



CARSON COUNTY, Texas (KFDA) - International investment company InfraTech announced today plans for a $2.7 billion data center in Carson County.
The center would be in addition to Fermi’s Matador facility near Pantex. InfraTech said in a statement that it would soon announce its first tenant and investors when details are worked out. The statement says the new data center will be on more than 5,000 acres.
Copyright 2026 KFDA. All rights reserved.
Updated, sorry guys the AI summary was trash. Here is the word for word with a few unrelated things taken out.
Fermi IR Call — Questions & Answers
Question 1
Question:
On Hillcore, what economics does Fermi retain on every megawatt ultimately sold to a data center tenant?
Rodrigo:
So here, we haven't broken out specific price per megawatt spread publicly, but think about this deal as—imagine, like a good way to imagine how this deal works is imagine if Excel told us, Excel, the local utility, I will give you 2.5 gigawatts more. You won't pay for anything. You are going to be able to lease more power land, more power shells, or more turnkey based on that availability of power.
So the way Fermi makes money on this is it allows us to lease data center or power land faster. And Hillcore, the way that they make their money is similar to the utility. They charge their price per kilowatt hour. It is a straight pass-through. Fermi will make money there, unless Hilcorp would have a base power charge that we haven't disclosed. But if Fermi is able to negotiate a higher power price, we'll make that margin.
But the reality of this deal is that this allows us to bring NOI faster from leasing, from rent activities, faster, and that is the goal of this deal. If we can make some money on some margin on the power, of course, you know, we'll always negotiate to make it as attractive as it can be for us. But the goal of this is allow us to bring NOI for leasing activities faster without Fermi putting any capex or operation capital towards that part of the power.
On top of that, we haven't disclosed that, but there's going to be a rent that Hilcorp pays on the footprint that they have on the site. But again, we haven't disclosed that, but that is not the goal for this deal. It's more about the economics that it unlocks for us.
So I'll pause there to see if you have any follow-up questions.
Me:
No, I think it's a great deal, especially with how much demand is out there right now and how much demand will be out there the next two or three years. Being able to capture some of that stuff early and sign contracts long term, even though they're going to be building up the power kind of separate from what y'all are building, going ahead and getting tenants leased to the power that they're building right now and signing up that demand while it's out there, I think it's genius. I think it's a great deal. I'm very excited that y'all are able to find something like that.
Rodrigo:
We see it like that, and again, we're incredibly excited with the partnership. And again, we'll, in the next, I don't know if today, but definitely this week, you'll see a release from us with a PD of Hakopo explaining this as well. But again, we just want to give more clarity on Hillcore.
If you don't mind, I'll go to the next one.
Question 2
Question:
Can you help bridge the 5.5 billion tensor wave phase one contracted revenue?
Rodrigo:
What I can tell you here is that these 6.5 billion on the 15 years is the base rent plus a fixed power charge. So there's not going to be economics on top of this 6.5 as it relates to variable power charge. That variable power charge is going to be a straight pass-through.
So what we have disclosed of the 6.5, that is the economics for Fermi on the deal, and we haven't also told the market the breakdown, the exact split of what is base rent and what is the power charge.
As I mentioned, you know, the tenant, it is very sensitive for us releasing this breakdown. We'll see. I cannot guarantee this, but we want to be more forthcoming with project economics. But after the—
Me:
The guarantee has been announced.
Rodrigo:
Gotcha.
Me:
Yeah, that'll be good. Yeah, yeah, it's, I mean, it's early. I, as much as anybody, love, like, you know, punching stuff in Excel, kind of seeing where the NOI will go over time and kind of trying to paint the picture. But again, it's still early days, and over the coming months and years, we'll be able to dial that stuff in a lot more.
I cannot give you a timeline, but definitely it's on the top of our list to give more clarity on economics. So, you know, if you write about this, that I expect you will, don't put a timeline. It's, you know, it's coming.
Me:
Okay, absolutely. Yeah, that'll be great.
Question 3
Question:
What exactly does the investment grade backstop guarantee, and does it run for the full 15-year lease?
Rodrigo:
The guarantee is signed to backstop the tenant contractual obligation under the lease. This is both the rent and the fixed power charge, not the variable energy pass-through.
Again, this is something that, as you heard from the call, we wanted to announce the deal. That was the important thing for us under 90 days. We understand that a very important thing for this to get built, especially when the tenant is not an investment-grade counterparty directly, is to announce who the backstop is.
