Greetings, and welcome to TeraWulf 2024 First Quarter Earnings Call. [Operator Instructions] As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Mr. Jason Assad, TeraWulf's Director of Corporate Communications. Thank you, Mr. Assad. You may begin.

Thank you, operator. Good afternoon, and welcome to TeraWulf's first quarter earnings call. With me today are Chairman and Chief Executive Officer, Paul Prager; Chief Operating Officer, Nazar Khan; and our Chief Financial Officer, Patrick Fleury. Before we get started, I'd like to remind everyone that our prepared remarks may contain forward-looking statements, which are subject to risks and uncertainties, and we may make additional forward-looking statements during the question-and-answer session. These forward-looking statements are subject to risks and uncertainties, and actual results may differ materially. When used in this call, the words anticipate, could, enable, estimate, intend, expect, believe, potential, will, should, project, and similar expressions as they relate to TeraWulf are such forward-looking statements. Investors are cautioned that forward-looking statements involve risks and uncertainties, which may cause actual results to differ materially from those anticipated by TeraWulf at this time. In addition, other risks are more fully described in TeraWulf's public filings with the U.S. Securities and Exchange Commission, which may be viewed at sec.gov and in the Investors section of our corporate website at terawulf.com. Finally, please note that on today's call, we'll refer to certain non-GAAP financial measures. Please refer to our company's periodic reports on Form 10-K and 10-Q and on our website for full reconciliation of these non-GAAP performance measures to the most comparable GAAP financial measures. I'd like to also inform investors of our new investor deck, which may be founded also at terawulf.com. We'll begin today's call with prepared remarks from Paul, Nazar and Patrick, then we'll proceed to Q&A. It's my pleasure to now turn the call over to TeraWulf's CEO, Paul Prager. Paul?

Thank you, Jason, and good afternoon, everyone. We appreciate your attendance today as we dive into our first quarter 2024 financial results. Just a couple of months ago, at our last earnings call, we highlighted TeraWulf's significant growth and accomplishments throughout the fiscal year 2023. We experienced robust organic growth at existing sites, achieved substantial debt repayment and bolstered liquidity. These achievements underscore the strength of our foundation and set the stage for the growth we are excited to talk to you about today. For those unfamiliar with TeraWulf, we specialize in energy infrastructure, drawing from over 3 decades of experience. Our primary focus presently centers on Bitcoin mining, where we employ sustainable practice utilizing zero-carbon energy resources, while also contributing to grid stability. Operating from our 2 principal data centers, Lake Mariner and Upstate New York and Nautilus Cryptomine, in Pennsylvania, a joint venture with Talen, we take great pride in sourcing 95% of our power from clean energy sources. Our expansive mining facilities currently have a combined self-mining hash rate of 8 exahash per second, powered by approximately 64,000 deployed miners with a fleet efficiency of 24.6 joules per terahash, we utilized 210 megawatts of infrastructure capacity. Notably, our miners consistently operate at an impressive 98% of installed nameplate capacity. We're actively expanding our mining operations at Lake Mariner with Building 4 scheduled to be complete at the end of June and Building 5 commencing construction. These expansions are projected to raise our total operational capacity to over 10 exahash per second by midyear with the fleet efficiency of 22.7 joules per terahash and subsequently to more than 13 exahash per second with a fleet efficiency of 20.9 joules per terahash. We have one of the most efficient miner fleets in the industry. As Patrick will elaborate, we…

