Keel Infrastructure Corp. (KEEL) Earnings Call Transcript & Summary

[Audio Gap] coming for Bitfarms presentation. Before we begin, I just want to remind everybody that we will be making references to forward-looking statements that might differ with actual results. We also refer to non-IFRS measures. If you're looking for definitions on both of them, please make sure to look at our MD&A on our website. Additionally, all financial figures are in U.S. dollars unless we stated otherwise. And with that, Liam.

Does this work, guys? This work, okay. This clearly does work. Okay. Welcome, everyone. My name is Liam Wilson, I'm the COO of Bitfarms. Today, I'll walk through our investment thesis, value proposition and key developments, including updates on our energy portfolio and site-specific advancements, all of which give Bitfarms a competitive advantage to capitalize on the surging demand for HPC and AI infrastructure. I would like to kick off today's call by outlining our market thesis, one that we believe differentiates us from our peers in our transition to HPC and AI. And that is that infrastructure is not a bubble. Since the invention of modern compute, the supply of compute has increased exponentially. As compute grows, so too does the data center industry that powers it. This is a trend that has a trajectory of over 20 years of exponential growth and an annualized growth rate of 8.8% behind it. This isn't a bubble. It's a reflection of a new paradigm that showed no signs of slowing down before AI. Now as AI rewrites the rules of how humans interact with computers, the demand for data center capacity is accelerating, but the demand for compute and infrastructure has reached an impasse. The exponential increase in demand for power can no longer be met at the pace the market demands. As a result, the lease rates for data center infrastructure, which have grown at an average rate of 3% over the last 20 years are now growing at an average rate of 12% since 2022. We expect this trend to continue. Turning to Slide 4. Infrastructure is a bottleneck. As manufacturers continually introduce more efficient chips and increase production every year, this trend continues to accelerate. Next year, NVIDIA alone is expected to be shipping between 10 and 15 gigawatts of GPUs. That doesn't include AMD, Intel, Qualcomm and others who are also producing their own hardware with over 100 gigawatts of chips expected to be produced by 2030. While the supply of compute chips continues to increase, the growth in data center infrastructure is happening at a much slower pace. It is not silicon nor capital that will be the real bottleneck for continued growth in HPC and AI, but power and infrastructure. Over the next few years, the gap between the amount of chips that are being produced and the megawatts and the racks available to plug them in and operate them will continue to widen. We strongly believe that as this dynamic continues to play out, the value and the economics will continue to move in favor of those who own the energy and data center infrastructure. We've watched this play out in the market with the contracts that have been announced in the industry to date. As we've moved further along this curve that's shown on the slide, those rates have continued to trend upward. Most of the contracts over the past few months have been sitting near $150 per kilowatt per month. As time goes on, this trend is expected to continue with analysts predicting a massive shortfall of nearly 45 gigawatts of power for data centers by 2030. Just recently, the CEO of Microsoft confirmed this shortfall when he publicly stated that they have GPUs they cannot deploy. We believe that over time, the companies who have allocated and will continue to allocate billions of dollars into compute will be increasingly economically incentivized to pay rising prices in order to deploy their compute faster and with greater certainty. Every day they do not deploy is a day of lost revenue. Their customers will simply move on to a competitor. Our investment thesis is clear and backed by decades of data. Our conviction is high backed by consistent incoming demand. We don't want to cap our upside by signing leases prematurely. Instead, Bitfarms plans to optimize and achieve higher lease rates and margins through the following 3 strategic actions: #1, prioritize infrastructure development first. By minimizing the time between signing a lease and generating revenue for a customer, we will minimize the discounts that would otherwise be applied to the lease rates and locked into multiyear contracts. #2, take advantage of the widening gap between supply of data center infrastructure and data center demand to lock in higher rates and greater margins under multiyear agreements. And three, while the industry is focused on NVIDIA GB200 and GB300s, Bitfarms plans to leapfrog NVIDIA's Blackwell architecture and lead the industry in developing infrastructure for NVIDIA's next-generation Vera Rubin GPUs across 99% of our 2026 and 2027 development portfolio. With Vera Rubin GPUs expected to begin shipping in Q4 of 2026 and the infrastructure requirements to support them largely incompatible with facilities designed for Blackwell GPUs, we believe Vera Rubin infrastructure will be in the greatest demand and shorter supply in 2027 and will command significantly greater economics. Turning to Slide 5. We are able to take this approach because we have a robust balance sheet to fund development and know the value of what we own. We have the largest portfolio of power in each of the regions in which we operate, none of which are in Texas and all of which are existing or emerging data center hubs. With consistent inbound demand for our sites, we have high conviction in the value of our unique energy portfolio, the demand for our power and our ability to develop next-generation HPC and AI infrastructure. We believe that not all megawatts are created equal. Our megawatts are strategically located in high-value areas that have multiyear wait lists to secure the power we have secured today. Our campuses are close to major metros and existing data center clusters have ample access to major fiber trunk lines and undersea fiber optic cables and temperate mild climates. While