Hello, and welcome to HIVE Digital Technologies earnings webcast for the second quarter of fiscal 2027. I am Nathan Fast, Director of Marketing and Branding at HIVE, and I will be moderating today's call. Before we begin, please turn to slide two for our disclosures. Today's presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995, forward-looking information under applicable Canadian securities laws. Any statement that is not a statement of historical fact is forward-looking. Words such as expects, believes, anticipates, and similar expressions identify these statements. These statements speak only as of today's date. Actual results may differ materially, and we undertake no obligation to update them except as required by law. For a full discussion of risk factors, please refer to our most recent filings at sec.gov and on SEDAR+.
We will also reference certain non-GAAP financial measures, including Adjusted EBITDA, adjusted net income, and free cash flow. Management uses these measures to evaluate operating performance and believes they give investors additional insight. They are supplemental only. They should not be considered in isolation from or as a substitute for GAAP results. Reconciliations to the most directly comparable GAAP measures are included in the appendix to this presentation and in our earnings press release furnished to the SEC on Form 8-K. On the next slide, it is my pleasure to introduce today's presenters, Frank Holmes, Executive Chairman, Aydin Kilic, President and CEO, and Darcy Daubaras, Chief Financial Officer. With that, I will turn the call over to Frank for a macro recap of the quarter.
Thank you, Nathan. As usual, I try to give my Chief Investment Officer as Executive Chairman of HIVE a macro recap and interject some of my macro commentaries as a money manager in what I have seen for the company and how it fits into the global scene. Before we do this, I always like to talk about VNA volatility, and it is so important for investors to recognize that different factors and over time changes with the volatility, but the S&P daily volatility is plus or minus 1%, and the 10-day is 2%. It is interesting because gold used to be the same as the S&P 500, and now it is much greater, and NVIDIA is also greater than what the S&P 500.
Bitcoin, as you can see, is three times the volatility of daily of what the S&P 500 is, and almost two and a half times, or sorry, three and a half times what the daily volatility is for the S&P. Strategy, Michael Saylor's incredible company, it even has a greater volatility. CoreWeave, because of its extremely leveraged balance sheet, it seems to get more volatility as the poster child in the AI business to use to say it has all kinds of financial problems. At this stage, I think so much is greatly exaggerated, but it still is pretty strong, the negative narrative on both tier 1 to tier 3 data centers, and I will give you more commentary on that.
HIVE itself, its volatility has expanded along with, you can see, with CoreWeave, and the 10-day gets exasperated even more so because of Bitcoin and the movements in Bitcoin. I will walk you through some of the numbers so that you can get better understanding of what the great value proposition. Next, please. Before we get into that, I want to thank the investors. I want to thank for being loyal with this volatility, with Invesco, Citadel Advisors, Millennium Management, Situational Awareness, and Biden Advisory for believing in our vision and recognizing the great value we provide to this industry. The team, most important is the quality and the depth and breadth of the team. As you can see, Aydin Kilic and Craig Tavares.
Aydin is the CEO for HIVE, and Craig is the President and Chief Operating Officer of BUZZ, which is our HPC AI factory strategy, in particular, sovereign data centers. Craig is doing a phenomenal job in Canada and also going back and forth to visit the team now, Johanna Thörnblad in Sweden to expand that footprint. Darcy Daubaras is the longest-standing CFO. Gabriel Lamas is our President, electrical engineer that's in Paraguay. Gabriel Ibghy, as you can see here, is General Counsel, and he goes between Germany and Canada. So we do function over many time zones and provide interesting skill set. The world's first publicly listed digital asset miner on the TSX, 2017. You can see last year with the big growth was our exahash and Bitcoin.
We're so happy we put that on the balance sheet and the income statement because it's given us economies of scale as Bitcoin went through a big correction this year. I'll give you more color on that in a minute. It's important to recognize that HIVE functions over nine time zones, five languages, and three continents. As you can see from Sweden to Vancouver to Montreal, and Bermuda is in there right in the middle, and Texas, the head office is here where I am. We continue to be very efficient as miners for Tier 1 data centers or bare metal AI business model. So HIVE is turning hydroelectric power into computing power. We've been doing this when we started in Iceland, then Sweden. The dam, the hydro dam is very close to our facility in Boden, Sweden, and the same thing in Canada.
In Paraguay, this photograph is an incredible, the largest dam in the Western Hemisphere. It is like 5 km in length, massive, and generates about 15 GW of electricity, half for Brazil and half for Paraguay. So Paraguay, with a small population of 7 million people, is predominantly exporting beef and produce. They feed about the world, about 70 million buyers of their products around the world, but now it's accelerating compute power. I'll give you more color in a second. Next, please. What's really important intellectually to grasp, because we find the frustration with regulatories and banks in Sweden or Canada, that we transform electrons from hydroelectricity into the value chain, what they like to call it, and we export compute power. We don't export beef or food or iron ore. We export compute power.
This is very significant because you can't export hydroelectricity 5,000 miles, but you can transform it and export. We export this compute power to Bermuda, which then goes to New York State for Bitcoin mining tier 1 data centers, where we're paid in Bitcoin. For bare metals, we get paid in USD. That is why, a big reason why we're in Bermuda. When the company first started, we started in Switzerland. Why Switzerland in 2017? Because it had the most advanced regulatory model and structure. We, and me as being co-founder and putting up capital and reputation, I wanted to make sure we're in the most regulated world, who had some guardrails up early, and Switzerland was much further ahead than other jurisdictions. In 2018, Bermuda came out with regulatory arms and basically a whole digital crypto set of rules.
From an accounting point of view, it made it much easier to go from Switzerland into Bermuda to be the comptroller, not in Zug, Switzerland. That's where we've been for many years now. It's a very important global design to be tax effective for investors and also for ourselves, and we end up paying a lot of taxes. We have paid a lot of taxes in Sweden, a lot in Paraguay, and a lot in Canada. What it does too is it allows us to be in a regulated world, and now the U.S. has set up with a regulatory regime for the crypto industry, and I think that the GENIUS Act is a very important step stone for America now.
HIVE uses only green energy, a low electricity cost, low temperature, except for when we go to Paraguay, it's much hotter, so we have a different type of chip in our tier 1 data center there. HIVE stock rises above the 50-day, falls below it, and now it's below it, and a lot has to do with the summer, was the Japanese yen that was going through it. We saw earlier this year in June, going into June with a call, the carry trade, and it first impacted Korean semiconductor massive sell-off that impacted U.S. stocks and impacted HIVE. You can see these external events such as Bitcoin pricing or the unwinding of Japan has had this incredible multiplying effect and amplification of volatility.
