You're in the middle of something and the computer just shuts off. No warning, no blue screen.
Most people land on “it must be overheating.” So we opened Intel's temperature documents and Microsoft's unexpected-restart document side by side. ⭐ They tell two different stories.
⭐⭐⭐ Intel's document gives an order.
“When a core exceeds the set throttle temperature, it will reduce power to maintain a safe temperature level.”
And the sentence right after it —
“If throttling fails, it will automatically shut down to prevent permanent damage.”
⭐⭐ Shutting down is second. Slowing down is first.
⭐ Which means a computer that shuts off because of heat passes through a stretch of being slow first. ⚠️ If it was fine and then simply died, heat is that much less likely.
⭐⭐⭐ Heat slows the machine before it stops it
Intel defines throttling like this.
“Throttling is a mechanism in Intel® Processors to reduce the clock speed when the temperature in the system reaches above TJ Max (or Tcase).”
“This is to protect the processor and to indicate to the user that there is an overheating issue in their system that they need to monitor.”
— Intel, “What Is Throttling and How Can It Be Resolved?”
| Stage | What the processor does | ⭐ What you notice |
|---|---|---|
| ⭐ Stage 1 | Reduces power, lowering clock speed (throttling) | ⭐ Sudden slowness — frame rates drop, work backs up |
| ⚠️ Stage 2 | ⚠️ Automatic shutdown, if stage 1 fails | ⚠️ It just dies |
| ⭐ “Reduce power to maintain a safe temperature → if that fails, shut down to prevent permanent damage” is the order in Intel's temperature document. | ||
⭐ Intel puts a number on the limit too.
| Term | Intel's definition | The number |
|---|---|---|
| Tjunction max | “The maximum thermal junction temperature that a processor will allow prior to using internal thermal control mechanisms to reduce power and limit temperature” | ⭐ Varies per product, usually between 100°C and 110°C |
| Tcase max | “Measured by characterizing a socketed processor with an integrated heat spreader for operation within an assembled system” | Per product |
| ⚠️ This does not mean “100°C is dangerous.” It is the point at which the protection starts working, and ⚠️ it differs by product. | ||
⭐⭐⭐ Microsoft's list, meanwhile, has no heat in it
Here is what Microsoft writes about unexpected restarts.
“An unexpected restart might be caused by a hardware device, its driver, or software.”
— Microsoft, “Troubleshooting Windows unexpected restarts and stop code errors”
⚠️⚠️ Three causes named, and heat is not among them. The six steps that follow are the same.
| # | The step as written | ⭐ What it separates out |
|---|---|---|
| ⭐ 1 | ⭐ Remove any new hardware | ⭐ The first suspect is “whatever changed recently” |
| 2 | Start your PC in safe mode | Brings up the minimum of drivers and software |
| 3 | Check the Device Manager | Looks for flagged devices |
| 4 | Check for sufficient free space on the hard drive | ⭐ The document's figure is 10% to 15% |
| 5 | Install the latest Windows 11 updates | |
| ⚠️ 6 | ⚠️ Restore Windows 11 | ⚠️ Last resort — “choose the recovery option that most closely matches your scenario” |
⭐⭐ Overlay the two documents and a fork appears.
| Just before it shut off | ⭐ Which document fits | What to do first |
|---|---|---|
| It had got noticeably slow | ⭐ Intel's — matches the thermal protection sequence | Intel's five checks (below) |
| No slowdown, it just died | ⭐ Microsoft's — hardware, driver, software | Whatever changed recently (step 1) |
| ⚠️ This fork is ours, from setting the two documents side by side. Neither document splits the problem this way. | ||
⭐ Intel's five checks (for self-built PCs)
What Intel lists, in order, when overheating is suspected.
| # | The item as written | ⭐ How to read it |
|---|---|---|
| 1 | Ensure the thermal solution is compatible and correct for this CPU | ⭐ Parts matching comes first |
| 2 | Verify the cooler is properly installed | A loose mount will not perform |
| ⭐ 3 | ⭐ Apply the appropriate amount of thermal interface material | ⚠️ “Appropriate amount” — more is not better |
| 4 | Check system fan operation | |
| 5 | Check air ventilation | ⭐ Space around the case and dust belong here |
| ⚠️⚠️ Laptops are not what this list addresses. Intel directs laptop owners to their OEM, since the OEM controls power and current limits. | ||
⚠️ On overclocking, Intel notes that altering clock frequency or voltage may void the warranty and reduce stability, security, performance and processor lifespan.