That is, you know, upcoming. We cannot build and we cannot project finance until we get this, so expect updates on that soon.
Me:
Okay, great.
Question 4
Question:
Once the backstop is executed, what conditions remain before the lease is fully financeable and Fermi can issue the full notice to proceed?
Rodrigo:
After the backstop is announced, we go and run to our project finance syndicate and to our bank partners to determine what is the best way to project finance this. The gating item here is the backstop. Once that happens, with the guarantee and the counterparty that is guaranteeing, we don't expect to have any issue for getting the financing need to build this project.
Me:
Okay, great. Yeah, I remember Anna always saying that she was kind of running those two things in parallel with each other. As they were doing their tenant conversations, they were also checking up to make sure everything was going to be financeable and that they were going to be able to get project finance. I remember her saying that she kind of runs those at the same time, so I figured it wouldn't be too difficult or it wouldn't be like surprising a bank or something that they got this contract, you know.
Rodrigo:
No, no, no. And again, you can imagine the opposite has also happened. Like the big banks that we're talking for the syndicate, they knew who we were talking with. They have an idea and understanding of who the backstop is. But once we announced the tenant, we have also received the amount to, hey, we're very interested in being part of that financing.
So this is the type of deals that Wall Street loves, right? There's an investment-grade backstop. There's off-takers. You know, this is not going to be hard. It's going to be time-consuming, of course, because again, this is a separate process, and we have tried to advance it as much as we were able throughout the negotiation.
But the bank needs to do their own diligence on the documentation and the power and all of that. But again, we expect this to happen, right? It's the bread and butter of large infrastructure for Wall Street.
Me:
Right.
Question 5
Question:
Has the investment thesis shifted from can you find customers to how fast can you build capacity for the customers you already have?
Rodrigo:
Yes and no. So that is a fair read, and we have the risk demand side with the tensor wave, but the broader pipeline is still very important for us.
We think that we have all the conditions and capacity currently to have tensor wave extend to the full 660 megawatts of the three phases of the project. But given that we also have, from the power that we own and control, 1.6 more gigawatts, and from the Hillcore, 2.5. So commercial activities don't stop until we lease all that power, and then we go and get more and we try and lease more.
So definitely our operational focus is on delivering the power on time and on budget, but our commercial focus is strong as ever.
Me:
Awesome. One quick question to plug in there. So I know the Hillcore power, the 2.6 gigawatts, I think y'all said that was expected to be 30 months on that construction, based on, you know, different factors. But was y'all's 2.2 gigawatts y'all have, what was the timeline y'all were thinking on? And I know it depends on tenant signing, project finance, but were y'all thinking 24 months on putting that 2.2 gigawatts, that construction?
Rodrigo:
That is all contingent on lease agreements being signed and project finance. There is, we understand there's currently a mismatch there in between where... When we have said that that full power that we control, especially the first 1.5 gigs, can be in place versus what's contracted.
So that is some feedback that we have received this week. We will be communicating to the market soon how those two different power ramps collide.
Me:
Okay, great.
Rodrigo:
So, you know, give me a little bit more time for that. We know where we need to be for the 660 for TensorWave, but for their, call it 900 more of the simple cycle 1.5, that is, of course, contingent on commercial agreements.
What I can tell you is a good place to see that is on our company presentation that we released for the Q. We have the slide that has the—
Me:
The power ramp.
Rodrigo:
The power ramp. Well, not the power ramp, the power, the generation portfolio, and you can see there time to power from contracted. That is the fair assumption of when can this be online assuming contracting. I know it doesn't answer the question because the catalyst for a full power ramp is commercial.
Me:
Right. Okay.
Question 6
Question:
Broad overview, what are the big things you're watching for now?
Rodrigo:
So... In this order, first the backstop, then announcing the project finance. It can be, you know, at the same time because they are unrelated, but second tenant, and then hitting constructions and first power.
Me:
Okay. Great.
Question 7
Question:
On revenue NOI per gigawatt for the 2.2 gigawatt build, would it be in line with the previously discussed $1.5 billion in revenue, $1 billion NOI per gigawatt? How does that differ across Turkey, PowerShell, Powerland?
Rodrigo:
This is part of the work that I've been talking with you through this call. When we did the S11 and we were talking about the PowerShell deal, those economics for PowerShell roughly stand.
