Bit Digital Inc

NASDAQ:BTBT

Earnings Call Transcript 2024-Q3

[Operator]

Hello, and welcome to the Bit Digital Third Quarter 2024 Earnings Conference Call. Good morning, good afternoon, and good evening, depending on where you are joining us from. Thank you for being here. We're just giving a few more moments for attendees to dial in, so, thank you for your patience. [Operator Instructions] Also, as a reminder, today's conference is being recorded.

I'll now hand the call over to your host, Cameron Schnier, Head of Investor Relations at Bit Digital. Cameron, the floor is yours.

[William Schnier, executive]

Thank you. Good morning, and welcome to the Bit Digital Third Quarter 2024 Earnings Call. Joining us on the call today are Sam Tabar, Chief Executive Officer; and [ Eric ] Huang, Chief Financial Officer.

Before we begin, I would like to remind all participants that some of the statements we will be making today are forward-looking. These matters involve risks and uncertainties that could cause our results to differ materially from those projected in these statements. I, therefore, refer you to our latest 20-F filing, yesterday's 6-K filing and our other SEC filings.

Our comments today may also include non-GAAP financial measures. Additional details and reconciliation [ and ] the most directly comparable GAAP financial measures can be found in our 20-F filing and today's 6-K filing, which are on our website. After our prepared remarks, we'll open the call up for questions. If you'd like to ask a question, please hit [ #1 ] on your keypad.

With that important note covered, I will now turn the call over to Sam to discuss our performance. Sam?

[Samir Tabar, executive]

Thank you, Cam. Ladies and gentlemen, thank you for joining us on the call today. Today, I will dive into our third quarter results, discuss some ongoing strategic initiatives and provide some color on what we envision as the future for Bit Digital. Eric will then provide more detail on our financial results, and then we will open the line for your questions.

The third quarter of 2024 and the last 12 months have revolved around laying the foundation for the future of Bit Digital. We've embraced a completely different mindset than when we previously had when we were strictly a Bitcoin miner. We now make decisions that we think will bear fruit over a longer time line. We prefer this mindset over the near-term fixation and macro sensitivity inherent to Bitcoin mining decisions.

In October, we announced the acquisition of Enovum, an owner, operator and developer of high-performance computing data centers, for approximately $46 million. This was a transformational deal for Bit Digital. Why? Because the transaction vertically integrated our HPC operations with an existing, fully operational and fully leased Tier 3 data center in a major city.

We added colocation as a new business and revenue line. We also added a strong mix of existing and prospective customers. Further, we gained a strong pipeline of expansion site opportunities and an extremely experienced team with a proven track record to lead the development process. The benefit of adding a highly experienced team cannot be understated. Developing Tier 3 data centers was not a skill we possessed in-house. We [ solved ] this gap through M&A, and now we have what we believe to be one of the best data center teams in the world. This is a huge advantage in terms of accelerating our time to market.

The [ Enovum ] team is adept at bringing new sites online in an accelerated time line. We are pushing forward with our development pipeline and expect to be able to add 8 megawatts by 2Q 2025 and reach a total of 32 megawatts by the end of the year. We will likely dedicate a portion of the near-term megawatts to Bit Digital's GPUs.

Our pipeline remains strong. We are conducting diligence on a location that has the potential to reach 40 megawatts by mid-2025 and an additional 100 megawatts in around 18 months. We're currently evaluating the viability of being able to bring this site online to meet a specific customer's time line, and we will only pursue this project with a firm customer commitment and our own assurance that we can satisfy the time line.

In aggregate, we have seen around 90 megawatts of incremental customer demand since our initial Enovum announcement last month. The demand is spread across 5 to 6 clients with various time lines. Right now, we are seeing that deployment time is -- we are seeing that deployment time is the key factor in turning indications of demand into firm commitments. We are in the final stages of securing the real estate for our 2Q '25 deployment and beginning to place equipment orders to meet that time line.

I'll now move to our GPU cloud business. We announced the term sheet with Boosteroid, the third largest cloud gaming provider in the world, on our last earnings call. Since then, we've executed an MSA, and we've placed an initial order of -- for about 300 GPUs. These GPUs are in the process of being delivered to data centers in the U.S., and we expect revenue generation to begin by the end of this month. This is just the starting quantity of GPUs. It is smaller than the initial quantity that we outlined as this deployment does not include the GPUs to be delivered to European data centers, though we expect that deployment to begin in the near term.