Thank you, Paul, and good afternoon, everyone. Before delving into the latest developments of WULF Compute, I'd like to provide some context regarding the surging demand for energy infrastructure. As Paul mentioned, our expertise lies in the power generation sector. Over the past 2 decades, we've been involved in the development, construction, financing, ownership and operation of power generation facilities, both domestically in the United States and in various overseas jurisdictions. Central to our operations has been our key understanding of power blocks, the scale that's particularly pertinent in today's landscape. With the advent of high-demand loads, such as HPC and AI, ensuring timely access to sufficient power has become a critical concern. However, access to power and energy infrastructure remains a significant constraint. These high-density compute loads have 2 primary objectives: first, to secure access to power at scale; and second, to do so within the next 12 to 24 months. Once these primary objectives are met, considerations then shift to securing low-cost and zero-carbon power solutions. Fortunately, the TeraWulf portfolio boasts substantial capacity to fulfill each of these objectives, as high-density compute loads gravitate towards areas with ample power resources, we're actively engaging with various counterparties to allocate our energy infrastructure accordingly. We are in the early stages of this transition and are thoughtfully analyzing the alternatives. Over the past 9 months, WULF Compute has been dedicated to developing the blueprint for our low-cost zero-carbon and scalable digital infrastructure. Traditional data centers are facing challenges in meeting the escalating rack density requirements of current and next-generation GPUs. To address this, our engineering team has collaborated closely with industry leaders such as NVIDIA to design air-cooled and liquid-cooled facilities tailored to GPU demands. The culmination of these efforts is the construction of our 2-megawatt facility at the Lake Mariner…

Thank you, Nazar. As Paul stated, our first quarter was strong across all financial metrics, driven by favorable business fundamentals and outstanding execution. In the first quarter of 2024, we self-mined 767 Bitcoin at Lake Mariner and our net share of mined Bitcoin and Nautilus was 284 for a total of 1,051 Bitcoin or about 11.5 Bitcoin point per day, a 10% increase over the 959 bitcoin mined in 4Q '23. We received an additional 6 Bitcoin in 1Q '24 and 13 bitcoin in 4Q '23 from profit sharing associated with the hosting agreement at Lake Mariner that expired in February 2024. Our GAAP revenues saw outstanding growth of 82% quarter-over-quarter, reaching $42.4 million in 1Q '24 from $23.3 million in 4Q '23. Our value per bit point self-mined this quarter, a non-GAAP metric that includes Bitcoin mined at Nautilus, averaged $53,750 per bitcoin for a total of $56.8 million as detailed and defined in our monthly operating reports and press releases. As a reminder, there is a key difference between our GAAP financials and the monthly operating reports and 2024 guidance. Due to our 25% ownership in Nautilus, the revenue, cost of revenue, operating expenses, depreciation and amortization at Nautilus are not consolidated into our GAAP financial statements. Instead, the financial impact of the Nautilus joint venture is reflected in the equity, in net income or loss of investee net of tax line item on the GAAP income statement. Our GAAP cost of revenue exclusive of depreciation for 1Q '24 was $14.4 million, a 61% increase over $8.9 million in 4Q '23. The quarter-over-quarter increase was principally due to a 37% increase in average operating exahash in addition to higher realized power costs at Lake Mariner in 1Q '24. Looking now at our gross profit, also exclusive of depreciation,…

Thank you, once again for joining us today. In summary, TeraWulf stands at the forefront of the industry with its best-in-class assets, lowest unit economics and unparalleled ability to scale our owned infrastructure. We remain deeply committed to delivering consistent growth and profitability while aligning closely with the interests of our investors. As we look ahead, we believe our infrastructure strategic importance is currently undervalued by the market. The foundation we've laid and the trajectory run, underscore our potential for continued success. We appreciate your continued support and confidence in TeraWulf. Together, we are poised to capitalize on the exciting opportunities to add uphold our position as a leader in capital efficiency and deliver sustained value to our shareholders.

[Operator Instructions] The first question comes from the line of Josh Siegler with Cantor Fitzgerald.

I just want to start by saying I really appreciate the breakdown of your unit cost here. It's very helpful for us, investors. For my first question, I was wondering now that we have passed to having, you've demonstrated this profitability now. Do you expect more financing channels to really open up maybe perhaps debt markets once you've proven that profitability post halving?

Josh, this is Patrick. Yes. Look, it's a great question. I think we've talked about this before. I'd like to see the debt markets kind of from a corporate level, so not the old miner finance model because that obviously didn't work very well. But from a corporate level debt perspective, as the business gets more mature and candidly, post halving, there's a separation of companies that are profitable and those that are not that are really set up more as lifestyle companies, as Paul mentioned, for management teams. Yes, I think I'd like to see the debt markets reopen. And you'll see in a slide that we put in our slide deck, I think it's Page 12, but we are providing a bridge of capital needs for the year. And we have assumed that we just take all of our free cash flow and pay down our debt. I do think there's a possibility. I've already received a bunch of inbounds, I mean, certainly given our business and the collateral that we put in the ground has grown over time and the debt has been reduced. Yes, I think there'll be opportunities to kind of turn that debt out. So it really will become a question of whether we want to keep it or not. But I think for us, as we pivot a bit more into the high-power compute and AI space, that space, the public companies trade at 20 to 25x EBITDA, and they all have 6x leverage because you've got long-term contracts, and it's a much more stable revenue business. So I could see us as we kind of get into that, really have 2 kind of silos, one, the Bitcoin mining business to the compute business, and those are very different financing beasts. So yes, I'll just leave it at that.

