I believe that's called a defective product --- and reminds me of a certain somewhat infamous series of Toshiba laptops which would overheat and shutdown if you spent too long in BIOS setup, because the fans were entirely software controlled.

There's zero reason for all this extra "thermal management" crap beyond corporate greed.

You can design hardware that's capable of dissipating the maximum heat output of every component simultaneously, but you'll end up with hardware that's significantly larger, heavier, and costs more to manufacture. Or you can design hardware that's capable of dissipating the expected heat output under a range of use-cases, and devolve policy management to the OS so it can (based on user preference) prioritise appropriately, allowing you to produce smaller, lighter machines that either cost less or generate more profit. Even if the motivation for cutting costs is to increase profit margins, users generally still seem to prefer the smaller, lighter hardware that works just fine under most circumstances.

That is them, not me.

No... That's what you, the manufacturer looking for a rationale to justify extracting more profit want.

What people want is a box that does computing within spec and without melting, sterilizing their laps/hands, or otherwise introducing malbehaviors they have to work around.

People want stuff that doesn't or is very difficult to break. When it does break, they want it to be a straightforward repair.

I had an HP laptop that did that in Windows or Haiku. I rewired the fan to run 100% all of the time and aside from losing some battery time the laptop ran fine. And when plugged into the external power supply I would run it all weekend (4-days) non-stop with no problems.

The EC may trigger an emergency shutdown if any monitored component (not just the CPU) ends up outside its thermal limits.

Indeed; I've replaced the laptop since then, so I don't think I have the log, but I believe it very explicitly logged it was the CPU temperature provoking it before actually shutting down.

Is this normally logged anywhere? You may have just saved me RMAing a laptop.

I had a computer with a broken GPU fan and it thermally shut off. There where kernel log entries about temperature warnings, though I don't remember if it mentioned which component. Systemd should preserve them before the system shuts off.