AI data centers are being announced and built at a pace the U.S. electric grid was never designed for — and the biggest constraint isn’t generation. It’s substations and grid infrastructure.

In this episode of The Clean Energy Edge Podcast, Russ Bates is joined by Ben Watkins, P.E., Vice President at ARM Group, to break down what’s really happening behind the scenes as AI-driven load reshapes the power system.

This conversation goes beyond headlines to explain why AI data centers are fundamentally different from traditional data centers, how extreme load volatility stresses substations and transmission systems, and why grid upgrades are becoming the pacing item for development across the U.S.

The key differences between traditional vs. AI data centers

Why AI training workloads create massive, near-instantaneous load swings

How substations actually work — and why they’re a critical bottleneck

Why upgrading one substation can trigger cascading grid impacts

New vs. existing substation upgrades and siting challenges

Supply-chain and equipment lead-time constraints (transformers, breakers, insulators)

Why engineering decisions are increasingly driven by timeline, not optimization

The role of batteries and fast-ramping generation in managing AI load

Why early engineering coordination is essential for successful projects

What developers, utilities, municipalities, and communities should understand before approving AI data centers

Ben works at the intersection of substation engineering, grid infrastructure, and large-scale project execution, supporting complex energy and industrial projects across the U.S.

ARM Group provides multidisciplinary engineering, environmental, and project support services for complex energy, infrastructure, and industrial projects nationwide.

🌐 Website: https://www.armgroup.net/

📧 Email: [email protected]