Following up on that, I did want to ask a little bit more about the HPC side of things for my second question, was notably, last you talked about high-performance compute was a couple of months ago. And in your deck, you outlined the 3 potential business models that you could pursue when you have the pilot program and works here. I was just wondering now that time has passed, if you're becoming more comfortable with one specific model over another in terms of what the future of HPC may look like for WULF?

Josh, it's Nazar here. Yes, I think as we've noted, our core expertise is really around energy infrastructure. And so we are getting more comfortable and we've been working to design the next-generation data center, given where GPU rack density is going, a lot of legacy data centers may not be able to most efficiently meet the power needs. And so we think from delivering a infrastructure that can be used for not just H100s but even kind of Blackwells and beyond, we're designing digital infrastructure that can meet those needs. So we're getting very comfortable that we can design, put up, and operate that physical infrastructure. So, kind of in that, if you look at page, I think it's Page 14 in our deck, that middle column really kind of what we think is a sweet spot between owning the infrastructure, bringing in customers who may pay for the GPUs, and having kind of a long-term ownership around that. Now that being said, this is kind of the full spectrum of the alternatives that are out there. And as we get further along into it, post our 2-megawatt pilot, we're finalizing the design for a 10-megawatt facility. If we see an opportunity, Patrick just mentioned financing or other things, to be able to facilitate kind of being participating in more of the chain, we'll do so. But currently, we're really focused kind of on that middle column and that's where we think our expertise lies. I don't know if Patrick could add to that.

This is Paul. The only thing I'd want to add is a little bit unique to WULF, which is their sites. Not everyone has the energy infrastructure for that kind of scale in terms of power. Not everybody has the water permit for that kind of volume of water and all the access to the contiguous land that one of these massive data centers want. That makes it exciting, but again they're all out shopping for this stuff which just means that the energy infrastructure value is going to continue to head north. I think Nazar said the other day on one of our calls, we're not looking for a trade, we're looking to build a business. So in that regard, I think our inclination is to focus on that middle column because I think it's just -- it's about returning maximum value to the shareholder as opposed to sort of having us work for Meta, Amazon or one of those guys. And so that's how we're currently thinking about it. But these are early days, Josh, and I don't think one needs to sort of make a commitment at this point as much as one needs to really study this market and understand it. That said, it's also very important to us that we remain mining Bitcoin. We're really good at it. I think we're best in class at it. And we believe in Bitcoin. So I think, what Patrick said earlier, which is we see ourselves possibly having at least a couple of buckets out there, 1 HPC and 1 mining Bitcoin. I think you'll continue to see that as a driver of our business philosophy.

Next question comes from the line of Lucas Pipes with B. Riley Securities.

Patrick, I guess I'll have to look up that industry classification system, but in more serious matters. We are hearing a lot about the tightness in the power markets with the AI growth. And Paul, you know power markets extremely well. So I wanted to hear your opinion on that. Is that true? Do you see that tightness in the power markets? And then if it is true, who gets that economic rent? Is it the utility, is it anyone with a PPA? And to what extent does geography matter in all of this? Appreciate your thoughts on this.

Nazar, do you want start with respect to HPC and I'll follow.

