Can AI Save AI Infrastructure? Cutting Energy, Water, and Wear in Data Centers

Data centers have always pursued energy efficiency through better hardware—smarter chillers, advanced cooling systems. But there's a ceiling. You can only make hardware so efficient.

Seven years ago, Jasper de Vries discovered the butterfly effect in data centers—something on a roof rippling through 300 billion sensor readings down to valves in server rooms. His company, Lucend, ingests that sensor data to generate operator recommendations. One facility cut power usage by 40% in a year, saving $4.3 million.

Yet here's what most of us miss about AI's big energy problem: we focus on operational energy use while Scope 3 emissions—the embodied carbon from manufacturing hardware—creates massive impact, so much so that Microsoft won't hit its 2030 climate targets because of its data center growth plans. With JP Morgan projecting $5 trillion in AI infrastructure buildouts by 2030, the need to bring embodied carbon under control is urgent.

Lucend's software addresses both challenges: it slashes operational energy while extending hardware life through predictive maintenance, reducing the physical wear that forces early replacement. Its technology is now deployed across over 50 facilities globally.

Show Notes

Guest: Jasper de Vries

Company: Lucend

For more low-carbon innovations now scaling—and the playbooks driving their market adoption—subscribe to the podcast plus our:

* Weekly Newsletter

* Climate Adoption Playbook

* Supercool on Instagram 

* Supercool on LinkedIn