HIVE Digital Technologies Q1 2027 Earnings Call Transcript

HIVE Digital Technologies (NASDAQ:HIVE) held its first-quarter earnings conference call on Monday. Below is the complete transcript from the call.

This content is powered by Benzinga APIs. For comprehensive financial data and transcripts, visit https://www.benzinga.com/apis/.

View the webcast at https://us06web.zoom.us/webinar/register/WN_LRcK8oGmRWuM33I40R6Cnw#/registration

Summary

HIVE Digital Technologies reported a 73% year-over-year revenue increase to $79.1 million, driven by its Bitcoin mining operations and expanding HPC and AI business.

The company announced $180 million in contracted revenue for its HPC business, with a significant five-year $350 million contract recently secured, indicating strong growth potential.

HIVE continues to focus on expanding its AI infrastructure, leveraging partnerships with Bell Canada for GPU clusters and exploring additional financing strategies to support further growth.

The company faced a $143 million net loss due to non-cash items, including a $85 million provision for a disputed Swedish tax liability, which HIVE plans to contest.

Management highlighted operational efficiency, with a lean cost structure and strategic hires, contributing to a positive adjusted EBITDA of $13.4 million for the quarter.

Full Transcript

Nathan Fast, Director of Marketing and Branding

Hello and welcome to today's webcast covering HIVE Digital Technologies' financial results for fiscal Q1 2027. My name is Nathan Fast, Director of Marketing and Branding at HIVE, and I'll be your moderator for today's call. Before we get started on slide two, I'd like to briefly note the disclosures for today's presentation. Except for statements of historical fact, this presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. Words such as "expects," "believes," and similar expressions identify these statements. Actual results could differ materially and we disclaim any obligation to update them except as required by law. For a full discussion of risk factors, please refer to our most recent SEC filings at sec.gov. In addition to discussing results that are calculated in accordance with GAAP, we will also reference certain non-GAAP financial measures including Adjusted EBITDA, Adjusted Net Income, and Free Cash Flow.

Management uses these metrics to evaluate operating performance and believes they provide investors with additional insight, and they're presented for supplemental purposes only and should not be considered in isolation from GAAP results. Reconciliations to the nearest GAAP measures are included in the appendix to this presentation and in the press release and Form 8‑K furnished to the SEC. On the next slide, I'm pleased to introduce today's presenters: Frank Holmes, Executive Chairman; Aydin Kilic, President and CEO; and Darcy Daubaras, Chief Financial Officer. I'd now like to hand the presentation over to Mr. Frank Holmes for a macro recap of the quarter.

Frank Holmes, Executive Chairman

Thank you. And so let's go a quick macro recap of what's been affecting stock markets and HIVE's stock price and valuations over this past recent quarter. Next please. So before we jump into those granular details, it's always important to understand the DNA volatility, and every asset class has its own DNA volatility. And this is to highlight that—especially we find so many people like to trade HIVE—it's because of that volatility. It is a non‑event over one day to go up or down 6%, and over 10 days 23%.

You can see that's a little more than CoreWeave. It's substantially three times more than what Bitcoin is, and four times more what Nvidia is when we look over six trading days. So it is just a factor of these macro forces pulling, with these announcements from what Bitcoin is doing on a daily basis to the announcements of our AI build‑out. So when we have a strong Bitcoin day and good AI sentiment, the stock has these big surges; vice versa, a negative day in the world of AI and Bitcoin down, and you get these downdrafts.

These appear to be the factors driving a lot of the sentiment. The leadership team has been up here before. Aydin, CEO, the best operator—and I'll walk you through why—in the data center business. And Craig Devar is our superstar builder—an incredible builder of data centers—to build out our vision in Canada in particular, which is the fastest track that we have, and the government's committed to it with having a minister for AI. Darcy Daubaras, our CFO, Gabriel Ibghy is in Europe, and Gabriel Lamas who's in Paraguay, and Johanna Thornblad is also in Europe.

HIVE operates over nine time zones in five languages, and we're able to manage that complexity and still be one of the most efficient Bitcoin miners. And our HPC AI data centers that we have functioning today in downtown Montreal, Stockholm, in Manitoba and British Columbia—our efficiency is always ranked by third party—it's the top of the stock. HIVE has always used green energy—Canada, Sweden—powered by low electricity cost, low temperature, fast internet connection, and what we've been doing to accelerate the AI build‑out is largely dark fiber.

Dark fiber in British Columbia—no, sorry—dark fiber in New Brunswick in Canada. And dark fiber is next to be laid in Paraguay. But we've been able to demonstrate on this ability to convert electrons and export to what they call compute power 5,000 miles away. So that's pretty impressive. And so we're pretty excited about what the opportunities down the road is from Paraguay. I want to thank all the shareholders. You can see Invesco, Citadel Advisors, Millennium Management, Two Sigma Investments, Charles Schwab.

Charles Schwab is probably the biggest in the retail component from registered investment advisors and the retail public. So thank you all for listening to the call. For being shareholders for the quarter ended June 30th, you can see that HIVE outperformed its 50‑day moving average. It had a big sell‑off along with the whole industry in July, and it was one of the biggest outpouring in the month of July by hedge funds. And a lot has to do—which I'll mention later on—about the carry trade out of Japan and the significance of this sort of domino effect, a contagion.

Japan wanted their money back because rates are going up, and it impacted first the Korean stock market, which was one of the best performers last year—up about 100%. It had a huge correction as particular technology and AI‑related stocks sold off out of South Korea. And we saw that dominant impact in America, and we saw a lot of AI—anything related—also being deleveraged. And I think that that sort of worst is behind us. And these sentiment factors, they swing back and forth.

What investors have to realize is, this carry trade of Japan is quite significant. In my years as a money manager, I've seen the swings back and forth globally of what it's done. And so it appears that the worst is behind us right now out of the Japanese economy. We're happy to see that HIVE outperformed Bitcoin at the end of June when you look over that time period this year, and I think a big part is much of our AI strategy and announcing the growth in the revenue from a million a month to $3 million a month, and now expanding that.

It'll be going to five to six to $10 million a month with these contracts that we're building. And we have machines that are increasing this year‑to‑date substantially over the year. The daily cash flow coming from our GPU chips, in particular the suite of Nvidia chips we have. I mentioned earlier that Japan carry trade unwind is a macro risk. I think the worst is behind us, but any rising rates in Japan will have a big impact. We've seen the administration come in to support Japan's currency even though rates were rising, and it appears to be sort of an aftermath of COVID‑19.

After three decades, inflation started showing up in Japan and rates started rising. They went for three decades of basically, a majority of the time, issuing zero cost of capital—or 10 basis points. Funds would borrow in Korea, in the U.S., Canada, England, and get U.S. rates from 4% up to dividend‑paying stocks and in speculative stocks. That was starting to unwind as Japan started seeing the rates rise and they want their money back. This is a real important phenomena for investors just to be able to follow the Japan carry trade because it can impact you.

You don't realize why the stocks are off or up or down during the day—there's no news—and it could be an unwind or reloading around the world. Well, it's been exciting to respect that. Last year we substantially increased our hashing power. In Paraguay in particular, taking overall complex from 6 hexahash to 25. That gave us economies of scale and that gave us the ability to deal with the drop in bitcoin pricing and the difficulty rising. And those are two real significant headwinds, especially in February of this year. We've been able to make money every month, even though, and I could share with you, had we not strapped on that additional power, we would be very difficult to be operating a bitcoin operation today.

But we have this key scalability at 2%. And then we've been deploying our bitcoin into expanding our AI Gigafactory vision. So Aydin's going to give you—and Darcy, our CFO—and Aydin Kilic, our CEO, is going to give you more granularity on these numbers. But underneath the hood, underneath all these non-cash charges, such as the depreciation that we have to take for our chips, and in particular the non-cash charges out of Sweden—which we'll have more discussion on as we go on—this ongoing battle with the Swedish tax agency which changed the rules from when we initially went there.

The interpretation, the laws haven't changed. It is their interpretation because they're very anti-crypto industry. And even though we try to explain that we're exporters of electrons: we take hydroelectricity, run them through an ASIC chip and we export that hashing power, compute power, to Foundry, which is a pool in the US which is SOC 1 and 2 compliant, and they pay us in bitcoin. In Bermuda, it's an export industry. And what's really exciting is that Paraguay's Central Bank gets that.

And it's important now for them in calculating what the GDP is and the contribution to their GDP of how we use a chip. And the AI business is very similar. That is, you take the electrons you run from the hydroelectricity of hydro, they go through your GPU chips and you can transmit that GPU power, compute power, to New York City. And we demonstrated that in a test—5,000 miles. Now, you cannot send electricity on transmission lines 5,000 miles, but you can convert them into compute power and send along fiber-optic cables along the bottom of the ocean 5,000 miles.