Texas is undisputably a great energy market and arguably the easiest market to grow and develop megawatts in the U.S., there are, of course, trade-offs. The trade-off to short-term development efficiencies is long-term operating inefficiencies. It is no secret that besides power, the primary challenge with data centers is cooling and cooling is becoming an increasingly more difficult problem to solve as energy density continues to increase with every generation of new hardware. Building and operating data centers in a hot arid climate like Texas as opposed to cooler northern climates like Pennsylvania, Washington and Quebec means more CapEx and OpEx for cooling. This isn't an opinion. It's just math and engineering. If we built our exact same data center for Panther Creek with the same design, equipment and materials in Texas, it would have a PUE of about 1.4 to 1.5, whereas in Pennsylvania, Quebec or Washington, it would be about 1.2 to 1.3. That means for every megawatt we are converting, more of those electrons are going to compute, which is the revenue-generating activity for customers as opposed to supporting revenue generation through cooling. Simply put, our megawatts are harder to get in higher demand areas, produce more value for customers and are worth more per megawatt. In Pennsylvania, we have the strategic foresight to acquire our 3 campuses and submit our energy applications in 2024 before the HPC and AI demand really came into play in the state earlier this year. This has positioned us with secured power at Panther Creek and Sharon and at the front of the queue with very well-advanced power applications at Scrubgrass. In Quebec, new power allocations are near impossible to get with numerous data center applications denied by the province in the last year. Bitfarms has 170 megawatts operating with some of the cheapest power rates for data centers in North America, and they are 100% renewable. 100% of these megawatts are currently being utilized for Bitcoin mining. And in the last month, we confirmed that we will be able to convert our Bitcoin mining megawatts to HPC and AI. This means our Quebec portfolio represents a unique and strategic opportunity to increase total data center megawatts in the province by 25% from about 700 megawatts today. while also fulfilling 2 strategic national and provincial objectives, the scaling back of Bitcoin mining megawatts while increasing HPC and AI infrastructure and data sovereignty. In Washington, we have 18 megawatts of secured power in the largest data center cluster on the West Coast with the cheapest power in the United States for data centers, which also happens to be 100% renewable. Because of this, the area has a 10-year waitlist for power. Everybody is looking to grow here, and it is nearly impossible to do so outside of secured megawatts like ours. This means that despite the relatively smaller scale of Washington, sites in the area are in high demand by both enterprise and hyperscalers alike. Turning to Slide 6. Now we'll go through our sites one by one and discuss the key developments at each of them. As we mentioned, Washington has a 10-year wait time for new power and the cheapest power in the United States for data centers. We are actively pursuing colocation for both hyperscalers and enterprise where we can capitalize on the long wait times as previously discussed. While our focus is on developing next-generation Vera Rubin infrastructure across most of our portfolio, we believe there are some compelling reasons to keep our options open with cloud as a potential monetization strategy at Moses Lake specifically. #1, GPU as a Service would enable us to capture the benefit of the lowest cost power for data centers in the United States for ourselves and generate what we expect to be above-market margins and returns for cloud. #2, the relatively small scale makes cloud at this site easier to execute and finance. We have more than enough liquidity to consider this site and strategy fully funded today and are in active discussions with leading GPU manufacturers on GPU sourcing and financing, which we believe could be done on very attractive terms. GPU financing could materially reduce CapEx requirements and enhance expected returns. #3, we expect that by demonstrating our ability to execute across the entire stack, we will also be better able to understand customer needs, provide better quality service and negotiate better leases at our other facilities. Lastly, but most importantly, despite being less than 1% of our total development portfolio, we believe that the conversion of just our Moses Lake site to GPU as a Service could produce more net operating income per year than we have ever generated with Bitcoin mining, providing the company with a strong cash flow foundation that would fund OpEx, G&A, debt service and contribute to CapEx as we wind down our Bitcoin mining business. I will now walk through the rest of our sites in a little bit more detail, starting with Panther Creek. Panther Creek is our flagship HPC, AI campus in Eastern Pennsylvania. As we have discussed previously, we have 350 megawatts of secured power with PPL. This power is contractually obligated to be delivered with 50 megawatts at the end of 2026 and 300 megawatts at the end of 2027. The site has sufficient acreage for the development of the entire 350 megawatts with capacity to go beyond that. Additionally, we have $200 million remaining on our project facility with Macquarie that is intended to finance Phase 1 of the project as well as long lead expenses for Phase 2. There is also the potential for expansion. Recently, there have been a number of developments that have given us line of sight to expand beyond the existing 350 megawatts of secured power capacity. We have received positive indication on converting our existing 60 megawatts -- our existing ISA of 60 megawatts to a firm ESA of 60 megawatts to expand power to 410 megawatts total and on a recent load study to expand power capacity to over 500 megawatts of gross capacity. Turning to Slide 8, please. Moving on to Sharon, where we have 110 megawatts of power secured by an ESA with FirstEnergy in PJM under development. We are currently operating 30 megawatts of Bitcoin mining on site but have