Cause
AI for Climate Action
energy reduction in Google data centers using DeepMind AI
9 reported stories
AI is doing three jobs for the climate: measuring what polluters never reported, forecasting weather and disasters days earlier than physics models could, and cutting the energy that machines themselves use. These stories follow the projects with published numbers, from Climate TRACE's emissions map to GenCast's weather forecasts.
Google DeepMind cut data center cooling energy by 40% using AI. The WIFIRE Lab at UC San Diego predicts wildfire spread in real time, giving firefighters hours of warning they never had before. Microsoft's AI for Earth program has funded over 900 projects tracking deforestation, biodiversity loss, and carbon emissions across 100+ countries.
Stories
Climate · 7 min read
Climate TRACE Tracks Emissions From 745 Million Sources the Polluters Never Reported
Climate TRACE uses satellites, open data and machine learning to track greenhouse gas emissions from 745 million sources and publish facility level data.
Climate · 8 min read
Spherical DYffusion Simulates 100 Years of Climate in 25 Hours
UC San Diego and Allen Institute researchers built a generative AI climate emulator that projects a century of climate in 25 hours, with no supercomputer.
Climate · 7 min read
MethaneSAT and Google AI Exposed Hidden Methane Leaks From Space
The Environmental Defense Fund's MethaneSAT, built with Harvard and paired with Google AI, shows oil and gas methane emissions far above official inventories.
Climate · 7 min read
NVIDIA Earth-2: The Digital Twin Predicting Extreme Weather Before It Strikes
NVIDIA's Earth-2 platform uses generative AI to simulate Earth's climate at kilometer scale, helping agencies like Taiwan's forecast typhoon landfall.
Climate · 7 min read
GenCast: How DeepMind Built a Weather Forecaster That Outperforms Supercomputers
Google DeepMind's GenCast outperforms ECMWF on 97.2% of targets, runs 15-day forecasts in 8 minutes on a single TPU, and is now open-sourced for the world.
Climate · 7 min read
FireSat Can Spot a Wildfire from Space
Google Research and the Earth Fire Alliance are deploying a satellite constellation that detects 5x5 meter fires and scans the entire planet every 20 minutes.
Climate · 8 min read
Europe is Building a Digital Twin of the Planet
The EU's Destination Earth initiative combines AI with supercomputing to build a digital replica of Earth's climate. Phase 1 was delivered in June 2024.
Climate · 6 min read
Google Flood Hub: Protecting 460 Million People
Google Flood Hub uses AI to deliver flood forecasts up to 7 days ahead in over 80 countries, with street level predictions for governments and communities.
Energy · 6 min read
Atlas AI: Solar Where It Matters Most
How satellite imagery and machine learning are helping governments and NGOs target solar installations for the 770 million people who still lack electricity.
Questions people ask
How is AI used for climate change?
AI tracks emissions from satellites and sensors, forecasts weather and floods faster than traditional models, finds methane leaks from space, speeds up climate simulations, and reduces the energy that data centres and grids use. Each story on this page covers one project with measured results.
Can AI solve climate change?
No single tool can. AI makes measurement, prediction and efficiency better, which gives governments and companies earlier warnings and harder evidence. The cuts in emissions still depend on what people do with that information.