Confirming this is still happening with the identical signature, five years later — I think that's the important part. iMac20,1 (27" 2020), Core i9, AMD Radeon Pro 5500 XT 8GB, macOS 26.6.2 (25G83), single internal Retina 5K display (no external monitor). Same fingerprint as the original report in this thread: Event: GPU Reset Graphics Hardware: AMD Radeon Pro 5500 XT AMDRadeonX6000_AMDNavi14GraphicsAccelerator state: ENABLED Restart Channel: 18 VMPT FirstPendingCB: PID = WindowServer VM Protection Fault events. Count: 0 …which then cascades into: panic(...): userspace watchdog timeout: no successful checkins from WindowServer (2 induced crashes) in 120 seconds WindowServer's main thread is stuck uninterruptibly (TH_WAIT|TH_UNINT) in IOKitWaitQuiet during re-init, blocked in AMDRadeonX6000 → IOAcceleratorFamily2 → IOGraphicsFamily. VM protection fault count is 0 and thermal pressure was Moderate, so this is a clean GPU hang on the paging channel, not a memory fault or a thermal event. In my case the trigger is specifically the display-sleep power transition (machine at rest, internal panel only), which is a different entry point than the external-monitor reports but produces the same channel-18 VMPT reset. This signature is documented on this same iMac20,1 / Navi14 hardware from Catalina 10.15.7 (this thread, 2021) through Big Sur, Monterey, and now macOS 26.6.2 — unresolved the entire time. Given that it has survived every OS generation, could someone from Apple confirm whether this is a known Navi14 driver issue or whether it indicates GPU hardware degradation on these units? I'm filing this with Apple Feedback Assistant now, with a full sysdiagnose + the crash reports attached (regression / hardware question). Happy to share the FB number, additional stackshots, or the tailspin with anyone from Apple who wants to pull it.