started development on an additional 80-megawatt substation, bringing the total available for HPC and AI use to 110 megawatts. We expect to have the full 110-megawatt substation online by year-end 2026. Similarly to Panther Creek, we'll be working to develop the campus for Vera Rubin GPUs, targeting site completion and revenue in the first half of 2027 for the full 110 megawatts of gross capacity. Turning to Slide 9. In Quebec, we have 170 megawatts of low-cost hydropower currently operating across multiple Bitcoin mining sites, almost all of which is within a 90-minute drive of Montreal. This is an incredibly attractive opportunity for hyperscalers who are following what's called a regional campus strategy. This is something that was pioneered by Amazon where smaller sites can be directly connected with a direct fiber infrastructure in order to reduce the latency between sites below 2 milliseconds, enabling many small sites to be connected together to function as one larger site. As I mentioned, it's almost impossible to grow organically in the province. And in October, we confirmed the ability to convert our Bitcoin mining infrastructure to HPC and AI with regulators and utilities in the region. With that pathway clear, we are accelerating our plans in Quebec. We will focus our development efforts on the city of Sherbrooke, where we have 96 megawatts robust fiber connectivity, a strong and developed local labor force and ample support from the local energy utility and municipality. We will be applying some of the standardized engineering and design plans completed for our Washington site to Sherbrooke in order to convert these facilities from Bitcoin mining into next-generation HPC and AI infrastructure adapted for Vera Rubin GPUs. Similarly to Washington, Quebec has a cool climate and some of the lowest cost energy in North America for data centers with strong unmet demand for GPU cloud in Montreal. Sherbrooke also represents a potential opportunity to scale up the cloud business in 2027 with VR200s, a strategy that we will evaluate as we work through the engineering and development plans for Sherbrooke. The remaining 74 megawatts of Bitcoin mining in province are earmarked for potential expansion in 2028, and we look forward to providing more detailed plans for Quebec in 2026. Turning to Slide 10. Last but certainly not least, we have our Scrubgrass campus in Pennsylvania. This is about 30 minutes away from Sharon, Pennsylvania on the western side of the state. With the exception of the new Panther Creek Phase 3 and Phase 4, which I spoke to a minute ago, this is the only power in our portfolio that is not 100% fully secured today, but this is a very, very exciting development opportunity for Bitfarms. We believe that this is the only campus outside of Texas for public miners converting to HPC and AI that has over 1 gigawatt of potential capacity. And while we have made great progress on developing the power story for this giga campus, there are still quite a few steps to be taken in order to contractually secure the power, which falls into 2 buckets. First, we have completed 3 conceptual load studies with FirstEnergy, starting with 250 megawatts and 500 megawatts and 750 megawatts. That's moving over to what's called a detailed load study with FirstEnergy, which would eventually be converted over to firm service in an ESA. Second, we have made substantial progress on evaluating the potential to add additional generating capacity on site. This could be accomplished by building a 3- to 4-mile pipeline from our campus to the second largest natural gas pipeline in the United States, the Tennessee Natural Gas Pipeline, which we have confirmed could supply up to 550 megawatts of natural gas, multiplying our generation capacity on site. While we're still in the early stages of evaluating how we would expand the generating capacity, and we'll provide more details as we progress. Combined, the 2 buckets could potentially provide 1.3 gigawatts of gross capacity. Additionally, there is very good fiber infrastructure in the area with over 8 fiber infrastructure networks nearby. And it is in close proximity to Pittsburgh and Cleveland as well as the other data centers, which are starting to pop up throughout the state. The earliest time that we anticipate we could have additional power at this kind of scale implemented at Scrubgrass is around 2028. Though this is a longer lead time campus for us, we believe that the forecast on power and demand for HPC and AI infrastructure, the timing for our giga campus will play in well with this cycle and our investment thesis as well as our continued other development plans. Turning to Slide 11 and sort of sum it up. We believe that we are incredibly well positioned to execute against our investment thesis in 2026 and 2027 and maximize long-term shareholder value. #1, we have a very unique portfolio of energy assets that we aim to fully convert to HPC and AI infrastructure. #2, we are exploring converting our Washington site to HPC and AI workloads and lead the industry in the development of next-generation data centers for NVIDIA's Vera Rubin GPUs. #3, we are actively evaluating a potential cloud monetization strategy for our Washington site, which we believe will be a meaningful driver of cash flows and could eclipse any BTC mining cash flows we have ever generated. #4, we are well capitalized to make our currently planned investments with financial flexibility that exceeds $1 billion across cash, Bitcoin in our Panther Creek project facility with Macquarie, all of which are going to fund CapEx as we continue to produce strong free cash flows from our Bitcoin mining operations that fund OpEx, G&A, debt service and contribute to CapEx with no planned minor CapEx. And lastly, we continue to execute on our U.S. pivot with the recent sale of our Paso Pe facility marking our full Latam exit, our transition to U.S. GAAP for 2024, 2025 results and the establishment of our New York City office and working towards a U.S. redomiciliation in 2026. We believe this would give us significantly greater index inclusion and meaningfully improve the institutional composition of our cap table. I now have the pleasure to hand the presentation over to our CFO, Jonathan Mir.