And someone can create models in New York City, at Columbia University, which has basically validated that exercise. So when you look at a bitcoin mining industry or you're looking at an AI Gigafactory, what investors have to realize is that you don't have an ATM, you don't have a bitcoin machine that spits out bitcoin. No, you really export out to a pool—that's bitcoin mining—or to an end user who wants your compute power and you can transmit that.

So this is phenomenal because you can take all this abundance of electricity in Paraguay and you can turn around and export by converting those electrons into compute power and they can be used all over the world. So it's very exciting what we are doing on that. But it doesn't stop this ongoing saga in battling with Sweden. And we believe that, based on law and our expert witnesses, we will go through this process and we believe that justice will prevail.

But what's important for you is to recognize underneath the hood is that the revenue quarter grew 10%. Net operating income grew by 86% because the beginning of this year, February, was a very big challenging month. So the year-over-year goes to show the significance of the scaling I mentioned at the very beginning. Taking revenue as year-over-year grew by 73% even though bitcoin fell by 40% and operating net income grew by 50%. So we're really thrilled of our positioning of having 2% of that global network and continuing to build out that dual engine.

In particular this year is our HPC, high-performance compute, strategy to build AI Gigafactories. And Craig Tavares will give you more granularity as we go on. But who's really going to carry that ball for you today is going to be Aydin Kilic. The team is very, very proud of closing another 130 million of exchangeable notes, zero cost of capital for interest payments. That money is not earmarked for Sweden. That money is earmarked for the fastest path to cash flow and revenue, in particular in our partnership with Bell Canada.

The countries that win the AI race won't just produce the smartest engineers; they’ll build the infrastructure to support them. And data centers are becoming as essential to economic growth as railroads were a century ago. And a quote I've been saying in speeches, and one of the things that in our Gigafactory being the biggest in Canada, I think what's really interesting is that it's really not taught in school in Canada to the degree. But I grew up there as a child, and Toronto was very proud of their medical breakthroughs and research.

It inspired me, and part of my journey of education was going into medical school before I pivoted to go into business and economics because I was just so in love with what was taking place in Toronto and Waterloo and then University of Western Ontario where the founding school is called Huron University. I sit on the board. I was always thrilled to know in 1921 insulin was discovered in Toronto. Pablum was developed in Toronto. The pacemaker, pioneered.

Polio vaccine in scaling—it was in 1955 when I was born—was done in Toronto. The first successful double-lung transplant, cystic fibrosis gene—I can go on. But it is rich with intellectual capital with two major schools, the University of Toronto and Waterloo University, which is like Canada's MIT. There are many other schools all around like the Schulich School is famous for business in Toronto, but there are many scientific research laboratories in that area, and Toronto is the financial capital of the country.

Ottawa is like Washington D.C.—it is the political capital of the federal government. But the universities are really in that Greater Toronto Area. So we did a campaign to try to educate investors in Canada as well as in the US about the University Health Network, SickKids Hospital. One of my friends is a doctor, a pediatrician, and his internship was in Toronto at SickKids; the Princess Margaret Cancer Centre; the Vector Institute where the Nobel Prize winner two years ago, Geoffrey Hinton, was at the University of Toronto.

This use of artificial intelligence to improve cancer diagnosis, drug discovery, medical imaging—it's very, very big. But what they do not have are these big Gigafactories, AI Gigafactories. And that is where HIVE is buzzing with activity to build out. So other things is to understand and appreciate Toronto more. Outside of the Maple Leafs or the Raptors basketball and the Blue Jays baseball. The telephone innovation was not too far from the city of Toronto.

The electron microscope in 1938. Deep learning 2006. Ethereum—next, please—was discovered by a student out of Waterloo University. What's also important in this visual is to show you the triangle of concentration of internet nodes and, in particular for AI and transferring of data. The big concentration is from the Toronto region up to Boston and down to Virginia. This is a visual to show you where Lake Ontario is and Finger Lakes. As you see is New York, and University of Toronto, and the HIVE Gigafactories right in between Waterloo University and University of Toronto.

The University of Toronto is much closer to where Toronto is in this map. But it gives you an idea that this is an important intellectual capital. Just like I mentioned before that you look at bio research—50% of bio research is done in Boston, San Diego. If you look at other types of gaming, there's clusters of software coders that show up. And when you look at cybersecurity, one of the biggest clusters is right here in San Antonio, Texas, with UT University which has the biggest cyber school, cybersecurity school, in the country with 10,000 students.

So Vitalik Buterin, the creator of Ethereum, went to Waterloo, AKA as I call it, the MIT of Toronto because of the software number of geniuses that come out of that school. So this is to give you an idea from a macro point of view which has happened this year. Semiconductors have had a big run and they started coming off with the correction as you can see here in July. And a lot of this had to do with the contagion from Japan to Korea to North America to Europe.

And it looks like that worst is behind us now. What's important when I look at—I hear, oh, it's a bubble. It's already... it's the worst ever. It's a bubble, bubble, bubble, bubble. And all these people are coming in with PhDs in bubblology. I share with you that it's far from that. And the amount of these collateral minerals you need—lithium batteries, not just for cars, but you need them for all these data centers. And you can see the drive for lithium and graphite and nickel and zinc and copper.

Copper is making an all-time high because a gigawatt electricity converting that to an AI factory like Texas is going to consume 50,000 tons—not pounds—tons of copper. Most people think of pounds of copper—no, 50,000 tons, and that's 2,000 pounds per ton. That's a significant amount. And when we look around the world it means that the supply is limited and it continues to have big demand globally. That's another sort of demand that I see that's going to continue with building out of AI centers.

And there's lots of political headwinds as election year, so it becomes on the agenda. But it's not going to stop this buildout. This is just looking at spending for you and you've seen this in other visuals. But it's just important to recognize the peripheral equipment for computers—those stocks have been on a tear. And it's not just here in America. It's a global phenomenon that anyone involved in the building construction of these data centers.

I mentioned earlier bitcoin, particularly the Central Bank of Paraguay, recognized that tier-one data centers are a big contribution to the GDP because it's a way for the country to export besides food and beef. They're able to export electrons and they export them with compute power, and it's significant contributing income to the country for its size. So I'm very pleased that the Central Bank has had this ability to really grasp the significance of this being an export industry.

And hopefully other countries will all of a sudden recognize that, like Sweden and Canada, Norway and Iceland, that it's an important component of converting hydroelectricity or any other type of electricity with GPU chips to be able to export that compute power. Just a visual for you to grasp: you cannot transmit electricity from Paraguay to New York, but you can send AI compute power, and those GPU chips is what we do and we've shown it, we've demonstrated, and that is a big breakthrough.

We think this is very significant what we've been able to demonstrate. And after we lay all the dark fiber necessary in Paraguay like we're doing in New Brunswick on the border of Maine, I think that the ability to move those molecules—electrons—into compute power will even accelerate. That's the team: 300 megawatts building out in Paraguay and expanding another 100 megawatts. So this will give you more granularity of other things we're looking at.

We are looking in Texas, Paraguay substation being built, owning the land that we end up purchasing in a separate entity as we repurpose the land and power in Sweden, and then the Gigafactory that's taking place in Canada, in particular in New Brunswick, and then this joint venture with Bell Canada rapidly expanding in British Columbia. So we are coast to coast and the biggest, technically, hyperscaler in Canada. That's the building—beautiful Scala center—that we bought in northern Sweden with the land and long-term power contracts.

And we have a very strong footprint in the community. There's the HIVE hockey arena in which we sponsor 12 kids learning how to play hockey. This is a draft, a visual for what we expect to see over the next couple of years in Toronto for the AI Gigafactory. This is really quite phenomenal that—I didn't give you more granularity—but the relationships that Craig has been able to build and accelerate with other key groups and entities. So I think it's really important to see that a couple years ago it was predominantly the relationship that we had with NVIDIA, and Aydin and myself had been and purchased a lot of NVIDIA chips.

But what we've seen with Craig taking that ball and running with it has really accelerated these other relationships and the thrust to be in these other countries that have sovereign data centers. So this is just to give you an idea. We're traveling all over the world all the time meeting with captains of industries like Michael Dell several times, President Peña, Jensen Huang. So we are meeting with very important people that have visions, and President Peña has an incredible vision for the country.

You can see Gabriel Lamas with his arm on President Peña's shoulder. Well, the big vision is to make that the biggest AI destiny center for Gigafactories in Latin America. So it's great to be with young executives like Santiago Peña, the President of Paraguay, that has this phenomenal vision, and we hope to be fast-tracking that process and growing with this country. Well, here's... I'll turn it over to Aydin, and that's Aydin throwing a touchdown pass to Craig Tavares, which I just mentioned.

