A heat wave, a poorly ventilated equipment room, a NAS tucked into a closed cabinet: heat shows up in the life of a storage device more often than people expect. Mechanical hard drives and SSDs do not react to it the same way, and the limits manufacturers actually publish are wider and more variable than the figures repeated everywhere online. This article explains what heat does inside each technology, which temperature ranges matter, how SSD thermal throttling works and which habits reduce the risk. The mechanical damage suffered by a hard drive that is already overheating is covered in our article The Dangers of an Overheating Hard Drive.
Heat does not act the same way on an HDD and on an SSD
A mechanical hard drive (HDD) has several heat sources: the motor spinning the platters, the actuator moving the read/write heads and the controller board. Heat expands parts, speeds up the ageing of electronic components and lubricants, and can disturb the very fine positioning of the heads above the platters. The outcome is not always a sudden failure. Read errors, repeated retries and slowdowns usually come first.
An SSD has no moving parts, but its controller and NAND flash chips also heat up, especially in the M.2 NVMe format under long write workloads. A hot SSD does not seize up: it protects itself by lowering its speed. Over the long term, sustained heat wears the cells faster and shortens the time an SSD keeps its data when it is left unpowered.
Temperature ranges published by manufacturers
There is no universal threshold. Each manufacturer publishes an operating range for every model in its datasheet or product manual. A few reference points taken from official documentation:
| Drive type | Published operating range | Source |
|---|---|---|
| 3.5-inch NAS HDD (WD Red Plus range) | 0 to 65 °C operating, -40 to 70 °C non-operating | Western Digital datasheet |
| 3.5-inch NAS HDD (Seagate IronWolf range) | 0 °C ambient up to 65 °C drive-reported temperature; sustained operation above 60 °C not recommended | Seagate product manual |
| Consumer NVMe SSD (Samsung 990 PRO) | about 0 to 70 °C | Samsung specifications |
Three clarifications avoid common misreadings. First, these ranges vary with capacity and model revision, and some references stop at 60 °C. Second, the value that matters is the temperature reported by the drive itself, not the room temperature: a drive reads several degrees above the air around it, and more still inside a confined enclosure. Third, change matters as much as level: the Seagate manual cited above limits the operating temperature gradient to 20 °C per hour, a reminder that a sudden swing from cold to hot is a stress in itself.
Throttling: why a hot SSD slows down
NVMe SSDs constantly monitor the temperature of their controller. When a manufacturer-defined threshold is reached, the firmware lowers clock speed and throughput to limit heat output. This is thermal throttling. Samsung, for example, calls its mechanism Dynamic Thermal Guard on current models. If temperature keeps rising anyway, the drive may shut down to protect its components, with a risk of losing data that had not yet been written.
The symptom is easy to recognise: a transfer that starts fast and collapses after a few dozen seconds, a large copy that never finishes, a machine that stutters during a backup. A slowing SSD is therefore not necessarily failing. It is signalling a cooling problem. Hard drives have no equivalent mechanism. They keep working at full load, so ventilation has to do the job.
What studies say about heat and failures
The question has divided researchers for years. Google’s 2007 study of a very large drive population (Pinheiro, Weber and Barroso) found no simple link between temperature and failures: the coolest drives were not the most reliable, and an effect of heat appeared only on the oldest drives, from roughly three years of age. In 2014, Backblaze analysed more than 34,000 drives from its fleet and found no overall correlation, with one notable exception: a family of 1.5 TB Seagate drives whose failure rate rose with temperature. In both cases, however, the drives ran in air-conditioned rooms, with Backblaze model averages between about 22 and 31 °C.
Two lessons follow. In a controlled environment and within manufacturer limits, moderate warmth is not the leading cause of failure. But these studies say nothing about a drive running at 50 °C or more in an unconditioned room for several days, which is exactly what a heat wave can produce. Manufacturers themselves state that sustained operation near the upper limit degrades rated reliability.
Heat waves, server rooms and NAS units: situations at risk
A period of intense heat stacks several factors: hotter outside air, computers and servers that work harder, and air conditioning that is sometimes undersized or down. These are the setups that most often cause trouble:
- a NAS installed in a closet, cabinet or equipment room with no air extraction, several drives heating side by side;
- a patch rack or small server room whose air conditioning switches off at night or on weekends;
- a laptop used on a soft surface or in direct sunlight with blocked vents;
- an external drive left in a car or near a sun-exposed window, even powered off, since non-operating limits exist too;
- an NVMe SSD without a heatsink, mounted under a graphics card that blows hot air right above it.
In a disk array, one more point deserves attention: the drives of a given NAS are often the same age and endure the same heat. An overly hot environment can therefore weaken several drives at once, which complicates the rebuild of an array after the first failure.
Ventilate and monitor: good habits
- Give the unit room: leave several centimetres free around air intakes and exhausts, and avoid stacking equipment that blows out hot air.
- Dust fans and filters regularly, with the unit powered off and unplugged; dust forms an insulating layer and reduces airflow.
- Check that fans spin at the expected speed and replace those that clog or wear out.
- Fit NVMe SSDs with a heatsink, especially for heavy write workloads (video editing, large copies, file servers).
- Plan for cooling continuity in an equipment room: an alert if the air conditioning fails, an ambient temperature probe.
- Monitor the temperature reported by the drive through SMART data (temperature attribute), using the operating system tool, the NAS software or smartmontools, and set an alert before the manufacturer limit rather than at the limit.
- Keep backups independent of the hardware concerned: heat is a risk factor, not the only one, as the effects of a power outage on a server or NAS also show.
A practical rule of thumb: if the reported temperature stays in the last ten degrees below the manufacturer limit for long periods, cooling is insufficient, even if nothing seems broken.
A powered-off hot SSD: data retention
An SSD stores information as electrical charge in cells, and that charge slowly leaks over time. The JEDEC JESD218 standard sets minimum unpowered retention times: about one year at 30 °C for a client SSD and three months at 40 °C for an enterprise SSD, applying to a drive that has reached its rated endurance. These are minimum requirements, not guarantees for a new drive. The takeaway is that heat shortens retention: an archive SSD kept in a hot attic is not a sound preservation plan.
A drive got hot: what to do
If a drive started clicking, disappeared from the system or showed errors after a period of intense heat, follow a few simple rules:
- shut the device down cleanly and let it cool down completely before any new attempt;
- do not power it on repeatedly and do not run file system repair or defragmentation: every attempt stresses an already weakened device;
- if the data matters, copy it to another device right away while the drive still responds, starting with the most valuable files;
- at the first abnormal noise (clicking, scraping) or if the drive is no longer detected, stop all handling.
For a hard drive that no longer responds or an SSD that has become unreadable, a laboratory analysis shows whether the data can be recovered before the situation gets worse.
In most cases heat does not destroy a device overnight, but it shrinks the margins: more slowdowns for an SSD, more errors and wear for a hard drive, more risk for a whole array. Respecting manufacturer ranges, watching the drive’s real temperature and organising airflow cost little compared with what the data inside is worth.