Can you flip forward one page? So I'll stand. I'm all right. So just by necessity, this is essentially our Q3 presentation. I'm going to finish this really fast, so we can do Q&A, which will be more interesting. But Q3, we did our convert that raised $590 million in proceeds. The pricing was really just terrific. We were the beneficiaries of great execution. We've got $200 million remaining on the project development facility at Panther Creek, and we're generating about $8 million a month from our Bitcoin operations. So the Bitcoin operations, if you think about Bitcoin as a company, let's say, 200 employees pro forma for being solely North America, about 50 of which are what you could call corporate. So that $8 million a month pays for all of those people. And then it's creating the capital to fund needed investment. The point is, right now, we've got $750 million of unencumbered liquidity and $200 million available on the Macquarie facility. This is important because if we are good custodians of the capital, it enables us to push Washington, Sharon and Panther Creek through to NTP. Certainly, one of -- whichever one gets done first opens up borrowing capacity to us for construction finance that lets us recycle our equity so that we have that available for whichever one is going to come online last. So we have a balance sheet right now, and we view having a strong balance sheet as a benefit to customers and shareholders. We know we can finance the development of the first 3 projects right off our balance sheet without external capital, which we have the flicker doing for us. So it's a -- you can flip forward. Flip, I don't want to go through all this with everyone.

