Hyperscale Power and the Pace of Change Part 3a

Hyperscale: Power and the Pace of Change — Part 3A

Inside the Machine: What’s Actually Driving the AI Buildout

Jon Judah grew up in Bath County. He now advises some of the largest technology companies in the world. He came back to Allegheny Mountain Radio to tell us what the industry looks like from the inside — and how fast it’s actually moving.

By Danny Cardwell  |  Allegheny Mountain Radio  |  Supported by the Listening Post Collective

In Parts One and Two of this series, we heard from two people whose job it is to keep the lights on. Bill Buchanan and Jason Carter described a demand unlike anything rural electric systems were built to absorb. Hyperscale data centers drawing 95 percent more electricity than the largest industrial loads they currently serve. A potential load in Shenandoah Valley’s northern territory bigger than the entire rest of their membership — from a handful of facilities.

That demand has a source. Artificial intelligence. The companies building it. The hyperscalers — the Amazons, the Microsofts, the Googles — and the businesses running complex technology stacks inside those warehouses. To understand what’s driving that demand, I went to someone who works inside it every day.

Jon Judah is a Senior Vice President at DEPT®, one of the world’s leading digital agencies. He teaches Digital Communication Strategy at Georgetown University. And he graduated from Bath County High School in 1995. That last detail matters more than it might seem. The person helping us understand the technology reshaping rural energy systems grew up in one of those rural communities. He knows what’s at stake on both sides of this story.

Where the Industry Actually Is

When Jon was last on this program in January of 2025, most businesses were still trying to figure out where AI fit. Fifteen months is a long time in this space. His answer, when I asked whether anything had changed, was more measured than you might expect from someone inside the industry.

About a third of his clients have genuinely turned the corner — restructuring internal teams, shortening timelines, rethinking what they even need from an outside agency. The other two thirds are still figuring it out. More informed than they were, but not transformed. He referenced research from Deloitte suggesting that roughly half of people now have access to AI tools, but only about a third of organizations are actually using them to change how they work.

“The depth is still kind of catching up to the breadth.”

— Jon Judah, SVP, DEPT®

More people have access to the tool. Fewer people have changed what they do because of it. That gap — between access and transformation — is the honest picture of where AI actually is right now. Not where the headlines say it is. Not where the investment dollars suggest it should be. Where it actually is.

Where the Money Is Going

Then Jon mentioned something that connected his world directly to the one Bill Buchanan and Jason Carter described. The capital being spent by the hyperscalers. The pace at which products and platforms are coming online. Microsoft CEO Satya Nadella put the infrastructure constraint plainly on the BG2 podcast last November.

“The biggest issue we are now having is not a compute glut, but it’s power — it’s the ability to get the builds done fast enough close to power. I don’t have warm shells to plug into.”

— Satya Nadella, CEO, Microsoft — 

Power. That single word connects Jon’s world to the cooperative world. The hyperscalers need it. The cooperatives have to provide it. And the members of those cooperatives — the people whose bills reflect whatever happens next — are in the middle.

Two Lenses

I asked Jon to look ahead at what the next year might look like. His framework was one of the most useful I’ve heard for thinking about what AI is actually doing to work. He described two lenses. First: empathetic work — work that requires a human being to connect with another person, to understand context, to bring something a machine cannot yet replicate. Second: rote execution — the canonical, repeatable stuff, the processes and steps that can be documented and handed off. His company, DEPT®, is building an internal initiative around this distinction. Not asking whether AI will replace people. Asking how it changes the way work gets done.

“It’s making a whole lot of stuff faster. I think it’s also opening up a lot of new questions, and there’s a lot of anxieties around that.”

— Jon Judah

In Part 3B, we hear what that looks like from the inside — for the companies doing this work, and the people who work for them.

Listen to Part 3A below.

Part 3B: Inside the Machine — Agents, Obsolescence, and What Comes Next

Supported by the Listening Post Collective  |  MacArthur Foundation / Internews

All five parts will be available at alleghenymountainradio.org

Allegheny Mountain Radio  |  Listening Post Collective  |  Danny Cardwell