Overall, we now expect our deployment with Boosteroid to reach around 10,000 GPUs through the course of 2025. The deployment cadence should be gradual, though there may be some lumpiness depending on the size of certain purchase orders. We are extremely proud to formally begin our relationship with Boosteroid, and we look forward to going alongside them in the years to come.

Moving to what we refer to as our anchor customer. Recall that we currently have a contract with this customer to deploy an additional 2,000 H100s, but we disclosed that our customer is considering upgrading the contract for newer generation NVIDIA chips. While nothing has yet been formalized, we expect our customer will likely upgrade the contract for [ Blackwell ]. Based on our conversations with NVIDIA and OEMs, we are confident in receiving a relatively early allocation of B200 and GB200s.

At this stage, timing for starting the next tranche of the contract is subject to when we will be able to receive those chips. We don't have much greater visibility beyond publicly available information from NVIDIA. But we do think that 2Q '25 is a decent bet on when we were able to -- when we'd be able to receive the Blackwell chips.

We made significant hires during the third quarter and beyond. Our GPU cloud business is now run by a world-class team. We hired an extremely experienced and proven Head of Revenue and an extremely experienced and proven CTO and multiple new hires on the engineering and business development side for the respective business and technology teams.

Our collective intelligence in the HPC industry has gone up exponentially. This is particularly beneficial given that the GPU-as-a-Service market has become more competitive since the end of 2023 when we launched this business.

In the current market, it's become increasingly difficult to pre-sell GPU capacity, similar to what we did with our first customer. The supply of H100s in the market has increased, and it has become more challenging to win deals without an available inventory on hand. The sales cycle has compressed from months to weeks in some cases.

We are evaluating pre-ordering more GPUs prior to having them contracted to better compete for the types of customers where deployment speed is the key determinant. However, we are more inclined to build inventories of the newest GPU models to reduce the risk of [ obsoletes ]. We are currently in the queue for GB200s and [indiscernible] 200s.

Our pipeline has continued to grow in tandem with the institutionalization of our sales process. We are now actively vetting and negotiating anywhere between 5 to 10 deals at any given time. The pipeline is dynamic in the current market. We have passed on various deals based on pricing or certain customer nuances and missed out on different deals due to not being able to meet deployment time lines.

We have several deals that are being finalized. Last week, we signed term sheets with 2 new customers and executed an MSA and purchase order with a third new customer. The MSA pertains to 64 GPU deployment that commenced last week on a month-to-month contract worth about $1.2 million of annualized revenue. We fulfilled that deployment using GPUs we had on inventory, having purchased 42 H200s or 336 GPU cards in October for around $9.7 million.

And another new customer term sheet is for 576 H200s for a 12-month period, representing $10 million of revenue. And another term sheet provides yet another new customer that provides for 512 H200s over a period of at least 6 months, representing $5 million of revenue over the initial 6-month term. The MSA has been executed with this client and an initial 2 server purchase order was executed, and those units have begun earning revenue. The remainder of this deployment is expected prior to year-end.

We are also actively negotiating several other deals that have the [Audio Gap]

We are also actively negotiating several other deals that have the potential to begin generating revenue by year-end.

In aggregate, we remain confident in hitting our $100 million run rate revenue target for our HPC business by year-end 2024.

In the rapidly evolving cloud computing landscape, many competitors have entered the market, viewing it simply as a matter of deploying capital, [ racking ] and stacking GPUs and then just handing over the access keys. This oversimplification has led to solutions that fall short on performance and reliability, resulting in a wave of cautious buyers who have been burned by underperforming platforms.

Bit Digital recognizes this challenge and sees it as an opportunity. We've assembled top-tier talent to build future-proof performance-first solution engineered to meet and exceed the needs of the most demanding machine learning and AI workloads.

Our platform is designed not only to address today's market complexities, but also to anticipate where the industry is heading. Our focus on a performance-first design means that buyers can trust us to deliver a reliable long-term solution built for even the most advanced applications. This commitment not only restores buyer confidence, but also fosters high retention rates, resonating with customers who are looking for a stable partner in their journey towards innovation.

By delivering on these promises, Bit Digital aims to be the platform of choice for machine learning and AI pioneers who require infrastructure as forward-thinking as they are. While we are in the early stages of this build-out, we believe we have the right framework to create a platform that will expand our revenue and margin opportunities, improve customer retention and reduce churn based on industry-leading reliability and performance.

