What else should I check for random PC freezes?

Start with the logs rather than buying parts. After a freeze, check Reliability Monitor and Event Viewer’s System log around that time. Note any display-driver resets, storage or controller warnings, hardware errors, or unexpected shutdown entries. A logged error is a clue, not proof, but it can narrow the search.

Also watch temperatures and CPU, GPU, and disk activity while the machine is under its usual workload. If disk active time spikes, record the process and transfer rate too. A high active-time reading with little throughput can point to a device waiting on I/O, but it does not identify the cause by itself.

A cursor that keeps moving while the rest of the desktop hangs makes me look at the graphics path and power delivery, though it is not a diagnosis. Try a clean graphics-driver reinstall, then test a game or video workload again. If you can temporarily borrow a known-good power supply, that is a useful comparison before spending money.

Check the GPU power connectors are seated and avoid changing several things at once. Keep notes on what changed and whether the freeze returned, otherwise it gets hard to tell which test mattered.

Keep the test setup boring for a while: one monitor, basic keyboard and mouse, and no optional USB devices. Try another wall outlet or location if practical, and compare wired and wireless networking. If the issue only appears with one connection or in one room, that is useful evidence, but it still does not automatically mean the house wiring is at fault.

Passing a memory test is reassuring, but it may not catch every issue that appears with a particular combination of sticks or settings. If the RAM is a mixed set, test one stick at a time in the motherboard’s recommended slot. You can also temporarily turn off any memory overclock profile and see whether the behavior changes.

For storage, check the drive’s health and firmware with the manufacturer’s tools, and inspect the SATA or power connection if it is a SATA drive. These are checks to narrow things down, not a reason to assume the SSD is bad.