But all through all this noise, we just continue to plod along and make sure that we're doing the most effective mining or bare metal expansion program. What's important for investors at the Central Bank of Paraguay now recognizes Bitcoin as an export. As I mentioned at the beginning, they export a lot of food for the world, and now we're showing up as exporting compute power is very important for the country. We're talking about more than a couple hundred million USD of value that goes through the central bank now. You can see on the map, this visual, this helps you to I mentioned earlier, the map was showing you how the compute power goes from Sweden and from Paraguay and from Canada to Bermuda.
All of this is really, I think, we are going to see going forward, where more central banks are going to start adding that as an important component of GDP. HIVE was the first to build Tier 1 data centers in Paraguay, and there is long-term Tier III AI factories. The rally, this rally is a massive event, it is like Formula 1 in Texas and Mexico and Miami and Las Vegas and all over Europe. In Paraguay, it is just car racing, and it is a big event. But us being the first is to build these unique data center construction. HIVE has led the way. Stargate is heading to Paraguay, where in that course there, that is opening. Crusoe is partnered with locals to do 100 MW going up to higher for a long-term energy contract.
HIVE has also been going through the formal process for our data center for HPC for the 100 MW. But we have been going through this sort of drama that has happened just recently in Paraguay, which everyone knows, regarding conflicts of interest, betrayal by the President of the country, who is now in jail. But the country shows that it has great, what is important for investors, great stability because it has the checks and balances of power and accountability. The institutions there have been able to create this stability, and I am so impressed with it. But in addition to characters that would happen in the sort of disconnected short term with Paraguay, which will resolve it, we come back, it does not matter.
There are always these bad characters. Bad characters, and it is how fast can you deal with it. Precious metals in China, 83 metric tons of gold-plated copper bars for collateral.
Again, it just does not matter where you go, there are these bad characters, and there has got to be checks out there to try to unravel this. Binance again, last year they blame it on Algo. This year, interestingly enough, just before the American elections, Binance gets hacked and Bitcoin falls 10% immediately. The real concern is it is not Bitcoin. Bitcoin as itself has not been hacked. It is the on-ramps and off-ramps. As more policies come embraced in America and other jurisdictions around the world, then I think things like this hacking into Binance by supposedly one of their engineers left his computer open. But does it really matter? There are just bad actors no matter where you go in any industry, and it is how quickly can you cut those losses, correct the problem, learn from the lesson, and push forward in life.
Recently, we all went to the All-In, that is Aydin and Craig and myself went to Los Angeles to the All-In Summit, and I think it is the number one podcast in the world. It comes out on Fridays, and it is a great analysis of technology, and it gives you lots of different opinions. If you are not listening to it, I highly recommend it. They bring on usually CEOs of companies, and they have different perspectives of what happened in the sweep just recently in this election cycle in America, and what does it mean to AI and data centers? But the blue midterm sweep puts sovereign data centers like BUZZ in focus.
I think this is extremely valuable because there's something that's happened in America with lots of FUD and deep concern that the blue swing across America, and one of the big concerns besides the Iran war lasting too long and Americans being really ticked off about it because every week they're spending $100-$200 more at the gas pump because of this war in Iran. That was a big swing, especially by the libertarian votes and people that sort of are swing voters between the Democrats and Republicans. They all swung from being Republican, that swing vote, in the Democratic Party. With that, there are some higher percentages in states like Texas, which is 28% of all the data centers being built and has the biggest infrastructure electricity, that there's a concern in this state.
I think that it lends itself that if anything, it's just going to delay data centers. This makes China very, very happy in the AI boom. I don't think it stops it, but it will delay the process of building out. What this allows is countries like Canada and Paraguay and Sweden is to be able, as they say in Canada and Sweden, skate faster. They have an AI minister. They have policies in place for AI and for data centers. So this, I think, is really a great way of looking at HIVE as the strongest position in these other jurisdictions with sovereign data centers. So HIVE's BUZZ HPC partners with ProCogia to deliver end-to-end sovereign AI across Canada and U.S. and Europe. Next, please.
We know that recently, the Kenny Business Development in Sweden have had meetings, and our country president, Johanna Thörnblad, has been back and forth visiting Sweden, where they're truly working together for cross-border. This is very different with Trump and his fight with tariffs is basically seeing a realignment of interest between Europe and Canada and sovereign data centers. So I think this is a really interesting time for the value of these assets we have. What's really important is the sheer size of BUZZ Gigafactory of 320 MW of electricity, sponsorship and support by the federal government and the provincial government. It's going through its normal process right now, some additional clearance to make sure that it's not going to harm the community, but really, it's a beautiful location, and it only enhances Canada's ability.
I think what we're going to see more and more in Canada, where Canada will be exporting oil and gas to the U.S. to a degree, or electricity, it'll be exporting to the U.S. compute power. What we've been demonstrating, we've been successfully doing from Canada, from Sweden, and from Paraguay. At one time when we started in Iceland, this is a rendering of what the potential is for our 320 MW GTA Gigafactory, smack in the middle between the University of Toronto and Canada's, what they like to call University of Waterloo, which is Canada's MIT. So we have the intellectual capital in the largest city in the country, but it's also a city that's been well-known, if you look at history, for R&D and for innovation. I often think it's because of the long, cold winters.
Gets dark at 4:00 o'clock and the sun doesn't come up till 9:00 the next morning, so people are really quite advanced at doing research in the winter. This is a rendering which we feel is a huge expansion opportunity over the next 12 months. HIVE and BUZZ HPC's partnerships with NVIDIA, with Simon Fraser University, as you can see here, with Bell Canada, with Dell Canada, NVIDIA, Tigo in Paraguay. We do have top of the food chain intellectual and Vector Institute, where Nobel Prize winners come out of. We are connected at the very top end of the intellectual food chain and business development in the ecosystem of AI Tier III data centers. Expanding our industry reach. We've been very, very busy this quarter. Conferences attended, 16 of them.
Just had an incredible event in Berlin with NVIDIA and diplomats and the head of the whole AI build-out and data centers attended our event, and that was just super charged with this, I think, just great people. Interviews, doing two to a day, it appears. Sometimes we do three in a day. Building lasting relationships in the quarter. We can see here that the President of Taiwan met with the Country President of Paraguay, Gabriel Lamas, in Taipei. Evan Solomon is the Minister of Artificial Intelligence and Digital Innovation under Prime Minister Carney. There's myself with Michael Dell. This is the second meeting we've had on brainstorming of how we build our companies. It's just an honor for me to be with Michael Dell that's given just billions of dollars for American kids for education.
