HIVE Digital Technologies Ltd. (HIVE) Earnings

Management highlights

### Strategic Transition and Dual-Engine Model - HIVE is a vertically integrated data center builder/operator, shifting from primarily Bitcoin mining to a dual-engine model where Bitcoin mining cash flow funds long-term growth of HPC/AI infrastructure, which is expected to eventually command higher market multiples than mining. - The company holds a global 860 megawatt total footprint, with 440 megawatts of active capacity currently consuming 395 megawatts after efficiency upgrades. ### Capital Markets Milestones - Closed a $115 million 0% coupon exchangeable senior note offering due 2031, oversubscribed with $500 million in total demand from 24 institutional buyers, introducing the company to a broader base of U.S. institutional investors. - The notes include a capped call structure that eliminates dilution up to a $4.92 share price, and delivered a net $105 million in proceeds after the $19.8 million cost of the capped call, providing low-cost funding for GPU expansion. - Completed a listing uplist to the Toronto Stock Exchange (Canada's main board), which generated significant institutional investor interest. ### AI Infrastructure Expansion - Announced the 100,000 GPU Gigafactory project in the Greater Toronto Area (GTA), a 320 megawatt Tier 3 data center on 25 acres of already acquired land (for $58 million), with 90% renewable power allocated and a planned closing-loop 0-water-use design. The project has a total projected CapEx of CAD 3.5 billion, is expected to be energized by the end of 2027 and go live in early 2028, with a projected $360 million in annual colocation ARR. - Hold an exclusive strategic partnership with Bell Canada to scale GPU cloud capacity across Bell's pre-built Tier 3 data centers in Canada, enabling a low-CapEx, fast time-to-market ramp. Bell provides colocation at 20% below market rates and shares a 5% revenue share for enterprise clients it refers, creating a large demand funnel for sovereign Canadian AI compute. - Signed MOUs for two large GPU clusters: 2,304 GB200s and 2,080 GB300s, which will add $70 million and $65 million in ARR respectively, bringing the total GPU cloud ARR to ~$170 million when deployed, on track to hit the $200 million full-year 2026 ARR target. - The total combined HPC/AI target ARR across GPU cloud and colocation is $660 million ($200 million GPU + $460 million colocation). ### Global Footprint Development - In Paraguay, HIVE is the largest electricity consumer, accessing surplus zero-cost renewable power from the Itaipu Dam, with a current 300 megawatt footprint and a long-term target of 1 gigawatt. The company has expanded its land bank near Yguazú substation for future growth, and completed a successful AI research initiative with Columbia University that validated latency and bandwidth for North American clients. - In Sweden, HIVE holds 40 megawatts of available capacity for future HPC expansion.

Guidance

- Total HPC/AI annualized revenue (ARR) target is maintained at $660 million, split between $200 million for GPU cloud and $460 million for HPC colocation, to be achieved through 2028 as sites come online. - Management maintains that the company is on track to deploy 11,000 total GPUs by the end of calendar 2026, hitting the $200 million GPU cloud ARR target by the end of the calendar year. - The Gigafactory project guidance is maintained: energization by the end of 2027, with commercial operation and cash flow starting in early 2028, delivering $360 million in annual colocation ARR. - No new CapEx is allocated to Bitcoin mining for the coming 12 months; all growth CapEx is dedicated to HPC/AI expansion. The Bitcoin segment is only being optimized to maintain low breakeven costs and steady cash flow. - Management's long-term valuation guidance sees HIVE reaching a $5 billion+ U.S. market capitalization as HPC/AI targets are achieved, based on a sum-of-the-parts valuation using peer-multiple ranges for colocation and GPU cloud.

