vor 6 Stunden
Summary HIVE Digital Technologies has pivoted from a struggling Ethereum miner to a rapidly scaling Bitcoin and AI compute platform, now targeting $140M AI ARR by year-end. Q3 FY26 saw hashrate surge nearly 300% YoY to 25 EH/s, gross margins jump to 34.5%, and BUZZ AI revenue reach $4.9M, yet shares remain discounted. Despite a headline $93.1M GAAP net loss driven by accelerated depreciation and non-cash adjustments, core operating metrics and guidance signal strong operational progress. HIVE’s valuation lags peers, with AI and mining assets underappreciated; execution, dilution, and hardware cycle risks remain key factors for rerating. HIVE Digital Technologies Ltd. ( HIVE ) is one Bitcoin miner I have steered clear of over the years among Bitcoin mining companies. My view over the years has been that HIVE has been the smaller-scale miner. HIVE has faced a number of structural headwinds in its operating model, including exposure to geographic policy risk. An instance was when the miner closed shop in Norway in 2021 after the government removed the electricity tax subsidy for crypto miners (became effective March 2019), which materially increased its power costs. The timing was particularly jarring; while peers were aggressively acquiring capacity to capture bull market margins, HIVE was paying to exit a flagship 1 GW project that had become an impairment on the balance sheet. HIVE originally bought that Norway site in 2018 with dreams of building a 1-Gigawatt data center. They never broke ground on the massive scale they promised, and the project was written down from $15 million to $0 before the sale . HIVE started as a miner mainly focused on mining Ethereum ( ETH-USD ), but as events turned out, Ethereum moved to a Proof of Stake consensus in 2022 during The Merge, and HIVE had to pivot aggressively toward Bitcoin mining. The miner’s earlier GPU-heavy strategy became an earnings drag; thus, the risk-reward profile has historically been weaker than the larger, more liquid, and operationally robust mining peers. Riot Platforms ( RIOT ), Hut 8 ( HUT ), and MARA Holdings ( MARA ) are a few examples that come to mind instantly. HIVE released Q3 FY26 earnings two days ago. Highlights from the earnings include nearly 300% YoY growth in hashrate to reach ~25 EH/s, record revenue of $93.1 million, with the BUZZ AI segment contributing $4.9 million to total revenue, the huge GAAP net loss of $93.1, as well as strong guidance for the coming fiscal year. Gross operating margins jumped to 34.5%, which is over sixfold YoY improvement. Though data from the earnings will make up the core of this piece, as my initiating coverage on HIVE, I feel the need to not dive into the Q3 earnings directly but to contextualize the journey so far and where the Q3 numbers put things for HIVE. HIVE’s Pivot From Mining Lag-out to AI Alpha I believe HIVE’s early misstep can be overlooked at this stage, and earlier investors can let bygones be bygones. The company has really taken steps to strengthen its balance sheet, scale its mining fleet and capacity, reposition its infrastructure strategy, and move away from chasing the latest shiny object to protecting downside. Virtually every miner has diversified in recent years to protect against downside risk, especially during crypto market downturns. HPC and AI expansion has been the next natural evolution for Bitcoin miners. Anyone who follows Bitcoin miners even slightly can attest to the rerating mining stocks have seen from HPC and AI workload deals in the past year. HIVE has taken the same route and, in 2023, began repositioning parts of its business toward HPC and AI workloads. HIVE already built its GPU infrastructure during the ETH mining phase, and these could be repurposed for AI training and inference, as these are also GPU intensive, likely needing similar cooling and rack setups even if most of the GPUs should be out of spec for current workload needs as these depreciate faster than a crypto trader's optimism (Nvidia H100 and future Blackwell GPUs are the engines for the sites). That vision drove the rebrand from HIVE Blockchain to HIVE Digital Technologies. There has been a further rebranding of the AI segment to BUZZ AI Cloud. I'd say a forward-looking business structure sets this up for a potential spin-off in the future as a fully owned subsidiary, just as with Bit Digital ( BTBT ) and WhiteFiber ( WFYI ). While HIVE’s pivot and rebrand are encouraging, the question is: Is HIVE living up to its new name? While most of HIVE’s revenue still