Operator: Our next question comes from the line of Mike Grondahl with Northland Capital Markets. Please proceed with your question.
Mike Pochucha: Hi, this is Mike Pochucha on for Mike Grondahl. Thanks for taking our questions. Maybe just on Jamestown, you talked about the kind of retrofitting for that Web 3 application customer. Can you talk about how difficult that? Is that more or less plug and play? Or how much has to go into that?
Wes Cummins: Yes, it’s not plug and play. We put some extra walls up a little extra filtration, but it’s point being, it didn’t take long. It took about seven weeks to do that. So if we needed to do more, more of that. It was one we needed. We were trying to find space for this. We wanted to put it online pretty quickly, because we wanted to wanted to get on this application fairly quickly. But two, I thought it was a nice exercise as far as can we retrofit some of this space? And can we run GPU CPU inside of it? And we can’t. So but it didn’t take that long. And it’s not, it’s not wildly expensive either.
Mike Pochucha: Got it. And then just Jamestown more broadly, you talked about the kind of 5 megawatts with the machine learning use case. There are a certain amount of either like space or bring in new power, like at a certain megawatt level, or become kind of a higher CapEx type of field to add more to Jamestown.
Wes Cummins: Yes, so if we want to expand more HPC or Bitcoin at Jamestown, we’re going to have to spend some additional CapEx on the electrical infrastructure, mostly running align, it won’t be prohibitively expensive. It’s not like a substation or anything like that. So, but we will have to spend some extra. But I do hope that we’re doing that in the second half of this year. Because that’ll mean that the HPC stuff is growing as fast as we expect it to.
Mike Pochucha: Thanks.
Operator: Our next question comes from the line of Kevin Dede with H. C Wainwright. Please proceed with your question.
Kevin Dede: Hi, Wes, thanks for taking my question. On the HPC side, it seems like you’re your go-to-market strategy is on power costs. I was wondering how your customers balanced that sacrifice or that gain for the sacrifice and latency? And it seems that even though demand is high for you, how, how do you market the fact that you’re pretty much still non redundant facility? I guess I’m whatever additional color you can provide to sort of set your, the applied offering versus say AWS or Azure?
Wes Cummins: It’s, it’s a great question. So there’s a couple of things that set us apart not just the power cost, it’s going to be the total cost of the infrastructure, right. So the infrastructure is built specifically for this, the applications do not require ultra-low latency. So the machine learning and AI applications just don’t require that nearly as much. Many of them are even interruptible, similar to how the Bitcoin mining is. We don’t expect them to be in or interruptible. But we can run it in that fashion. We’re going to give a lot more details as the year goes by. But we have a partnership was formed with a software company, a software development company that has a specific software that really is necessary to run these type of machine learning applications in our style of datacenter.
Ultra efficiently. And so, I think this is going to come down to just a game of cost of compute, and we’re playing to be the lowest cost compute provider. We get to put these facilities in North Dakota. We’ve low cost power, again, for purpose built for this, the facility is designed for this. And we’re doing it around wind power. And the other big component here, Kevin, we’ve talked about is that we’re going to use air cooling for a vast majority of the cooling here because of the climate. And that would that there’s a couple of things. The electricity usage at most data centers is about 50% for the compute and networking etcetera, and 50% for HVAC. And so if we can take that down to you know HVAC being less than 10%, or even lower than that, that’s going to be another significant cost saving.
But it’s also what some people refer to as green computing or truly green computing where we’re using primarily renewable energy. We’re using air cooling instead of using electricity to burn — for HVAC. And so from that perspective, which the customers we’re talking to, they absolutely care about that aspect of it as well. So if you’re, if you’re a significantly lower cost, truly green computing and provide performance that’s on par with anything else they can they can sign up for, for the applicant for these specific Apple applications, I think we’re in a really good position to compete and the kind of the conversations we’re having tells me that we’re on the right track here.
Kevin Dede: Okay. On the Bitcoin mining side, Wes, we’re looking at a half price that’s, I think, 25%, of where it was through the last downturn, or the last winter. I guess, listen, it seems to me that you’re pretty fully booked up. But it just also seems to me that you would be getting some pressure from your customers, on helping them manage their profitability. And I know, you alluded to it a little bit, with Bitmain, and Marathon and just doesn’t seem that you’re getting that kind of pushback. But it’s also a little bit difficult, at least from the outside to get my arms around that. Can you add anything to how you’re managing those negotiations?