But as you can see, on a per megawatt basis, the Turkey, of course, is more profitable. Now, this is part of the work that we need to do and relate to our investors on NOI economics and cost of what a typical full Turkey would look like.
Again, this is, you know, directionally, Turkey is the highest revenue per NOI and NOI per gigawatt, but it's also more capex and capital intensive.
Me:
Right.
Rodrigo:
But what I can tell you on the Turkey right now is that the power and battery is going to be between $3 to $4 per megawatt. And on the data center, turnkey is going to be between 10 and 12 dollars per megawatt.
Me:
Okay.
Question 8
Question:
Will the annual shareholder meeting be in person or online?
Rodrigo:
We haven't finalized that publicly to know if it's going to be in person or virtually. We'll confirm logistics as soon as we can. What I can tell you is that we are going to have that before the end of October, the shareholder meeting.
Me:
You said before the end of October?
Rodrigo:
Yes.
Me:
Okay.
Rodrigo:
So, you know, we're mid-August. We have a month and a half. So, again, we want to be fair to our investors in case it is in person, so they can, you know, make their plans to be either in Amarillo or Dallas, where we select to have that, or New York.
New York is also a—it can be New York, especially based out on we're going to be spending so much time in the next month in New York with the project finance that it might be the best for management to do it from there.
Me:
Yeah.
Rodrigo:
But, you know, we'll communicate that as soon as we're able.
Me:
Okay.
Rodrigo:
But New York is definitely a possibility to your detriment, Kevin.
Me:
Yeah, that's fair.
Rodrigo:
But, you know, having lived there, I also have, you know, a good network there, so it's always fun to be in New York. But yeah, we'll let you know when we have more information about that.
But again—
Me:
We have a month and a half to, to, well, two months and a half, right, to the end of October.
Rodrigo:
But again, we want to be fair to our shareholders and give them ample time to organize themselves, either if it's virtual or if it's in person.
Me:
Okay.
Question 9
Question:
What are you most excited about over the next six months?
Rodrigo:
I believe that we have all the ingredients here to continue to be very successful on commercial activities. That is what I am the most excited about.
But going back to question six, there are catalysts that will happen in this quarter. We haven't given a timeline, but, you know, the backed-up project finance and start constructing, that is, that, you know, has me very, very excited. But the most, of course, is, you know, bringing in more tenants.
Me:
Gotcha.
Question 10
Question:
For the 222 megawatt tensor wave phase one, how much capital does Permian expect to invest? How much will Permian fund itself, and what stabilized NOI yield should investors expect?
Rodrigo:
That is, again, going back to what we have talked internally and with our investors and sell side, we plan on being more forthcoming with this.
What I can tell you now is three to four dollars on power and batteries per megawatt, and ten—sorry, million dollars. I wish there was dollars.
Me:
Man, it'd be cheap.
Rodrigo:
24 million dollars on power and batteries per megawatt, and 10 to 12 million dollars on data center turnkey.
But again, as we have been talking the last 25 minutes, this is something that we do owe our investors on deal economics all the way to the NOI, EBITDA tranche, and the costs. So again, expect us to be more forthcoming in the coming months.
Me:
Yeah. Yeah, I did mine just based off of the numbers y'all's CFO, Masson, I forget his name.
Rodrigo:
Yeah.
$6.849 Billion annually
Just my estimates, but this could be a massive company in the future if they execute from here. Feel free to tell me if you would have done anything different with the estimates/ math/ NOI multiple, but either way if they are successful in getting tenants for their power, financing and building the power infrastructure and data centers, the numbers are huge.
If you slap a 15x NOl multiple on the $6.849 billion, you get $102.73 billion dollar company. Roughly $160
a share.
I'm not saying I think it will trade at $160 a share in three years. I'm just punching estimates into Microsoft XCEL right now. But it's interesting.
Edit: I changed my NOI multiple to 10x instead of 15x.
That’s very conservative, and you still end up with a $100 stock in 2029/2030 if they execute.
Chat GPT
Yes. I checked the SEC record, and Riot is a very useful precedent for thinking about what Fermi could do Thursday morning.
Riot's 191 MW AI deal was first publicly disclosed today, August 10, 2026, essentially as part of its Q2 earnings package.
