HIVE Digital Technologies Ltd. Q2 FY2025 earnings call

Frank Holmes provided a macro overview, highlighting the impact of President Trump's win on the crypto market, global debt issues, and Bitcoin's performance. Aydin Kilic detailed the business operations, with revenue from the Bitcoin mining unit and HPC unit. HIVE is a pioneer in green energy Bitcoin mining, operating in multiple time zones and languages. The company has plans for a 100-megawatt expansion in Paraguay using green hydroelectricity, has upgraded ASICs to enhance efficiency, and focuses on conservative capital allocation and operational excellence with low general and administrative (G&A) expenses. It also discussed its AI strategy with HPC revenue and future expansion plans.

For the quarter ended September 30, 2024, HIVE Digital Technologies generated total revenue of $22.6 million. The Bitcoin mining business unit contributed $20.8 million, which is approximately 92% of the total revenue, and the HPC business unit contributed $1.9 million, accounting for about 8% of the total revenue. The company achieved a gross operating margin of $1.1 million in the quarter and an adjusted EBITDA of $5.6 million.

The company aims to increase exahash and double its global footprint. By summer 2025, it targets 12.5 exahash with improved efficiency. With Bitcoin price rallies, annualized mining margin is expected to grow. For example, by calendar Q1, the annualized mining margin will grow by $70 million, by calendar Q2 by $95 million, and when Paraguay is complete by summer 2025, it will grow by $150 million. HPC revenue is expected to grow, with annualized rates around $9 million to $10 million initially and potential for larger contracts in the future.

Risks include regulatory uncertainties due to the SEC's anti-crypto policies affecting capital formation. Bitcoin price volatility, changes in mining difficulty, and potential market fluctuations in demand for HPC services are also risks.

Q: Congrats on breaking ground in Paraguay. For my first question, I was hoping you could share an updated capital allocation strategy as it pertains to HPC initiatives. And curious to hear how that's weighing on the back of management's mind on -- with the recent rise in mining economics?

A: Hey Bill, it's Aydin here. Thanks for that question. So we've been focusing on the infrastructure side of the HPC business earlier this year, a couple of months ago, we noted a pipeline of conversion of existing crypto mining assets, data centers that we own. And so we've kept it high level, and we've been focusing on cash flow. So right now, we are working on a few things that we haven't specifically commented on yet. So that's about as much as I could say right now. But the economics for both business units are both very attractive. And that's -- as far as HPC, that's sort of the extent of what I could say right now. And just stay tuned for more announcements. But we are gearing up for next-generation compute as the data center requirements to power NVIDIA's next generation of GPUs will be a cut above what exists today on the market.

Q: And just as a follow-up to that, can you offer more color in terms of what the outlook could be for gross margins on the HPC side? Has your targeted margin profile changed at all?

A: No. So I think what's going to happen is the latest hardware -- like everybody knows that the H200 is out, and that's the next version of the H100. It has twice the amount of vRAM. So really, we see the indicative pricing for H100s and H200s in the $2 to $2.50 per hour range. And so when Blackwell comes out next year, how will that affect the existing H100, H200 pricing and what will Blackwell rent out at. And so if you look back to when the H100 came out and the A100, the A100 had really good staying power even to this day. And so if you think of the Melting Ice Cube analogy of dollar per GPU per hour pricing, it decreases a lot slower than in crypto mining. So I think for the most part, you could keep your margins as an analyst similar as new gear comes online. If you wanted to be more bullish, you could project perhaps upwards of $3 an hour for Blackwell, but that's just narrative and commentary. I think because Blackwell has substantively better performance than the H200, a lot quicker flop speed, even more memory, et cetera. So it will -- I think the market will determine what they'll pay as a premium for that improved compute performance. I hope that's helpful.

Q: Your goal kind of being 2% of the Bitcoin mining network, is that like a firm percentage? I guess I'm looking at your target of 12.5%, that's less than the 2% of the current network hash and I would assume network hash will be a bit higher this time next year. So I guess how should we think of your long-term targets as a percentage of total network hash?

A: Hey Brett, yeah, it's Aydin here. So that's a great question. I'm a very mathematical guy, and I could open Excel sheets and we could look at the time variant nature of Bitcoin mining network. But what we really try to do and what I've trained myself to do is just to provide simple run rate metrics that are easy to walk away. I'm very much aware that 2% could represent much more than 12.5 exahash, but 2% is our target. And so what we've got secured, construction has broken ground, ASICs allocated, etcetera, we have a clear path to 12.5 exahash. So that's what we are putting in the presentation and in our projections. So for all the analysts, it's really easy for them to have concrete numbers of where we'll be. But of course, there's always more opportunity to expand beyond the 12. And our target really is to be at 2% of the network, right? So yes, I'm aware that, that would constitute more than 12.5, and we certainly have plans to grow beyond the 12.5. 12.5 is a concrete number. Again, the site has broken ground. We're building 100 megawatts. We've just ordered 6,500 Canaan to upgrade our existing fleet. So I hope that answers your question.