Last year I rented a car and spent three weeks driving across the U.S. I wanted to see for myself what all the buzz was about. I met fleet operators, logistics teams, data center people. Amazon, UPS, Avis, FedEx. Everyone I sat with had electrification plans. Charging timelines. Rollout targets. Real budgets behind real deadlines.
But every single conversation ended in the same place.
Not "how do we build more chargers" or "which vehicles should we buy." The problem was simpler than that. And harder. How do we get enough power to the site?
The grid can generate the electricity. That's not the issue. The issue is that connecting new load takes years, not months. The interconnection queue in the U.S. is six years deep. And these operators need power now, not in 2032.
The edge is booming. The grid isn't.
Edge data centers are projected to be a $40-110B market by 2030. 5G needs local compute. AI inference can't tolerate a round trip to Virginia. Hospitals need real-time processing. Autonomous systems need millisecond response times. All of it is pushing compute out of centralized facilities and into cities, rooftops, basements, parking structures.
These locations were never wired for this kind of load.
Hyperscale data centers get to pick their location. Rural Oregon, west Texas, wherever the grid is open and cheap. Edge doesn't get that option. The whole point is being close to the user. Which means going exactly where the grid is already maxed out.
Diesel doesn't work here
The old playbook was always "drop a diesel genset on site." I heard this a lot. But look at what an edge site actually needs and diesel falls apart.
These are rooftops. Commercial buildings. Hospital campuses. A diesel generator runs at 82-86 dBA. That's a noise complaint before you finish install. It weighs over 3,000 kg and needs a crane to place. It only works as backup, not as continuous grid-parallel capacity. And good luck getting a tenant to renew their lease next to a diesel exhaust stack while the building owner puts out an ESG report.
Diesel was built for occasional outages at big industrial sites. Not for always-on compute at millions of distributed urban locations.
What I kept hearing
After that trip I came back and we started doing formal interviews. Over 60 conversations across 12 verticals. Edge colocation operators, telecom teams rolling out 5G, hospital networks, logistics companies. The ask was the same everywhere.
They need something compact. We're talking 15 square feet, not an industrial yard. Quiet enough for a hospital campus. Deployable in weeks, not years. Running on clean fuel. And smart enough to manage remotely when you've got hundreds of sites to worry about.
Nobody had a good option. They were either waiting on the grid and watching the market window close, or deploying with inadequate power and hoping nothing breaks.
The bottleneck was never building the power. It was getting it to the site on a timeline that matches the business.
Where this is going
There are over 100 edge operators globally. No category leader on the power side. The big equipment OEMs are chasing hyperscale contracts. They're building for 50-200 MW facilities with two-year deployment cycles. Meanwhile, the edge needs 200 kW to 1 MW at millions of sites, deployed in weeks.
That's a massive gap. And it's only getting wider as 5G, AI inference, and data sovereignty regulations push more compute to the edge.
Whoever figures out how to deliver compact, quiet, grid-parallel power at the speed edge operators actually move is going to own a very large market. The technology to do it needs to exist at the intersection of power electronics, multi-fuel flexibility, and software-defined fleet management. That combination doesn't exist in the market today. Not from the diesel guys, not from the fuel cell companies, not from battery storage alone.
But it will. And soon.
If you're running into this wall at your sites, I'd like to hear about it. Drop us a line.