The Cost of Intelligence: Generative AI's Energy Footprint

Generative AI has seen explosive adoption in recent years—but behind the sleek interfaces and impressive outputs lies a growing concern: energy consumption. Some estimates suggest that the environmental cost of these models is staggering. But how accurate are these claims?

In this talk, Danni Dromi and Kasper Groes take a closer look at the energy footprint of generative AI. We'll explore what we do know, what remains unknown, and why the lack of transparency from major tech companies makes it hard to assess the real impact. Drawing on real-world energy measurements from open-source models, we’ll shed light on where the numbers come from—and where they might be taking us. Our goal is to move beyond speculation and toward a more grounded, constructive conversation about the climate impact of AI.

Bio: Danni Dromi is an independent AI consultant and founder of Fine Chariot, with over 10 years of experience in AI and technology. He has led projects for clients including Novo Nordisk, Rockwool, and Teknologisk Institut, and regularly delivers talks and workshops on generative AI across industries. Danni is part of the board of Danish Data Science Community (DDSC).

Kasper works as a scientist and LLM engineer and is co-founder and board member of Danish Data Science Community (DDSC). He's a nuanced source on the energy and climate impact of AI and his work has been widely used in media like Washington Post, The Economist, Børsen and Mandag Morgen. On his own blog, his writings about green AI has been read more than 60,000 times.