Processor temperatures alarm people more than they should. Modern chips are designed to run hot and to protect themselves, and the numbers that look frightening are frequently normal.
What is normal
Idle: roughly 30 to 45 °C, depending on room temperature and cooler.
Gaming: roughly 60 to 80 °C.
Sustained all-core load: 80 to 95 °C is normal for many current processors, particularly higher-power models.
Some processors are designed to run at their thermal limit and boost until they reach it. Seeing such a chip sit at 95 °C under sustained load is the design working, not a fault.
What actually goes wrong
Hitting the thermal limit and reducing clock speed. The machine is safe; you are losing performance you paid for.
Temperature climbing steadily at idle, which suggests the cooler is not making proper contact.
A sudden change from previous behaviour, which usually means dust, a failed fan, or a cooler that has come loose.
Causes worth checking, in order
Cooler mounting. The most common cause by a wide margin. Uneven pressure or a protective film left on the base of the cooler. If temperatures were bad from the first start, this is where to look.
Thermal paste. Too little leaves gaps; far too much insulates. A rice-grain amount in the centre is enough for most processors.
Case airflow. A cooler can only move heat into the case air. If that air is already hot, the cooler has nothing to work with.
Dust. On an older machine, a fin stack packed with dust is the most likely explanation for a gradual change.
Power limits. Some boards default to unrestricted power for processors that are specified more conservatively, which produces much higher temperatures for a small performance gain. Setting the manufacturer specified limits in firmware often drops temperatures substantially.
Why modern processors run at their limit deliberately
Older processors had a fixed clock speed and ran well below their thermal limit. Modern ones do the opposite: they raise clock speed until they meet a limit, then hold there.
That limit might be temperature, total power draw, or current. Whichever comes first is where the chip settles. A processor sitting at 90 °C under sustained load has usually found its thermal limit and is running exactly as designed.
This is why the old instinct that a hot chip is a chip in trouble no longer applies. What matters is whether it is reaching the clock speeds it should, not what number the temperature sensor reports.
It also means better cooling on such a chip buys performance rather than safety. A cooler processor boosts higher for longer; it was never in danger either way.
How to tell whether the cooler is at fault
The useful measurement is not the peak temperature but how quickly it gets there.
Start a demanding sustained load and watch. A processor that jumps from idle to its limit within a few seconds, before the heatsink has had time to warm through, is usually not making good contact with the chip. Uneven mounting pressure and a protective film left on the cooler base are the two most common causes.
A processor that climbs steadily over a minute or two and then settles is behaving normally, whatever the final figure.
The other useful check is idle temperature relative to the room. A machine idling 15 to 20 °C above room temperature is fine. One idling 40 °C above it has a mounting problem.
Compare against the maximum operating temperature the manufacturer publishes for your specific processor, not against a general figure. That limit varies between models, and a chip specified to run at 95 °C is not in difficulty at 90 °C.