As businesses embrace the transformative power of AI, it’s essential to consider both the opportunities and the environmental responsibilities that come with innovation.

Our mission is to connect organizations adopting AI with partners and solutions dedicated to minimizing the environmental footprint of digital transformation.

We help companies:

• Understand the real-world impact of AI on energy and water resources.

• Connect to a network of businesses focused on sustainable practices, including advanced heating and cooling for data centers.

• Take practical, positive steps - like investing in eco-friendly infrastructure or carbon offset programs - to ensure AI adoption is both powerful and planet-friendly.

Join us in building a future where technological progress and environmental stewardship go hand in hand.

Resource Use Estimation Formula

While exact numbers vary by task and model, recent research and expert analysis provide several benchmarks for understanding your personal AI footprint:

• Electricity:

  • A single ChatGPT query is estimated to use about 3 watt-hours (Wh) of electricity—roughly the same as powering a standard LED light bulb for 20 minutes.

• Water:

  • Water use is highly variable, but some studies suggest that generating a 100-word AI response could use about 500 milliliters (ml) (one standard bottle) of water for cooling, though this figure is debated and depends on data center technology and location.

Context:

• For comparison, the average US household uses about 30 kWh of electricity per day, and a single Google search used to use about 0.3 Wh - AI queries are roughly 10 times as energy intensive.

• Water use is especially dependent on the cooling technology used by the data center; new methods like immersion cooling are helping to reduce this footprint.

  ZDNet did a nice story on this complex topic. You can find it here.