I think that's it.

That's it. All right. So let's have a real conversation. So I came to Bitfarms to focus on 3 things. It's capital allocation, capital formation, raising the money and then capital structure, right, ensuring that we've got a capital structure that maximizes return on equity without creating risk around overleverage. And again, one of the strengths of the company beyond its sites is we have a balance sheet right now that lets us continue developing through Notice to Proceed, NTP, right, the day you can turn shovels and all of your approvals and permits are available on the 3 sites. We'll raise capital before we get NTP in all 3 sites, I'm certain. But that's a very comfortable place to be in as a developer. Generally speaking, we expect to be at NTP for Washington, Sharon and Panther Creek sometime in the second half of 2026. So that will be an important catalyst for us economically. Once we're at NTP, you can then -- you're on the best terms possible to negotiate a lease and then finance against that lease. And those projects would produce revenue late 2027. But once you have a signed lease, it's basically -- it's an NPV into the stock for the value of that asset. And then we'll continue to develop Scrubgrass and opportunistically other sites, if they are at least as good as our existing site in the same market. So we're not going to enter new markets. We're not going to deal with modest quality sites, we would -- but we will be aggressive if we think we have the opportunity to find a good asset for a modest amount of our liquidity. In terms of how we plan on financing the business, debt and equity, so to start pretty conventional for energy infrastructure finance, you use basically project finance down at the asset level to get the assets built and borrow maybe 60%, 65% LTV. So that's downstairs, each individual silo, you'd start out borrowing some modest amount of development capital. As you get pretty close to being ready to go, that's available in the bank market. That lets us recycle equity for the next project. That's great. And then once you are at NTP with a signed lease, you enter into construction capital. So you don't have to do that off your own balance sheet. Every project, let's just say -- I mean we can put some debt maybe upstairs at the public company bank loan or something like that and run at 70% debt to cap-ish. We're going to figure out what the right number is as we go along, but that's fairly conventional for an infrastructure company with contracts that are 15 years long as opposed to having an enduring monopoly or 30-year contracts. So the next question is, well, where do we get the equity, right? So if I spend -- if we spend $1 billion on a project, it is a mathematical certainty that we need $333 million of equity. So that's either coming off our balance sheet. It's coming from the sale of minority interest equity downstairs at a project, sell 30% of a project to an infrastructure fund or it comes upstairs with common equity at the holdco, use some gentle amount of ATM. That's completely unattractive to us right now, I can assure you. And we'll probably do some converts over there. I happen to think of converts as just being debt. So there's a limit to how far I'm going to push that. But just to be clear, people need to -- I want to always make sure we're transparent with investors, so they understand how much equity needs to accompany every project. And we're going to look to find ways to reduce that. The more we can lever as long as we feel comfortable with our ability to repay it, the higher the returns to equity, maybe we decide 70%, could be 75%, or a little bit higher. But fundamentally, creating risk from financial leverage is not what our investors will pay us for. Investors are going to pay us from being good owners and operators of infrastructure assets and delivering infrastructure-like returns. And then if we are able to continue developing beyond this enormous pipeline, perhaps they'll pay for some growth as well, but we have a lot of growth to finance to start. So that's like the basic financing strategy. I mentioned the equity point, just so no one yells at me when we start selling equity because I told everyone that we're going to do that. But that should be obvious to anyone who is an infrastructure investor. By the way, it's like that's a good sign if you're in an infrastructure company because it means you're growing. And the 3 risks that we really think of what are the 3 ways to break a developer, one, run out of money during the development process, right, before you get to NTP. So you really got to be careful with your capital and recycling it before the time; two, overlever yourself, right, get to a point where you can't manage through variations or just surprises in the rest of the business affects a lot of developers. And three is built on spec, building on spec being the greatest sin in the developer space. And that's just not part of our business plan. You -- by definition, you're developing on spec, but you don't start spending real money on spec. And that's fine because once you have a lease, all the financing you need becomes available to you. So that's just a quick financial overview and turn it over to Q&A.

Can you just talk about where you are in the process of commercializing your sites, talking to tenants [indiscernible] potential tenants, whether you've got inbound after your pivot, you clearly done a lot of work on kind of deciding to make this pivot. But [indiscernible] can you just talk through interest levels, whether you're talking to neoclouds or hyperscalers.

