I asked GPT 5 to extrapolate this to a Max Claude code monthly usage, two sessions a day, just business days. It guessed that would be 21k prompts a month. Google’s proxy numbers give a monthly footprint of ~5.1 kWh, ~5.53 L water, and ~0.64 kg CO₂e for heavy CC use.
That's equivalent to doing less than two miles driving(CO2), one toilet flush (water) and about three dryer loads of laundry.
And then what kind of reaction will we have when inference and work is less intensive environment wise for AI to do vs a human? That may even already be the case. See here: https://grok.com/share/c2hhcmQtMw%3D%3D_60f4890d-711b-4331-9... credit: HN user pama
That's equivalent to doing less than two miles driving(CO2), one toilet flush (water) and about three dryer loads of laundry.