So you and I spent a lot of time together in Abilene, Texas. Side of Stargate, which people here aren't familiar, is OpenAI's big data center with SoftBank and Oracle and Crusoe. You're the developer. So I just wanted to start there. If you haven't seen the premiere episode of The Circuit, we go to Abilene and you can see lots of beautiful shots of data centers rising from the red dirt. But I just wanted to start there and talk about, even in the last two months, the progress since then, because you're moving very quickly.
Yeah. No, it's great to see you again, Emily. Even, you know, you look great without a hard hat on, too, but. Ah, yeah. So, you know, things have been moving very, very quickly in Abilene. You know, at this point, I think I was like, looking at the just like the number of hours that I've been worked. I think it's we're roughly working 200,000 hours per week, 200,000 hours per week by 5,000 people. It's going to get to 5,000. But today's like around 3,500 and total in aggregate. In the last year, we've worked about 3 million man hours. So it's like a ton of work and, you know, hard work labor and effort's gone into sort of making this a possibility. But, you know, the first two buildings which, you know, we announced last summer, you know, are approaching for completion and the, you know, the following six are kind of being stood up as we speak. So a lot of the area that we went through that, you know, they were doing a lot of the earthwork there are now steel structures. They're building three kind of has the wall panels on it. And we're moving, we're continuing to move very, very quickly.
So this is not only like maybe the biggest data center ever, but the fastest built data center ever. I mean, you guys started last summer. Yep. And when's it going to be powering our chat? Yeah.
So. Ah, so I don't know that the exact justice. I mean, people have told me it's the largest data center financing ever. We announced a $15 billion facility between Blue Owl as our project capital partner and JPMorgan as the credit on that on the deal but we, you know, as far as speed goes that's something that, you know, Crusoe's, you know, very focused on how do we actually accelerate innovation by bringing the infrastructure needed to power it to market very, very quickly. And so sort of the numbers that I was told was the fastest 100 megawatt or greater data center had ever been built was roughly two years. The first two buildings, which are each of our megawatts, is roughly 300 days. And then the next six are a little over 200 days. So I think it's about 215 days per building. So all to say things are moving quickly.
So President Trump announced this the day after the inauguration with Sam, with Masayoshi Son, with Larry Ellison. You didn't necessarily know it was coming. So take me back to that day when Abilene was uttered by Donald Trump.
So, yeah, you know, definitely like caught us by surprise. But, you know, we were sort of happy to be affiliated with the project. I was actually had a dinner with a bunch of folks in Switzerland. And, you know, I sort of had my phone off and I came out of the dinner and I had like 15 missed calls and like 25 text messages from all these people saying, hey, is this your project? Like, you know what? What's going on here? Like, you know, I just watched this announcement in the White House about this thing called Stargate. And I was like, wow, I don't even know what's going on. So it definitely was an interesting moment that really catapulted the project into the limelight. But, you know, we've been just thrilled to be a part of it and, you know, are excited to build this very, very quickly.
So in the episode, we also interview Sam. When we interview Masa, you're kind of our tour guide. How did you and Crusoe end up in the middle of this massive deal?
Yeah, I think, you know, I started Crusoe about seven years ago and our goal with the business was really taking this energy first approach to high performance computing. And, you know, I've been an AI researcher, you know, mostly focused on finance for about a decade before that. And I came to appreciate just how much energy a lot of these larger deep learning models used when you sort of were training them at scale. And so, you know, our philosophy was always kind of taking this energy first approach. And, you know, as ChatGPT took off and as the demand for AI infrastructure really, you know, went parabolic, people were very, very capacity constrained. There wasn't enough energy in markets like Northern Virginia. And so, you know, Crusoe was well positioned for this. And in that we had sort of lined up a lot of large quantities of interesting power sources in nontraditional datacenter markets like Abilene, Texas. And our goal was really how do we instead of trying to build the next data center in Northern Virginia, how do we actually find power so that we can build really big facilities in markets like Abilene? And, you know, Abilene is very unique in the sense that there's a ton of wind that's been built there on the back of production tax credits. And wind developers were incentivized to come there and they were sort of getting paid to produce clean power. And one of the big problems is actually they don't have enough demand or transmission to get that power out of that market. And so, you know, it was sort of a perfect fit for a large data center. And, you know, we sort of, you know, were well-positioned to capitalize on this. And we had some interesting conversations with some of the partners and everybody sort of locked arms. And we said, let's go build the world's biggest AI data center.