Aydin Kilic is our electrical engineer, he's our CEO, was originally our President and Chief Operating Officer and saw the building and construction of New Brunswick and helped dramatically build out Paraguay for HIVE Digital Technologies. And I think it's important to listen to him give you the story, and also he's been very much involved in Wall Street in raising capital for this growth we have for our Gigafactories. Here you go, Aydin.

Aydin Kilic, Member

Thank you, Frank, for the excellent strategic overview. Now let's get into an executive summary of the quarter to date, our recent accomplishments, and some exciting things in the pipeline starting here. Financial highlights for the quarter period ended June 30: 79 million of revenue, 24 million of gross operating margin, and 15.2 million of earnings from operations, which is revenue less cost of goods sold less corporate G&A. We do have a $143 million net loss which is really driven by two non-cash items.

This is depreciation of about 54 million and this roughly $85 million provision for tax liability in Sweden. This is actually a long-standing issue that we've previously disclosed over the last two years, going back as far as 2023, related to the Bitcoin mining business in Sweden and VAT tax treatments for basic imports in the Bitcoin mining business. So we are appealing it, we are contesting it, we do not plan to pay, and that is why it is a non-cash charge on the financials.

There's further disclosure in our earnings press release, but I really want to focus on the operations of the business. We did make 15 million this quarter on a non-cash basis, or if you interpret through adjusted EBITDA, 13.4 million. And of course the depreciation—look, we have a very aggressive two-year depreciation cycle for those ASICs. That means over eight quarters ASICs will get depreciated to zero. Well, guess what, we imported about a quarter billion dollars of Bitcoin mining equipment, the containers and ASICs to Paraguay last year.

So that depreciation is still hitting us quarter over quarter. And we have four-year depreciation for the GPUs and of course as we have GPUs from the past and bring online more, that's really where that non-cash number comes from, quarter over quarter and year over year. I'm very pleased. It was a strong quarter—growth: 10% in revenue quarter over quarter, 74% growth year over year, again driven by the expansion in Paraguay. And moreover, if you look at the earnings from operations, again that figure, 15.2 million, that's up 86% quarter over quarter.

Again last quarter, Jan, February, March, we saw the downturn on February—mining economics were the lowest they've been. We navigated that still with the business generating 8 million on a cash basis this quarter. April, May, June, that number almost doubled to 15.2 million. And year over year that number's up about 50%. So again, having a lean and mean corporate G&A, we've made a lot of strategic hires to scale the HPC business. But we really want to focus—does the business fundamentally on a cash basis make money.

So what is your revenue? What are your direct operating costs or your COGS? And of course your corporate G&A. And you can find this nicely summarized on the page 20 of the MD. But here it is graphically represented. This is a breakdown of Bitcoin mining versus HPC revenue. Currently our HPC revenue is represented by the GPU cloud revenue. 7.1 million for the quarter represents about 10% of the total revenue this quarter and 72 million as Bitcoin mining, about 90%.

But I want to point a couple things out. So the last four quarters, the previous four quarters, we're doing roughly 5 million a quarter—we're at that 20 million ARR ticket. Well, in May, that B200 cluster went online in Bell Canada's Winnipeg facility and that got us to 7 million of actual revenue for the quarter. So you're seeing that growth and that number is going to continue to grow and we're going to explore why. But just looking at the quarterly analysis, Bitcoin mining revenue grew 7% quarter over quarter, 77% year over year, again having Paraguay fully at scale.

And of course as I mentioned, HPC is growing. So if you look at that 7 million realized for the quarter of HPC revenue, that's about 28 million ARR if you annualize it. Let's go to the next slide. Well, that means by Q2 we exited the quarter at 28 million ARR. We're actually doing 35 million ARR today because we're about 97,000 daily HPC revenue. But here's the big news, everybody. We have hit a massive milestone—so excited to share. Our team's done a tremendous job and we are at $180 million of contracted revenue because we just announced a new five-year deal which adds 70 million ARR to our HPC business unit.

And this is new news. This is brand new. We just announced it this morning. This is in addition to the Cohere deal which we announced a few months ago. So you now have $180 million of combined active and contracted revenue. We're very excited. This is a five-year contract that we just announced. These are GB300s. 2,088 GB300s are going to go to the Bell Merit facility. These will be delivered in point Q4. It is a five-year contract for $350 million total contract value, therefore 70 million ARR.

Super excited. We funded the acquisition of these GPUs using proceeds from our June convert where we did the $130 million convert, 0% coupon. Lenovo is our partner for the OEM on these GPUs. And this gets us to that sweet, sweet $180 million number. And by the way, we're still targeting 200 million ARR for the GPU cloud business with before. We still have more GPUs to bring online in the pipeline but we're at critical mass now, ladies and gentlemen, so very exciting.

And of course this will go into the Bell Merit facility which will be closed-loop, liquid-cooled, ultra-low PUE. I was actually just there last week on a site tour. The facility is looking phenomenal. That Merit facility will house both the Cohere cluster of GB200s and this new GB300 cluster. And this is with an investment-grade global technology giant that we signed this deal with. So that's fantastic. And they're actually putting a 10% down deposit of total contract value—about 35 million.

So that's very exciting. If you zoom out and look at the total contract value now of the cluster deals we signed, we're up $600 million of GPU cloud TCV signed this year. This year the Buzz team has done a tremendous job. Craig Tavares, Mark Volk, Mario Sergi—they've all been doing a phenomenal job. Gabriel igb—really, the whole executive team, him, of course, Darcy—everybody has been working around the clock to make this a reality. And of course, you know, it all really started with Frank's vision when we ordered those Nvidia GPUs back in 2021 to pivot from Ethereum mining to HPC cloud.

So to see here today that we've got $600 million in total contract value signed really this year and our market cap is about 800 million—you see how attractive we are. And so I think that the stock should be due to re-rate with this fantastic news. And these are long-term contracts. We told the street when we were doing our converts to fund this growth that we were going to focus on long-term three- to five-year contracts which is exactly what we've done.

So the Cohere contract—three-year, 225 million TCV—and this new GB300 contract with the global tech giant, investment grade, is a five-year deal. So very exciting stuff. And of course that is a snapshot, a picture there, of one of our actual clusters. So this is a snapshot of the two convert deals we did this quarter in April and in June—collectively almost a quarter billion dollars, zero-coupon debt, five-year bonds. We did purchase a capped call for each to minimize dilution.

So the capped call conversion premium—$4.92 for the April bond and then $8.53 for the June bond. Those capped calls very attractive as you minimize dilution and of course having that zero-coupon interest. So we are delivering on our promises. The proceeds from these notes were to go towards the acquisition of these GPUs which, by the way, the GPUs we still—so we put a healthy down payment down and tried to minimize our cost of capital all the way around and then have financing for the rest, and targeting about 20% to 30% down per GPU cluster and the balance to be financed.

So this is just an overview for all the analysts and entities out there: if you want the granular details of the three deals we have now—so three deals in Bell AI Fabric data centers in the Winnipeg facility and of course the Merit facility—those are going to get delivered and deployed in Q4 this year. The GB200 cluster for Cohere and the GB300 we just announced. So we talked about that landmark $180 million of contracted revenue. Here's the breakdown again.

This is a really handy slide if you want to pause on this slide for all the analysts and enthusiasts out there—shows you that we have the approximately 5,500 GPUs online today doing 35 mil ARR. And then the two large clusters coming online in Q4 which brings us to 9,800 GPUs contracted or active, bringing us to 180 mil ARR target. And again, by the way, our year-end—we're still targeting 200 million by Q4. We still got some more bullets in the chamber, so to speak.

But, you know, for the most part we've done a tremendous job. It's August and we're already at 180 million of that 200 million target that we had. So very proud of the team. It's been a tremendous quarter and I think it's going to be an amazing year. Let's go to the next slide. We also announced an LOI for our Odin site. This is an HPC colo lease. This is for 45 million ARR. We announced it in June, 25 megawatts of IT load. So I want to pause here and let this sink in.

We are at $180 million of active and contracted revenue today with our GPU cloud business. If you add this $45 million ARR HPC colo deal, which we will announce the total deal size, how many years, et cetera, when we announce a defender, which we hope to announce before the end of September, then it puts us at $225 million ARR contracted HPC revenue with cloud and colo. And I think that's really exciting because our target was 200 million for the end of the year.

So we're blowing past that target ahead of schedule once this deal is announced formally. But I just want to put that into context. So really exciting stuff. Nothing's slowing down. We just announced this $350 million five-year GPU cloud deal, the GB300s, and we've already got another bullet in the chamber. So really exciting. Let's hop to the next slide because I want to underscore the value proposition. The value proposition: if you look at having a diversified sum-of-the-parts valuation, we've still got the Bitcoin mining thing—about 750,000 a day of revenue, 275 million ARR.