We've -- so firstly, thanks for your question. I think the inbound that we've received certainly over the last, call it, 3 months is inbound that we did not foresee coming our way, particularly around the Sharon facility and also Moses Lake. Moses Lake is this 18-megawatt facility that kind of people scoffed at and frankly, we did internally as well when we thought about it. But the reality is that, that 18 megawatt sits in the West Coast version of data center alley and a lot of the hyperscalers up there seem to be quite keen on Washington, our facility in Washington has essentially excess power. And 18 megawatts, they could pick it up, why wouldn't they? So that's -- we're seeing some inbound there. The same I could probably say for Sharon, too. Sharon was 110 megawatts. Initially, we were told that there was no way that any hyperscaler would look at Sharon. And now Sharon seems to be quite attractive in the market. The facilities outside of Sharon and Washington are not at a point yet where I think we could actively market those facilities. We need to get closer to what Jonathan refers to as the NTP, Notice to Proceed. Once we do get closer to that NTP date, I think the turnaround will be quite quick. But I would say that at the moment, certainly, Sharon, certainly Washington, Quebec is becoming more and more attractive to some of our Canadian counterparts up there that are looking to power.

PC.

I'm sorry.

PC as it gets closer.

And Panther Creek as it gets closer as well. That's starting to heat up at the moment.

Just to amplify Liam's comments. So we are encouraged by the nature and number of inbounds we received. As a practical matter, one would not want to enter into a lease negotiation until you are ready to proceed with the project at NTP because customers are going to heavily discount you over execution risk of actually being able to start turning shovels. And so then the question becomes how deep into the actual -- the phase between being ready to go and having -- being ready to energize, do you want to go? What's the optimal time to sign a lease. If capital weren't a concern, you get the highest pricing the closer you were to the date of energization. That's just like because whoever is in the market at that time will pick the project that could deliver tomorrow as opposed to the one that could deliver in 18 months. So what we need to do is balance maximizing lease value, where our thesis is it will increase over time because of demand and increase by site as you get closer to a tangible completion date. That gets balanced against, one, our balance sheet because you can only build so far off your own balance sheet because the financing isn't available for a full build until one has a lease. And then two, our shareholder expectations. Shareholders are looking to see us get things moving and developed with good customers. So this is what we spend and will spend a lot of our time on is constantly evaluating where are we on lease pricing versus our balance sheet and shareholder expectations and other factors. We're not going to let the perfect get in the way of the good, but we think we can do -- we think patience will let us do well. And if you look at -- we think patience will let us do well and help us maximize the value of the lease. And this is one of the reasons we put so much value on the balance sheet is because we are not under pressure to market a lease early. And the general expectation is that Washington, Sharon and Panther Creek should all be at NTP, Notice to Proceed by the end of '26, which means you can start marketing certainly proximate to that date. And then it's a question of how do you see -- where do you see the most value.

You mentioned kind of gearing for Vera Rubin. Is there going to be a big CapEx difference to build site specifically for that versus maybe what some peers are doing and they're kind of benchmarking like $9 million to $11 million per megawatt.

$9 million to $11 million sounds optimistic, I would say. I haven't seen $9 million to $11 million. We're not budgeting $9 million to $11 million. We're not far off there. But if your question is, will it cost more than $9 million to $11 million?

[indiscernible] talking for colo too, not for the GPU...

It's up, but only slightly.

Yes. So we wouldn't underwrite $9 million to $11 million, nor would we guide investors to that. If we can, we will. We just wouldn't do it now. So what's interesting, and I think answers your question is to -- is NVIDIA, right? The reason NVIDIA creates a new chip architecture is because the value of the increased compute is more than the incremental cost versus their prior generation chip. So in the HPC market where $1 of compute, you know, compute is compute, its pure commodity, whoever has the lowest marginal cost will sell all their capacity first and then you'll have the next participant and the next participant. It looks like a generation stack in an ISO. And so productivity is the sort of the -- it's just the inverse of marginal cost. That is to say with every successive generation of NVIDIA chip, you can produce more revenue at a given site. And the chip by definition, will generate more benefit and increased revenue than the increase in cost from buying a new chip and constructing it. Because if it didn't, NVIDIA wouldn't have any customers.