The thoughtfulness and the stewardship of he and his wife are over the top. I just applaud for his leadership. I really think you got to think of Prime Minister Carney has reversed a lot of Justin Trudeau's policies, which were really anti-growth, anti-capital investing. A lot of global macro economists look at GDP per capita, and they look at GDP, what percentage of money of the economy went into infrastructure and innovation and R&D. Under Justin Trudeau, it fell way behind, and that impacted the relative GDP per capita of the U.S. to Canada, which were always in the same trending line. It also impacted Europe.
What's really important for investors is that Prime Minister Carney has showed tremendous leadership of reversing a lot of these, and the same political party recognizing with his experience from Goldman Sachs partner to the head of Bank of Canada and head of Bank of England and Chairman of Brookfield, what creates wealth. It's about capital investing long term that unlocks jobs, sustainability, and growth. Building industry partnerships, as you can see, many pictures of the team traveling the world. Land and power, that's building the pipeline. 320 MW Gigafactory in Toronto region, 32 MW in Boden, 100 MW in Paraguay, and Pecos County working on 500 MW. Next, please.
I like to try to relate to some other things as a macro investors for investors to understand that there's some macro events taking place, and they're still not over under the Trump era, and that is the Monroe Doctrine of looking at the Western Hemisphere. There's many reasons for opening up, and you've seen a huge swing in South America going to the middle on the election process from the far left. What are some of these reasons? Well, a lot of the strategic minerals which China is cutting America off, and Japan, they're in Latin America. When we look at population of consumers, we're talking about one billion.
The idea of not being pulled into wars in Europe or into Asia of really north-south traffic, and I know from other products I have, the north-south bound traffic for the airlines industry, known as digital nomads, is huge. It's absolutely huge in the winter season of people flying from Canada and the U.S. that fly south to Mexico, to Argentina, wherever there's strong internet. They like to go to Colombia, they go to Medellín, and they learn the local cultures, and they really use Airbnb. It's really quite fascinating to see this opening up and focus. Rather than going across the Pacific, the north-south traffic is going to only increase. Next, please. The reasons for that is this what's called the Ring of Fire. What we see here is Bolivia with all of its lithium.
In Chile, also with lithium in Argentina, what's known as the big lithium triangle. We need that. We need the lithium batteries, the backup for HPC data centers, and for Tesla vehicles or EV vehicles, you need lithium. This is another big factor of north-south in addition to oil in Venezuela, which is turning around their GDP dramatically since Maduro's been taken out of his corruption and dictatorship in that country. AI data centers are driving demand for lithium and other critical minerals. Next, please. We can see this with copper prices. We're hitting all-time highs because of the build-out and electrical grid. Copper, what they call the Ring of Fire, is very important. It comes from Peru and Chile, the big copper porphyry deposits, and they also come from in the U.S., and up into Canada, into the Rockies.
With that, we try to break down here for you, land is $1 for every $100 spent of AI CapEx. This is sort of a breakout that what does it cost by Louis Gardner, and I think it's just helpful to recognize that the CapEx cost for a megawatt for HPC data center used to be 33x what a tier 1 data center is for Bitcoin mining, and now it's pushing more than 40x . Most of that is really the new power structure and the cost of the NVIDIA chips has gone up dramatically. What I want to highlight to investors is that, I'm going to turn it over here to Aydin, but for the financials is to recognize that revenue was up about 3% even though Bitcoin was off 27%.
So that means in the past year, we built economies of scale, and our margins are still very healthy and robust. I think the other part is in US GAAP and mark-to-market, during this quarter, we started buying some better, more energy-efficient ASIC chips. Mark-to-market says, well, if Bitcoin is down, then the value of those ASIC chips is down, and before the next halving, they are down. So we still say, well, they are great prices to lower or become more efficient, and with that, we bought some more of these chips. Well, that just says your depreciation, even though we have an accelerated depreciation for our chips, has not been as great, and that is how inexpensive these chips have fallen. So we thought it was very prudent to go and start buying these chips on a weekly basis to upgrade our data centers. What does that do?
It creates, on a mark-to-market rulemaking, a non-cash charge of $31 million. So it is recognizing, following US GAAP, that you really have to drill through and take a look at what was the revenue, and then what was our Adjusted EBITDA or operating income from that. You can see that it is very, very healthy, especially when you compare to our peers, and we compare to ourselves quarter-over-quarter and year-over-year. So Darcy has covered this, and Aydin is going to probably give you some more interesting granularity with it.
Thank you, Frank, for that excellent macro strategic overview of the industry. Now, let us get into some highlights from the quarter and the executive summary. It was a strong quarter for us, approximately $85 million of total revenue, of which we had about $30 million of gross operating margin. Now, netting off our $10.7 million of G&A, our earnings from operation in the business came out to $19.5 million. So what I like to think of as corporate margin on a cash basis, what did the business produce? Now, along the bottom, if you look at the Adjusted EBITDA of $24.8 million, that stands alongside a net loss of $78.7 million, which is measured against $102 million of depreciation and non-cash impairment and other non-cash charges.
By the way, of that $102 million, about $86 million is really just depreciation and a non-cash impairment relating to ASICs and hashprice, which applies to any miners following US GAAP operating ASICs. Let us double-click on those operating margins. So again, we have the two engines, Bitcoin mining and HPC Cloud. So the total revenue, $85 million with 36% gross operating margin, was really made up of $76 million of Bitcoin mining revenue, where gross mining margin was 34% for the quarter, and the HPC AI revenue of $8.8 million actually did a very nice 51% gross HPC margin, resulting in a $4.5 million gross HPC margin. Now, a couple things to point out. One, we realized our $35 million HPC annualized recurring revenue target, so that is great. Also, it is worth pointing out that we are still scaling this business.
As we bring online new GPU clusters, our margins will continue to climb on HPC business. Obviously, as you are scaling the business, bringing on more resources, you are having additional overhead with that growth. Our newer deals are targeting about 80% EBITDA on GPU Cloud, but we will get into that shortly. Let us go to the next slide. Going into the winter season here, as of September 30th, $139 million of cash on the balance sheet and $22.6 million of Bitcoin on the balance sheet as well, for a combined $160 million of liquid assets, cash and cash equivalent on the balance sheet. With the volatility in the market lately, it is good to have that strong balance sheet, which we are protecting. Our mandate this quarter is to protect our balance sheet.