Today we've got the 180 million signed contracted revenue, including 35 million active. And then of course, once we bring online that Bowden HPC colo deal, what you can look at here is if you look at the multiples that our peers are trading at. So on the cloud, looking at the CoreWeave, Nebius, Irons of the world—it's about a 5.5x based on two-year forward revenue multiple. And if you look at the HPC colo, your Wolfs, your Cyphers, your HUTs—that's about almost 11x multiples.

So you can apply those multiples and do a sum of the parts and then you sort of have a nominal valuation based on where our peers are, such as Marin, Queen's Park, that still have large hash rate online. The composite sum is a $2 billion enterprise value. And again we're hoping to announce this Bowden HPC lease before the end of September. So I think that there's a lot of real exciting near-term value growth. In fact, I think with the announcement again of this GB300 deal, you know we're due to re-rate as we've massively grown our ARR to that 180 million contracted number.

But this is the case for a $2 billion enterprise value near term based on where peers are trading. So recap. I know it's a lot going on. We huddle every day over nine time zones. We operate in three continents—in Europe and North America and South America. And we've got 35 million ARR active today. Again, that really golden 180 million contracted GPU cloud revenue ARR. And then that lease which was announced in June for Bowden. And again we also announced the Gigafactory in May, so that was technically in this Q2.

So just giving you a quarterly recap. It's a very dynamic quarter. So very excited. Here's a bit of a double click on this Boden data center LOI. So it's 32 megawatts of utility load, which would be 25 megawatts of IT load. We have a FIT test complete. We have a Tier 3 data center builder that actually does some Tier 4 as well. It's very prevalent in the Swedish market. They've been at it for nine months. This isn't something that we just picked up last week, guys. We just announced the LOI, so that's why it's new news. But this is a legacy site.

We've been operating here since 2018. This is the GPU supersite where HIVE Digital Technologies was mining with 130,000 GPUs, Ethereum, in the Ethereum mining heyday, almost 6% of the network at its peak. And what the catalyst is, the catalyst was that we got approval from the Boden municipality to buy this building because this was a long-term lease with the municipality. They actually own the building, and we just got approval to buy it. And that was the catalyst that suddenly made it an HPC colocation conversion candidate.

And so that was the big news in June. And here it is by the numbers, what it works out to. And that colo rate, by the way, is about 150 bucks a kilowatt a month, so very strong, very comparable to some primary US markets that we've seen our peers sign data center deals for colo throughout the US. So very exciting. Now once we announce a definitive deal, we'll have the total contract value, we'll have the length of the lease, but right now on an indicative basis, you know, that's 45 million ARR HPC colo, and it's with an investment-grade Swedish offtaker as well, a telco company.

So we've talked about the Gigafactory, the crown jewel of our Canadian assets. We announced this in May, and of course this slide was in our last quarterly update, but here it is again just as a frame of reference. Expect a lot of news as we advance this site. It's very exciting. We expect to energize the site end of '27 and have compute come on early '28. Our dual-engine strategy: we are at approximately $850,000 today, and that comes from $750,000 a day mining Bitcoin.

We've got 24 exahash operational; it's actually 25.3 installed. We just optimized with firmware and downclocks and machines. Again, we are in bear market economics. So how do you maximize your profit? So that's running off about 275 mil ARR. And then on, of course, buzz with the HPC is about 97,000 a day. You add that up, there's about 850 grand a day of total revenue. And if you look at our COGS and SG&A, so total cash cost to run the company globally—again, I always like to go to page 20 and 21 of MD&A, got a nice chart.

It's my favorite page in the MD&A, which is long and it's a great document for all the analysts and enthusiasts out there. If you add our total operating costs, about $63 million for the quarter, 690 grand a day. And so that's just a good number I like to have at the tip of my fingers, because if we're doing 850 grand of daily revenue on a cash basis and a global operating cost—everything, going to conferences, paying for data centers, electricity, you name it, salaries, everything—it's about 690 grand a day.

That means we're printing, we're doing about over 150 grand a day of profit. So it just gives you a sense of the health of the business, and then again you have all these non-cash charges and accounting treatments, etc. I'm an engineer, I just want to know as a businessman, fundamentally, are we making money on a cash basis? Yes. Let's, let's look forward. So once we get to that $180 million contract and those go live in Q4. Once those go live in Q4, it's really easy math: $180 million of GPU cloud business, about $500,000 a day. So what does that mean? Well, I mean hashprice can vary, but let's just assume mine economics are similar in the next three to four months, which isn't that far away. We're in August right now. That would be 750 grand a day on Bitcoin mining and 500,000 a day on HPC, which includes our GPU cloud revenue.

So that's about 40% of global revenue is what the HPC business is going to constitute, with Bitcoin being 60%. So again, this quarter it's 10% AI and 90% Bitcoin mining. Within the next few months, that's going to be 40% on the AI and 60% on Bitcoin mining. So really exciting outlook on the dual-engine strategy. Zooming out, global power footprint. We have 860 megawatts globally. That includes 440 megawatts active with Tier 1 sites and, of course, the additional 420 megawatts, which is 100 megawatts in Iguazú that we're building out a substation, of course the Gigafactory in the Greater Toronto Area.

So what you can see on this slide—to try to make it easy for the viewers—is the sites that we've highlighted in green are sites where we own the land. These are sites that are candidates for conversion to Tier 3. And if you tally that number up, it's about a 440-megawatt pipeline between Canada and Sweden for our conversion to Tier 3, which is very exciting, and by the way there's on top of that Iguazú as well. So it's very exciting, and I think that if you look at the value proposition and what that looks like on a revenue basis on our two-year vision.

Let's go to the next slide. So on the left GPU Cloud, on the right, HPC colo. So again we are at $180 million contracted today. We went over that earlier in the presentation. We still have some room in Quebec. Let's put another 500 B300s in. That'll add $20 mil ARR. So our end-of-year target is actually $200 mil ARR in the GPU cloud business, and we're so close to that, given that we're at $180 million contracted now. And then on the right-hand side, if you look at the portfolio of sites, the Toronto site, both of our Boden sites, of course New Brunswick and the Gigafactory.

On an HPC colo basis, 325 megawatts of critical IT load would generate, at prevailing lease rates—you know, $150 a kilowatt in Boden, $130 in New Brunswick, so on and so forth. Gigafactory is about $160 a kilowatt. If you do the math, what that works out to accomplish is about $500 million of HPC colo revenue with these sites being developed over the next two years. So on a total basis between AI Cloud and HPC, $700 million ARR. And that's very exciting.

Now, by the way, any one of the sites on the right—HPC colo—you could stand up GPUs as well. We have that optionality. If we did that, those sites on the right would be able to accommodate over 120,000 GPUs. That's very exciting. Of course Big Boden's already been earmarked for HPC colo, but just to give you a flavor, the other sites—Brunswick, GTA, Gigafactory, and the Toronto Airport site—we still have the optionality if we want to send it more cloud.

But to keep it simple, right now we've done the research, we've had talks with parties that are interested in colo at any of these sites. That number, on prevailing market rates: $500 million target ARR, plus of course the $200 for GPU Cloud—$700 million is the number to take away from this slide. That's the vision. Plus, of course, the Bitcoin mining purview. Now, where will Bitcoin mining be over the next two years? Of course it depends on hashprice.

But if it's where it's at today, that puts us collectively at almost a billion dollars of combined revenue. So very exciting. As we continue to grow, we want to see the stock re-rate, secure the lowest cost of capital convertible. We've seen high-yield bonds being used to finance the construction of data centers. A lot of our peers have done that. We are exploring that as well. Of course the convertible bond market, we did two of those deals this year.

It has worked out very well for us. And, of course, leveraging vendor finance for the GPUs as GPUs are emerging as an asset class. So we've got a very methodical and forward-looking strategic cost of capital plan to realize this growth. Let's go to the next slide. On a sum-of-the-parts, $5.2 billion is the implied enterprise value. If you look at having $200 million of GPU cloud revenue, $500 million of HPC colo revenue, and again you apply those industry multiples—this is actually a base case.

We're not even using our peer multiples. We're discounting a bit. We're going 3.5x on cloud and 8x on colo. And then, of course, you have the Bitcoin mining business. Puts it at a $5.2 billion base case. And you can see here how you get to that number during the sum of the parts. If you actually go to where our peers are trading and you use the same multiples, the mid-range case is actually $7 billion market cap USD as we scale towards those revenue targets we presented.

So that's very exciting. And on the upside case, you know, depending—some peers trade at higher multiples—as high as $8 billion. But really, to be conservative, we say the base case is $5 billion. And again, these are sites we own. This is land and power that's secured, or sites that are operational that are converting/catalyst to Tier 3. And, of course, the growth in our GPU cloud business. A summary of the land and power. We are advancing on that.

