Google and Amazon have seen their carbon emissions increase by 16% and 18% in 2025. The main cause is the massive energy demand from data centers powering artificial intelligence. This growth jeopardizes both companies' climate commitments and, in turn, hinders the global fight against climate change.
The energy cost of training an AI ⚡
Each query to a large language model consumes up to ten times more electricity than a traditional internet search. To maintain response speed, companies deploy thousands of GPUs running 24/7. This consumption forces the use of fossil fuels when renewables cannot meet demand. The result is a carbon footprint that grows at the same rate as the sophistication of the algorithms.
AI will save the planet (but first it warms it up a bit) 🔥
Now it turns out that for an artificial intelligence to recommend the most eco-friendly route by car, we first have to burn coal in a data center. It's like asking a plumber to fix a leak by turning on all the faucets. While tech companies promise a green future, their electricity bills and emissions are skyrocketing. Ironies of progress.