GPU Hotspot / Junction temperature reaching 105C while Core temperature is only 70C
HWiNFO shows a 35C delta between Core and Hotspot, causing fans to scream at 100% emergency speed.
Internal desktop parts run on harmless 12V DC voltage and cannot shock you. Simply unplug the wall power cable first, then touch the unpainted metal case frame to ground static electricity.
Common Symptoms & Tells
- core reports 68C-72C, but Hotspot / Memory Junction reaches 105C-110C
- Fans violently spike to 100% (3500 RPM) while playing demanding games
- throttles clock speed down by 300MHz to prevent thermal breakdown
Interactive Fix Checklist
0 of 4 steps completedSafety power-off and discharge
Shut down the PC, turn off the power supply rocker switch (O), and unplug the AC power cord from the wall outlet. Press the front power button once while unplugged to discharge residual capacitor energy before touching any internal components.
Check GPU mounting screw torque on the back retention bracket
The X-bracket or 4 spring screws behind the core provide mounting pressure. If screws loosened over time, heatsink loses contact with the silicon center. Gently snug down the 4 center spring screws in an X-pattern.
Repaste GPU with viscous, non-pumpout thermal paste
If the delta between core and hotspot is greater than 20C, thermal paste has dried out or pumped out. Disassemble cooler, clean with 99% IPA, and apply a thick viscous paste (Gelid GC-Extreme, Thermal Grizzly Kryonaut Extreme, or PTM7950).
Apply GPU undervolt in MSI Afterburner
Press Ctrl+F in MSI Afterburner to open the voltage/frequency curve editor. Flatten the curve at 900mV (0.900V) at your stock boost clock. This reduces power consumption by 50-80W with zero loss in FPS, dropping hotspot temps by 15C.
Still having trouble with your PC?
Describe what is happening to your AI Technician. We will tailor the fix to your exact hardware setup.