⭐ If it slows down but never shuts off, the slow-computer article is the right one. If only the connection drops, see when the internet is not working; for running costs, the electricity bill calculator. ⭐ How heat acts on a device's lifespan is covered from maker documents in the phone battery article.
Questions people ask
It runs above 90°C — is that dangerous?
⚠️ Intel's document cannot answer that. All it says is that “the maximum junction temperature limit varies per product and usually is between 100°C-110°C.” ⭐ You need your own CPU's figure from its specification, and ⚠️ that figure is where protection begins, not where damage begins.
Will cleaning out the dust fix it?
⭐ Item 5 on Intel's list is “check air ventilation,” so the direction is right. ⚠️ But items 1 to 4 come before it — cooler compatibility, mounting, thermal paste, fan operation. ⚠️⚠️ Opening a laptop is not advised.
Could it be the memory?
⭐ Microsoft names “a hardware device” first and makes step 1 “remove any new hardware,” so if you changed the memory recently, reverting that follows the document's order. ⚠️ The document contains no memory diagnostic procedure of its own.
Does shutting off without a blue screen mean something different?
⚠️ Neither document draws that distinction. Microsoft's page treats blue screens, black screens and stop codes as one group. ⭐ One line is worth keeping in mind though — “In most cases when Windows 11 restarts unexpectedly, the problem is resolved by the restart and no further action is needed.”
Sources
- ⭐⭐⭐ Intel official article — What Is Throttling and How Can It Be Resolved? (checked August 2026). Source for the definition of throttling, “to protect the processor and to indicate… an overheating issue… they need to monitor,” the five checks (compatibility, installation, thermal interface material, fan, ventilation), the direction to contact the OEM for laptops, and the overclocking caution.
- ⭐⭐⭐ Intel official article — Information about Temperature for Intel® Processors (checked August 2026). Source for the Tjunction max and Tcase max definitions, ⭐ “when a core exceeds the set throttle temperature, it will reduce power… if throttling fails, it will automatically shut down to prevent permanent damage,” and “the maximum junction temperature limit varies per product and usually is between 100°C-110°C.”
- ⭐⭐ Microsoft official support — Troubleshooting Windows unexpected restarts and stop code errors (checked August 2026). Source for “a hardware device, its driver, or software,” the six steps and their order, the 10% to 15% free space figure, “in most cases… resolved by the restart,” and the closing “choose the recovery option that most closely matches your scenario.”
- ⭐ Checked directly. “There is a stretch of slowness before it shuts off” is our reading of two Intel sentences in sequence, and the fork built on “did it slow down first?” is ours. ⚠️ No document splits it that way.
What we could not verify
- ⚠️⚠️ How often each cause applies. No document gives a share for heat against power against memory. The fork above divides by symptom, not by frequency.
- ⚠️⚠️ Judging the power supply. None of the three documents opened here covers PSU capacity or age, so none of it appears in the body.
- ⚠️ AMD processors. The temperature material is Intel's only. AMD's figures and behaviour were not checked.
- ⚠️ Memory diagnostics. Not in the documents, so not included.
- ⚠️ Korean data. There are no Korean statistics in this article.
As of August 2026. Temperature and throttling come from Intel's official articles; the causes and six steps for unexpected restarts come from Microsoft official support; and the fork drawn across the two is marked as ours. ⚠️⚠️ If hardware failure is suspected, do not open the machine — take it to the manufacturer or a service centre. For laptops in particular, Intel itself points you to the OEM.