The third quarter was challenging for our mining business. It was the first full quarter following the April halving event and hash price reached all -- new all-time lows. This coincided with a period in which our electricity costs increased due to normal seasonal patterns. Seasonal increases were most acute in Texas and to a lesser extent, in Upstate New York. These factors [ coalesced ] to produce underwhelming mining margins for the quarter. But Bitcoin prices have since rallied about 50% compared to the 3Q average.

Also, our variable electricity costs have subsided from summer highs. However, network hash rate continues to grow, which will partially offset the benefit of higher Bitcoin prices.

During the quarter, we were informed that Coinmint, our largest hosting provider, was acquired by a third-party, and we received notice of termination for our hosting contracts at those sites.

At the end of the third quarter, we were using approximately 46 megawatts from Coinmint sites. Notably, we are running more than 10,000 S19 type miners with Coinmint. Those units run at an average of 31 joules per terahash, which is inefficient in the current market and are close to the end of their operational [ lives ].

For context, our fleet-wide efficiency was approximately 28 joules per terahash. So the S19s were a considerable drag on our fleet efficiency. We plan on selling those units and replacing the hash rate with newer generation machines. By doing so, we can replace the lost hash rate with around 50% less megawatts. We have already signed term sheets for more than enough hosting capacity to replace that lost hash rate.

We've been very cautious about fleet growth this year, but see this time as a golden opportunity to high-grade our fleet and improve our mining margins. That said, mining CapEx will remain a relatively small portion of our company spend given our plans to develop our data center pipeline and increase our GPU fleet. We're currently assessing different hosting and miner purchase opportunities through the lens of maximizing cost efficiency and compressing payback periods.

We plan to return to around an operational exahash of around 3 in the first half of 2025.

The price of ETH was down 24% at the end of the third quarter 2024 quarter-over-quarter, resulting in unrealized losses on that position. The price of ETH has since recovered and is up 20% quarter-to-date. We hold a significant portion -- sorry, we hold a significant position in both Bitcoin and ETH.

We are not Bitcoin or ETH maximalists. We are shareholder maximalists. We believe that both Bitcoin and ETH will increase in price on a structural basis over time. We think the broader market is starting to understand that Bitcoin and ETH serve different purposes and carry different value propositions. We're still early in terms of awareness, but we expect ETH will continue to benefit from a rising public understanding of its underlying utility. We remain bullish on diversification.

I'll now hand over the line to Eric, who will discuss our financial results.

[Erke Huang, executive]

Thank you, Sam. I will now discuss certain financial results for the third quarter of 2024. Total revenue was $22.7 million, a 96% increase compared to the prior year. The revenue increase was driven by the contribution of our HPC businesses and partially offset by decreased mining revenue.

Our Bitcoin production decreased 59% year-over-year to a 165.4 Bitcoins. Despite more than doubling our hash rate growth, the impact of having coupled with higher network difficulty led to decreased production. Bitcoin mining revenue decreased 11% from the prior year to $10.1 million with higher Bitcoin prices partially offsetting the decrease in production.

Our HPC services business recognized $12.2 million of revenue during this quarter. As a reminder, this business began generating revenue in early 2024. And we earned a 161.9 Ethereums from native staking rewards during the quarter, representing approximately $450,000 in revenue.

Our total cost of revenue was $15.5 million compared to $8.8 million in the prior year, and the increase was driven by an increase in our active mining fleet, higher Bitcoin network difficulty and the start of our HPC business.

Gross profit increased by a 162% from the prior year to $7.2 million. Our HPC services business was the main driver of gross profit, contributing $6.7 million towards the total, while ETH staking added $435,000 and mining contributed $112,000.

The total gross margin expanded 800 bps from the prior year to 32%, with the addition of HPC offsetting lower mining margins related to increased Bitcoin network difficulty and reduced block rewards.

General and administrative costs were approximately $13.7 million compared to $4.9 million during the prior year's quarter. Third quarter 2024 G&A expenses includes $5 million of stock-based compensation and $4.8 million of professional and consulting expenses. The increase related to the prior year can be partially attributed to expense related to our acquisition of [ Enovum ] as well as expenses related to hiring new employees for our HPC business.

Depreciation and amortization expense was $8.4 million for the third quarter compared to $3.6 million last year, with the increase driven by a larger mining fleet and our GPU fleet that was deployed in early 2024.

