues a PC owner can face. Fortunately, most causes of overheating have fairly simple solutions that can keep your GPU running cool and performing at its best.