CPU Overheating on Windows 11: Software Checks First

CPU overheating in Windows 11 — fans running at full speed constantly, the system slowing to a crawl mid-task, unexpected shutdowns during gaming or video rendering, or task manager showing 100% CPU with temperatures above 90°C — is almost never a mystery once you measure actual temperatures and identify what’s driving the load. Most fixes fall into two categories: reduce unnecessary software load, or improve the physical cooling path. This fits into the wider topic we cover in our Complete Guide to Fixing Windows, Browser, and Software Errors.

Get actual numbers before doing anything else. Download HWMonitor or Core Temp (both free) → run while doing whatever triggers the problem → note the “Max” temperature for each CPU core. If cores are hitting 90°C or above under sustained load: overheating is confirmed. Below 80°C even at full load is normal on most consumer hardware. This single measurement tells you whether you’re dealing with a temperature problem at all or something that just looks like overheating.

What’s using the CPU — stop diagnosing heat before diagnosing load

High CPU temperature means something is pushing the CPU hard. The first question is always “what?” — not “how do I cool it?” Ctrl+Shift+Esc → Task Manager → CPU column sorted high-to-low. A single process at 30%+ while nothing obvious is running is your answer.

Common Windows 11 CPU hogs you might not expect:

• Microsoft Compatibility Telemetry (CompatTelRunner.exe) — runs for hours after Windows updates, inventorying your system. If it’s at the top after a recent update: wait 2-4 hours, it usually finishes
• Windows Search Indexer — rebuilds the index after adding large numbers of files; check in Task Manager, it shows clearly
• Antivirus full scan — most run on their own schedule, often at inconvenient times
• Windows Update delivery — downloading and installing updates in the background

If the process is a legitimate Windows operation: let it finish rather than killing it. It will eventually complete and CPU returns to normal. If it’s an unknown process or third-party software: investigate further before assuming a cooling problem.

Power profile — the fastest thermal fix

Settings → System → Power & battery → Power mode → change from “Best performance” to “Balanced.” High performance mode runs the CPU at maximum clocks even at idle, generating heat unnecessarily for any workload that doesn’t actually need it. Balanced lets the CPU clock down during light tasks, dropping idle temperatures by 5-15°C in most cases.

For laptops specifically: Control Panel → Power Options → your active plan → Change plan settings → Change advanced power settings → Processor power management → Maximum processor state → change from 100% to 99%. This disables Intel Turbo Boost/AMD Precision Boost, which significantly reduces peak temperatures with minimal real-world performance impact for everyday tasks. Keep it at 100% for gaming or heavy compute work; 99% otherwise.

Dust is almost certainly part of the problem

If the machine is more than 12-18 months old and hasn’t been cleaned: dust accumulation inside the heatsink fins is reducing airflow. The CPU generates the same heat but the cooler can’t move it out as efficiently — temperatures rise, fans spin faster to compensate, but can’t fully catch up.

Desktop: unplug, open side panel, use compressed air (held upright, in short bursts) on the CPU heatsink, GPU heatsink, case fans, and PSU vents. Hold fan blades still while spraying — over-spinning them can damage the bearings. 15 minutes of work can drop CPU temperatures by 10-20°C.

Laptop: more complex. Compressed air into exhaust vents helps somewhat. For a thorough clean, the bottom panel needs removing (varies widely by model — check iFixit for your specific laptop). Even without disassembly, blowing into intake and exhaust vents with short bursts of compressed air improves airflow meaningfully. A full laptop cleaning at a repair shop runs £30-50 and is worth it on a machine that’s 3+ years old and running hot.

Thermal paste — for machines 3+ years old

The thermal interface material between the CPU die and heatsink dries and degrades over 3-5 years. When it does, heat transfer efficiency drops noticeably — even a perfectly clean system runs hotter than it should. Symptoms: temperatures have gradually crept up over months with no change in workload.

Replacing thermal paste: desktop CPUs are accessible after removing the cooler (usually 4 screws). Clean off old paste with 90%+ isopropyl alcohol. Apply a pea-sized amount of new paste (Arctic MX-4 or Noctua NT-H1, both around £7-10) in the centre. Reinstall the cooler. On laptops, repasting usually requires significant disassembly — many people have a repair shop do it (£50-100 including parts).

Our guide on Windows 11 CPU usage diagnosis covers identifying which specific process is responsible for high CPU load, and our Windows freezing troubleshooting covers the thermal throttling symptoms that present as freezing rather than shutdown. For CPU-specific temperature specifications and thermal design point values for your chip, Intel’s processor documentation and AMD’s product pages list TDP and max operating temperatures for every current model.

Fan speed and BIOS curves

Many motherboards use conservative fan curves by default — fans don’t ramp up significantly until the CPU is already at 70-75°C. The CPU spends time at elevated temperatures while the fans spin slowly, rather than the fans proactively cooling before temperatures climb.

Access fan controls: reboot → enter BIOS (F2 or Delete during POST) → look for “Fan Control,” “Q-Fan Control,” or “Smart Fan Mode” → adjust the fan curve so fans begin ramping at 50-55°C instead of 70°C. The trade-off is fan noise under load, which most people accept in exchange for lower temperatures.

Software tools for fan control without BIOS access: Fan Control (free, open-source) or Argus Monitor work well on Windows and let you create custom curves for case and CPU fans. AI Suite (ASUS), MSI Center (MSI), and similar manufacturer tools offer the same for their specific motherboards.

Startup programs and background load

Applications that run at startup contribute to baseline CPU load. Every running background app has a thread keeping it alive. With 20-30 startup items, the cumulative baseline CPU load can be 10-20% even when you’re not actively doing anything — and that 10-20% generates heat continuously.

