You’ve already tried the standard checklist. Updated drivers. Verified files. Restarted Steam. Maybe even reinstalled. And Star Wars Zero Company still dumps to desktop after the splash screen or during the first firefight. The crash reporter dialog pops up, sometimes with “Low Level Fatal Error,” sometimes with “GPU Crash Dump Triggered,” sometimes with nothing useful at all.
The basic fixes work for a decent chunk of users. But if you’re reading this, you’re past that. You need the non-obvious root causes — the ones the generic troubleshooting guides skip because they assume a clean, predictable system.
We’ll go straight there.
Why the Standard Fixes Fail for Zero Company
The Shader Cache Trap
Star Wars Zero Company compiles shaders on first launch — a process that can take five to fifteen minutes depending on your hardware. If the compilation is interrupted (a background app kicks in, a power state shift, a Windows update check), the cache lands in a corrupted state. The game then tries to reuse those bad shader chunks on subsequent launches, triggering a CTD before you even see the main menu.
The non-obvious fix: clear the shader cache using Windows Disk Cleanup and also delete the game’s local cache folder manually. On Steam, that’s typically under %localappdata%Star Wars Zero CompanySavedShaderCachecodecodecodecodecodecode. Delete the folder, then reboot. First launch will recompile everything — let it finish. Do not alt-tab. Do not open Discord. That single process is the most common cause of repeated startup crashes that survive a game file verify.
Overlay Interference Patterns
Most players know to disable overlays. But the advice usually stops at “turn off Steam overlay.” The reality is that Zero Company is sensitive to multiple overlays stacking. Discord’s overlay, Nvidia’s in-game overlay, AMD’s software overlay, and even Windows’ own Xbox Game Bar can conflict with Unreal Engine’s rendering pipeline on certain driver versions.
The fix most guides miss: disable overlays system-wide, not just in the launcher. For Steam, go to Settings > In Game > uncheck “Enable the Steam Overlay while in-game.” For Discord, Settings > Game Overlay > toggle off. For Nvidia, open the GeForce Experience overlay settings and disable both the overlay and the share feature. Then restart the game. If the crash stops, re-enable overlays one by one to isolate the offender. Often it’s Discord’s overlay — it hooks directly into the rendering thread.
Targeted Fixes for Specific Error Codes
Low Level Fatal Error – Not Always a GPU Issue
The “Low Level Fatal Error” dialog in Zero Company is a generic Unreal Engine catch-all. It usually points to a hardware miscommunication, but the specific cause varies. In many cases, it’s not the GPU itself — it’s the GPU driver failing to allocate memory because the system’s paging file is too small.
Here’s the nuance: Zero Company can allocate more memory than your physical VRAM when using high textures and ray tracing. If the paging file on your system drive is set to “System Managed” and that drive has less than 20GB free, Windows may not expand the page file fast enough. The game requests memory, the driver tries to allocate, fails, and the engine throws a fatal error.
The fix: Set your virtual memory to a custom size. On a system with 16GB RAM, set the initial and maximum paging file to 32768 MB (32GB) on the same drive as the game. On 32GB systems, 49152 MB (48GB). This gives both Windows and the Unreal Engine enough breathing room to handle transient allocation spikes. Reboot after changing the setting.
GPU Crash Dump Triggered – The Mixed VRAM Scenario
This specific error message appears when the game’s renderer detects an unexpected GPU hang and generates a crash dump for analysis. It’s common on systems with multiple GPUs — laptops with integrated Intel graphics plus a discrete Nvidia/AMD card, or desktops with both an iGPU and a dGPU.
The obvious fix is to force the game to use the high-performance GPU in Windows Graphics Settings. But that alone doesn’t always work because Unreal Engine 5 can still try to offload certain compute tasks to the first available GPU it enumerates. The advanced fix: use the Nvidia Control Panel (or AMD Adrenalin) to set the preferred processor to the dGPU globally for the game executable. On Nvidia, go to Manage 3D Settings > Program Settings > add ZeroCompany.exe > set “Preferred graphics processor” to “High-performance Nvidia processor.” Then also go to Windows Settings > System > Display > Graphics > add the game > Options > select High performance. Two layers of enforcement.
System-Level Root Causes Most Users Overlook
Power Delivery and Transient Spikes
Star Wars Zero Company is a heavy load on the GPU — it can draw power in sudden spikes, especially during shader compilation and cutscenes. If your power supply unit (PSU) is borderline for your system’s total draw, or if your GPU is connected via a single daisy-chained PCIe cable, a transient spike can drop voltage briefly below the GPU’s threshold. The card resets, the driver crashes, and you get a CTD with no warning.
This is hard to diagnose without monitoring tools. Use GPU-Z or HWInfo to log the “GPU Voltage” and “GPU PerfCap Reason” during the crash. If you see “Pwr” in the perfcap reason and voltage dips below 1.0 V during spikes, your PSU likely lacks sufficient headroom. Minimum recommended PSU for a RTX 3070 and above running this game: 650W. For a 4080 or higher, 850W. Use separate cables from the PSU for each power connector — no daisy chains.