We've talked about everything except Paraguay. So we have completed the civil work. We've talked about that all summer. We just had two 80 MVA transformers installed. And by the way, in June we announced the proof of concept between New York and Ciudad del Este. So that is the cherry on top, I would say. Stay tuned for updates there. But just a little bit of eye candy for everybody. Let's go to the next slide. Here is one of those 80 MVA transformers that were just delivered and dropped on the site last week.

So again, we were doing the civil work. This is heavy civil infrastructure, guys. We're trenching. There's large concrete pads that go deep underground, etc. Of course you've got cables and so forth. This is a 200-megawatt substation in Iguazú that is going to add an additional 130 megawatts of utility load from these two 80 MVA substations, which will allow for 100 megawatts of IT load. This is just at Iguazú, in the backyard of the Itaipu hydro dam.

It's a 1.2-gigawatt substation regionally that we feed off of, which directly feeds off the Itaipu dam, which is a 14-gigawatt dam. So yeah, just letting you know a lot of things happening and progressing in the background. So again, we are data center builders and developers and operators. So we do everything from substation construction, maintaining substations, of course building the data centers and operating them. So

Frank Holmes, Executive Chairman

Steady progress on all fronts around the world. Just a little bit of context. So Iguazu is very close to São Paulo. And why São Paulo? São Paulo is where the hyperscalers in Latin America are mostly serviced. There's a data center park in São Paulo, NVIDIA's Latin American headquarters in São Paulo. So São Paulo is kind of the hub for Latin America, but most of the data centers there are 15 or 10 megawatts. So this will be, we believe, based on our market research, the largest AI factory in Iguazu with 100 megawatts of critical IT load.

Design development is underway as well. We're completing a basis of design. So stay tuned for updates. This is a little bit of a geographic snapshot of how close we are to the region that is currently serving all of Latin America. Again, with that proof of concept we did earlier this year, and we're in talks with different groups. There is a lot of international interest in Latin America as the market emerges. Let's go to the next slide. Looking at the last 12 months, here you've got the revenue and here you have earnings from operations.

So 330 million of revenue in the last 12 months. I would say that is very impressive. Almost a million dollars a day we've done. And again, that's navigating the downturn. You know, in calendar Q1 of this year, for the entire market, it was tough. But you know, we persevere, we optimize, and on a cash basis we still made money every quarter. So again, that earnings from operations is revenue minus cost of goods sold, minus corporate SG&A. So we have done 80 million of earnings from operations in the last 12 months, which I think is very impressive.

And again, we've seen bull markets and bear markets in the last 12 months. And by the way, we're showing five quarters here. I realize that it's just so you can do the year-over-year comp as well. But if you look at the last 12 months, that's what those sums are. But let's actually zoom out and look at how the industry has done. So I think what's overlooked—everyone is so hyper-focused on the next year or two. That's great. We've got a two-year target of 700 million ARR.

We're 180 million contracted revenue now in GPU cloud. So that's great. That is great. I think it's about getting the story in front of peers. By the way, we should be Russell 2000 qualified by end of this year. We of course file in US GAAP now, principal executive offices in San Antonio. We're very much aware of the importance of having a strong presence in the US capital markets. Let's look at the actual revenue we've done amongst this peer group here.

And you've got $2 billion and $10 billion companies represented here. We've done more revenue than all of our peer groups: 331 million in the last 12 months. Some of our peers have done half of that, which is interesting in some cases. Bitfarms' revenue is actually trending down. We get it. People are focusing on HPC conversion and colo, and we are too. But I think it's important to point to a track record. We stood up 300 megawatts in six months in Paraguay.

We've got 9,800 GPUs now contracted, 5,500 active. So we're not only pointing to the growth, but we're actually doing it today. And so I think it's very noteworthy to point out a strong track record of accomplishment is a good indicator of future success, at least in our opinion. At HIVE Digital Technologies, we get it. It's all about megawatts, powered land. So if you look at what we've got secured in our pipeline, when you compare that to our peers, we have a very healthy 860 megawatts.

And so of course, you know, you've got the hots of the world with 2 gigawatts. But you know, outside of them, you know, our pipeline is in line. So I think it's really just framing this in context for the Street. I think HIVE Digital Technologies is an incredibly attractive value proposition now. You know, we've got some really smart money in our cap table and we look forward to growth, and our team is working hard around the clock to build value for our shareholders.

This is just a quick crib sheet for you. As Bitcoin price fluctuates—$60,000, $70,000, $80,000—this is just what the mining margin is after electrical costs. Using the indicative 5-cent opex, you can see hash price varies versus Bitcoin price, etcetera. So right now we're sort of in the 36% range given where Bitcoin is—36% to 40% margin on that $750,000 daily revenue that we're at right now. So just a handy reference slide, and I'm going to turn it over to Mr. Darcy Daubaras, the longest-standing CFO in the industry since 2018. Darcy, thank you so much. And the team working tirelessly. It was a super solid quarter. Over to you.

Darcy Daubaras, CFO

Thank you, Aydin. I'll take the next few minutes to walk through HIVE's financial results for the first quarter of fiscal 2027. This was a strong quarter from an operating perspective. We delivered significant year-over-year revenue growth, improved our gross operating margin in dollar terms, returned to positive adjusted EBITDA, and substantially strengthened our liquidity position. At the same time, our reported GAAP net loss was significantly impacted by several non-cash items, most notably a provision associated with the ongoing Swedish tax matter, which I'll discuss in more detail.

Before getting into the financial results, I'll briefly highlight our capital structure at June 30, 2026. HIVE had approximately 271 million common shares outstanding together with approximately 3 million warrants, 2.6 million options, and 16.7 million restricted share units. Our shares continue to trade on the Toronto Stock Exchange and NASDAQ under the symbol HIVE, as well as on the Colombian stock exchange under HIVECO. Turning to our first quarter financial highlights, there are several numbers I want to emphasize.

HIVE generated $79.1 million of revenue compared with $45.6 million in the same quarter last year. Bitcoin mining remained our largest contributor, generating $72.1 million of revenue, while our HPC and AI business contributed approximately $7 million. Importantly, our gross operating margin increased to $24.2 million compared with $15.8 million in the prior-year quarter. We also generated positive adjusted EBITDA of $13.4 million. Our reported EBITDA was negative $86.3 million and our GAAP net loss was $142.9 million.

However, there is an important distinction between the operating performance of the business and the reported GAAP loss this quarter. The net loss included an $84.7 million non-cash provision related to regulatory liabilities associated with our ongoing Swedish VAT dispute. It also included $53.7 million of depreciation, $7.1 million of share-based compensation, and the impact of fair value adjustments. While those items are appropriately reflected in our U.S. GAAP financial statements, they are important to consider when evaluating the underlying performance operationally of the business. We ended the quarter holding 190 Bitcoin in treasury. Stepping back from the individual line items, three numbers really summarize the quarter for me. First, revenue of $79.1 million demonstrates the increased scale of HIVE's operations. Second, adjusted EBITDA of $13.4 million returned to positive territory after negative adjusted EBITDA in the fourth quarter.

And third, we produced approximately 1,004 Bitcoin equivalent during the quarter. Taken together, these metrics demonstrate the operating leverage we're beginning to see from the investments we've made in our global infrastructure. We continue to balance growth in our core Bitcoin mining operations with the development of our higher-value HPC and AI infrastructure business. That operating performance is supported by a substantially stronger liquidity position.

We finished June with $208 million of cash compared with approximately $23 million at March 31. In addition, we held approximately $11.2 million of digital currencies, $10.9 million of investments, and $18.9 million of receivables and prepaids. Total current assets were approximately $280 million compared with current liabilities of approximately $143 million. The increase in cash primarily reflects the financing activity completed during the quarter, including our exchangeable senior note offerings and proceeds from our ATM program.

These financings have provided HIVE with significant liquidity as we continue investing in our Bitcoin mining infrastructure and increasingly focusing on our HPC and AI growth initiatives. Our objective remains to maintain financial flexibility while deploying capital into opportunities that we believe can generate attractive long-term returns for shareholders. Turning from the balance sheet back to operations, gross operating margin showed meaningful year-over-year improvement.

We generated $24.2 million during the quarter compared with $15.8 million in Q1 of last year. That's an increase of approximately 53% year over year. This is particularly noteworthy given the substantial increase in the scale of our operations during the past year. Our basic loss per share was $0.54 compared with earnings per share of $0.19 in the comparable quarter. Again, the current-quarter loss per share reflects the significant non-cash charges recorded during the quarter, particularly the Swedish regulatory provision and depreciation associated with our expanded infrastructure base.

The year-over-year comparison really demonstrates the increased scale of the business. Revenue increased from $45.6 million to $79.1 million, representing growth of approximately 73%. At the same time, gross operating margin increased from $15.8 million to $24.2 million, an increase of approximately 53%. As a percentage of revenue, gross operating margin was 31% compared with 35% in the prior-year period. So while the percentage margin moderated somewhat, the absolute dollars of gross operating margin increased significantly as we expanded the scale of the business.