Sure. Lucas, it's Nazar here. So as we look at the landscape, there's a mismatch. The size of the demand and the timeframe for which the supply is available. So with infrastructure, if you have enough time and enough money, things can be solved. And right now we're at a point in time where the near-term demand helps strips the available supply. And that's particularly the case if you want scaled access to power at a low cost and it comes from a zero-carbon source. Now there's not that many places where you can find all of those variables available to you. And so as we think about it, kind of the owner of that infrastructure or the owner of that right to procure or use that power should be able to capture the rent. And so what we see happening right now is if you look at kind of at one end of the spectrum, the hyperscalers are flush with cash and they're out there optioning up as many of these sites as they can. And whether they ultimately use the capacity or not, that's not their concern for today, they'll worry about that tomorrow. But in the near-term, they're just optioning it all up. And so, again, from our perspective, we've got 300 megawatts of incremental available capacity yet to be developed. And so we think fundamentally we're sitting at the constraint point where again, there is not enough supply to meet the demand and so we should be positioned to capture a fair amount of that rent. And how we do so and I think that those business models are still evolving and so the other thing that we're seeing is the traditional model for how data centers operate it. They went into a territory, they called up Dominion, said we need 40 megawatts. Dominion said we'll serve you. That may occur. We think that will occur -- will still occur, but we're also seeing the demand sources going directly to procuring that supply for themselves, and whether it's Microsoft announcing with Brookfield saying, build out 10.5 gigs of capacity for us, or what Amazon did at the Talen site. So again, those places where you have access to that power and you can deliver that, then we do think that those will be best positioned to kind of capture the value that's coming from that demand.

The only thing that I would add to it is, look at there being some competition then to a Bitcoin miner because he's got to compete with a whole different level of fellow in these hyperscalers for sites with power and not all sites are created equal. Likewise, I think Patrick mentioned earlier today, he said if you've got a good business model but bad power, well, you're hosed because at the end of the day, you have bad power contracts, it's a bad power contract. You'll never right size your costs. And so we have good power. We have good location, but green power, but good power in that it's really inexpensive. So I think you have to look at some of the shops in Bitcoin mining that really have great power agreements and are transparent by the way, about them. And that's where you ought to be making your bet.

Really appreciate the color. And, yes, that provides a lot of context. Quick second question. Just in terms of that $9.6 million towards an additional 30,000 miners, how should we think about kind of that by itself from a timing perspective and then kind of longer term or rather beyond that whereas you had -- as it relates to growth.

So, Lucas, I think that whole contract, right, is for 7 exahash total. 5,000 of that we pointed out in the call has been paid for and will be delivered at the end of this month. And then really I think, as you -- we haven't committed beyond what you see on Page 12 of our deck, and perhaps Nazar can speak to this, but I'll just mention it because, part of it is we think that model is changing. I mean, even over the past few weeks, we've gotten some calls from folks that are discussing concepts like giving the miners to us for free with a profit share. And so I think we're hesitant to kind of commit to anything beyond what we've kind of laid out. You know, we'll have 13 exahash of infrastructure ready to rock and roll this year. We'll be at 10 exahash by July 1. So that's -- I realize, Lucas, that's not as specific as I typically like to be, but there's a lot of movement in that market. So that's kind of how we think about it. I don't know, Nazar or Paul, if you want to add anything there?

Yes, just to add to that, Lucas the contract is for 25,000 remaining machines at $16 per terahash and that $9.6 million represents a 10% deposit on those machines. And so as Patrick said, that we've -- in that contract, we've effectively taken down 10,000 of those machines to cover the Building 4, which will come online here, as Patrick said, by July 1. And then the remaining miners are on option available to us. And for us, it's really about capital efficiency. So as we are thinking about continuing to grow the business and allocating the infrastructure most effectively, that capital kind of question comes in. And so as Patrick said, we're looking at a number of different ways to skin that cat, and it ranges from purchasing more miners at $16 per terahash, some sort of profit share, and even kind of something in between which would just be effectively being able to replace hashboards and machines. So you're significantly limiting your capital cost by using the existing box that exists for the miner and swapping out hashboards with the next-generation hashboard. So there's a whole spectrum of things that we're actively looking at in terms of how we will populate those next buildings that are coming up.

And is the deposit transferable? So if you chose to go for it with another option, how should we think about the deposit?

No, I mean the deposit for those machines Lucas, so we'll either -- but as you know that deposit kind of works down every time we take a delivery of machines ratably right? So, yes, so the deposit's down to secure the price and the future delivery and then we'll kind of work that down over time as we take machines.

Next question comes from the line of Joe Flynn with Compass Point Research and Trading.

