GPU Crash dump Triggered means your graphics card stopped responding for long enough that Windows reset it, and Satisfactory shut itself down instead of freezing. The message comes from the game’s DirectX 12 renderer, but the fault almost always sits underneath the game: the driver, the card’s clock speeds, or the power and memory feeding it. The usual fix is a clean driver reinstall, a renderer swap, and backing off any overclock. Here is the order worth working through.
What the error is actually telling you
The crash text points at a function called D3D12RHI::TerminateOnGPUCrash(). Unreal Engine calls that in one situation: the graphics device disappeared in the middle of drawing a frame.
Windows runs a watchdog called timeout detection and recovery. If the graphics card cannot finish a task or hand it back within two seconds, Windows decides the card is frozen. It resets the graphics driver, clears everything out of video memory, and hands the game a device removed error. Satisfactory cannot keep drawing on a device that no longer exists, so it writes a crash dump and quits.
In the log the line usually reads DXGI_ERROR_DEVICE_REMOVED with Reason: DXGI_ERROR_DEVICE_HUNG. Hung is the useful word there. The card was still present. It just stopped answering.
Satisfactory is named in the message because Satisfactory was on screen. The message describes what happened to your GPU, not a bug in the game.
If you are getting blue screens as well, that is the same fault escalating. When Windows sees five or more of these hangs and recoveries inside a single minute, it stops trying to recover and bug-checks the machine on the sixth.
Why Satisfactory triggers it when other games do not
Satisfactory runs on Unreal Engine 5 and defaults to DirectX 12. Update 1.2, which went live on 2 June 2026, moved the game to Unreal Engine 5.6.1. DirectX 12 keeps a long queue of work sitting on the card and leans on Nanite and Lumen, the engine’s geometry and lighting systems, so your GPU runs at heavy, steady load rather than in short bursts.
A grown factory makes that harder still. Thousands of belts, machines, signs and lights inside one view all get pushed at the card together, and none of it lets up while you stand there.
Then there is memory. The recommended spec asks for 8 GB of video memory, with 4 GB as the floor. A 4 GB card running high textures in a dense area has very little room left over.
The crash can also hit on the main menu, before any save loads. That is worth noticing, because it rules out factory size as the whole story and points straight at the driver or the renderer.
Fix it in this order
Change one thing at a time, then play for at least half an hour before you decide it worked.
- Clean install the graphics driver. Not an update on top of what is already there. Use the clean install option in the NVIDIA or AMD installer, or run DDU in Windows safe mode and install fresh afterwards. If the newest driver crashes, install the one before it. On NVIDIA, the Studio branch is worth a try too: same hardware support, tested harder, released less often.
- Switch renderer. Add -vulkan to the game’s launch arguments. In Steam, right-click Satisfactory, open Properties, and type it into the Launch options box on the General tab. In Epic, use the three-dot menu next to the game, choose Manage, and tick Additional Command Line Arguments. If Vulkan misbehaves, swap it for -dx11. DirectX 11 has been deprecated since Update 8 and loses Nanite and Lumen, so treat it as a test rather than a home. If Vulkan runs clean, stay on Vulkan.
- Go back to stock clocks. That includes a factory overclock on the card. Reset MSI Afterburner to defaults and close it fully, since it hooks into the driver even when you are not looking at it. Clear any tuning preset in AMD Adrenalin. An undervolt counts as a change too, so start from stock.
- Close the overlays. Discord, Steam, RivaTuner and the NVIDIA and AMD overlays all inject themselves into the game’s rendering path. Any one of them can be the piece that tips a marginal setup over. Turn them all off for the test and add them back one at a time later.
- Cap the frame rate. Set the in-game limit to your monitor’s refresh rate. An uncapped menu will happily run a card flat out for nothing. While testing, turn off upscaling and frame generation, and set anti-aliasing to TAA.
- Turn off hardware-accelerated GPU scheduling. Go to Settings, System, Display, Graphics, then Change default graphics settings, toggle it off and reboot. It moves scheduling work onto the GPU, and on some driver and card combinations that is where things come apart.
- Clear the shader caches. Run Disk Cleanup and tick DirectX Shader Cache, then delete the cache files under %LOCALAPPDATA%\FactoryGame\Saved. A cache built by a driver you have since replaced can feed the card work it no longer understands.
If your card has 4 GB or 6 GB of memory, add one more step: drop textures and global illumination a notch and turn off any hardware ray tracing option. Running out of video memory looks identical to a hang from the outside.
On a laptop, check which GPU is running the game
Laptops usually have two graphics chips: a small one built into the processor and a faster dedicated one. If Windows hands Satisfactory to the built-in chip, it will struggle and can hang under load.
Open Settings, then System, Display, Graphics, add Satisfactory to the list, open Options and pick High performance. NVIDIA Control Panel can do the same per program under Manage 3D settings. It is a two minute check, and it explains a fair share of laptop crashes that happen straight away on launch.
