NVIDIA RTX 50 GPUs have been reported to exhibit thermal throttling that reduces gaming performance, a condition that modders have traced to the removal of a GPU hotspot temperature sensor from the cards.

When the RTX 50 Blackwell series launched, NVIDIA changed its temperature reporting to emphasize average GPU temperature and omitted the hotspot sensor. Diagnostic tools such as HWMonitor, HWiNFO, MSI Afterburner, and GPU‑Z began displaying only the average temperature.

Brazilian modder Paulo Gomes used a modification to re‑enable the hotspot measurement on RTX 50 cards. The hotspot sensor recorded temperatures around 107 °C while average temperatures remained in the 70‑80 °C range, indicating that the GPU was reaching its thermal limit despite the lower average reading.

After replacing the thermal interface material, hotspot temperatures dropped to approximately 100 °C, and gaming performance returned to expected levels, suggesting that poor cooler contact or insufficient thermal paste was the root cause.

NVIDIA has not announced plans to restore the hotspot sensor in consumer drivers, but its data center GPUs include a comprehensive diagnostic suite that monitors hotspot and airflow conditions.

The findings suggest that users who rely solely on average temperature readings may miss critical thermal information, and that software updates or continued modding could address the issue.