We have navigated two Bitcoin halvings and Ethereum Merge, and this is the AI bear market that we have seen with all the recent pullbacks and the calamity, almost $4 trillion in market value wiped out in the last couple weeks. Again, this is not our first rodeo. HIVE has been a publicly traded company. In fact, the longest standing of our peer group of HPC and AI infrastructure builders that have pivoted from Bitcoin mining. Keeping a strong balance sheet, conservative G&A, although we are growing the business, is our solution to scaling and, of course, growing cash flow and realizing return on invested capital. Let us go to the next slide. If you look at our operating margins today, some really great news.
Not only did we reach that $35 million ARR for the quarter, but as of today, we are actually at $40 million of annual recurring revenue for HPC. What that means is that we are doing over $110,000 daily revenue, which is fantastic. If you apply those same HPC margins of 50% approximately from the quarter to our current $40 million, it is about a $20 million annualized gross HPC margin today. On the Bitcoin side of the business, it was $300 million ARR for the quarter, but even if you look at mine economics and production since the period and to today, it works out to a very similar figure. Applying that same 34% gross mining margin, it is about another $100 million of annualized mining margin.
I just want to give the street and all the investors out there a snapshot of where we stand today, November 9th, since the quarter end. Again, really exciting that our HPC revenues continue to climb due to strong demand and, of course, GPU hourly prices. Let us go to the next slide. Global power footprint, 860 MW. We have updated this slide to include all the sites, and there is really two categories. There is Bitcoin mining sites and there is Bitcoin mining sites that are being converted to Tier III. We have highlighted them in green and orange, and you can sum up that across all the sites that have been identified for conversion to Tier III, we have a 530 MW Tier III conversion pipeline globally, which is very exciting.
Of course, Big Boden, the Gigafactory site in the Greater Toronto Area, their New Brunswick site, are big contributors to that pipeline. And of course, 440 MW is active, currently mining Bitcoin. We are really excited about the growth that we have ahead. That is just a good global snapshot because land and power is the input to the system. Whether you are selling rack space or standing up GPU Cloud and selling tokens, really the input to the system is having powered land. We have got a lot of it. As I have said in the past on earnings calls, we had been land banking by substations, and this, of course, is the fruit of that labor. Let us go to the next slide. Some great news out of Ontario, actually, on the home front in Canada.
We saw that blue wave sweep the U.S. midterms, and there was massive headwinds and negative sentiments surrounding AI, and we saw collectively $4 trillion of market cap from American tech stocks really get wiped out over the last couple weeks. There is a massive breath of fresh air, and candidly, what we think is an extremely strong tailwind for HIVE and BUZZ as our data center pipeline across Canada totals 400 MW of utility load. Most notably, our Gigafactory site and our Toronto Airport site sit proudly in Ontario. Premier Doug Ford announced a playbook in August where he said that they are really choosing to build the economy of the future that will generate billions of economic growth. They want increased tax revenues and create thousands of new jobs. We think Premier Ford, he gets it. We salute you, Premier Ford.
His Minister of Economic Development, Vic Fedeli, said the province will make certain to give priority to data centers that use a closed-loop system, which I do note all of our data center designs for our sites, notably the Gigafactory, Toronto, and even Grand Falls, have always been based on a liquid-cooled closed-loop system. We have talked about this before, so it is very comforting to see the strong support from the office of the Premier. Moreover, they doubled down on this because the data center playbook came out in August, which is a few days ago. In fact, Ford's office stated that they will be approving the first swath of new AI data centers, quote, "sooner rather than later." Again, a breath of fresh air amidst a lot of global turmoil, at least in the U.S. markets. So, we are very excited about the outlook in Canada.
Let us go to the next slide. So, this is a recap of all the sites that we have underway, their status. Again, starting with the Big Boden site, that one is the most near term. We have got the LOI executed. We will be providing the streets update as we get closer to that definitive agreement. Really, it is an exercise not in rushing, but getting the right deal done for our shareholders. It is an IG off-taker, and so one must contemplate how much deposit you are getting on signing, what parental guarantees look like, and all of the things that will affect your cost of capital when you are going to finance it.
We're very much aware of the headwinds in the industry right now, so we are being cautious and protecting our balance sheet while also seeking the lowest cost of capital by strongly negotiating terms that behoove our shareholders and our stakeholders. That's the quick update on Big Boden, and then, of course, the rest of the sites. This is more of a handy reference slide. We can actually flip through one by one. Let's go to the next slide. Here's a beautiful visual of the Toronto Airport site. This is not a Bitcoin mine. It's a proper data center that we acquired in September of last year. Again, we've talked about this site being very attractive for military applications.
We've had some reverse inquiries from Canadian government, and budget for this will be about $40 million, which we're funded for, thanks to the financings we've done earlier this year. At 5 MW of IT load, this would do $8 million annual recurring revenue as colo, or as GPU Cloud, you'd fit about 2,000 GB300s, which at today's prices would easily throw off $80 million ARR. The operative word is optionality. It's a bite-sized site. It will be a showcase marquee, and we welcome to have visitors as the site progresses, and we want to showcase the progress. Stay tuned for updates. Let's go to the next slide. The Gigafactory, of course, a crown jewel. Again, how could we not mention this site, especially with the recent positive words from Premier Doug Ford and the Province of Ontario?
Again, as we said, our design has always been a closed-loop cooling, zero water usage design. Using liquid cooling, you have loops where, whether it's glycol fluid, you've effectively got a system whereby thermal energy is transferred through closed-loop piping to chillers that sit outside of the building. Thermal energy gets expelled that way, and so no water is actually consumed despite all the FUD. It seems that a lot of the protesters, it's ideology-driven. I don't know anybody that actually builds data centers, whether it's in the industry, that's protesting them, because they know that these things don't drain energy. When you're consuming electricity, you are an industrial client of utility companies. Do we shut down aluminum smelters? Do we shut down every other industry all across North America that uses megawatts of power? No, because these drive revenue and taxes for the utility companies.
No one's putting a gun to the head of the utility company. I think that there's a lot of misunderstanding. It's really FUD. As a Bitcoin mining company, we understand FUD, fear, uncertainty, and doubt. Anyways, just to remind folks that this is a project that we engage with the region, as well as the municipality, on sustainable planning design, as well as engineering infrastructure upgrades. That means widening roadways, improving civil infrastructure for the 25-acre parcel where we assembled land for this future site. Again, closed-loop cooling, zero water design, as is called for in NVIDIA reference design. NVIDIA doesn't make an open-loop system that just continuously drains water. Again, I just think there's a lot of misunderstanding. We just want to educate the public. Let's go to the next slide. New Brunswick campus, of course.
