Issue link: https://iconnect007.uberflip.com/i/1538540
22 PCB007 MAGAZINE I AUGUST 2025 You see them in some out-of-the-way places like Des Moines, Iowa, or Prineville, Oregon. Some of the power shortage problems that we're having are caused by these data centers being in the hin- terlands instead of being concentrated in the hubs. Everybody wants to be green, so they're contract- ing with the power companies so that when they build a data center, they're utilizing green power, like solar and wind. It seems like nuclear is coming back into play in a strong way. Nuclear power is on everyone's lips. There are something like 90 companies thinking about using nuclear power. That's because wind and solar won't meet the demand. With AI driving this, the demand is as large as ever. This is true, though, the AI demand on the power grid is strange. There are two types: the training of the AI and the actual production. The power signature is unlike anything we've seen before, and it's causing a bit of a problem with utilities when they hook up. They can suddenly shift power from full load to no load in practically milliseconds. That's how these things operate. They operate in small, tiny batches of computation, and the on-off switches are quite bipolar. Flicking them on and off can wreak havoc with load control and normal utility. Going back to the nuclear issue, many people are toying with the idea of small modular nuclear power plants. The problem is that people really don't know how to run those things, and there's a huge licensing issue. Only a few companies have actually acquired licenses to build and operate these things. I see a really slow learning curve with nuclear power. Meanwhile, the established nuclear power plants—the ones currently feeding our grid—have a really good idea. The land surrounding nuclear power plants is empty, and we could invite data center companies to build their campuses there. We would be able to feed them power ahead of the meter. That's a great idea. Mike, You've been an electri- cal power engineer for 40 years. That's a lot of wisdom gained and lessons learned. What advice would you pass on to a young engineer? Power engineering has been kind to electrical engineers, more so than electronics or computer science. You can become obsolete in electronics, medical applications, or software because it changes so quickly. But power engineering has changed a lot less over time, so the things you learned in school are still relevant 20 or 30 years later. I would recommend that a young engineer who is interested in electrical engineering look into power engineering, especially in things like how to make the grid more stable. Our grid is pretty fragile? It's horribly so and it's getting worse.