Segment performance

HIVE operates two core business segments under a dual-engine strategy: Bitcoin Mining and High Performance Computing (HPC/AI BUZZ). 1. **Bitcoin Mining Segment**: For full fiscal 2026, the segment was the company's largest revenue contributor, generating approximately $275 million of the total $300 million full-year revenue (≈91.7% of total revenue). It produced 876 Bitcoin in Q4 2026, with 25 exahash installed capacity and 23 exahash average operational capacity during the quarter. After strategic efficiency upgrades, the optimized hash rate reached 24.6 exahash with a breakeven improvement to 16 joules per terahash. The segment generated positive cash flow and margins even during periods of Bitcoin price volatility, with projected daily profit of ~$300,000 at a $70,000 Bitcoin price, ~$400,000 at $80,000, and ~$500,000 at $90,000 (based on $0.05 per kWh electricity cost). 2. **HPC/AI (BUZZ) Segment**: For full fiscal 2026, the segment generated approximately $25 million of total revenue (≈8.3% of total revenue). In Q4 2026, the segment generated $4.6 million in revenue, with annualized revenue reaching $35 million as of quarter-end. The first 504 Blackwell B200 GPUs cluster in Winnipeg (Bell Canada site) is live, under a 2-year $30 million contract that covers the full capital cost of the cluster. As of June 2026, the segment has 5,500 operational GPUs globally, with contracted total annualized revenue (ARR) of $35 million.

Risks & headwinds

- High inherent volatility: HIVE's stock price is significantly more volatile than broader markets, Bitcoin, and even other AI/blockchain infrastructure firms, which can create short-term price dislocations. - Global supply chain constraints: Delivery lead times for critical HVAC and electrical infrastructure for data centers have extended from 20 weeks to 50-60 weeks, which could delay project timelines. - Power and infrastructure permitting risk: Large-scale data center projects require multi-year regulatory, grid study, and civic permitting processes, which can create delays even after power allocation is secured. - AI market bubble risk: Prominent short sellers have publicly called the AI infrastructure boom a bubble, which has created short-term market sell-offs that can impact share price. - Bitcoin price volatility: The company's near-term cash flow is heavily dependent on Bitcoin prices, and sustained downward price movement could reduce the funding available for AI expansion.

Analyst Q&A

• Q: Can you elaborate on power security and CapEx funding for the 320 megawatt GTA Gigafactory, and share updates on Paraguay's AI expansion? /

  A: Full 320 megawatt power allocation is already secured for the Gigafactory after more than a year of negotiations with provincial regulators, with over 90% expected to come from renewable sources. The project will follow standard large data center funding models, with long-term leases to offtakers enabling bond funding similar to peer projects. For Paraguay, the Columbia University AI research trial has been completed successfully, with its work submitted to the prestigious ICML conference, validating latency and bandwidth between Paraguay and New York. HIVE has expanded its land bank near the Yguazú substation for future AI expansion, with industry peers already expressing interest in the region.

• Q: What is the expected CapEx for BUZZ AI over the next 12 months, and are there plans to deemphasize Bitcoin mining? /

  A: The two large GPU clusters under MOU are approximately $170 million each, with 20% down payments ($~34 million per cluster) required for 80% LTV third-party financing, which aligns with the $115 million in proceeds from the recent convertible note offering. An additional 1,500 GPUs to fill out the Winnipeg site will cost ~$90 million, also eligible for third-party financing. No CapEx is allocated to Bitcoin mining; the segment will remain a core cash flow generator, with ongoing efficiency upgrades to maintain low breakeven costs and steady cash flow to fund AI expansion, so there are no plans to sell or deemphasize the business.

• Q: Is the Bell Canada data center infrastructure ready for upcoming GPU deployments, and what gross margins can be expected for the GPU cloud business? /

  A: The Winnipeg site is already live with the initial 504 GPUs, and the British Columbia Merit site was ready for occupancy as of April 1, 2026. The first large new cluster can be deployed immediately at the Merit site, with the second cluster coming online later this year, aligning perfectly with the deployment timeline laid out in the presentation. The partnership with Bell allows fast, low-CapEx scaling while HIVE completes longer-term buildout of its own sites. With the below-market colocation rate negotiated with Bell, the expected EBITDA gross margin for the GPU cloud business is over 75%.

• Q: Have you started marketing the Gigafactory to potential tenants, and what types of tenants are you targeting? /

  A: The recent Gigafactory announcement only disclosed the secured land, power allocation, and core project details. Specific discussions with tenants and disclosure of targeted tenant types will come in future updates. Directionally, the site is well positioned for both Canadian enterprise and government sovereign compute tenants, and HIVE's partnership with Bell Canada provides access to Bell's large existing network of Canadian enterprise clients that are seeking sovereign AI capacity.