This week on Cleaning Up, host Michael Liebreich sits down with Varun Sivaram and Steve Smith to explore one of the most urgent, and overlooked, challenges of the AI revolution: how to power it without breaking the grid.

As AI demand explodes, hyperscale data centres are emerging as massive, inflexible loads, rivaling entire cities. But do they have to be a burden on the grid?

This conversation dives into a groundbreaking trial led by Emerald AI in partnership with National Grid and NVIDIA—demonstrating that data centres can dynamically adjust their power consumption in real time using software.

Key insights include:

How AI data centres could reduce grid stress instead of increasing it
The concept of “flexible demand” and why it’s a game changer for AI data centres
Real-world trial results
Why “speed to power” matters more than cheap electricity in the AI race
How software, not infrastructure, could help unlock billions in grid capacity
The hidden flexibility inside AI workloads (and why not all compute is equal)

From kettle spikes during football matches to lightning strikes on the grid, this episode reveals how intelligent systems can respond in seconds, turning a looming energy crisis into a massive opportunity.

Leadership Circle:

Cleaning Up is proud to be supported by its Leadership Circle. The members are Actis, Alcazar Energy, Arup, Copenhagen Infrastructure Partners, Cygnum Capital, Davidson Kempner, Ecopragma Capital, EDP, Euroelectric, the Gilardini Foundation, KKR, Mitsubishi Heavy Industries, National Grid, Octopus Energy, Quadrature Climate Foundation, Schneider Electric, SDCL and Wärtsilä.

Read more:

The Emerald AI/National Grid white paper: https://www.ngpartners.com/stories/emerald-ai-whitepaper
The $60 Billion Plan For Europe’s Largest AI Data Centre | Ep235: Robert Dunn: https://www.youtube.com/watch?v=juAyLAUmU3w