Maybe last one for me. How do you think about the strategy for GPU as a service versus colo [indiscernible] shape up for that?

Yes. If you -- if one is focused on maintaining a strong balance sheet that gives you great line of sight to getting through development on a portfolio like ours, that's pretty valuable. I think that the bar would have to be very, very high to invest a healthy chunk of our liquidity in chips, which even if you can finance have only so much life in them. And I think with a plan, I'd say, under all circumstances, we're going to be an infrastructure company, full stop. So the question is, we have the option to do something a little bit different in Washington, but that's 1% of our portfolio. So then we start thinking about how much liquidity that would take, financing available. It feels like -- I'd say it's more likely than not that we -- Washington becomes a standard colo HPC data center. There are a lot of other data centers in the region. They grab up the incremental megawatts quickly. But we're keeping the option open. We might find economic circumstances and say, you know what, it actually makes sense to do this. But for us, this is all measured by like are we creating shareholder value, right? So if you are in the GPU as a Service, the returns expected on that by definition, your cost of capital has got to be higher, right? It's a riskier business. So like are you actually getting the returns commensurate with the higher risk if you enter into that position. So we like -- we're not going to do things unless they are accretive to shareholders, that is to say the use of our equity, the returns on our equity are always going to be designed to better the cost of our equity, right, create value.

Any other questions?

How have you -- it sounds like you're in the process now [indiscernible] your existing portfolio. Have you started to look at additional sites? And how is that process going? How have you stacked up [indiscernible] potentially provide a continual source of additional power [indiscernible]?

We have, and Jonathan has underneath him an incredibly strong corporate development team, the same corp dev team that I think -- I believe has been at Bitfarms for the last 3 or 4 years. They have not slowed down even through the acquisition of Stronghold, they're looking probably at more sites around that period than they have in their entire careers at Bitfarms, and that hasn't slowed down. We just need to find what's exactly right for us as well. We have a very sweet spot. We love that our facilities are on the Eastern Seaboard of the United States and Canada. We think that that's incredibly advantageous. And these sites over here don't pop up every now and again. We could go down to Texas and go and grab 2 gigawatts of promised capacity. We could do that tomorrow. But for us to get stuff in Pennsylvania, Ohio, PJM, where we like to see our facilities, it's a little bit more complicated, but the team is [indiscernible] opportunities daily.

Yes. Are you seeing others look at this market -- to your point, it's -- I'm sure others are looking, but it seems like there's a lot of activity right now. So [indiscernible] seem to be somewhere else [indiscernible] that point or am I correctly assessing...

I'll let Jonathan -- you get it.

Yes, I think. So we get -- we're constantly getting inquiries around pieces of land. Here, the circumstance, we want to be aggressive, and we want to earn excellent returns for our shareholders. If someone came to us with a site that was in our desired location, Pennsylvania being the simplest, and whose quality was at least as good as our existing sites, that is to say time from ownership to getting it online was at least as good as, we would allocate some amount, 5%, maybe 10% of our liquidity to that site, because why not? What we won't do is start spending more than that on sort of real speculative land or other enterprises. And what I always find so interesting about this industry is people, okay, you got the 3 sites, it's $10 billion of CapEx, where are you getting to do next? A $10 billion, you can build an entire LNG export terminal. The largest single project in the United States in 2020, which is not the dark ages, was $5 billion. The amounts of capital that need to get spent in the space are just astonishing. And so the casual nature of a $10 billion pipeline and well, what are you doing for me today, I find interest -- look, that's market expectations, but I think people should -- and I think this is a risk issue that is not talked enough about by participants in the space. If you take on a $10 billion project or $20 billion project, that's the largest thing that's ever been built in the United States, except for the Vogtle nuclear plant down in Georgia. You build an entire chip fab for $20 billion, and there are people trying to build those. So there are going to be some painful lesson learned about how to deliver projects on time and on budget at the scale of CapEx. We intend to learn as many as we can and think really hard about risk mitigants at every stage.

Makes sense. Thank you.

No worries. Guys, we actually have to run. We've got another meeting. Does anyone have any last questions? No. Okay. Thanks so much for your time.

Thank you. Really appreciate it.