When it is not the graphics card at all
Three things imitate a GPU crash well enough to send people shopping for a new card.
Memory. XMP and EXPO profiles run your RAM faster than the CPU’s memory controller is guaranteed to handle. Usually that works. Usually is the problem. Turn the profile off in BIOS, run at the default speed for a day, and see whether the crashes stop.
Power. Modern cards spike hard for a few milliseconds at a time. A tired or borderline supply can dip under that and trip its own protection. Check that every 12V connector is fully seated, and run separate cables to the card rather than one daisy-chained lead.
Heat. Core temperature is not the whole picture. Memory on the card runs much hotter than the core, and a dusty heatsink or aged thermal pads will show up as random hangs long before the core reading looks alarming.
The pattern of when it crashes narrows it down fast.
| What you see | What it points to |
|---|---|
| Crashes in several games at stock clocks | Driver, power, or hardware |
| Only Satisfactory, everything else fine | Renderer or driver bug, try Vulkan |
| Crashes on the main menu before a save loads | Driver or shader cache, not factory size |
| Started right after a driver update | Bad driver, roll it back |
| Freezes and blue screens as well | Memory or power |
| Only after twenty minutes or more | Heat, or instability under sustained load |
How to read your own crash log
Press Win+R and paste %LOCALAPPDATA%\FactoryGame\Saved\Logs, then open FactoryGame.log. The crash sits near the end of the file. Full crash folders, one per crash, are in %LOCALAPPDATA%\FactoryGame\Saved\Crashes.
Two things are worth finding. The reason code after DXGI_ERROR_DEVICE_REMOVED tells you how the device went away, and DXGI_ERROR_DEVICE_HUNG is the common one. The header line also records IsUsingMods, which decides your next step.
Then cross-check in Windows. Open Event Viewer, go to Windows Logs, then System, and look around the time of the crash. An entry saying the display driver stopped responding and has recovered confirms the watchdog fired, which means the card really did hang rather than the game losing track of it.
Mods, updates, and where to report it
If IsUsingMods says Yes, run the game clean before anything else. Mods are built against one specific game build and break when that build changes. Back up your saves from %LOCALAPPDATA%\FactoryGame\Saved\SaveGames first, then remove the mods and test.
Make sure the game itself is current. Patch 1.2.4.0, build 502094, went out on 11 August 2026. Verify your files through Steam or Epic while you are in there.
Still crashing on a clean driver, stock clocks and no mods? Post it on the official Q&A site at questions.satisfactorygame.com with your log, your exact card and driver version, and a list of what you have already ruled out.
Is the card actually failing?
Run something else heavy at stock clocks: a different DirectX 12 game, or a graphics stress test for twenty minutes. If that crashes too, the problem has left Satisfactory’s territory.
Visual corruption before the crash is the clearer warning sign. Flickering blocks, wrong colours or garbled textures usually mean the memory on the card is going. So does a crash that repeats at the same moment after a fresh Windows install. A card that only misbehaves in one game, on one driver version, is far more likely to be fine.
Frequently asked questions
Is GPU Crash dump Triggered a sign my graphics card is dying?
Usually not. It normally means the driver or the card’s clock speeds are unstable, not that the hardware is broken. Worry about failing hardware when several games crash at stock settings, or when you see visual corruption before the crash.
Should I use Vulkan or DirectX 11 in Satisfactory?
Vulkan. DirectX 11 has been deprecated since Update 8 and drops Nanite and Lumen, so it costs you visual quality. Use -dx11 only to work out whether the crash follows the renderer.
Why does Satisfactory crash when my other games run fine?
It runs Unreal Engine 5 on DirectX 12 and keeps the card under heavy, steady load, especially in a large factory. That is enough to expose a card or driver that is only just stable everywhere else.
Does raising TdrDelay in the registry fix it?
No. It only gives the card longer before Windows steps in, so it hides the hang rather than stopping it. It also delays the safety net that keeps a stuck GPU from locking up your whole desktop.
Where do I find the Satisfactory crash log?
Press Win+R and paste %LOCALAPPDATA%\FactoryGame\Saved\Logs, then open FactoryGame.log. Individual crash folders sit in %LOCALAPPDATA%\FactoryGame\Saved\Crashes, each with its own copy of the log.
Conclusion
GPU Crash dump Triggered is Windows resetting a graphics card that stopped answering, and Satisfactory stepping aside rather than freezing your desktop. Work the list in order: clean driver install, switch to Vulkan, return to stock clocks, cap the frame rate, then look at memory and power. Change one thing at a time. A good result is a full session with no crash and nothing left that you feel the need to keep adjusting.
Related Post:
What the DaVinci Resolve GPU Configuration Warning Means (and How to Fix It)
“System Is Out of GPU Memory” in Blender: Why It Happens and How to Fix It
GPU Utilization at 100%: When It’s Normal and When It’s Not
What Is a Good GPU Clock Speed in MHz?
Fix GPU memory leak [Forge/Fabric]: what it does and how to install it
