Help with critical AM5 error: machine_check_exception (0x9c)

Hey everyone,

I'm dealing with an intermittent issue that has now evolved into a complete boot failure, and I'd appreciate some help confirming the diagnosis.

System Specs

• CPU: AMD Ryzen 5 7600
• Motherboard: ASUS TUF Gaming B650M-E WIFI (latest BIOS installed)
• Memory: 32GB XPG DDR5 (running at stock 4800MHz in BIOS)
• Storage: Kingston 1TB NVMe SSD + SanDisk USB drive (Windows installer)
• Graphics: Tested with a dedicated GPU installed in the primary PCIe slot

The Problem

The PC initially started throwing BSODs with the MACHINE_CHECK_EXCEPTION (0x9C) error. At first, it would still reach the Windows login screen before crashing, but the issue has progressively worsened.

Troubleshooting Already Performed

1. Q-LED Check

During POST, the motherboard Q-LEDs cycle normally (Red → Orange → White) and eventually stop on the Green BOOT LED.

2. CPU Reseat and Cleaning

The CPU was removed, thoroughly cleaned (old thermal paste completely removed), and reinstalled. Memory modules were tested individually in the A2 slot.

3. BIOS Behavior

The system consistently posts, displays video output, and remains completely stable inside the BIOS. It correctly detects the CPU, all memory, and storage devices. DDR5 memory training completes successfully.

4. Exact Failure Point

The system crashes the instant it attempts to load any operating system:

• Existing Windows installation on the SSD
• Windows Safe Mode
• Fresh Windows installer from a bootable USB drive

In all cases, it freezes almost immediately and throws a MACHINE_CHECK_EXCEPTION BSOD.

My Current Theory

At this point, I'm leaning toward either:

• CPU silicon degradation/failure (an internal defect that only manifests when executing more complex instructions under OS load), or
• A motherboard power delivery/VRM issue affecting the CPU under load.

Since the behavior gradually worsened over time—from reaching the Windows login screen to failing before any OS can load—I'm currently about 85% convinced the CPU is failing.

Has anyone experienced this exact behavior on the AM5 platform? Is there anything else worth testing before I start the RMA process for the CPU?

Thanks in advance.