There are two simultaneous pieces:
This was posted on the Fermi website on the news section a few weeks ago. Just noticed it today for the first time.
Looks like they moved more offices to the front and there is maybe 30 vehicles at the front office right now. No gas turbines yet. Nothing else noteworthy.
Okay, I’m not trying to overanalyze, but I get the feeling that maybe Fermi has a deal nearly complete but is just planning to finalize in a way to where they can disclose it on the 13th. Curious to hear any thoughts on that.
Tenant within the next 10 days according to IR, strap in
Chat GPT
This page is far more revealing than it first appears. It is Exhibit A – Performance Conditions for a Performance Restricted Stock Unit (PSU) Award. Unlike Anna Bofa’s separate Deal 1 Award (2 million RSUs before July 31 vs. 1 million after), this is a multi-factor executive incentive plan tied to Fermi’s 2026 execution.
1. Tenant contracts dominate everything (70% weighting)
This is extraordinary.
It means the Compensation Committee determined that tenant execution is by far the most important objective for 2026.
Construction, power procurement, and safety combined account for only 30%.
That tells investors exactly what the Board considers success.
2. The Board expects tenants—not just one
Notice the payout structure.
1 tenant = only half credit
2 tenants = target
3 tenants = maximum
That suggests the Board isn’t simply hoping for a single lease eventually.
Their compensation framework assumes:
one tenant is the minimum acceptable commercial success,
two is what management should realistically accomplish,
three is an exceptional year.
That is an aggressive set of goals.
3. Construction metrics are also aggressive
Construction is worth 20%.
Targets include:
GE 6B turbines
45% = threshold
90% = target
100% = maximum
Siemens SGT-800s
35%
65%
85%
F-Class turbines
10%
20%
40%
Procurement
Necessary long-lead equipment for all F-Class units operating in simple-cycle by December 31, 2026:
50% completion
100% completion
This lines up with recent announcements:
turbines arriving in Houston,
Primoris,
TSK,
balance-of-plant work.
4. Safety matters—but only 10%
Safety carries 10%.
Measured by TRIR.
If there is a fatality:
payout = 0%
That’s a strong incentive but relatively small compared with commercial execution.
5. The stock-price modifier is fascinating
Below Exhibit A:
If the 20-day VWAP ending December 31, 2026 is:
Above $21
Performance score × 1.25
Meaning management earns 25% more shares.
At or below $10
Performance score × 0.75
Management loses 25% of earned shares.
This creates alignment between:
operational execution
shareholder returns
6. This is different from Anna’s Deal 1 Award
The two awards work together.
Deal 1 Award
Immediate incentive:
2,000,000 RSUs if first qualifying lease by July 31
1,000,000 RSUs after July 31
That is a binary milestone.
PSU Award
Measures overall 2026 success:
tenants
construction
safety
stock price
These are separate compensation mechanisms.
7. Why this matters for investors
The Board voluntarily chose these goals.
They could have used:
EBITDA
financing
permits
hiring
revenue
Instead they centered compensation around tenant contracts.
That tells you where management’s efforts are expected to translate into value.
8. An interesting observation
The target is two tenants, not one.
If the Board believed only one lease was realistically achievable in 2026, they likely could have made one tenant the 100% payout target. Instead:
1 tenant = threshold (50%)
2 tenants = target (100%)
3 tenants = stretch (200%)
That may indicate the Board viewed multiple tenant agreements as achievable when this compensation framework was established. However, it’s important not to over-interpret it: compensation committees often set aspirational targets designed to motivate executives, so these goals do not guarantee that two or three tenants will ultimately be signed.
$GOOGL expects only a small portion of revenue from existing TPU system sales agreements to be recognized in 2026, with the vast majority coming in 2027.
Due to supply constraints, it will expand the use of third-party capacity in Q3 as a temporary bridge, which is expected to create modest near-term margin pressure.
Plenty of deliveries coming in, cars in and out, equipment and materials are really starting to pile up. Crane is at the site, not the big dawg crane but a pretty big one.
“It is our expectation that you should measure us on delivering on these five key points: a secure and binding tenant agreement, that we maintain capital discipline to support liquidity, that we hire our next CEO, that we deliver power at our project site and that we explore strategic partnerships for accelerating data center and power deployment on our site. Those are the five commitments and deliverables that we’re focused on for the next 90 days without distractions. That’s how you should measure us.”