Skip to content

How much electricity does running AI locally use?

Less than most people expect. Apple measured an M5 Ultra Mac Studio at a maximum of 385 W from the wall, so even at full load around the clock it uses about 281 kWh a month, roughly $56 at $0.20 per kWh. A single NVIDIA RTX 5090 is rated at 575 W, and an H100 SXM at up to 700 W, before the rest of the server.

Real use is lower than these maximums because machines sit idle between requests.

Lower power also means less heat and noise: a Mac cluster can sit in an office instead of a server room.

Maximum power draw
HardwareMax powerMonthly energy at full load
Mac Studio, M5 Max200 W (measured)146 kWh
Mac Studio, M5 Ultra385 W (measured)281 kWh
NVIDIA RTX 5090 (GPU only)575 W (rated)420 kWh
NVIDIA H100 SXM (GPU only)700 W (rated)511 kWh

Sources

  1. Apple: Mac Studio power consumption
  2. NVIDIA: GeForce RTX 50 Series announcement
  3. NVIDIA: H100 Tensor Core GPU
  4. exo on GitHub (README)

Updated

Planning local AI for your team?

Tell us about your models, users and hardware. Exo Labs helps teams size, build and run private AI clusters.

Talk to Exo Labs →Download exoCheck your hardware on local.ai ↗