The update here is we have now expanded the site. We have closed as recently as September, all the parcels of land to the east and to the south. This is now a 64-acre campus. We are very excited. 15 MW of IT load, and stay tuned for updates. This forms part of our Canadian sovereign compute thesis. Our dual engine strategy, approximately $300 million ARR, whether you look at the fiscal Q2 or really production and economics since Q2, sits squarely at 300, actually slightly above since the quarter. Notwithstanding the recent downtrend in Bitcoin price over the last few days, we actually saw really strong hashprice in the month of October. In the HPC GPU Cloud, we are now at $40 million ARR, again, about $110,000 a day. In fact, a bit north of that on some days. Very exciting.
On a combined basis, our dual engine is about $340 million ARR. This is just a recap, a handy recap, especially for the analysts out there. We signed three GPU Cloud deals this year, totaling $600 million of total contract value. That $40 million ARR that we are at today, if you look at the bar chart at the bottom, the Cohere deal gets us $115 million by adding $70 million ARR. Again, that is, sorry, $75 million ARR, and that is a three-year deal. The IG five-year deal for $350 million adds $70 million ARR, and that gets the contracted total GPU Cloud revenue to $185 million. Again, just a handy recap of deal terms, hardware, et cetera. If you zoom out and just look at our HPC revenue roadmap, well on track to hit that $200 million ARR target.
Expect an update on the 504 B300 GPU deal. I will tell you that we have demand through this and beyond on the GPU Cloud side. We are going to talk more about that shortly, so stay tuned and expect us to provide updates on more GPU deals. Of course, as I mentioned earlier, we are advancing the Big Boden. We actually had a site tour with a potential lender last week in Boden. So again, focusing on getting the best cost of capital, advancing that process, making sure we are looking out for our shareholders, and negotiating deal terms that really, this will be a long-term deal. So we want to make sure that we are doing the right thing today, and securing strong cash flows. So stay tuned for updates on Big Boden.
On a combined basis, if you looked at all the sites on the right as HPC/Colo, 325 MW of IT load, that would be $500 million ARR. If you look at a GPU Cloud on the left, that would be $200 million ARR. Both would target about 80% EBITDA, $700 million ARR combined, over $500 million of projected EBITDA based on our HPC revenue roadmap. On that note, once we hit our $200 million GPU Cloud target, targeted for end of this year, going into 2027, with that $200 million GPU Cloud annual recurring revenue, plus the Bitcoin mining revenue, what is really exciting, it actually puts us at a $500 million combined ARR target for early calendar 2027.
Now, of course, Bitcoin mining revenues vary with hashprice, but still, with what we have seen recently and the growth and deployment of our GPU Cloud clusters, over the coming months, we are really excited to see that growth, because this also means more Adjusted EBITDA and more cash flow for the business. As I mentioned, we have outsized demand beyond what we projected. This is an illustrative slide to show the street if, for example, we took the Toronto Airport site and allocated it as GPU Cloud, that would have our total GPU target revenue to $285 million. Again, we have demand. This is an illustrative slide, and you would see that the HPC/Colo target still is about $495 million on the right. It actually increases the combined HPC and GPU Cloud target to $780 million. Again, there is a lot of volatility.
We have seen commentary on GPU financing and credit market crunch, et cetera. We are protecting our balance sheet. Our goal is to navigate the headwinds. Stability is key, maintaining positive cash flow. This really just shows what we have got demand for near term, and we will provide the street more updates. Really, our focus is to build out the Toronto Airport site for 2027 and have that go live before the end of the year. Again, just so you can understand the metric of, if you took a colo site and moved it to GPU Cloud, the rule of thumb is every 5 MW of IT load is 2,000 GPUs, and 2,000 GPUs at current prices is about $80 million ARR. Current prices, we are seeing GPUs are pricing in the $4.50 range and even higher.
Again, this is just an indicative slide for people to plot the economics. Optionality is the key word. We have talked about annual recurring revenue, but on an Adjusted EBITDA basis, if you look at the margins of the, I should say the higher margins, from the new GPU Cloud deals, it would put collectively our Adjusted EBITDA at a potential target of $350 million, with the Boden lease HPC/Colo deal being executed and with the GPU Cloud deal growing to $280 million ARR. Strong outlook for us to grow into as we navigate these market cycles. Again, we are not necessarily in a rush this spring, but rather protect the balance sheet, and only do what we think are the best deals for our shareholders. Let us go to the next slide. Speaking of funding, again, just a quick recap.
We did the $245 million of zero-interest coupon convertible notes this year. After capped calls, net proceeds were $200 million. In addition to that, we recently announced $120 million GPU loan, and funds are with an escrow agent from the lender. This is a committed GPU loan. Again, deal terms, 80% loan to cost, three-year term, 9% interest, and those funds get released to the GPU OEM milestones, because this is the anatomy of a GPU deal. These proceeds were to fund the Cohere deal, which again, was a three-year, $225 million total contract value deal. The hardware, the CapEx was $165 million. Nowadays, GPU clusters of this size, 2,000 GPU cluster of GB300, are close to $200 million plus.
Again, we locked this in much earlier in the year, so $129 million of financing on the $165 million of GPU cluster. This is a straight line. We will own the GPUs outright. The loan balance is $0 after the 36 months. It is co-term with the offtake agreement of three years, which is Bell Canada. If you do not know about Cohere, right now they are doing a $35 billion Series E backed by the Government of Canada, Bell, and even NVIDIA. Again, we own the GPUs outright at month 36, and the resale value is actually upside. This is SPV level debt. When you look at, on an annualized basis, $75 million of revenue with about $49 million of debt service, it is very, I would say, well-structured financing for our first large cluster deal.
Again, we want it to cash flow, we want to make sense. Yes, revenue is important, Adjusted EBITDA is more important, and cash flow is very important as well. We are happy to announce this, and we will continue to navigate the market and scale our business very intentionally by sourcing what we think is the lowest cost of capital. Just a recap. The last 12 months of revenue, approximately $329 million for HIVE, which is more than a lot of our peers. Again, our dual engine strategy continues to generate revenue and cash flow. Very healthy, 860 MW of secured power. You see some of our peers between 800 to over 2 GW of power. Our two-year forward HPC ARR stands very respectable, $700 million in our base case.
Really, we think we have got very strong optionality as we continue to execute on leases for HPC colo and GPU Cloud, and of course, growing at re-rating and valuation alongside our peers. Let us go to the next slide. On the near-term basis, with the GPU Cloud deals that are contracting $185 million, including our active $40 million revenue, and the HPC deal at Boden, $45 million ARR, using current multiples, even though the market has pulled back, using current multiples, it would put us at about a $1.4 billion valuation. We think that, notwithstanding the headwinds in the industry, we are continuing to execute, so we think that there is a lot of upside here, especially if you look at where our peers are trading.