Adjusted EBITDA was negative $21.8 million for the quarter compared to negative $2.9 million last year. Adjusted EBITDA for 3Q includes $21.9 million unrealized losses on digital assets as well as certain costs for our [ Enovum ] transaction and new hire that we would expect to be [ nonrecurring ]. We expect normalized going forward G&A to be more along the lines of prior quarters with slightly upward adjustments for additional headcount.

GAAP earnings per share was a loss of $0.26 for the quarter compared to a loss of $0.08 in the prior year with the change in digital asset prices being the primary driver of the decrease.

Turning to our balance sheet. We held approximately a $106 million of cash and restricted cash as of September 30, 2024, and our digital assets [ recognition ] was worth approximately a $118 million. Total assets amounted to $376 million at the end of the quarter, while shareholders' equity was $315 million.

Our balance sheet remains debt-free, but we continue to evaluate potential debt financing options to accelerate growth of our HPC business. The [ Enovum ] transaction and planned CapEx related to data center infrastructure opens new possibilities for financing that we are actively exploring. CapEx was approximately $1.6 million during the quarter. The CapEx was used for [ deposits ] on certain HPC equipments.

As a note, this will be the last quarter that we file as a foreign private issuer, and we will soon transition to a domestic filer and plan on filing a 10-K when we publish our 2024 annual results.

I'll now turn the call back to Sam for closing remarks.

[Samir Tabar, executive]

Thank you, Eric. We are extremely excited about the direction of Bit Digital. The company has been completely transformed relative to where we were just a few years ago. We believe we are on a credible path towards becoming one of the preeminent players in the HPC industry.

We entered the HPC market in 2023 with our existing team at the time and relied on our core competencies to launch that business. We continue to believe that there are overlapping skill sets between Bitcoin mining and HPC and that they can be synergistic or complementary businesses under the same umbrella.

That said, the overlapping skill is like comparing the basic skill sets of a taxi driver and the Formula 1 driver. Both Bitcoin mining and HPC require capital, but deploying capital is just the beginning of the HPC life cycle. HPC isn't just about winning customers. It's about keeping those customers happy and retaining them.

Unreliability is a critical issue in the HPC market, and it creates churn. Churn creates opportunity for a customer-first business model, which is what we are driving towards.

We learned quickly that we need to augment our talent base to really compete in this industry. The GPU-as-a-Service market has become more competitive, and we needed to recruit top talent to effectively compete. We solved that need by hiring a Head of Revenue and CTO in addition to other key team members across the sales and technical functions. We then acquired one of the best HPC data center teams in the business through our acquisition of Enovum.

The groundwork is now in place for us to become a formidable player in both the GPU cloud and HPC data center business. We are only scratching the surface on the operational synergies we can achieve between our GPU cloud business and our HPC data centers business. There has never been a more exciting time to be a part of Bit Digital.

As for our mining fleet, we are planning to opportunistically high-grade our fleet to lower production costs and enhance our competitive position. However, it's clear to us that simply maintaining market share requires significant capital given the explosive growth in network hash rate.

Bitcoin mining is an [ arms ] race. Essentially, companies like us that are investing heavily in HPC can't compete with pure-play miners on an exahash basis given that capital is finite. But we are comfortable with this dynamic. We view HPC as a better long-term investment for our company and are comfortable viewing our mining business as a noncapital-intensive call option on higher Bitcoin prices.

It doesn't seem realistic to us for a company to maintain material market share in mining while also developing a highly competitive HPC platform. Eventually, hard choices on where to invest capital will need to be made.

The predictability of HPC cash flow can be helpful in smoothing out the volatility inherent to Bitcoin mining. HPC provides stable revenues. Digital asset offers [ spicy ] asset appreciation. Stable and spice is the perfect balance for our company. This was on display during 3Q when our HPC business helped offset a challenging quarter for our mining business.

However, an unrealized loss on our digital asset position caused our headline results to appear underwhelming. This dynamic in our digital asset position looks like it could reverse during the fourth quarter based on current asset prices.

Our perception is that the market doesn't ascribe proper sum-of part value to a company that has both an HPC business and a digital asset business. We suspect that separating the 2 businesses may result in a more appropriate valuation for each discrete business line, and that is something we are continually studying internally.

We believe our HPC business is primed for an exciting and prosperous future and that we have exposure to the most promising and exciting asset class of the 21st century with our digital asset position. If the market continues to underappreciate that concept, we may proceed with the separation of the business lines.