Virtual Memory and Paging File Sizing
We touched on this with the Low Level Fatal Error, but it deserves its own section because the fix is simple and high-impact. Zero Company can use large memory pools. If your system has 16GB of RAM, the game plus Windows plus a browser can easily consume 14GB. When the paging file is too small (or system-managed on a near-full drive), Windows starts swapping to disk frantically, causing stutter and eventually a hard crash.
Set a fixed paging file size. For a 16GB system using an SSD, set initial and maximum to 32,768 MB. For 32GB, set to 48,000 MB. Use an SSD — do not put the paging file on an HDD. And make sure the drive where the paging file lives has at least 20GB free space after the allocation.
Advanced Verification and Repair Sequence
Game file verify is the go-to, but it only checks a hash against Steam’s manifest. It won’t detect corrupted shader cache files, damaged saved profiles, or broken local config files. A more thorough sequence:
- Delete the local preferences file:
%userprofile%AppDataLocalStarWarsZeroCoSavedConfigWindowsGameUserSettings.inicodecodecodecodecodecode. The game will regenerate defaults. This fixes crashes caused by a corrupted change to an unsupported graphics option (e.g., forcing DX11 when the game needs DX12).
- Run the DISM and SFC system file checker steps from the source (as administrator:
sfc /scannowcodecodecodecodecodecode, thendism /online /cleanup-image /restorehealthcodecodecodecodecodecode). Restart.
- Perform a clean GPU driver installation using DDU (Display Driver Uninstaller). Boot into safe mode, run DDU, clean for both driver and GeForce Experience. Then install the latest driver fresh. Do not use “Express Installation” — choose “Clean Installation” if the driver installer offers it.
- Temporarily disable any hardware monitory utilities: MSU Afterburner, RivaTuner, HWInfo, ASUS AI Suite. These tools inject DLLs into the graphics pipeline and can trigger the crash reporter.
Launch Parameter Experiments (DX11 vs. DX12)
One source suggests using - dX11codecodecodecodecodecode, - dX12codecodecodecodecodecode, or - windowed nobordercodecodecodecodecodecode as launch options. The nuance: Zero Company is built on Unreal Engine 5. It runs natively on DX12. Switching to DX11 often introduces visual artifacts and lower performance — but if your GPU is a Pascal-era card (GTX 10 series) or older, DX12 route inefficiencies can cause stutter and CTD. For those cards, - dX11codecodecodecodecodecode is worth trying. For RTX 20/30/40 cards, stick with DX12. The - windowed nobordercodecodecodecodecodecode option helps if the crash occurs when alt-tabbing from fullscreen — it prevents the engine from re-initializing the swap change on focus loss, which sometimes triggers the GPU crash dump.
The specific launch string I recommend for troubled systems: -dX12 -windowed noborder -noborder -USEALLAVAILABLECORES -notexturestreamingcodecodecodecodecodecode. The last two flags force the engine to use all CPU cores and disable texture streaming, which reduces disk I/O spikes that can cause hitching and crashes on slower SSDs. (Note: -notexturestreamingcodecodecodecodecodecode increases VRAM usage — only use if you have 8GB+ of VRAM.)
FAQ
Why does verifying game files not fix the crash on Steam?
Verification only checks for missing or corrupted base game files. It does not touch the shader cache, config files, or saved game data. If the crash is caused by a corrupted shader cache or a broken config value, verification will show nothing changed.
Should I disable overlays permanently?
Temporarily for testing. Once you identify which overlay causes the crash (often Discord), you can either keep it disabled for this game or update the overlay software to the latest version. Sometimes an overlay update fixes the conflict.
Can a corrupted Windows system file cause Zero Company to crash to desktop?
Yes. The Unreal Engine depends on system runtime libraries like Visual C++ redistributables and DirectX. A corrupted d3d11.dlcodecodecodecodecodecode or dxgi.dlcodecodecodecodecodecode can cause launch-time crashes. Running sfc /scannowcodecodecodecodecodecode and then reinstalling the DirectX runtime from Microsoft’s website is a non-obvious fix.
Is there any benefit to limiting the framerate to prevent crashes?
In some cases, yes. If the GPU crashes under full load in menu scenes where framerates hit 200+, capping the FPS to 60 or 120 with RTSS or Nvidia Control Panel’s “Maximum Frame Rate” can reduce transient power spikes. This is rarely the primary fix, but it can stabilize an unstable system while you diagnose the root cause.
A Final Nuance: Frame Generation and Windows 24H2
One edge case we’re seeing more reports of: Star Wars Zero Company crashing on systems running Windows 11 24H2 with Nvidia Frame Generation enabled. The 24H2 update changed how the GPU scheduler interacts with certain Unreal Engine versions. If you’re on the latest Windows release and using a 40-series card with Frame Gen on, try disabling Frame Generation in the in-game graphics settings. The visual difference is minor compared to the stability gain. This is not a game bug — it’s a scheduling conflict that Nvidia and Microsoft are still working out. Until a driver or patch lands, the non-obvious workaround is to toggle Frame Gen off.