The macro zoom out, if you look at all of our sort of growth roadmap, the $500 million target for HPC colo, or the 325 MW by IT load, and then $200 million on the GPU Cloud, using peer multiples again, even though the market has pulled back and there are more conservative multiples, puts us at a $3.9 billion implied valuation to grow into as we build out our sites over the next two years. On that note, I am going to turn it over to Mr. Darcy Daubaras, the longest-standing CFO in the industry. Next slide.
Thank you, Aydin. For our investors, I will walk through HIVE's second quarter financial results for fiscal 2027 that is ended September 30, 2026. This quarter presents two important stories. Operationally, we delivered higher sequential revenue, increased gross operating margin dollars, and significantly improved Adjusted EBITDA. At the same time, our reported US GAAP results reflect the challenging economics of Bitcoin mining and a significant non-cash impairment of equipment. I will explain both, because understanding that distinction is important to assessing our performance. Before reviewing the quarter, I will briefly touch on our capital structure. At September 30, HIVE had approximately 276 million common shares outstanding, 3 million warrants, 2.6 million options, and 15.6 million restricted share units. Our shares trade on the Toronto Stock Exchange and Nasdaq under HIVE, and on the Colombian Exchange under HIVECO.
Turning to our financial highlights, total revenue was $84.9 million, including $76 million from Bitcoin mining and $8.9 million from HPC and AI services. Gross operating margin was $30.2 million, and Adjusted EBITDA was positive $24.8 million. These represent meaningful improvements from the prior quarter. on a US GAAP basis, however, EBITDA was negative $22.1 million, and the net loss was $78.7 million. The net loss includes a $31.2 million non-cash impairment of equipment, as well as $54.9 million of depreciation, $8.6 million of share-based compensation, and a $7.3 million loss from changes in the fair value of derivatives. The impairment deserves some explanation. Under US GAAP, we assessed our Bitcoin mining equipment following lower mining revenue per unit of computing power and reduced market values for ASIC equipment. These factors lowered expected recoverable cash flows and residual values.
Our assessment determined that the asset group's carrying value exceeded its recoverable value, requiring this $31.2 million non-cash impairment. As mentioned, this is a non-cash charge, but is a real adjustment to the carrying value of our assets. We finished the quarter with 269 Bitcoin in treasury. Three measures summarize the operating quarter. Revenue was $84.9 million, Adjusted EBITDA was $24.8 million, and we produced approximately 1,070 Bitcoin equivalent during the period, keeping that side of the operations extremely strong. Together, these metrics show the operating contribution of our expanded platform, while our US GAAP statements also recognize the depreciation and non-cash impairment associated with the capital-intensive nature of Bitcoin mining. Looking now at our balance sheet, we ended September with $139 million in cash, $22.6 million in digital currencies, $18.6 million of receivables and prepaids, and $11 million of investments.
Current assets totaled $213.4 million versus current liabilities of $137.8 million, leaving working capital of approximately $75.6 million. That liability balance includes an $84.5 million provision for regulatory liabilities related to the previously disclosed VAT matter that occurred in the prior quarter. Cash declined from approximately $208 million at June 30 as we continued investing in our infrastructure and growth plans. We report a $12.2 million of operating cash flow for the first six months of the fiscal year alongside substantial investing activity. We remain focused on maintaining financial flexibility and allocating capital carefully across Bitcoin mining and HPC and AI infrastructure. Turning back to operations, gross operating margin was positive $30.2 million compared with $42.4 million a year ago, a decrease of approximately 29%. Basic loss per share was $0.29 compared with a loss of $0.07 last year.
The larger reported loss reflects both operating economics and those significant non-cash charges, including the equipment impairment recognized this quarter. Gross operating margin excludes depreciation, so it offers a different view of performance from GAAP earnings. Looking at that, the year-over-year comparison highlights the pressure on mining economics. Revenue was $84.9 million compared with $87.3 million in the prior year quarter, a decline of about 3%. Gross operating margin declined from $42.4 million to $30.2 million, and the margin percentage decreased from approximately 49% to 36%. The 10-Q identifies lower mining revenue per unit of computing power as one of the factors affecting our mining equipment recoverability assessment. This is an important reminder that revenue scale alone does not determine profitability in a changing Bitcoin network environment. Sequentially, we made meaningful progress. Revenue increased from $79.1 million in Q1 to $84.9 million in Q2, or approximately 7% increase.
Gross operating margin increased from $24.2 million- $30.2 million, or approximately 25%. That represents an improvement in gross operating margin dollars despite the challenging industry backdrop. Based on the reported amounts, gross operating margin was approximately 31% of revenue in Q1 and 36% in Q2. The improved operating contribution helps explain the increase in Adjusted EBITDA, even though it does not eliminate the effect of non-cash expenses on US GAAP results. From an earnings perspective, I want to distinguish our operating performance from the accounting charges affecting the reported result. Adjusted EBITDA was $24.8 million compared with $31.5 million in Q2 last year. Our GAAP net loss was $78.7 million compared with $15.8 million in the prior year quarter. The most significant new item this quarter was the $31.2 million non-cash equipment impairment.
Under US GAAP, we evaluate long-lived assets for recoverability when circumstances indicate their carrying amounts may not be recoverable. Lower expecting mining economics and ASIC resale values triggered that review. In addition, depreciation totaled $54.9 million compared with $38.3 million a year earlier, reflecting the expense associated with our asset base. Share-based compensation was $8.6 million, and the derivative fair value loss was $7.3 million. These items are reflected in GAAP earnings, while Adjusted EBITDA provides a supplemental view using the reconciliation in our reported results. I would emphasize that the impairment is not a cash outflow this quarter, but it should not be dismissed. It reflects a reassessment of the recoverability of our mining assets. Finally, the sequential earnings comparison shows the improvement in operating contribution. Adjusted EBITDA increased from $13.4 million in Q1 to $24.8 million in Q2, an increase of $11.4 million, or approximately 85%.
At the same time, our GAAP net loss narrowed from $142.9 million in Q1 to $78.7 million in Q2. It is important not to treat that reduction in net loss as a like-for-like improvement in operations. Q1 included the substantial non-cash Swedish regulatory provision, whereas Q2 includes the $31.2 million non-cash impairment of equipment. In fact, the 10-Q in our reporting results reflects a small $0.2 million reversal of regulatory provision expense in Q2 due to foreign exchange movement. Taken together, the quarter demonstrates better sequential revenue, higher gross operating margin dollars, and stronger Adjusted EBITDA. It also demonstrates the need for disciplined investment decisions and careful assessment of asset values as mining economics evolve. As always, we will continue to focus on operating efficiency in our Bitcoin mining and disciplined development of our HPC and AI infrastructure business. With that, I'll turn it over to Nathan Fast.
