Constant random reboots (4 crashes in 40 min) — started ONLY after running Arch Linux as my sole OS and then switching back to Windows. Is this 100% a hardware issue or could it be software/BIOS related?

Post

I'm losing my mind. My PC started randomly rebooting mid-use yesterday — 4 hard crashes in 40 minutes. No BSOD, no warning, just instant black screen and then reboot. It happened while I was sitting in the Marvel Rivals lobby with a YouTube video playing in the background. Not even in a match. The critical detail: these crashes started ONLY after I ran Arch Linux as my sole OS and then wiped everything to go back to Windows. The PC was completely stable before switching to Linux. This wasn't a dual-boot — Arch Linux was the only OS on the machine. I then wiped the drive completely and did a fresh Windows 11 install on August 1st. I can't afford to replace hardware right now, so I really need to know: is there any chance this is a software/firmware/BIOS issue from the Linux switch, or is this definitely hardware failure? MY SPECS * CPU: Intel Core i5-12400F (6C/12T, LGA1700) * GPU: NVIDIA GeForce RTX 3070 (Driver 610.88 / 32.0.16.1088, dated July 22 2026) * RAM: 16GB DDR4-3200 CL16 (2x8GB Goodram IRDM X, IR-XR3200D464L16SA/8G) * Motherboard: ASUS PRIME H610M-K D4 * BIOS: American Megatrends Inc., Version 4004 (dated June 3 2026) * Storage: WD Blue SN580 2TB NVMe (reports Healthy in Windows) * PSU: Bundled with Cooler Master CMP510 case, 650W 80+ Bronze (no-name OEM unit, not a standalone Cooler Master PSU) * OS: Windows 11 Pro 25H2 (Build 26200), fresh install on August 1 2026 * Case: Cooler Master CMP510 Mesh ARGB BIOS AND POWER SETTINGS * Power Plan: Balanced (switched from Ultimate Performance, didn't help) * Fast Startup: Disabled (turned it off, didn't help) * Hibernate: Disabled * Crash Dump: Small memory dump (MiniDump) * Auto Reboot on crash: Enabled * Disk Partition Style: GPT * EFI System Partition: 200 MB (standard) * BIOS was reset to factory defaults (Load Optimized Defaults) — didn't help WHAT THE EVENT LOGS SHOW Every single crash follows the exact same pattern: stornvme Event 11 — "The driver detected a controller error on \\Device\\RaidPort2" volmgr Event 161 — "Dump file creation failed during dump creation" (BugCheckProgress: 0x00040049) Kernel-Power Event 41 (Critical) — "The system has rebooted without cleanly shutting down first" EventLog Event 6008 — "The previous system shutdown was unexpected" No minidump files are being created. The system dies so fast it can't write them to disk. There was also one WHEA corrected error (Event 17) on a PCI Express Root Port (PCI\\VEN\_8086&DEV\_460D&SUBSYS\_86941043&REV\_02, Intel PCIe Root Port), error source: Advanced Error Reporting. Some of the crashes also show winlogon.exe initiating power off with Reason Code 0x500ff, which Windows doesn't have a description for. CRASH TIMELINE (August 9 2026) * 17:31 — Unexpected shutdown * 17:33 — Unexpected shutdown (2 min later) * 17:42 — Unexpected shutdown (9 min later) * 18:12 — System-initiated power off (Reason Code 0x500ff) After the 4th crash I shut it down for the night. The next day it ran for a few hours without crashing while just browsing, no gaming. THE LINUX TIMELINE PC was running Windows 11, completely stable, zero issues Wiped Windows entirely and installed Arch Linux as the sole OS (not dual-boot) Used Arch Linux for a while, system was stable Decided to go back to Windows Wiped the entire drive and fresh installed Windows 11 Pro on August 1 2026 Crashes started on August 9, about a week after the fresh install WHAT I'M WONDERING * Could running Arch Linux have changed BIOS/UEFI settings (Secure Boot, CSM, power management, PCIe settings, ASPM) that didn't get reverted when I switched back to Windows? * Arch Linux fully owned the EFI partition and bootloader (GRUB). Could there be remnants even after a full wipe and fresh Windows install? * Could Arch have changed NVMe power management firmware settings (APST) that persisted across the OS wipe? * Is this the bundled PSU dying? It's a no-name 650W OEM unit that came with the CMP510 case, and the RTX 3070 can spike to 300W+ transient. But it was working fine before. * The crashes only seem to happen under GPU load (gaming). Light browsing doesn't trigger them. WHAT I ALREADY TRIED (NONE OF THESE FIXED IT) * Reset BIOS to factory defaults (Load Optimized Defaults) * Disabled Fast Startup (powercfg /h off, confirmed HiberbootEnabled = 0) * Switched power plan from Ultimate Performance to Balanced * Applied various recommended BIOS settings Crashes still happen after all of the above. WHAT I HAVEN'T TRIED YET * Running sfc /scannow and DISM /RestoreHealth * Checking if GRUB remnants are in the EFI partition * Running memtest86 * Stress testing with OCCT (PSU stress test) * Updating/reflashing NVMe firmware via WD Dashboard * Clean reinstall of Windows (again) WHAT I NEED FROM YOU Is this definitely a hardware (PSU) problem, or could it plausibly be software/firmware related given the Linux to Windows timeline? If software-related, what should I try first? Has anyone seen the stornvme controller errors + Kernel-Power 41 combo caused by something OTHER than PSU failure? Could a BIOS reflash or CMOS clear fix this? I really can't afford a new PSU right now. If there's any chance this is fixable without buying hardware, I want to exhaust those options first. Thanks in advance.

This signal has not been scored yet.