I had a question related to the Nautilus option to expand 15 megawatts. Do the terms of that option change at all with the Amazon acquisition or does it require like a sign-off? And then ultimately, how should we think about timing of getting that capacity online if there's no change?

Sure. The Amazon is a new landlord to the agreement we have. It's the agreement we have, so there's no change in the terms and conditions of the agreement. We are currently working to finalize the designs for that facility. We're targeting a 2025 online date for that. Currently we're focused on completing Building 4 and we've got a pretty quick turn on delivering Building 5 at Lake Mariner. And so we've slated the Nautilus expansion for 2025. And so we are working with a landlord there to finalize the design for the building that we put up at that site.

And Patrick, if you kind of touched on this already, but I guess in this low-cost hash price environment, I mean, how should we prioritize, like between paying down debt and CapEx opportunities to push some CapEx out if needed, the flexibility?

Yes, sure. I mean, I think what we try to be, again, our philosophy as everyone knows on this call, is being highly transparent, which many of our peers purposely do not. And so we are going the other route. So on Page 12, we've literally laid out everything that we -- the management team has committed to for this year. And so you can see on that slide, we've got around $45 million, $50 million left to raise over the next 6, 7 months. So that, just to put into context, and the reason why we put Slide 11 in our deck is that absolutely pales in comparison to even if -- again, even if we just did equity for that $45 million, $50 million that we're talking about and we used all of our free cash flow generated during this year to pay off the debt. It's just, Page 11 puts it in context for you. I mean, some of our peers issued over $500 million on the ATM in the first four months of the year, okay? And not only that, but they generated virtually no profit. I mean, two of our largest peers generated close to $10 million and $50 million being close to 50% and 150% larger than our company. And we generated $32 million of EBITDA. And so I think we are happy to own and commit to what is on Page 12 here, because we are massively and accretively growing the business using equity as an eye dropper. These other companies on this Page 11 use equity with a fire hose. And so not only that, but -- and by the way, my [half-thought] you'll see on the top of this page, CleanSpark is highly profitable. So using your ATM to fund highly…

Next question comes from the line of Mike Grondahl with Northland Securities.

Kind of referring to Slide 14, WULF Compute, or the HPC opportunity. Now you're kind of focused on that center column. Paul, are you thinking of this as a business tens of millions or hundreds of millions, and then maybe as a follow-up, what do you think this business will look like in 3 years to 5 years?

So, Naz, why don't you start, and then I'll come in and give my thoughts on where we're headed.

Sure. Mike, we think we can grow this business to hundreds of megawatts. And so, our 2-megawatt pilot and the 10-megawatt building that we're designing is more geared towards an engineering exercise to ensure that the product that we have available meets all of the specifications and demands of people that will use that infrastructure, not just today, but again, future proofing a bit for at least the next couple of iterations of GPUs that may come out. So we believe again, the availability of power at scale and low-cost and zero-carbon is tight. And so we firmly believe that over time this is a business that we could build into 100 megawatts plus. With the 300 megawatts that we have today, we think we could build that up. We are also looking at a number of other sites as well which could further add to the capacity to build that business up. So again, we think we are trying to build a business here that's really focused in on that. As we kind of fast forward 2, 3, 4 years from today, again this page also lays out somewhat of how we see the marketplace playing out. You have your hyperscalers again, who have the capital and the ability to throw down a lot of cash and try to control their own destiny. So you see them doing that in a number of different places. Again, I gave the example earlier of Microsoft and Brookfield, 10.5 gigawatts of power that will feed directly into Microsoft's needs. And so there's only a few names in the world that can afford to do that and they are going to do that. At the other end of the spectrum, you have the newer AI companies developing, 1 or 2 or 3 of them will likely grow from where they are today, so it's becoming something fairly significant. And you have enterprise customers in between. And so we really believe that for other than kind of the hyperscalers, the availability of what we can offer is going to be tight-limited and we're going to be a natural and complementary customer -- or partner for them. So that's where we're positioning ourselves and we do think that over time this can become a real business.