For now, we are laser-focused on prudent capital allocation. We currently have a wide menu of options for capital deployment compared to where we were as a company 18 months ago. Capital is finite, and we fortunately have 3 distinct options to choose from for deploying capital at any given time.

Funding needs for our data center business enables us to be more selective in terms of the GPU contracts we pursue and vice versa. These new business lines have opened an array of financing options that weren't available to us previously as a Bitcoin miner. We're actively assessing these options and looking to understand what the most cost-effective sources of capital we can reasonably pursue.

Specifically, we're looking at potential commercial mortgage financing options for our data center build-out. We're also looking at various vendor-based financing and leasing options on the GPU side.

Cost-effective and non-dilutive growth financing is the goal. We continue to evaluate more traditional corporate credit options and were recently close to a deal with a prominent counterparty, but decided to hold off until we had a fuller understanding of the financing options available to us in the wake of the Enovum deal. We will provide -- we will continue to prioritize long-term value creation for all stakeholders as we evaluate options for sustainably growing our business.

Lastly, to reiterate Eric's point from earlier, this will be the last quarter that we file a 6-K. We will convert to domestic issuer status and our next annual filing will be in the form of a 10-K. We believe that this move should incrementally help our stock become more accessible for certain investors.

With that, I would like to open the line for some questions.

[Operator]

[Operator Instructions] We will take our first question from Mike Grondahl with Northland Securities.

[Mike Grondahl, analyst]

It's Mike. Your pipeline on the GPU rental business and with Enovum is pretty robust. But can you help me, Sam, you're doing about $4 million of revenue a month in that business, the GPU rental business. How are you getting to like $8 million at the end of the year? I think that gets you to your $100 million annual run rate target roughly. Can you kind of help me with that bridge?

[Samir Tabar, executive]

Yes. So it's late November. We reaffirm that by year-end, our revenue target will hit a $100 million. I'll discuss why. First of all, we're pretty confident. And we're expecting the term sheets that we announced today on the call to commence prior to year-end. And combined with Enovum, that gets us to about 90% of the way there, and we still have a very busy sales cycle. As shown this week, deals can come in bunches, and we now have the institutional bandwidth and processes to manage sales.

Deployment time lines will effectively determine whether we reach the run rate in December or January. But currently, we're almost 90% there with the term sheets and the MSA that we announced today, including the Enovum contribution.

[Mike Grondahl, analyst]

Okay. And if I remember right, the Enovum contribution was like $1.6 million in the fourth quarter. Is that number still accurate?

[Samir Tabar, executive]

I believe it's about $7 million by the year-end. Eric, who's on the line, our CFO, can you confirm that?

[Erke Huang, executive]

Yes, that's a $7 million ARR going forward, starting the fourth quarter of this year.

[Mike Grondahl, analyst]

Okay. I guess I was asking about the fourth quarter of '24, the contribution from the acquisition. I had in my notes $1.6 million.

[Samir Tabar, executive]

Yes, but we look at it from an annualized -- go ahead. [indiscernible]

[William Schnier, executive]

You are right. It's -- $1.6 million is the aggregate sort of revenue that we recognize, but the December run rate would just be slightly higher. So $1.6 million annualized versus $7 million, there's not a huge delta. It's just that certain contracts passing towards [ MD ] are higher, so a higher run rate in 1Q.

[Samir Tabar, executive]

Yes, I was giving the annualized number. Yes.

[Mike Grondahl, analyst]

Okay. And then secondly, Sam, can you spend a minute -- Coinmint basically gave you 90 days on 36 megawatts. It sounds like you've replaced that. How should we think about average exahash 4Q, 1Q and 2Q '25 for your Bitcoin business?

[Samir Tabar, executive]

Yes. Well, as mentioned, we will get to 3 by the first half of next year. We were running about 1.4 exahash with Coinmint. And -- but again, we plan to return to around 3 exahash by, I think, around end of February, likely at the cost of around $15 million to $20 million, depending on the contracts.

But just as a reminder, Coinmint cost of production was slightly above our company-wide average on a year-to-date basis. And mining will remain a small part of our capital deployment going forward. We have signaled our intention to maintain a small fleet, and this represents maintenance CapEx to maintain the size and make the fleet more efficient, lower production costs and increase our margin of safety from a production cost perspective in a low hash rate price environment.