This is an important measure for us because it demonstrates our ability to generate positive operating contributions on a substantially larger revenue base. Sequentially, the trend is also encouraging. Revenue increased from $71.8 million in the fourth quarter to $79.1 million in Q1, an increase of approximately 10%. More importantly, gross operating margin increased from $17.5 million to $24.2 million, or approximately 38% quarter over quarter. Gross operating margin as a percentage of revenue improved from 24% to 31%.

So sequentially we saw improvement in revenue, operating margin dollars, and the margin percentage. That combination is a positive indicator of the underlying operating performance of the business as we enter fiscal 2027. From an earnings perspective, it's important to distinguish between our underlying operating results and the impact of several significant non-cash items. Adjusted EBITDA was $13.4 million compared with $44.6 million in the same quarter last year.

The year-over-year decline reflects a number of factors, including the changing economics of Bitcoin mining and the increased operating cost base associated with our expanded global infrastructure. Our reported U.S. GAAP result moved from net income of $35 million in the prior-year quarter to a net loss of $142.9 million this quarter. Again, the most important point when interpreting that result is the magnitude of the non-cash items. The quarter included the $84.7 million regulatory provision associated with the Swedish VAT matter, together with $53.7 million of depreciation, as well as share-based compensation and fair value adjustments.

The Swedish provision reflects our accounting assessment following the adverse Court of Appeal adjustments. We continue to pursue the available legal avenues in Sweden. Accordingly, we believe adjusted EBITDA provides investors with an additional perspective on the underlying operating performance of the business alongside our U.S. GAAP results. And finally, looking at earnings sequentially provides another useful perspective on the quarter. Adjusted EBITDA improved significantly.

We moved from negative $9 million of adjusted EBITDA in Q4 to positive $13.4 million in Q1, an improvement of more than $22 million. That improvement is consistent with the stronger revenue and gross operating margin performance we discussed on the previous slides. Our U.S. GAAP net loss increased from $76.3 million in Q4 to $142.9 million this quarter. But again, the comparison is heavily affected by the $84.7 million non-cash Swedish regulatory provision recorded in Q1.

For that reason, we believe it's important to look at both the U.S. GAAP results and the operating metrics when assessing the quarter. Overall, we entered fiscal 2027 with a larger revenue base, improving sequential operating margins, positive adjusted EBITDA, and a substantially strengthened liquidity position. That financial position provides us with flexibility as we continue executing on both sides of HIVE's strategy: operating our existing Bitcoin mining business efficiently while investing in the growth of our HPC and AI infrastructure platform.

With that, I'll turn the presentation back over to Nathan.

Nathan Fast, Director of Marketing and Branding

Thank you, Darcy. That concludes the presentation portion of today's call. We will now begin the question and answer portion of our call. Analysts on the line, if you could please click raise hand when you're ready with your questions, we will begin to choose and ask you to unmute. Our first question comes from the line of Joe Voffi from Canaccord. Joe, feel free to unmute. Proceed with your question.

Joe Voffi, Analyst at Canaccord

Hey guys, good morning. Great progress in the business, especially this new cloud deal with the investment-grade tenant. Maybe we could double-click on that. I know it's a five-year deal. Could we get some perhaps initial thoughts on, if you've run some IRRs on the GPU investment, what kind of returns potentially you're getting there, and maybe some additional thoughts on capex here to fund the build-out? I know you've got a deposit, I know you've done, you know, some of your convert deals.

Just kind of what the rest of the financing stack might look like here, and I have a quick follow-up. Thanks.

Aydin Kilic, Member

Hey Joe, this is Aydin. Great question, thanks for tuning in. It was good to see you last week at the conference in Boston. So, you know, having that five-year term locked in, we put in the press release. Capex is about $185 million for the GPU cluster, of course with the InfiniBand and so forth, really just to have an NVIDIA reference architecture designed for the 2016 GPUs. And so if you kind of do the math indicatively, model of course it'll be delivered and deployed in Q4.

So we're expecting EBITDA to land in the 75% to 80% range. And so if you do the math, you're paying off the GPUs completely in about three years, and the balance of that two-year term is free cash flow. So you effectively bake in a 1.6x return, roughly speaking, on the GPUs, and then you own them outright after the term. And now we've also financed the GPUs, so we're putting a portion down, about 20% down the cost of the GPUs, and so we can provide subsequent market updates with the financing terms.

But really the inaugural release was to let the street know that we've now hit that $180 million ARR target, well on our way to the $200 million target end of year, show the street that we were delivering and deploying the promises from our converts in April and June to use that capital as down payments to lock in these GPUs. The financing actually doesn't kick in towards when the GPUs ship. And so as we get closer to the deployment dates, et cetera.

But POs are secured for the GPUs, and that means they go into production, which is so critical. And that's really what locks in. So having the capital from our converts, the $245 million collectively that we raised, allows us to have those GPUs, POs accepted, hardware going into production. And then of course you've got a shipping date, which is important and critical for the offtaker, for the client, so they have assurance that the infrastructure—and by the way, the data center is ready.

I was actually just at the Bell Merit facility two weeks ago doing a diligence visit with one of the lenders, and I'm going again tomorrow for another trip. So things are really moving forward. And yeah, I hope that covers your questions.

Joe Voffi, Analyst at Canaccord

Deal. Sounds like, you know, it's well on its way. And then down in Paraguay, I know there's a lot of things, the substations are going in for that additional build. You know, the benchmark testing's been done. What should we be looking for down there as kind of a next step, you know, in the evolution of that power portfolio? Thanks a lot.

Aydin Kilic, Member

I would say the next thing to look forward to is just updates as we work through the basis of design. I'm actually planning to go to Paraguay at the end of September. I've got a trip to New York planned third week of September, and then I'll actually stop in São Paulo. I plan to visit Nvidia down there and then go straight to Asunción, do a site tour. So we're engaged with a design-build firm that's built a lot of data centers in the region. We're working through a basis of design.

So I think as we work through that process, we'll be providing updates just the same way that we put a photo of the substation being deployed last week. So we'll keep the street updated with progress as we work towards that. That substation we expect to be energized towards the end of this calendar year in parallel, of course, working through that basis of design. And that's all I'm going to say for now. I think really we like to let people know as a multinational company, we have progress in different jurisdictions and a lot of the growth, like the revenue growth, is happening in Canada this year with the deployment of these GPU clusters.

And of course we'll work on the conversion of the Toronto Airport site, the New Brunswick site as well. I was actually just in New Brunswick last week. After Boston, I flew to New Brunswick to Fredericton to meet provincial government and NB Power as well to talk about our vision there to make the New Brunswick site in Grand Falls the largest AI token factory in the Maritimes, which I think will be of national significance. And of course that complements the gigafactory in Ontario and the Greater Toronto Area.

So I would say stay tuned for updates coming out of Canada as we advance the conversion of those data centers. And then of course the Big Bowdoin site, I would say that's actually furthest along because we had that LOI signed with an offtaker as well. And so the next step there is really to watch out for the definitive agreement to be announced and then that plays into the growth of our ARR targets as well, which was highlighted in my section. So I would say keep your eyes peeled for updates on Big Bowdoin Canada and then Paraguay is just the icing on top and we'll update the market, but focus on Sweden and Canada for now.

Great work.

Nathan Fast, Director of Marketing and Branding

Thank you, Joe. We'll keep the Q&A moving. Next to Chris Brendler from Rosenblatt.

Chris Brendler, Analyst at Rosenblatt Securities

Hey, thanks, Nathan. Good morning, folks. Congrats on results here. Nice to see the progress and, you know, tough market for bitcoin, but you guys are executing pretty well. My first question is on the high-performance compute business. Given all the progress there, I was hoping you could give us a little color on what you are targeting for a gross margin in the Q. It looked like it was like 44% this past quarter. I know as these contracts ramp up I think they'll become less impacted by the service fees and potentially go a lot higher than 44%.

Any thoughts on the target gross margin for the HPC business after signing these contracts? Thanks.

Aydin Kilic, Member

Yeah, that's an insightful question, Chris. So our fleet right now is comprised still of—we've got about $20 million of our $35 million active revenue coming from Hopper Series GPUs, a combined total of 844 H200 and H100s. And then the balance is about 4,200 A-series cards. So the A-series cards are legacy; they've been running on cloud since 2023, which is, you know, a testament to their fortitude. But as older-generation cards, those rent out for $0.40 a GPU-hour.

Now keep in mind an A40 only uses 400 watts. It's still doing almost a dollar a kilowatt-hour. But still, of course, as we bring on GV200, GV300s, these are more profit-dense per watt. And also, of course, with scale you get economies of scale. So those two factors—newer-generation GPUs coming online at scale—both are indicative drivers for that margin to go up. And again, as we've built the business, of course you need to have the foundation in place to scale.