comes from Bitcoin mining, as the latest earnings for Q3 FY26 show, I believe HIVE is close to a breakout in AI-driven ARR. And beyond that sales scale, the quality of the sales is just as high. All these are being achieved through deliberate execution, which I believe HIVE is finally matching with operational delivery. Take, for instance, the $30 million, 2-year contract signed just last week for their first Nvidia B200 cluster, then the fully contracted $15 million ARR boost slated to hit the income statement as early as next month, which will add to a growing BUZZ AI segment that is targeting a massive $140 million ARR by year-end as they scale to 11,000 GPUs. Despite this near-term ARR guidance, HIVE has remained mostly muted. I think HIVE hasn't fully regained the trust of the market yet (since it's been years of missed milestones and slower-than-expected scaling); that's why investors remain cautious, and the stock still saw a ~5% drop following earnings on Tuesday. Because the guidance is strong, core metrics have shown improvement, and even the huge GAAP net loss was driven by $57.4 million in shortened depreciation schedules for certain assets from 48 months to 24 months, which front-loaded the depreciation figures. The GAAP loss was further driven by non-cash revaluation adjustments, where HIVE marks hardware or Bitcoin holdings to market. HIVE held 481 BTC on its balance sheet as of Q3 end. Based on the Q3 numbers and the market reaction, I think the market is still punishing HIVE for its past operational missteps, possibly giving us an opportunity to Buy its future at a discount. Progress and Performance of the HIVE AI Pivot Investors with a keen eye for operational detail know the importance of hashrate growth and energy arbitrage for a mining business and would see significant industrial scaling going on on the grounds at HIVE. In the past year, hashrate has grown by nearly 300% and now stands at 25 EH/s. While HIVE has been a dilutive engine by ballooning share count from ~130 million as of late 2024, which was Q3 FY25 when HIVE had a hashrate of just ~6.0 EH/s and a share count of roughly 130 million. Share count now stands at ~253 million, but hashrate has grown faster than the equity issuance while also self-funding AI CaPex. Meaning dilution has been accretive, and on a per-share basis, hashrate per share is now up about 114% compared to a year ago. For clarity, here's the math on the accretive hashrate expansion. As of Q3 FY25, HIVE’s 6 EH/s divided by 130M shares = 0.046 GH/s per share. Now Q3 FY26, hashrate of 25.0 EH/s divided by 253M shares = 0.098 GH/s per share. Implying that even though you own a smaller percentage of the company now because of dilution, each of your shares is backed by more than double the computing power it was a year ago. This accretion is a direct result of low energy costs the company has secured in Paraguay, where all of the ~25 EH/s capacity currently runs. HIVE’s management strikes me as pragmatic risk-takers unafraid to explore uncharted territories. I mentioned in the opening how the geographic exposure was one of the main risks in the early days and how that led to the closure of the Norway site; management has maintained a similar global playbook to date. The risk with this is that unlike miners who set up camp in the U.S. where frameworks surrounding miners curtailment and taxation are being codified and somewhat better known to investors, planting flags in unchartered jurisdictions comes with sovereign policy risks. HIVE’s 300 MW capacity in Paraguay draws directly from the second-largest hydroelectric plant in the world. HIVE is expanding its renewable footprint from 440 MW to 540 MW by Q3 2026. And will be locking in more hydro-powered capacity in Paraguay at sub-5 cent rates. In Sweden, HIVE has secured power rates at sub-2 cents per kWh, which are price levels that are nearly impossible to find in the continental US without heavy curtailment agreements. Is HIVE Being Valued Right? With all the numbers, growth, and guidance, I think the market is valuing HIVE with an eye of skepticism at a ~$530M market cap; the BUZZ segment, which is on track for $140 million ARR, seems to be left out entirely of the current valuation. Enterprise Value (EV) BTC Treasury Value at $67,000 per BTC Mining Infrastructure Value (minus BTC holdings) Current Hashrate Value per EH/s (Ex-Bitcoin) MARA Holdings $2.88 Billion $3.57 Billion -$690 Million 57.4 EH/s ($12.0M) Riot Platforms $5.49 Billion $1.21 Billion ~$4.28 Billion 45.0 EH/s $95.1 Million