We're still not interested in owning mining infrastructure, vertically integrating our mining, given the volatility of Bitcoin, that remains our position.

[Mike Grondahl, analyst]

And did you say 3.0 by the end of February or by midyear?

[Samir Tabar, executive]

Well, it's somewhere between conservatively end of -- by the end of the first half of the year and potentially by the end of February. So it's going to be somewhere in that range.

[Mike Grondahl, analyst]

In that window. Okay. And then lastly, I don't know, since July 1st or the end of June, how much did you guys pull off the ATM, maybe number of shares up until last Friday in total dollars raised?

[Samir Tabar, executive]

Well, I think we issued 38 million shares on the ATM since the end of June. We have massive CapEx plans and the equity issuance has been the bridge gap. Fortunately -- look, we need to ensure that we have enough capital to get the ball rolling in our data center build-out and to provide for near term GPU ramp.

And we've been careful on the credit side. Being risk adverse has always been in our DNA. So getting [ acclimatized ] to new industries where more risk is appropriate and can be underwritten confidently is something that we've -- basically it's been our North Star.

As mentioned on my call, we're very, very close to a massive debt deal, but we didn't feel like we had a firm enough handle on all the available financing options. And the deal would have made alternative options difficult, and we're [ wary ] -- very wary of maintenance covenants. So we are looking at commercial real estate loans. We're evaluating other financing avenues. We're just being very thorough as we evaluate financing options. Our North Star is to not dilute, rather go on the debt side. We will get there.

[Mike Grondahl, analyst]

Got it. And just as a follow-up, the 38 million shares, do you have an average price or dollar amount raised? And then just, what is the pro forma new share count fully diluted?

[Samir Tabar, executive]

I don't have that information in front of me. I can try and get back to you perhaps after the call.

[Operator]

We will take our next question from Kevin Dede with H.C. Wainwright.

[Kevin Dede, analyst]

Sam, you went through a couple of the new HPC deals that I think totaled to about $20 million on an annual basis, 1 MSA and 2 signed. Could you just kind of give us a little more detail on that again because it wasn't really clear to me? Apologies.

[Samir Tabar, executive]

No, that's okay. Perhaps the confusion is I'm using the annualized run rate number as opposed to the [ absolute ] contribution. Cam or Eric, why don't you drill down a little bit on giving the information on client 1, Boosteroid, client 3, client 4 and client 5 and how that gets us to $90 million.

[William Schnier, executive]

I mean, I could just sort of walk through the term sheet. So there's 3 total term sheets, 2 of them have converted into a -- the [indiscernible]. [ 2 ] have already moved to the MSA, so there's one that hasn't. One term sheet is for 576 H200s for 12 months. That's around $10 million of revenue. The other is 512 H200s for a period of at least 6 months. The 6-month period represents around $5 million of revenue, annualized gets to around $10 million. The MSA for this has been signed [ and ] -- [ from ] initial purchase order, a small one, went through. We expect the remainder of that to be through year-end.

And the third is a month-to-month contract. That's term sheet and MSA. That's 54 GPUs, representing $1.2 million of annualized revenue. It's around $21 million to $22 million ARR total across those 3. [ 2 ] to 3 are already at the MSA stage. One, we still have to, in that MSA. And based on what we have on hand, I mean, it's a total of like 1,160 H200s across those 3, we already had 336 on hand. So it's basically solving for an additional 800.

[Kevin Dede, analyst]

And you might look at vendor financing for that as you have in the past?

[William Schnier, executive]

Yes, perhaps. I mean, I think for the initial -- for the aggregate GPUs, just to solve for that, probably in the range of $25 million [ sale ]. [ Like ] we certainly could look to other options or take it down on the balance sheet based on the cash, still working through that. So we'll provide those details when we finalize the respective deals.

[Kevin Dede, analyst]

Great. And what about the power side? I think Sam alluded to having data center optionality in the U.S. versus building Enovum. Is that sort of the way to think about it, because Enovum won't be ready to handle that capacity at this juncture?

[Samir Tabar, executive]

[indiscernible] Go ahead.

[William Schnier, executive]

Is it for the 4Q deployments? Some of those will be going to Iceland with the capacity we already have. And there's -- I think the Enovum capacity issue you were referring to, just like building additional by 2Q and the note that we might take some of that capacity for our own, this would -- that would pertain to separate contracts, not the 3 term sheets mentioned today.

