[Official FAQ] Corrections and follow-up questions welcome.

LLM inference can hold high GPU/CPU utilization for long stretches, stressing power delivery and thermals differently from bursty games.

Recommendations:

• Leave PSU headroom; respect cable/connector specs—cheap units are a hazard.
• Watch sustained temperatures and throttling; “mystery slowdowns” are often heat.
• Clean laptop fans; elevate notebooks under long jobs.

Safety first with any power mods. Share undervolt/power-limit findings carefully.