Thank you, Darcy. That concludes the presentation for today. With that, 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 are ready with your questions. We will begin to choose and ask you to unmute. Our first question comes from the line of Bill Papanastasiou. You can go ahead and unmute and proceed with your question.
Yeah, thank you. Good morning, gentlemen. Congrats on all the progress with the HPC business. For my first question, I would like to touch on the letter of intent for the long-term colocation contract at the 32 MW Boden site in Sweden. Can you just provide a high-level update in terms of how far you are in discussions, what is remaining to get that over the finish line, and just an overview of what the customer profile looks like? Thank you.
Hi, Bill. Yeah, this is Aydin. Thanks for the question. So, I tried to address this in my section of the presentation, whereby I was referring to negotiating the terms of the final agreement that really looking out for our shareholders. There are terms that when you are finalizing an offtake agreement for a colo deal, termination rights, deposits, all of these different key factors that the street looks at, that you would look at to evaluate the risk profile of the deal, and moreover, the lenders. We could have signed the deal weeks and weeks ago, but really, as we are working on structuring the finance for that deal, this is where we have been very involved and even had a consultant for the lender come out and do a site visit last week. I mentioned that in my section as well.
Lots of progress, lots of work happening behind the scenes. So, it is advancing very well, and it is an investment-grade offtaker that we have alluded to. We will provide more details, of course, on term of the agreement as well as the profile of the offtaker. But really, it is an exercise in getting the best terms that allow us to seek lowest cost of capital and financing, which we think right now, especially in markets where we have seen the headwinds and the pullback, cost of capital is important. So, we do not want to rush into a deal just for the sake of reporting new top-line revenue. We want to make sure that along with that, we have got good cash flow and a good financing structure to support the growth.
Appreciate that. As a follow-up, I know this is still very fresh, but just curious how the team's expansion strategy is evolving. I recall the U.S. was a potential geography for expansion in the past. Have you guys parked that idea following the recent midterm results? If you could rank which regions are most attractive today, that would be helpful. Thanks.
Certainly. We've always had an eye on the U.S. We're always looking at sites throughout the U.S. It's got to be the right opportunity, of course. That's what I'll say there. I think that the political headwinds in certain states obviously are a signal. Signals must be read, and plans must be made accordingly. I would say to rank, Canada right now is standing in the lead. I addressed the playbook issued by the Province of Ontario and the administration of Premier Doug Ford, and his office doubling down just a few days ago, releasing a statement that they're going to be approving the first swath of new data centers sooner rather than later. I think that stands in stark contrast to a lot of the rhetoric coming out of different states in the U.S.
By the way, I really empathize with the rest of the industry. I think that all of us winning is good, and obviously provides a unique value proposition for HIVE shareholders and investors at this point. We're grateful that the winds at our back in that regard. I do hope that broadly, cooler heads prevail and people realize that AI is really going to unlock an immense amount of value. I think just kind of on a tangent, the concept of Jevons paradox is just because the cost of a commodity goes down, it doesn't mean that you use less of it. You expend the same cost, and you consume more. We'll have more economic output. I think that's the way it is, an abundance, not a scarcity mindset. At any rate, we're really bullish on Canada right now, obviously.
Of course, Sweden having this LOI signed is very optimistic. As Frank addressed, there have been some political sea changes in Paraguay, but we've been operating there for a long time. Operationally, it hasn't affected us. As a new administration comes into place, we think that, as Frank mentioned, there's calamity that happen everywhere. It's how well are the institutions set up to course correct. We believe that Paraguay is a very stable jurisdiction with good governance. We're also bullish on Paraguay as well. It's so rich in natural resources, very smart cabinet of ministers that are very much pro-data center and pro-AI. I think that consensus still remains strongly rooted, and we'll see how things unfold, but we're very confident that Paraguay is still going to be a fantastic horizon for data center building for AI.
Great to hear that color on Paraguay. You answered my final question. Looking forward to all the progress in the current quarter, and I will go back in the queue. Thank you for answering the questions.
Thank you, Bill. For our next analyst question, we will go to the line of Chris Brendler from Rosenblatt. Chris, please unmute yourself and proceed with your question.
Hey, thanks, Nathan. Thanks for taking my questions. Congrats on all the progress. I just wanted to ask on the Bitcoin mining side, even though it has been a little bit of a rough run, you have seen improvement recently in hashprice, and your results also suggest sort of really strong cost controls. I just wanted to ask from a macro, sort of top-down perspective, given what is going on on the HPC side these days, are you sort of interested in potentially expanding more on the Bitcoin mining side, or are we still prioritizing capital for the HPC side? Also, what is the latest on your plans in Paraguay? I know that you just talked about it, but I was wondering, I think last quarter it was potential for high-performance compute data centers in Paraguay. Right now, I think it is just Bitcoin mining.
Is the plan, sort of related to my first question, still to try to switch, or is there any interest in expanding Bitcoin mining at this point? Thanks.
Yeah, I tried to address this in my section as well. Right now, it is protect the balance sheet and navigate the market that we see ahead. I think different people on the street have different opinions. This pullback, is it short term? Is it medium term? But cash is king, and so we will be deploying cash for assets that will generate near-term revenue and cash flow. I think the other point that we also tried to address in the press release was that ending the quarter with $139 million of cash on the balance sheet, while also successfully meeting the capital commitments for the first large GPU cluster deal that we announced with Cohere, meaning that we had secured the $129 million of debt.
Really, instead of talking about how much money we can spend and how much we can grow, we are trying to show what we have got contracted, how we are financing it, and that we have got a steady hand to get through any market uncertainty and deploy the GPUs for the commitment. As for your question on Bitcoin mining, it is really a function of ROI. You can go back years and years and years. Frank, Darcy, and I always talk about return on invested capital. If the price of an ASIC is attractive enough where you are going to realize a net ROI after electrical costs of 12 months, and of course, not in a bull market, like you have got to measure that in a mid or even bear market, then that is a buy.
But right now, our capacity is full globally, and so we have got 25 exahash, which I think we operated an average of 23.8 exahash this quarter, with firmware downclocking, et cetera, to increase ASIC efficiency, optimize joule per terahash, fleet efficiency. So, we would not really be expanding per se. Opportunistically, as Frank noted, from time to time, if there are small orders that are very attractively priced, I mean, we are talking three bucks a terahash, then doing a small order to replace some older machines in existing infrastructure. But that is just more being opportunistic and not, like we are talking maybe $100,000 spend. So, nothing significant. I hope that answers the question, Chris. Anything else you would like to cover?
