Ecological AIThe AI that preserves
our water
ChatGPT uses 500ml of water for one email. ALFY hybrid? 10 to 100x less. ALFY local? Zero.
10-100x
less water than cloud
The Problem
Cloud AI consumes billions of liters of water
500ml
per ChatGPT query
700K L
to train GPT-3
+34%
Microsoft water 2021-22
Why so much water?
Cooling
Datacenter GPUs heat up enormously and require constant evaporative cooling.
Evaporation
80% of water used evaporates and is lost forever. It doesn't return to the cycle.
The fundamental distinction
Electricity
Solar, wind, nuclear... Electricity has renewable solutions deployed at scale.
Problem being solved
Fresh water
Millennial aquifers, FINITE resource. When an aquifer dies, it's IRREVERSIBLE.
Not substitutable
"Electricity gets solved. Water doesn't."
Transparent comparison
| Mode | Water per query | Savings |
|---|---|---|
ChatGPT / Cloud AI | 50-500 ml | — |
ALFY Hybride | 5-50 ml | 10x moins |
ALFY 100% Local | 0 ml | 100% économie |
Impact calculator
10ALFY
1,620 m³
of water saved per year
0.6
Olympic pools
Switch to green AI
Every deployed ALFY is less pressure on aquifers. Make the ecological choice without compromising performance.
Request a demo