What I'd like to add is a couple of things. One is, B.B. King has a great song, he used to sing it, don't make you move too soon. These are very early days in the high-speed compute business. And I think it would be prudent for us to really look at all the ways that we can take advantage of our megawatts, of our energy infrastructure. Because that's really our advantage, right? The sites, the water permit, but it's the access to low-cost carbon free energy that is what is so enabling on a long-term basis to building a business at WULF. So I think that is why we're located -- that's why we're focused on that center column right now because that's where we think it's likely to see the best value for WULF shareholders. That said, we're constantly, I mean, this is all we talk about every minute of the day, trying to figure out how do we get the most to our shareholders out of our energy infrastructure. I think on a long-term basis, you would see WULF having 2 divisions if you will, high-speed compute and Bitcoin mining. We like Bitcoin mining. We are very constructive on the price of Bitcoin. We are the most efficient at it and we don't think it would be prudent to just sort of switch gears and become a high-speed compute company. Particularly at this point in time, there's so much to learn. We think the right thing to do is to sort of continue our analysis, focus on the center column, do what we're doing in terms of the 2-megawatts and then the 10-megawatts, and then continue to be dedicated to Bitcoin mining at the lowest possible cost. And so that's where I think we're headed. I think that will also avail us additional opportunities in terms of our cap structure because I think that, again the high-speed computing space much more, it's much kinder on an institutional level to a business when you think about lending and the kind of terms that are available to you. So I think that's where we'll end up.

And then 1 follow-up. I'm trying to connect Slide 12. One of the uses of the $47 million of capital is WULF Compute and it says $5 million. How does that blend into Slide 14 in the financing line for the middle column, it says equity initially. Can you just speak to the financing and what equity initially means in that center column and how far $5 million goes?

Yes, Naz, do you want to take that first part and then I'll take the second on the [ $5 million ]?

Sure. The 2-megawatt pilot that we're building is largely most of that $5 million budget, so that's the full design engineering construction cost associated with that 2-megawatt pilot. That 2-megawatts of capacity will be available in early August and we're in discussions with a number of different counter-parties around utilizing that capacity. So that's where the bulk of that goes and that budget also includes the engineering and design work for the 10-megawatt building as well. And again, that 10-megawatt building is designed to be scalable, and so whether we build 10-megawatts or 20-megawatts or 30-megawatts, it's really just scaling up that design. And so that's what's included in that budget there for the compute business.

And then, Mike, to answer your question, so I think it's helpful to distinguish between perhaps the build-the-suit, which is the right column, and then the co-location, which is the middle column. So in the build-the-suit model, which I think we've talked a little bit about before, that's a single company, big public company, Magnificent 7-type stock. And what we've learned as we have gotten down that rabbit hole is that's -- to borrow Paul's phrase, although B.B. King wasn't a -- he was far from a one-hit wonder, but that's a bit of a one-hit wonder, meaning we think we can create a lot of value doing that, but then it's an one-time event creates a lot of value, and then we're working for that entity basically for the next 15 years. Whereas if we pursue the co-location, and by the way, sorry, just back it up, Mike, that's obviously very financeable because that corporate, right, is an A-plus credit, and so generally speaking, you can get 80% to 90% financing at 250 basis points to 300 basis points behind the corporate credit. And so that's different if you go to co-location, the middle column, it's essentially the same business, but instead of the sort of Magnificent 7-type stock taking all the profits, because as we know those companies are all very profitable, WULF gets to effectively bring in third parties to help us manage that, because that's our businesses in cloud compute. But we can leverage third parties that aggregate enterprise customers and then finance off that. And so initially, that is equity, but as that business matures, it is also financeable, up to kind of 80-ish percent, but it's kind of baby steps because you've got to aggregate the customers and then there's sort of safety in numbers once you have a bunch of customers in there, as opposed to just financing off of the back of the corporate credit rating of 1 customer. Does that make sense?

Next question comes from the line of Bill Papanastasiou with Stifel.

I appreciate the transparency in this slide deck and those funny analogies on dilution. For my first question, I'm just curious to hear whether there's been an uptick in the number of subscale Bitcoin mining peers knocking on your door following the halving event, just given the dynamics on mining economics. Do any of these peers have compelling assets or operations? That would be of appetite to TeraWulf today. Or will the focus remain on existing growth plans at Lake Mariner and the HPC-AI business?

