Hi everyone,

I’d like some opinions on an intermittent freezing issue I’ve been dealing with and whether it makes sense to start an RMA now or wait.

Summary of the issue:
– The freezes started only after upgrading the GPU to an RTX 5070 Ti (MSI Gaming Trio OC)
– Before that, with an RTX 3060 Ti, the system was completely stable
– The freezes happen randomly, mainly while gaming (Far Cry 6, Manor Lords, Satisfactory)
– When it happens, the entire system hard-freezes:

• frozen image
• audio fully stops (no looping sound)
• keyboard and mouse unresponsive – The PC does not reboot on its own; I have to hold the power button to shut it down

Current system:
– i7-10700F
– ASUS TUF B460-PLUS
– 32 GB DDR4 3200 MHz (2x16)
– XPG Core Reactor 750W Gold PSU
– Windows 11 fully updated
– NVIDIA drivers up to date (clean installs)

Monitoring & logs:
– Continuous monitoring with HWiNFO (sensors + logging enabled)
– Logs captured during freeze events
– GPU/CPU temperatures, clocks and power consumption always within normal ranges
– No signs of thermal throttling or power limitation

Windows logs:
– Kernel-Power events after the freeze (forced shutdown)
– Some Hyper-V related events, but nothing directly pointing to a GPU fault
– No clear or consistent error that explains the freezes

Tests performed:
– Stress tests (FurMark): no freezes
– With XMP enabled, freezes tend to occur sooner.
– With XMP disabled (JEDEC speeds), freezes still happen, but usually take longer to appear.
– This behavior does not occur at all with another GPU (RTX 3060 Ti), regardless of XMP state.
– BIOS and drivers updated
– Swapped back to RTX 3060 Ti:

• System becomes fully stable again, no freezes at all

Important notes:
– The issue only happens with the RTX 5070 Ti installed
– No overclocking or undervolting on GPU or CPU
– Sensor data does not show any obvious thermal or power-related problem

Possible explanations I’m considering:
– Intermittent GPU hardware defect
– Compatibility issue between the RTX 5070 Ti and an older platform

Future plan:
– I plan to build a new platform (AM5, PCIe 5.0) around March and test the GPU there

Main question:
Would you start the RMA process now even without a clear error in logs,
or wait to test the GPU on a newer platform first?

Any technical opinions or similar experiences would be greatly appreciated.
Thanks!

Sorry for any English mistakes, it’s not my first language.