Scaling Laws: Identifying the Myths and Facts of AI's Environmental Impact with Gavin McCormick 

In this episode of Scaling Laws, host Kevin Frazier sits down with Gavin McCormick, founder of ClimateTrace and former leader of WattTime, to dissect the complex environmental impact of AI and data centers. McCormick challenges the widespread myth that AI and large-scale computing are inherently bad for the planet, arguing instead that the real story lies in *how* and *when* energy is used. Drawing on his work tracking emissions via satellite and AI, he reveals that while data centers do consume massive amounts of electricity—some rivaling entire nations’ usage—their environmental footprint depends entirely on whether that power comes from fossil fuels or renewables. Companies like Amazon, Microsoft, and Google are leading the charge by investing heavily in wind and solar, often outpacing government mandates. However, outliers like Elon Musk’s XAI, which powers its Colossus 2 with fossil fuels, highlight the uneven landscape. McCormick emphasizes that the key to progress isn’t regulation for regulation’s sake, but outcome-oriented policies grounded in science and transparency. He also shares how ClimateTrace’s unique, decentralized coalition model—where nonprofits, tech firms, and governments co-create open data—has become the world’s largest emissions database, proving that collaboration, not coercion, drives environmental innovation. The episode concludes with a powerful message: doing the right thing for the environment doesn’t have to cost more—it just requires better data, smarter incentives, and a willingness to measure what matters.