[Kevin Dede, analyst]

The GPUs required for Boosteroid aren't the high-end NVIDIA type. Am I right in assuming that? Can you give us some...

[Erke Huang, executive]

That's right.

[Kevin Dede, analyst]

Cost ballparks on that?

[William Schnier, executive]

Yes, they're lower...

[Erke Huang, executive]

Go ahead.

[William Schnier, executive]

Those are gaming GPUs and specifically made by AMD. So certainly, a huge cost differential. But I mean, for that initial 300 GP deployment, it's a very small number.

[Kevin Dede, analyst]

Okay. And obviously, you're very confident that contract accelerates in deployment. When do you think you'll be able to signal to the Street on how that looks?

[Samir Tabar, executive]

You're referring to our target run rate or something else?

[Kevin Dede, analyst]

Yes. No, the Boosteroid GPU deployment getting you to your target run rate with that.

[Samir Tabar, executive]

Yes. We're pretty confident. I mean, like just last week, 3 deals in a row. So we still have November and December -- half of November and December. We're pretty confident we'll -- right now, we're almost at 90%.

[Kevin Dede, analyst]

Yes, I understand that. I'm more curious about Boosteroid specifically, because there seems to be pretty big targets there. I just was wondering about your confidence there and when you might be able to give us more insight on that?

[Samir Tabar, executive]

Yes. I mean there has been -- yes, the MSA there has been signed. There's been a purchase order of about 300 GPUs. Another deployment is expected in Europe this year. It's going to be a gradual build-out, and it's going to be lumpy based on orders. It could also be lumpy based on orders and market conditions. But based on our order book for other customers, a gradual and slow deployment cadence would frankly be the best scenario for us just in terms of capital deployment.

So margins are very high. They're -- We don't have the precise [ IRR ], but it meets our threshold for this sort of contract. Our CapEx for the first deployment is about $2 million.

[Kevin Dede, analyst]

Right. But that doesn't include your -- I guess, your -- the payments you need to make to secure power and space racks, right?

[William Schnier, executive]

That was in the form of pass-throughs to the client.

[Kevin Dede, analyst]

Okay. On Coinmint and your hash rate changes, I understand you're not going to take the S19s, but you'd look to something like maybe an S21 where for the same power you use half the machines and get the same hash rate. Is that the right way to think about it?

[Samir Tabar, executive]

I think so, yes. That is the right way.

[Kevin Dede, analyst]

Okay. And you mentioned something about $20 million to $25 million. Would that be the amount you need to spend to get to your 3 exahash the way you're looking at it now?

[Samir Tabar, executive]

Yes. I would love Eric -- for Eric to confirm. But just before that, I just want to mention that we are the exclusive partner to Boosteroid, just worth mentioning. But going back to your question, Eric, do you want to confirm that number?

[William Schnier, executive]

I mean, [ that's ] [indiscernible] $20 million to $25 million. Just -- if it's about an exahash, you can look at market prices, but probably closer to [ $10 million to $15 million ] for that deployment, Kevin.

[Kevin Dede, analyst]

Congrats on the HPC growth.

[Operator]

We will take our next question from Brian Dobson with Clear Street.

[Brian Dobson, analyst]

So you had mentioned potentially separating the Bitcoin mining business, and that's understandable. I guess, at its core, do you see that potential move as an effort to make the company's new focus easier to understand for investors?

[Samir Tabar, executive]

Yes, absolutely. We have 2 different conversations going on these days, one with retail, one with institutions. They both want different things. We continue to study it. It seems a very viable way to unlock the value of the HPC business. The valuations and multiples on that are much higher than Bitcoin mining.

This also has massive benefits from a financing perspective. So, some banks are reluctant to deal with anything crypto. So this makes it a much cleaner story for lenders if we were more pure-play HPC company than having that crypto business, which limits our financing options. But again, from a valuation perspective, it is -- night and day, it does make sense to, frankly, pursue this option sooner rather than later.

[Brian Dobson, analyst]

Yes, that makes a lot of sense. And I like your description of the business as stable and spicy because at this point, I think it's fair to say that HPC is the core of the business, the future of the company going forward. So I guess --

[Samir Tabar, executive]

Absolutely.

[Erke Huang, executive]

We don't disagree.

[Brian Dobson, analyst]

So on that note, as you're looking at Enovum, the acquisition there, as you [ peer ] into that pipeline, which is attractive that you brought that over, what are you most excited about?