Task Manager → Startup apps tab → disable anything you don’t need immediately at login: Discord, Spotify, Steam, Epic Launcher, Microsoft Teams (if you don’t need it immediately), Adobe Creative Cloud, gaming launchers, VPN clients. These can be launched manually when needed rather than sitting idle in the background. Settings → Apps → Startup for the same control in a slightly different interface.

Undervolting — advanced but effective

Most CPUs ship with slightly more voltage than strictly required for stable operation. Reducing this “offset voltage” by 50-100mV lowers heat output and power consumption with no performance loss at all — you get the same clock speeds with less heat generated.

Tools: Intel XTU (Intel CPUs), AMD Ryzen Master (AMD CPUs), ThrottleStop (Intel laptops). Start conservatively: -50mV offset → run a stress test (Cinebench R23, 10 minutes) → if stable, try -75mV → repeat until either stable margin is found or you hit an instability. When unstable (crash or BSOD): back off 25mV and call that your stable undervolt.

Note: Intel 12th generation and newer CPUs (Alder Lake, Raptor Lake) have undervolting locked by default via Platform Controller Hub security features. If XTU shows the voltage control as greyed out: undervolting isn’t available on that chip without a BIOS unlock (which may not exist for your board). AMD Ryzen and older Intel chips have much better undervolting flexibility.

CPU stress testing — verify the problem is heat

If you suspect thermal throttling (where the CPU reduces its own clock speed to prevent damage from heat) but can’t confirm visually: run a CPU stress test with HWMonitor open. Cinebench R23 multi-core → watch core temperatures climb → watch if CPU clocks drop when temperatures exceed 90°C. If clocks drop as temperatures climb: throttling confirmed. If temperatures stay below 90°C and clocks remain steady: thermal throttling isn’t the performance issue, and you should investigate other causes.

For a real-time view of throttling: HWiNFO64 → Sensors section → “CPU Package Power” and “CPU Throttling” rows → the throttling percentage shows directly when enabled. This is more precise than watching temperatures alone.

Laptop cooling pads

Laptop cooling pads with active fans can drop CPU temperatures by 5-10°C and are worth using for sustained heavy workloads (video rendering, long gaming sessions, compilation tasks). At £20-40 they’re cheap insurance against thermal throttling on laptops that run hot under load.

More important than any cooling pad: ensure the laptop has airflow underneath. On a desk, a prop of even 2-3cm allows the intake vents to breathe freely. On a bed or soft surface, intake vents are completely blocked — this alone can cause 20°C temperature spikes. A cheap laptop stand (£10-15) or the original box propped under the rear is enough.

Dust cleaning and power profile change together resolve most everyday overheating complaints in under 30 minutes. Thermal repasting extends the fix if the machine is older. Undervolting and fan curve adjustment are the next steps for machines where those don’t bring temperatures down far enough.

When GPU heat contributes to CPU overheating

In compact systems — small form factor desktops, gaming laptops, all-in-ones — the GPU and CPU share thermal space. A hot GPU raises the ambient temperature inside the case, which the CPU cooler then has to work harder against. If both CPU and GPU temperatures are elevated and the GPU is running a heavy workload: improving GPU cooling helps CPU temperatures too.

Check GPU temps in HWMonitor alongside CPU. If GPU is sitting at 80-85°C during gaming and CPU is also high: the case has a heat problem overall, not just a CPU problem. In a desktop, add a case exhaust fan or improve airflow routing. In a laptop, the GPU and CPU usually share a single heatsink/heatpipe system — the fix for both is the same (dust cleaning, repaste).

Hardware cooler upgrade

If a desktop CPU is regularly hitting 85°C+ under normal workloads with dust cleaned and fresh paste applied, the stock cooler is insufficient for that CPU’s heat output. Stock coolers are designed to keep the CPU within spec under the manufacturer’s testing conditions — not under sustained real-world loads that exceed those conditions.

Aftermarket air coolers: Cooler Master Hyper 212, DeepCool AK400, Noctua NH-U12S. These typically drop temperatures 10-20°C compared to Intel/AMD stock coolers and cost £30-60. AIO liquid coolers (Corsair iCUE, NZXT Kraken, Arctic Liquid Freezer) drop a further 5-10°C and are quieter under load.

For AM4/AM5 (AMD Ryzen): many aftermarket coolers are compatible with a bracket swap. For Intel LGA1700 (12th/13th/14th gen): ensure the cooler specifically lists LGA1700 support — it changed from LGA1200 and older coolers don’t physically fit without an adapter.

Monitoring temperature over time

For a clear picture of how temperature behaves: HWiNFO64 → Sensors → check “Log to file” → run for an hour of normal use → open the CSV log → see how temperature correlated with different tasks. This takes the guesswork out of “is it always hot or only during specific things” and makes the fix obvious.

Regular temperature monitoring is worth doing once a month on any machine more than 2 years old. Gradual temperature creep over 6-12 months is the earliest warning sign that dust or paste is degrading — catching it at 85°C and addressing it is much better than waiting until it hits 95°C and starts crashing. HWMonitor runs silently in the background with minimal overhead; keep it open occasionally and glance at the max column.

Case airflow for desktops

The direction of airflow through a case matters. Standard good practice: front intake fans pull cool air in from outside, rear and top exhaust fans push hot air out the back and upward. Negative pressure configurations (more exhaust than intake) can create turbulent flow that’s less efficient than balanced or slightly positive pressure (more intake).

Common airflow mistakes: all fans pointing the same direction (cancelling each other out), cable management blocking intake fans, CPU cooler facing the wrong direction (should exhaust toward the rear or top of the case). Correct airflow with fans already installed by verifying they’re pulling air in or pushing it out in the right direction — look for the embossed arrow on the fan frame showing airflow direction. Our guide on Windows 11 Battery Draining Fast covers an adjacent issue.