Absolutely. Yeah, that makes a ton of sense, Aydi n. Thanks so much for that detailed answer. Yeah, I guess the other question I had was on the HPC cloud business. Nice to see the sequential increase in revenue there. Also saw a sequential increase in the gross margin, and I just want to just remind us, right now we are sort of low 50s level. What is the sort of targeted run rate gross margin in that business?
80%. The new GPU cluster deals that we are bringing online, the GB300s, the GB200s, those are targeted approximately 80% EBITDA, and that is operating within a colo facility with Bell Canada. I hope that is helpful.
Yeah, so we have quite the progress this quarter, but as I guess as the top line and more machines come online, especially those 300s come online, then you will see closer to 80%. Got it. Okay. Thanks so much.
Yeah, because right now, the 51% that you see not only represents, is measured on the $35 million, which was booked for the quarter. So, as that $35 million grows to 200 mil ARR, all that incremental revenue is at 80% EBITDA, gets you very much closer. But moreover, as I alluded to in my section, as we are scaling the business and having costs and resources in place to allow us to scale, that also addresses part of the margin. So once we are kind of at critical mass, that also supports the margin improving.
Great. Thanks, Aydin. Appreciate it.
Yep.
Congrats.
Thank you, Chris. One final question this morning from the line of Fedor from B. Riley. Fedor, please proceed with your question.
Thank you very much, and good morning, everyone. I have a couple of short questions on funding. First of all, it is mining cash flow. You used it to fund HPC build-out, but hashprice is quite low, although with some improvement. If Bitcoin stays around $70,000-$80,000 per coin, how much HPC CapEx can mining actually fund? And with current environment, with current stock price, at what share price would you stop using the ATM to finance the equity portion of HPC build-out? Thank you.
Well, we did the converts, as you know, earlier this year, and so we have got a strong cash position, the $139 million. By the way, on top of that, 22.6 million Bitcoin, so just over $160 million of liquid assets. Capital commitments for the GB200 order satisfied. The GB300 order, we announced earlier in August on the five-year deal with an IG offtaker. We have put deposits in place for that as well, and so we will spell out further details when the financing on that is announced. But what I am giving you guidance on, Fedor, you always ask very nuanced, detailed questions, I appreciate it, is that our significant capital commitments have been satisfied for the most part, and we have also included a budget to complete the Toronto 1 data center for $40 million US to bring on that five megawatts of critical IT load.
So really, our focus is to deploy the clusters that are currently contracted, if you refer to the HPC revenue roadmap slide, AKA the money shot slide, and then build out the Toronto 1 data center. And separately from that, in addition to that, we are looking to undertake a financing process for the Boden site, of course, with a signed lease. We will use that likely to secure corporate bonds, and so not to dip into the balance sheet to finance Boden. So I hope that is helpful. And we have always been very conservative and judicious on the use of the ATM. We have never sold on a down day, and so with where the stock now is for us, and I mean, really the whole industry, yeah, it is not conducive, we think, to use the ATM when the stock is at these levels.
Fortunately, we have the balance sheet and cash flow that allows us to get our contracted revenue online. Notwithstanding where the street is, we always believe fundamentals, once the revenue is live and the cash flow is live, and we are at that $200 million target for the end of year, which by the way, we will be providing updates in very due course. That last tranche, we are $185 million active and contracted revenue as of today. That last remaining tranche of $20 million ARR, that cluster of B300s, 504 GPUs that is on that revenue growth slide, we will be providing an update on that very soon as well, and that will get us squarely at the $200 million target. Really, it is getting that cash flow online. With the margins, once we are at that $200 million, the company will have more cash flow.
But again, we want to focus on protecting the balance sheet and seeing how the market recovers before committing further capital expenditures. Cash is king right now. Hope that helps.
That is super helpful. Thank you. Just to follow up, regarding the GB200 facility, I think it is a good template, just asset level and size their contract. Can you finance that colocation size the same way before tenants sign, or does New Brunswick or Toronto financing depend on landing a lease first?
The nice thing about the colo sites, it is a bit of an easy button because we do not have to finance them. They are built by Bell AI Fabric, and we are a tenant. There are nominal deposits, couple million bucks, but other than that, it is very CapEx light and quick time to market. That was our strategy to get to our $200 million targeted ARR on the GPU Cloud business. Toronto is a nice bite-size, Toronto airport site, nice bite-size 5 MW IT load, $40 million CapEx is very digestible for us off balance sheet.
Then, as mentioned, for Boden, it would be corporate bonds, as we have seen our peers in the sector, the Cipher and Hut 8 of the world, financing with corporate bonds instead of HPC offtake. That is what we have provided guidance on, and we will update the street accordingly on New Brunswick.
We have a lot of interest. There is a lot of interest for New Brunswick. I mean, it is kind of a good question that you are asking because it spurs me to comment, both for GPU Cloud and HPC colo. I tried to thread in the operative word of optionality into my presentation. We know that sentiment varies on the street. Some investors are very bullish on just locking in long-term investment-grade HPC colo. Others believe the higher reward of GPU Cloud, albeit being more CapEx-intensive long term, is more beneficial. Of course, that sentiment changes with risk on, risk off. Stay tuned for updates on New Brunswick, but we have got a ton of demand on that site. As I did address, we expanded that campus to 64 contiguous acres. You might have heard me refer to land banking by substation.
We have done that in Paraguay, and we have done that in Ontario, and we have done that in New Brunswick. It is building out the site footprint while going through design, development, optimization, permitting, et cetera. Of course, working with utility companies, making sure that there is buy-in. They are very excited, of course, to have an AI data center conversion from a Bitcoin mine. Myself and Craig Tavares, we flew out to Fredericton in August of this year, actually, for a meeting with the Office of the Premier and NB Power. They were very excited about it. It will be the largest AI data center in the Maritimes. Stay tuned, but those are kind of the functional updates on it. We will provide guidance on the type of offtake we decide for New Brunswick and correspondingly, the financing structure to support that growth.
Thanks very much, and continue best of luck.
Thank you, Fedor. With that, we will conclude our Q&A session and our Q2 fiscal 2027 earnings call. Thank you to Frank Holmes, Aydin Kilic, and Darcy Daubaras for sharing their insight this morning. Thank you to our attendees and our analysts for joining. We look forward to speaking to you again soon.