So some of that cost basis is somewhat fixed. And of course as the scale grows, you know on a relative basis that fixed cost diminishes. So hence you can expect margins to improve substantially.

Chris Brendler, Analyst at Rosenblatt Securities

Okay, great. Thanks, Aydin. That's good color. My second follow-up question is for, open to either Frank or Aydin. Is the bitcoin mining business actually doing pretty well? Like, you've gained hash rate share. The gross margins there also improved sequentially despite a pretty tough environment for bitcoin. We've seen hash price stabilize. Network hash rate has also come down from the peaks. Just wondering how you feel about the bitcoin mining at this point and any updated sort of big-picture thoughts on bitcoin.

I've been a long-term believer myself and just sort of waiting for the next cycle. Is that kind of what you guys are thinking about as well, or are we in a new paradigm for bitcoin? I'd love to hear your thoughts. Thanks.

Aydin Kilic, Member

Yeah, so we see hash price sort of stabilize around the $31 per petahash per day level. We saw lows as low as $27–$28, which was not long-lived, but they were for a week or two, and we see difficulty adjust. So it seems that the prevailing floor of this bear market is just above $30 hash price. So, you know, what does that mean? Well, in our case in Paraguay we've got, I think last time I checked, about 19.5 exahash because we've optimized with firmware and we've had—earlier in this year we had some S21 XP orders go down and replace some of the BuzzMiners.

So we're actually at 19.5 exahash in Paraguay. That's new-generation gear that on a blended average is below 15 joules a terahash. So we've got a large amount of new-generation hash rate. Of course Paraguay's got very attractive power cost too. So that just forms part of that cash flow engine and dual engine strategy. Of course, with the Clarity Act, if that gets pushed through in September that could be a catalyst. But I think the street's really looking for a catalyst for what's going to be the next breakout for bitcoin price.

But 100% of our growth is in HPC and AI. So again we have new-generation here in Paraguay that's performing phenomenally well. Very, very close to 100% uptime. The hydro infrastructure we deployed has worked very well for the climate there. We've made some modifications to it. So yeah, it's there, it's performing exactly what we expected, and it's throwing off cash flow as we focus on growing the rest of the business. So I think that we'll see. You know, bitcoin always tends to come around.

Is it going to stick with a cyclical four-year cycle? Will the Clarity Act cause a big breakout? We'll see. We'll see.

Chris Brendler, Analyst at Rosenblatt Securities

Great. Thanks again and congrats. You have any results? Thank you.

Nathan Fast, Director of Marketing and Branding

Thank you, Chris. For our next analyst question, we'll pass the mic to Mike Rondel from Northland. Mike, the floor is yours.

Mike Rondel, Analyst at Northland

Hey guys. Thank you. Two questions. One, Aydin, is there anything significant left to get the definitive agreement with Bowdoin, or do you just need a little bit of time there? And then secondly, could you just talk a little bit about demand trends and pricing trends on both the GPU and the colocation side? Thank you.

Aydin Kilic, Member

So the lease for Big Bowdoin, that process has been advancing. What I can say is that, as I alluded to in my presentation, a lot of our peers are using corporate bonds—either investment grade or high-yield bonds—to finance the construction of these data centers. And so we see that as an attractive path to raise capital. And moreover, we're in talks with two lenders on that accord. You want to carefully, carefully structure that lease so that the terms are favorable and you could at least strive towards an IG-grade bond instead of a high-yield bond, and so really just to lower your cost of capital.

So it's an active process whereby we are, I would say, refining. I don't want to give the street, obviously, an exact date, but we're well on our way. The process has been well underway and it's really just fine-tuning those nuances in the agreement. But yes, we do have a draft that's gone back and forth and so just stay tuned. I'd love to give the street an update sometime in September on that definitive agreement. And in terms of demand, I mean I alluded to in previous presentations we had a B200 cluster on a two-year contract—the first one we deployed in Canada and the Winnipeg—504 GPUs and we rented those at $2.90 an hour.

And, you know, I think six to nine months before I rented a big deployment of B200s at $2.20 an hour. So what that tells you is that there's increasing demand in the markets and that tells you that you've got continued— I think CoreWeave came out last week and said they had GPUs from 2020 that they booked out to 2029. And by the way, like, you know, no one was really doing GPU cloud in 2020. They would have been mining Ethereum. We know the CoreWeave guys well and are big Ethereum miners, and so our A40s—we ordered those in 2021.

And the margins may not be as fat, but they're still cash flowing those things. So I think the demand is great. I think that you've got frontier labs that are always going to want the latest and greatest hardware from Nvidia, and then you're going to have other labs, other AI natives, that are more than happy with second-generation gear, and then people that are just using it for inference that just want the lowest cost per token. They're happy to use GPUs from early Hopper even and prior generation.

So we've seen demand very strong, and even in our current deployments we have a lot of stuff we're working at—and I'm trying to be mindful with my words here—but we're seeing tremendous demand for new potential deployments of GPUs beyond what we forecast in the earnings presentation today. We're not stopping at $200 million ARR. When we hit that number for GPU cloud, we see the demand taking us well past that number, very far past that number. We just wanted to give the street some very realistic targets that we were going to hit and blow past. So yeah, we're seeing tremendous demand. I would say that if we were to bring online another cluster of 2,000 GPUs, we have numerous parties—some that we have existing agency with—that would happily rent that out on a three-year or longer contract. So tremendous demand. And so we're very bullish right now just based on the quality of the offtakers, the economic terms.

In some cases we're seeing dollar-per-GPU price even higher than what we've seen previously. So it's definitely a rising-tides environment, which is great for us and our peers in the sector. Yeah.

Mike Rondel, Analyst at Northland

Hey, thank you guys.

Nathan Fast, Director of Marketing and Branding

Thank you, Mike. Thank you, Mike. Time for a few final questions here. Next we'll go to the line of Bill Papa Nastasio. Bill, floor is yours.

Bill Papa Nastasio, Analyst

Yeah, good morning. Thanks for taking my questions and congrats on the deal announced this morning. Aydin, the team has landed a number of attractive deals standing up GPU clusters. Maybe you can walk us through how management is thinking about weighing colocation opportunities compared to these GPU clusters, given the power portfolio. Thank you.

Aydin Kilic, Member

Yeah, that's a great question, Bill. So what we forecast is we've got a target of 10,500 GPUs that we plan to stand up, and that is with our partnership with Bell Canada AI Fabric. And of course we've got the Merritt, British Columbia facility and then the Winnipeg facility, and then our existing sites in Quebec and Sweden. And so after that we still have a pipeline of about 400 megawatts in Canada between New Brunswick, the Toronto Airport site and the Gigafactory site.

So that's about 400 megawatts of utility load. So if you look at that, I could tell you right now that the Gigafactory can do 100,000 GPUs, liquid-cooled GP300 spec type of GPUs, and New Brunswick would be able to do 20,000. It's 50 megawatts of IT load. One of those clusters is roughly—one of those clusters of 2,000 GPUs is roughly 5 megawatts. Just for all the analysts, it's helpful, taking notes, etc. So you could do the math and you could say, okay, so that's 120,000 GPUs.

Now we could phase New Brunswick and we have a two-phase design for New Brunswick now. So we have a very long stretch, a very long pipeline, and we did cite that in our press release that we have runway for over 120,000 GPUs for the sites that we own. So I think that it's looking at what is the value proposition. So if you do the math, on a dollar per megawatt basis, pick an indicative number, say $150 a kilowatt a month for HPC colo. What that works out to, if you do the math, 1,000 kW 12 months is $1.8 million a year of recurring revenue for HPC colo.

And you see our peers are trading at roughly 10x that multiple. So if you looked at it on an enterprise value basis, it's $18 million per megawatt of enterprise value. Cloud, on the other hand, you do $14 million per megawatt per month. So a cluster here—we just announced $70 million ARR. Five megawatts, $14 million ARR. And depending on the multiples, right now it used to be 5x, I think it's trending closer to about 3.5x. So, you know, three and a half times 14, you're well upwards of $50 million enterprise, or about $50 million enterprise value per megawatt.

So if you're in a megawatt-constrained environment, you can get maximum profit density and, based on multiples, enterprise value by going GPU cloud. That being said, the street is also clearly rewarding people that are signing long-term fixed agreements because they're looking at the total contract value. If you sign a 15-year offtake agreement, the TCV on that—you know, we've seen some of our peers sign deals, the biggest almost $9 million for 300 megawatts.

So we can take a similar approach for the Gigafactory, and that's why we sort of represented it such. We give a base case—hey, $360 million ARR if Gigafactory was HPC colo. If you did a 15-year on that, it's a $5 billion contract. So we will evaluate what we think is the best opportunity and bring those to light. But we're very much aware of the economics, and it's not like we're done looking for land and power as well. I mean we're constantly on the hunt.

