Definition
Thermal throttling is a protective behavior built into mining ASIC firmware: when a chip's junction temperature approaches its safe limit, the control board lowers clock frequency to cut heat output. The machine keeps running, but at a reduced hashrate. It is the silicon's way of trading performance for survival, and on most Antminer-class hardware it begins somewhere around the 80–85°C chip-temperature range, with hard shutdown reserved for more severe overtemp events.
The silent revenue leak
Unlike a fault that trips an alarm, throttling is easy to miss because the miner still reports as online. Yet a unit fighting poor airflow can shed a meaningful slice of its rated hashrate continuously, quietly eroding daily output. This is one reason the gap between nominal and actual hashrate often traces back to heat rather than hardware failure. A miner that benchmarks fine cold but sags after an hour of running is almost always throttling.
Causes and long-term cost
Common triggers include high intake-air temperature, dust-clogged heatsinks, a failed or undersized fan, dried-out thermal paste, and aggressive over-tuning that pushes voltage and frequency past what the cooling can absorb. Beyond the immediate hashrate loss, sustained high temperatures accelerate electromigration in the chips, stress solder joints on the hashboard, and shorten the machine's service life. Throttling is therefore both a symptom to diagnose and a warning to act on.
The fix is almost always better heat removal: cleaner airflow, lower ambient temperature, fresh thermal interface material, or a move to immersion or hydro cooling. For the broader picture of how power feeds into heat, see the power distribution unit entry.
See efficiency by model on the ASIC efficiency frontier.
In Simple Terms
Thermal throttling is a protective behavior built into mining ASIC firmware: when a chip’s junction temperature approaches its safe limit, the control board lowers clock…