CleanSpark $2.42 Billion $905 Million ~$1.51 Billion 32.0 EH/s $47.1 Million HIVE Digital $522 Million $147 Million ~$375 Million 25.0 EH/s $15.0 Million The above table juxtaposes established miners who are still majorly anchored on Bitcoin production revenue only and have not fully pivoted to other revenue streams. While still generating mainly Bitcoin revenue, a peer like MARA has basically become a Bitcoin treasury play, making MARA’s valuation now dependent on its Bitcoin stash as the market seems to have discounted the mining business. So I'll be ignoring MARA for this valuation and focus on Riot and CleanSpark ( CLSK ) Based on the figures above, we can see that when the Bitcoin stash is taken out of the valuation, the core mining exahash is valued much lower for HIVE compared to the other miners. And this is even more glaring considering that HIVE already has AI revenue from the BUZZ unit. At $4.9 million as of Q3 FY26, that puts the BUZZ segment at a ~$20 million annual revenue run rate. If we strip that out, you find that HIVE’s mining infrastructure is even valued much lower, around $11 million to $12 million per EH. If HIVE’s total hashrate would be valued at a conservative $25 million per EH (which doesn't even match peers), only the core mining capacity would be worth over $600 million. Data by YCharts At $4.9 million revenue quarterly current and a ~$20 million ARR, the BUZZ business unit itself would be worth $200 million to $300 million on a standalone basis, considering the AI infrastructure and GPUaaS peers typically trade above 15x EV/sales multiple. Applied Digital ( APLD ) and CoreWeave Inc. ( CRWV ) are good examples. That already makes up around 50% of HIVE’s current market value. And the fact that the ARR is projected to scale to $35 million in the very near term when the B200s are deployed next month, and a $140 million ARR once the 11,000 GPU fleet is fully deployed by year-end, a rerating is on the horizon in the near term. Risks While the valuation disconnect here seems huge, I believe it exists for a reason. To buy HIVE here, an investor has to be comfortable with these specific dilution and execution risks. Firstly, HIVE still relies on its ATM equity program to fuel the AI infrastructure build-outs. Even with record revenue, HIVE still filed a prospectus supplement late last year for a new funding round, and that $300 million funding round is still active. If HIVE over-dilutes before the $140M AI ARR is realized, undervalued shares today will be worth a smaller slice of a bigger but potentially slower growth business. The AI business is a high-stakes arms race. I've mentioned how 48-month depreciation was shortened to 24 months and contributed to the huge GAAP net loss in the Q3 results. HIVE is currently betting big on the Nvidia B200 cluster it is deploying next month. If spot prices for AI compute drop due to a GPU glut or if newer chips (like the Blackwell Ultra) make the B200s look like legacy hardware too quickly, the 24-month depreciation will move from a strategic choice to a financial necessity. Takeaways Based on the latest results and numbers, the HIVE we know today is no longer the small, GPU-heavy Ethereum miner the market remembers. HIVE has scaled to an impressive Bitcoin hashrate, secured low-cost power, and is now building out an AI compute platform that could shift its revenue mix over the next couple of months. The opportunity HIVE presents is clear. If AI ARR ramps toward the $140 million target and mining expansion continues without runaway dilution, the stock’s mispricing relative to its industrial footprint and revenue stream will become very glaring in a way that the market would likely not overlook anymore. The risk is equally clear. Execution must be tight. Depreciation risks, hardware cycles, GPU pricing, and equity issuance all matter and must be well-timed. A misstep in AI deployment or over-dilution before revenue matures will punish shareholders. For Q3, HIVE is a victim of its own accounting honesty. By aggressively writing down its fleet over 24 months, it handed the bears a scary $91 million net loss headline. Yet, under the hood, the company seems to be doing just fine, generating $32.1 million in gross operating margin and $5.7 million in adjusted EBITDA. I believe HIVE’s re-rating will not come from just narratives like IREN ( IREN ) and TeraWulf ( WULF ) have enjoyed. It will come from sustained AI revenue growth and proof that dilution creates ongoing per-share value. Until then, this discount exists for a reason.