**Title**: Energy in the North - Sam Enoka
**Date**: March 25, 2026
**Participants**: Amanda Byrd, Sam Enoka

00;00;00;12 - 00;00;09;02 [Sam Enoka] Our deployment will actually result in helping put downward pressure on rates for electrical rates for the entire community.

00;00;09;02 - 00;00;26;15 [Amanda Byrd] This week on energy in the North, I continue my conversation with Sam Enoka, founder and CEO of Greensparc, a data center infrastructure company based in San Francisco. Greensparc has one data center in Alaska, and I began the conversation by asking Sam where his first installation is.

00;00;26;15 - 00;01;34;03 [Sam Enoka] our first project is in Cordova at Humpback Creek, and we're currently looking at expanding and building another deployment in Cordova. That would be roughly six times the size of the first one. So the first one was only 170kW. The next one that we're building will be about 700kW. So we've been in the planning stages with Cordova Electric, for several months. And to make sure that what we develop in Cordova is a win for Cordova Electric, it's a win for their ratepayers because our deployment will actually result in helping put downward pressure on rates for electrical rates for the entire community. And it'll be a win for us because we get to work with innovative utility manager like Clay Koplin. And we get to be, in a beautiful community like Cordova. So it's a win for us. It's a win for the community and the ratepayers, and it's a win for the system and the data.

00;01;34;03 - 00;01;40;25 [Amanda Byrd] We've talked about the energy sizing, but the data that is, I don't know, being generated or stored or whatever at these data centers. Is that cloud data?

00;01;40;25 - 00;02;42;21 [Sam Enoka] It is. So, the reality is that right now, in the world of AI, customers that need that technology are global. There are customers that can't wait for our, our new Cordova expansion, to be energized because the line is already forming. For the customers that want to use those Nvidia chips that we'll be plugging in. So that's, you know, that's wind in our sail for you know, every, every community that we examine, if we can land some AI infrastructure there, we know that there are customers somewhere in the world that want to use that infrastructure. So it's great to know that we have customers lined up and waiting no matter where we land. If it's domiciled in the United States, we have a long list of customers that can't wait to get access to it.

00;02;42;21 - 00;02;58;08 [Amanda Byrd] And so 170 kilowatt hours and 700 kilowatt hours. How much data are we talking about?
Unknown


00;02;58;08 - 00;03;35;16 [Sam Enoka] That is about, at 700kW and, and that's probably on the hottest day in the summer in Cordova for us. In the all of our infrastructure will be running. That'll be 192 of thelatest generation of Nvidia chips. So B300, that that will stack in there, in a room about the size of a small garage. So it's about 20 by 20. And those chips are all fit into a total of seven cabinets. So space wise, not very big. Very compact, but the size of the cluster, might make it one of the largest air clusters in Alaska, I think, and it will be, a cluster big enough to service, a medium or even a large enterprise that has, you know, AI projects, somewhere on their on their project board.

00;03;35;16 - 00;03;39;03 [Amanda Byrd] So gigabytes?

00;03;39;03 - 00;03;39;25 [Sam Enoka] Oh, terabytesfor sure.

00;03;39;25 - 00;03;42;21 [Amanda Byrd] And then how much data does an Nvidia chip store

00;03;42;21 - 00;04;44;02 [Sam Enoka] It doesn't necessarily store the data, as much as it processes the data. So there's, a heavy component to an AI deployment is where the data lives. And so we'll, we'll be building, we'll be bringing a lot of storage to go along with, all those Nvidia chips. but you can think of each generation of Nvidia chip really being about anywhere from, you know, 20 to 70% faster than the generation prior. So these chips are getting you know, faster and faster every generation. And they're consuming, an incredible amount of power - almost one kilowatt per chip, almost. In some cases. So these are not the chips that are running in, our MacBooks or our laptops or, iPhones, if you will. But, but these are the, the workhorses of the, you know, new AI models and, and technologies that are in the market today.

00;04;44;02 - 00;04;52;06 [Amanda Byrd] Sam Enoka is the founder and CEO of Greensparc, and I'm Amanda Byrd, chief storyteller for the Alaska Center for Energy and Power at UAF. Find this story and more at uaf.edu/acep.