[Samir Tabar, executive]

About Enovum?

[Brian Dobson, analyst]

Yes.

[Samir Tabar, executive]

I mean... Yes. I mean, I would love to -- just as a thought, I will answer that question preliminarily, but I do want to mention that we have Billy, the CEO of Enovum on the line. I'd love for him to pipe in, no pun intended. But with respect to our pipeline development schedule, that's one of the things I'm extremely excited about. We kept the total pipeline constant at 288 megawatts in our deck.

We noted that we see an additional 100 megawatts based on conversations from the utility at a site we are reviewing, but we won't include that in the pipeline for now. If capital was unlimited, we'd break ground on it today, but we're not ready to formalize that site quite yet. We're still studying time lines internally. And we are planning on moving ahead with our plans. And Billy, I would love to perhaps hear from you on that. Billy, the floor is yours.

[Billy Krassakopoulos]

Good morning, everyone. Yes. So, I mean, it's just been just over 30 days since the acquisition. So I mean, we're formalizing procurement processes. We have a good line of sight on major long lead time items. What excites us the most is bringing all of this capacity online to meet the customer demands that we have for 2025. We're evaluating a different amount of -- large amount of projects, sites. It's a bit of a dynamic process right now. But, like I said, we have a very good line of sight, and we're looking -- we're very anxious to deliver.

[Operator]

We will take our next question...

[Samir Tabar, executive]

I also have -- Yes. Just to mention, as the next question comes on, I also have our Head of Revenue on the line on the GPU-as-a-Service business in case there are questions relating to that. So who's next?

[Operator]

Yes, we will take our next question from Joe Gomes with NOBLE Capital Markets.

[Joshua Zoepfel, analyst]

It's Josh filling in for Joe. So, you guys were talking about just obviously, the HPC side of the business. And I wanted to see kind of just an update just kind of on the site pipeline? You guys mentioned on the call earlier, but have you guys really [ secured ] any power outside the Montreal area? You guys were talking about going to Western Canada, Ontario, maybe the British Columbia, but can you kind of expand on that?

[Samir Tabar, executive]

Billy, why don't you take that on? The floor is yours.

[Billy Krassakopoulos]

Sure. Like I mentioned, we had a couple of projects on the line for Western Canada, Ontario. In the last couple of weeks, we've had an -- a site in the United States come across our table that we're actively vetting right now. So it'd be the first project that we would do in the United States, but we're well aligned and ready to go to execute that.

[Joshua Zoepfel, analyst]

Okay. Perfect. And you guys talked about obviously having the 3 exahash maybe by the end of February potentially and then going into maybe the end of the second quarter, middle of the year. With Bitcoin kind of trading at where it's at, maybe is there a chance that we might see like active hash rate higher than that, like you guys maybe activating idle miners?

[Samir Tabar, executive]

Well, I think the Bitcoin price going up is kind of sometimes misleading because there is another -- the other part of the equation which is hash rate difficulty, and that obviously has an effect on the margins of Bitcoin mining -- of the business of Bitcoin mining, which makes it a little bit more complicated than just the price of Bitcoin.

We think that capital is finite, and we have huge ambitions, and we're frankly executing on those ambitions on our GPU cloud and data center business. And so, that's where we have to really kind of put the capital. But we still remain at an incredible call option on digital assets. And as I mentioned earlier, our company is stable and spice. We have the stability of HPC revenues, and we have the spice of our digital assets position, both in Bitcoin and Ethereum.

[Erke Huang, executive]

Yes. And this is Eric speaking, if I may add, with the most recent [ halvening ] happened in April. So we see this increase of Bitcoin price as opportunity to upgrade the fleet. So what we're going to do is we're going to sell some of the older models rather than keep them running and replace them with newer generations miners because we put those older models as a way of storing Bitcoin.

When the Bitcoin prices are higher, we can sell them at a relatively higher price. And with the [ halvening ] just happened, we wouldn't necessarily have [ halvening ] in the next [ 3.5 ] years. So that gives us more stable, predictable revenue for this cycle. So upgrading and liquidating older models will be the strategy we're taking for the next couple of quarters.

[Operator]

We do not have any further questions in the queue.

[Samir Tabar, executive]

Great. Okay. I guess that concludes the call. We look forward to engaging with our stakeholders on any further questions. And thank you very much for the call today, everybody.

[Operator]

This concludes today's call. Thank you for your participation. You may now disconnect.