Listen, I think we have to focus on execution of our business plan and that's what we're focused on. We have even when Nazar mentioned earlier, like pushing Nautilus expansion out to 2025, it's because we have such a low cost to build at Lake Mariner that it just makes sense for us to focus on the careful and considerate execution of our business plan with our existing sites. That said, everybody in the space is talking to everybody in the space. And as you know, we have a management team that's been together for a very, very long time that are really focused on energy infrastructure. And as Patrick went off of that a little while ago, really focused on even a per exahash at the smallest dilution to the shareholders. So culturally finding somebody that would be a right fit for us, it's probably not easy. And we're very focused as well on the carbon-free element to our energy. So I think that sure, everybody is chatting. Are there many likely dates? I don't think so. I think we need to focus on just execution, paying down our debt, and run our way. I think the institutions, when they come to the market, are coming to us. They're appreciating the level of transparency and the actual performance that TeraWulf is delivering under Nazar and Patrick. And I think that we just need to continue to get that message out to the real world. I mean, if you again take a look at that Slide 11, we're talking about EBITDA per exahash and then you take a look at the massive dilution on the right side of the page versus the hardly any on the left side of the page. And then when you put in that third column where you look at stock-based comp, I mean, it really is very troubling and you'll see that there just are a lot of companies out there that just, they shouldn't be out there And it doesn't make sense for us to sort of join with them because we're all about making money through the appreciation of the shares of our stock. We're aligned with our investors. So that's sort of how I look at it. Cultural is a big issue with us because again focused on energy infrastructure and been around a long time.

I appreciate that color. Shifting gears, just hoping to get some more color on the debt repayment plans. Previously there were talks about potentially refinancing once the indebtedness dipped below $100 million. Should we assume that that's completely off the table and whole repayment's going to happen before 2024? Any color you can provide there?

Bill, it's Patrick. So, like I said, on, you'll see in our updated slide deck on Page 12, we've laid out the capital bridge for 2024. Our free cash flow this year will take care of all of our debt, and then we've got some commitments out there for Building 5. Also laying the foundation, you'll see on that page, we've got a bar chart that says L&D site electrical. That's some electrical work that we're doing on the site that will allow us to expand into 300 megawatts very quickly, so some significant line expansion, transformers, upgrades that we're doing. So I think, yes, I mean, look, if the financing markets get more affordable or if we want to sort of combine that debt maybe with a debt financing at the compute business, yes, I mean, there's a lot of options on the table that would then decrease the capital need there. But I think for now our view is let's just -- the biggest asset of our company is the cash flow generating ability because our power costs are so low and our profitability is so high. So let's take advantage of that, get rid of the debt and then start to look to the future.

Congrats on the strong operating cash flow.

Next question comes on the line of [ Marcelo Rossi ] a private investor.

I understood Nautilus runs on $0.02 per kilowatt power, but earlier in the call, somebody mentioned there was seasonality power cost adjustments. Can you elaborate on that? What percentage of the power is subject to seasonality?

Marcelo, it's a good question. It's Patrick Fleury. So we have 210 megawatts in operation right now, 50 megawatts at Nautilus, that is fixed power at $0.02. And then the 160 that we have at Lake Mariner, which is soon to be 195 when Building 4 comes online July 1, that power floats. And we realize generally Zone A power prices in the New York [ ISL ] at that plant. And so you can see if you look on Page 20 of our updated deck, there -- that power generally at Lake Mariner has been around $0.04. Last year, for the entire year of 2023, we realized they blended $0.032 at both sites, and that was roughly $0.02 at Nautilus and I think $0.036 -- $0.038 at Lake Mariner. And so, yes, they're typically in the first quarter, our power's a little higher in the shoulder months, which tends to be the second quarter and the fourth quarter. Those tend to be really low price periods at Lake Mariner because there's no real demand there. The demand is really a function of heating and cooling days, so it's mostly in the winter, where it gets real cold for a couple weeks and then in the summer where it's real humid for a couple weeks. But that's the beauty about being 30 miles east of Niagara Falls. So, yes, so we haven't changed our guidance for the year. It's still that we expect to realize $0.035 and like I said, we feel good because we realized $0.032 last year.

