Fire up a monitoring tool during a gaming session and you may see two GPU temperatures that disagree by 15-25 C. One is labeled “GPU Temp” (edge), the other “Hotspot” or “GPU Junction.” Knowing which one to watch — and what numbers are normal — is the difference between panicking over nothing and ignoring a real cooling problem.
Two sensors, two jobs
The edge temperature is an average-ish reading from a sensor near the edge of the GPU die. It is the friendly number most tools show by default. The hotspot (junction) temperature is the single hottest point anywhere on the die, sampled from a grid of on-chip sensors. Because it reports the worst spot, it is always higher than the edge reading.
| Reading | Typical load range | Concern threshold |
|---|---|---|
| GPU edge temp | 70-85 C | >87 C sustained |
| GPU hotspot/junction | 85-100 C | >105 C or big delta |
| Edge-to-hotspot delta | 10-20 C | >25 C = mounting issue |
| GPU memory (GDDR6/7) | 70-90 C | >95 C |
The number that really matters: the delta
A high hotspot on its own is not alarming — modern mobile GPUs are rated to run their junction near 100-105 C. What matters is the gap between edge and hotspot. A healthy gap is roughly 10-20 C. When that delta climbs past ~25 C, it usually means the thermal interface between the die and heatsink has degraded: pump-out of paste, a dried-out phase-change pad, or uneven mounting pressure. The heatsink is cool-ish (low edge) but one spot on the die is baking (high hotspot).
Why thin laptops show bigger deltas
Slim chassis run higher mounting tolerances and often use thinner thermal material, so a 18-22 C delta from the factory is common and fine. It is the change over time you track — if a machine that idled at a 12 C delta new now shows 28 C after a year, that is your cue.
What to do about a high delta
First confirm with a repeatable stress test (10-15 minutes of a demanding game or a GPU stress tool) and log both temps. If the delta is genuinely large and paired with throttling, options in order of effort: undervolt the GPU to cut heat at the source, clean dust from the fins, and — for older machines — repaste with a quality compound or, better on laptops, a phase-change pad like PTM7950 that resists pump-out. On a machine under warranty, let the manufacturer handle a repaste.
How to log the delta over time
The single most useful habit is recording your edge-to-hotspot delta when the laptop is new and healthy, then rechecking every few months. In HWiNFO, log a 15-minute gaming session and note the maximum GPU edge and GPU hotspot values. Subtract them: that baseline delta (often 12-20 C) is your reference. Save it in a note with the date.
If a later check shows the delta creeping up by 8-10 C or more, the thermal interface is degrading — usually paste pump-out on an older machine or a phase-change pad that has aged. Catch it early and a repaste restores temps; ignore it and you get progressively worse throttling as the die’s worst spot bakes while the heatsink barely warms. Because laptops run tight tolerances, a modest factory delta is normal, so resist the urge to repaste a brand-new machine chasing a number. Track the trend, act on a clear upward shift, and let the delta — not the scary raw hotspot value — guide your maintenance.
FAQ
My hotspot hits 100 C — is my GPU dying?
Almost certainly not. Mobile GPUs are designed to operate with junction temps near 100-105 C and will throttle to protect themselves before damage. Watch the delta and whether clocks are collapsing, not the raw hotspot alone.
Which tool shows hotspot on a laptop?
HWiNFO exposes “GPU Hot Spot Temperature” on most NVIDIA and AMD mobile GPUs. GPU-Z and MSI Afterburner may only show the edge temp depending on the model and driver.
Bottom line
Edge temp is the everyday number; hotspot is the die’s worst-case. Don’t fear a 95-100 C hotspot by itself — track the edge-to-hotspot delta over time. A widening gap, not a scary absolute number, is the real warning sign that your laptop’s cooling needs attention.