So I think that it's going to be a really exciting year, but I think that it's a lot more challenging and it requires more capex. But we've got the pedigree to do the GPU cloud business. I was at the AMD keynote that Lisa Su gave. Craig and I flew down to San Francisco a few weeks ago, and when you're actually there in the ecosystem at the industry conferences and you're seeing the amount of demand, the big players that are coming online that are lining up for the next generation of GPUs, it is very remarkable.

So I think that the GPU cloud business remains very, very, very interesting and I think the capabilities it'll unlock in the industry will continue to drive demand.

Bill Papa Nastasio, Analyst

Appreciate the color. Thanks for answering that question.

Nathan Fast, Director of Marketing and Branding

Thank you, Bill. Next question from the line of Brett Knobloch. Brett, what was yours?

Brett Knobloch, Analyst

Hi guys, thank you for taking my question on the GTA site. At what point do you guys have to make a decision to start to break ground and build the data center for that to be kind of ready for service in 2028, and then do you need to kind of decide whether that's going to go cloud or colo when you break ground, or how you would fund the buildout of that—maybe pre-signing a tenant? Yeah, just any thoughts on GTA and timing? Thank you.

Aydin Kilic, Member

Yeah, so we will provide the street updates on that project. We've got our basis of design, and so I think as you work through the process we'll have more, I think, collateral. Again, there was so much excitement when we announced it in May—just really announced that we secured the land and power—and we've got the process well underway from design and permitting. So I really just say stay tuned for more updates on that. And then in terms of cloud versus colo, it kind of goes back to the last question I answered where, if you just use a nominal prevailing market rate for HPC colo, that site would be over $360 million ARR.

But that site, our design right now is three phases of 80 megawatts of critical IT load in our basis of design. And so when you phase that we could have a government tenant in there doing colo. We can have a hyperscaler, maybe have three or four different clients. And that's not to say, you know, we can have a section of IT as GPU cloud, but I think that as we advance along that project there might be a component of it that we've funded—purchase of some long-lead items—using perhaps we did another financing down the road.

But right now we've got the other more near-term projects that we're going to be bringing to market, and I think I don't want to prematurely speak on what financing strategy we'll take. Obviously corporate bonds are very attractive. Typically you're going to want to have a signed offtake agreement for that before you go to market to get a corporate bond. And so again, just the jurisdiction of that site, the amount of demand that we've seen—even in our partnership with Bell Canada, a lot of their enterprise clients, etc., Fed gov—there's a lot of demand.

So we see really just finding the right mix of offtakers, be it HPC or cloud. And when you're raising capital you get that lead order and then it all kind of follows from there. And so we would undertake—but it's a three-phase design if that helps. And we expect the site to be energized end of '27 with compute live in early '28.

Brett Knobloch, Analyst

Awesome. Thank you. I really appreciate it. Congrats on the results.

Aydin Kilic, Member

Yeah, thank you.

Nathan Fast, Director of Marketing and Branding

Excellent. Thank you, Brett. I have time for two more total questions. Let's hear one from Mike Colonnese from H.C. Wainwright. Mike, floor is yours.

Mike Colonnese, Analyst at H.C. Wainwright

Hi, good morning guys. Thanks for taking my question, and congrats on all the progress on the HPC deployments. Great to see. So, you know, my question is really on capex in the second half of the year and timing to funding. Obviously you guys have two large GPU clusters that are set to come online over the next couple quarters. You're doing some design work at a few of your owned and operated data center facilities. So what are you guys estimating for the total capex lift through the second half here and then expected timing to secure the funding required to pay down some of these chips?

And I think you mentioned the goal is to pay 20 to 30% of the purchase price and then look for funding for the rest of it, if I heard that correctly. So any additional color on capex would be helpful.

Aydin Kilic, Member

Yeah. So the scale, the scale in the near term—so getting to our end-of-year target to $200 million ARR on the GPU cloud—is through our partnership with Bell Canada, and the AI Fabric facilities are colo. So the virtue of that partnership, just to recap for everybody, we are co-locating as a tenant in the Bell AI Fabric data centers. There's one in Merritt, BC, and one in Winnipeg. So they've given us a very attractive colo rate below market. And the other virtue of that partnership is that their enterprise customers who are looking for sovereign AI compute—HIVE is the exclusive partner on that accord. So we're building and deploying the GPU clusters in their facilities. So we've seen, for example, Cohere come in as a client and tenant through that partnership. So that was tremendous. Now one of the other benefits is capex-light. So, being that it's colo, we didn't have to shoulder the capex to bring this compute online.

So we think—and as evidenced by the relative valuation side—there's still a lot of upside very near term when you look at where we would be with a big Bowden lease signed. And now that we've got our two big GPU contracts announced as we promised the street—that was our mandate in April and June. So what I'm getting at is I think there's room for the stock to re-rate, and as the stock re-rates then you can kind of look at whether we use equity for financing some long-lead items for some of these data centers.

But again we've seen the corporate bond strategy, investment-grade bonds of course, being the method of choice to finance the construction of the data center conversions that we have in the pipeline. So really when you say, well, how much capex do you need to get through to your end-of-year target of $200 million ARR? As we put in this press release, that cluster, the Nvidia reference architecture, worked out to about $185 million. So if you're putting, say, 20% down on that, it's just under 40 million bucks.

And then the rest you get GPU vendor financing with a blue-chip lender, and so you're targeting single-digit lease-to-own. What I could say is, directionally, usually the terms of these GPU finance is less than the term of the contract. But again the actual GPU finance kicks in before the GPUs deliver because typically there's a big payment associated with that. And so we'll kind of provide the street more updates on IRR, etc., the closer we get to the deployment date.

But really, yeah, the capex is just what I described—it's the down payment requirement, roughly—like I'm giving you indicative figures here—is the down payment requirement on the GPU cluster. And, you know, there's nominal deposits related to the Bell AI Fabric colo, single-digit millions, which were paid up a long time ago. So that's what's really exciting. I think that the capex to convert the small Toronto site is about $40 million to bring that to HPC Tier 3 liquid-cooled, and that would be able to stand up another 2,000 GPUs or do colo.

We've had—reserving freeze on that site just based on where its jurisdiction is. That one's only a 7 megawatt utility load, 5 megawatts of IT load. But again, just based on its location, it's very attractive. So that's kind of the near-term stuff. And I think we put in the deck that it's a $200 million capex for the big Bowden conversion to get to 25 megawatts of critical IT load. Again we've got that LOI signed, definitive in the wings, and those are sort of, I would say, the most near-term capex figures for you, if that's helpful.

Mike, did I answer the questions?

Mike Colonnese, Analyst at H.C. Wainwright

Yes, I have. Great color. Appreciate that.

Nathan Fast, Director of Marketing and Branding

You bet, Mike. Final question this morning comes from the line of Stephen Glagola from KBW.

Stephen Glagola, Analyst

Hey, thank you for the question and congrats on all the cloud progress. For the $84.7 million Swedish tax liability, can you help us understand the likely timing of any cash payment there and what avenues remain available to mitigate or defer that obligation, and how management intends to fund that liability if it becomes due?

Aydin Kilic, Member

Thank you. Yeah, hi Stephen. We don't plan on honoring that liability is the takeaway. We addressed it in the press release. I would refer you to that as well as Darcy's section. And, you know, in our opinion the treatment of tax is uneven. It's, you know, they gave Northern Data a hard time too. So this is not exclusive to us. It's just how the FTA was such a huge fan of bitcoin mining. And so, you know, we've paid 50 million of taxes already from our normal course of operations and we have advisors in the country and we've looked at other remedies in terms of appealing and contesting and even going a step above to the European Union.

And again that commentary is detailed in the press release. I would refer you to that but really we do not think that that's a good use of shareholder capital to pay down this, in my opinion, egregious tax claim. And we've paid all the taxes in the normal course of business. And again, this is related to VAT on ASICs, so sort of a fossil relic from the past, if you will, that had long been contended for. You know, really it goes back to 2023 and it's always been there.

It's been in our disclosures for the last couple years. We haven't paid it, we don't plan to pay it and it's just something that we're going to continue to appeal. So that's really it.

Stephen Glagola, Analyst

Thank you.

Nathan Fast, Director of Marketing and Branding

Thank you, Stephen. Thank you to all of our analysts. That concludes our Q and A session and our Q1 fiscal 2027 earnings call. Thank you to all of our shareholders and the HIVE Digital Technologies community for joining. Look forward to speaking to you again soon.

Disclaimer: This transcript is provided for informational purposes only. While we strive for accuracy, there may be errors or omissions in this automated transcription. For official company statements and financial information, please refer to the company's SEC filings and official press releases. Corporate participants' and analysts' statements reflect their views as of the date of this call and are subject to change without notice.