Let's talk about how tightly the arms are locked. Because, you know, we did talk to a lot of people in Abilene. Residents were mixed. I mean, some people don't really know for sure what it is. They just know this big data center has been built. We talked to a lot of city officials who agreed to give up a lot of tax revenue in order for this data center to be built there. However, nothing would be there otherwise. And so in their view, it's still bringing in millions of dollars that they wouldn't have. I got some calls after the piece aired from folks who are really worried about the power and water demands that this will put on the area. You know, going back to the Texas power crisis in 2021, where like hundreds of people died during blackouts, if the power goes out in Texas, what happens? Does Stargate and ChatGPT stay on and do the residents lose their power?
Let's see here so you know where to start. So the first thing I'll say is that, you know, from a tax revenue standpoint, you know, the city of Abilene's been a great partner to us and Taylor County is a great partner to us. And, you know, as this facility gets built, we're going to make up about a third of the overall tax budget. So we are a very large contributor to the tax revenue in the city and about 25% of tax revenue in the county. So we are, I think we're the largest single tax contributor to, you know, the local economy. So from that perspective, you know, I feel like we're definitely contributing quite a bit, you know, very meaningfully in terms of our, you know, the overall tax revenue picture. You know, from the power and water perspective. You know, I think we talked about it quite a bit when we're on site. You know, I'll start with the water on the water side. You know, we actually purposely designed it as a closed loop architecture in the data center so that we, you know, the difference being a closed loop and an open loop system is an open loop water cooled system is where you have cold water feeding in. And then you're sort of, you know, creating hot water through the cooling process and then you're just basically disposing of that hot water and you have a closed loop system that actually goes through these chillers that basically takes the hot exhaust water and then cools it back down and then recirculate it through the system. So while each building requires a million gallons of water, you only have to fill it one time and then it's only one check because people are like, that's total B.S. There's no way. Yeah, I don't know what to tell them, but that's all right. That's just how it works. You know, there's like probably moments of topping it off a little bit. But this is like every few years. It's not like it's a very, you know, incremental amount. Now, on the power side, you know, we've gone through all of the appropriate processes to basically get, you know, the ERCOT approved power load. All of the load studies have been done across the grid to basically ensure that this load is approved. And, you know, it's not going to disrupt the grid in any sort of meaningful way. And then we've also invested in backup generation on site. So we've invested and we actually invested in gas generation as a backup resource, which is much cleaner than diesel. We've invested in all of the, you know, most aggressive mechanisms to eliminate pollutants. We've added something called SCRs to our gas driven stands for selective catalytic reducers that basically eliminate almost all of the NOx and the carbon monoxide, the two biggest contributors to smog, so that when we do have to run on the gas turbines, it really mitigates a lot of the local air quality issues that may be associated with, you know, that fossil based generation.
Now, I want to talk, you know, a little more broadly about you just raised money. You've got a big cloud business. There's, you know, CoreWeave which recently went public, which is not totally apples to apples, but, you know, you definitely do overlap. You know, can you talk about sort of the bigger picture for Crusoe amidst this data center boom that is not just happening in the US data center boom but around the world.
Sure, you know, Crusoe's an American business first and foremost, but we do have a global footprint. You know, we've taken a very vertically integrated approach to artificial intelligence, recognizing that, you know, everything is really had to be reinvented from, you know, the way you source power to the way you design the data center, to the way you build and operate and manage these clusters with, you know, physical infrastructure as well as high performance software. So from that standpoint, we've sort of taken a full stack approach. You know, certainly there are elements of where we're core. We definitely appear to Crusoe and I think we've been, you know, sort of rooting on their stock prices. You know, it serves as a great comp to us and really proves out the success of the business model. But, you know, I think our biggest footprint is in America, but we are, again, taking that energy first approach. We've made massive investments in Iceland where you have, again, abundant, low cost, clean geothermal and hydropower to power these large clusters and GPUs that we serve customers both in the US as well as in Europe.
