Why scientists should lead the shift away from AI mega data centres Publicly available AI models and local infrastructure can reduce AI’s environmental footprint while giving researchers greater control over the tools they use. By Cassidy K. Buhler, Fernando Pérez & Carl Boettiger

https://www.nature.com/articles/d41586-026-02451-2?utm_source=Live+Audience&utm_campaign=2923aae61a-nature-briefing-daily-20260811&utm_medium=email&utm_term=0_-33f35e09ea-50432164 Researchers who work with AI tools must pioneer the use of open-weight models that can be run on personal computers, write three data scientists. These models don’t rely on data centres that have enormous energy and water demands, and so can make the use of AI in research more sustainable. Educators can facilitate this change by teaching students how to determine when local models are preferable to cloud-based systems and how to develop applications that run efficiently on local hardware, the authors write.