“We all need to do more.”

As governments and businesses accelerate AI adoption, concerns around energy and water demand are intensifying. Microsoft leaders Melanie Nakagawa and Juan Lavista Ferres highlight new research showing that advancements in AI models, datacenter operations, and hardware could improve energy efficiency by 8–20x.

The findings suggest that, at scale, AI systems can become significantly more efficient than previously understood—enabling continued growth while reducing the resources required to support it and reinforcing a more sustainable path for AI expansion.

Read the full blog here.