Next question comes from the line of Bob Evans with Pennington Capital. Please go ahead.

Just to follow up a prior question, just to get a sense of scale that you're thinking for the HPC-AI opportunity. I know you're looking at the middle part of the slide, but again is this a hundreds of millions of revenue opportunity or give us some sense of scale, at least what we should think this should ramp to. I know you're trying to put various clients into this project, but just give us a sense of revenue potential.

Yes, look, I think, Bob, as we kind of put on the page, right, I mean, if we took 300 -- our 300 open megawatts at Lake Mariner, right, it would be, you can multiply 300 times 1.3 and 1.5, right? So it's $300 million to $450 million. So I think Naz answered earlier, like we expect it to be in the hundreds of millions. You know, I think at this point we're not committed to dedicating, as Paul mentioned, all of that capacity to HPC-AI because we're very good at Bitcoin mining as well. But we're also infrastructure people and there's a big business that sits outside of the public company as well. So I think I would just say stay tuned. There's lots to come this year on that.

Okay. And then is the right way to look at it on the cost per megawatt and revenue per megawatt? I mean, if we just take the midpoints of both of those ranges and divide it out, does that mean it's -- because my understanding is that this business might be kind of a payback period of a couple of years, but if I do that math it might look like it's longer than that. I'm just trying to understand the right way to think of cash generated and paying for itself?

Yes, so that's a really good question, Bob. So we are -- and look, I want to caveat this because we put down a lot of numbers here on the page, on Page 14, and you'll see that it says, the title of the page is Potential Illustrative Business Models. And I just want to point out Bob that we've been -- we're a hardworking smart management team, but we've been studying this business for 9 months. We've been in the private power business for 30. So when you're in your lane for 30 years and you get into a new lane, there's a lot of work to do before you get -- you don't want to get run over. So this is our initial learning. But I think generally speaking, the cloud business right, so the far left column, much more revenue, sorry, much more capital intensive business, but quicker paybacks because people are desperate for GPU as a service type business. We're more focused, again, I think our sweet spot, as Nazar pointed out, we're infrastructure people and we also don't necessarily, we are comfortable taking the GPU technology risk yet that the far left column would require. So the co-location business, yes, to your point, it doesn't have the sort of 18 months to 36 month payback. It's more of a -- I would say, 3-year to 8-year payback period. But that's still pretty good. And that business, like we said, we think these multiples down here on the bottom of the page are very conservative, but our view is that should be a 10 to 20 stable multiple business over time. So we think we can create a lot more value with less capital, focusing on that middle column versus either the left or the right.

Okay. So if I was going to simplify it, we'd say, you believe you can build a multi-hundred million dollar business with, call it 70% plus EBITDA margins that should have a higher EBITDA value multiple than a lot of other options out there.

Yes, correct. And the only thing I would add is less capital intensity as well.

I would just say one other thing. As Nazar mentioned in this presentation, we've got these 2 megawatt at WULF Compute, and we're putting up a 10-megawatt facility. So we are -- we're not set, Bob on what we are going to do here. We're set on taking the experience of WULF Compute to and putting up this 10-megawatt building and continuing to refine our analysis to come with the right program for TeraWulf, utilizing both the existing infrastructure that we have and the access to large, massive other infrastructure that we would have sort of outside of the public vehicles. So that's the point. The 2-megawatts and the 10-megawatts are us dipping our toe in the water in the prudent way to ensure we come with the right answer on behalf of our shareholders.

I get that. And I agree with you and that obviously things work out to anywhere on this slide, it's a higher multiple, more predictable business model. So I agree.

Thank you. Ladies and gentlemen, we have reached the end of question-and-answer session. I would now like to turn the floor over to Jason Assad for closing comments.

Thank you, operator. That concludes today's conference call. Thank you for joining us. Please note this has been one of our highest [spending] calls. So thank you for the continued and growing support. I'd also like to again remind you about our new investor deck, which may be found on our website at www. terawulf.com. Much like everything we do, it was proposed in the spirit of transparency. We hope investors will find it insightful. If you have any questions or need anything, please reach out directly as always and please remember to follow us on social media. Operator, you may end the call.

Thank you. This concludes today's teleconference. You may disconnect your lines at this time. Thank you for your participation.