HIVE Digital Technologies Ltd. (HIVE) Earnings

Management highlights

### Strategic Positioning & Dual Engine Growth - Pursuing a dual engine strategy combining profitable Bitcoin mining and fast-growing HPC/AI cloud and co-location businesses, with a goal to rebalance revenue to 60% Bitcoin / 40% HPC by Q4 2026 - Operates across 9 time zones in three continents (North America, Europe, South America) with a focus on low-cost green energy locations (Canada, Sweden, Paraguay) - Raised a total of $245 million in zero-coupon convertible notes this quarter, with capped calls to minimize shareholder dilution, to fund GPU expansion - Ended the quarter with $208 million in cash, a substantial increase from $23 million at the end of Q4 FY2026, providing strong liquidity for growth ### HPC/AI Business Milestones - Signed $600 million in total GPU cloud contract value TCV so far this calendar year, consisting of long-term 3–5 year contracts with investment-grade clients - Has ~5,500 GPUs online today, growing to ~9,800 contracted/active GPUs after Q4 2026 deployments - Announced a letter of intent (LOI) for a 45 megawatt HPC co-location (COLO) lease at the legacy Bowdoin, Sweden site, which would bring total contracted HPC ARR to $225 million if finalized - Completed a successful proof of concept transmitting compute power from Paraguay to New York City over 5,000 miles, demonstrating the viability of exporting low-cost hydropower as global compute - The 200-megawatt substation in Iwazu, Paraguay is under construction, with major hardware already delivered and civil work progressing, expected to be energized by the end of calendar 2026 ### Financial Operating Performance - Total Q1 revenue: $79.1 million, up 10% quarter-over-quarter and 73% year-over-year - Gross operating margin: $24.2 million, up 53% year-over-year to 31% of revenue - Adjusted EBITDA: $13.4 million positive, up from negative $9 million in Q4 FY2026 - Reported GAAP net loss of $142.9 million is driven entirely by non-cash items: $53.7 million in depreciation and an $84.7 million provision for the Swedish tax dispute; underlying operating cash flow is positive - Daily operating cash profit of ~$160,000, with total daily revenue of ~$850,000 and total daily operating costs of ~$690,000

Guidance

- Expects to hit the original full-year 2026 target of $200 million ARR for GPU cloud business ahead of schedule, having already reached $180 million in contracted ARR by mid Q2 FY2027 - Expects the 45 MW Bowdoin co-location LOI to be finalized into a definitive agreement by the end of September 2026, which would push total contracted HPC ARR to $225 million, exceeding the original full-year target - Expects the new GB300 GPU cluster to be delivered and deployed at the Bell Canada Merit Facility in Q4 2026, bringing the HPC segment to 40% of total company revenue by the end of 2026 - Two-year base target of $700 million total ARR for combined GPU cloud and HPC co-location: $200 million GPU cloud ARR + $500 million HPC co-location ARR across the company's development pipeline - The crown jewel Greater Toronto Area Gigafactory project is expected to be energized by the end of calendar 2027, with compute coming online in early 2028 - Expects the 75-80% EBITDA margin range for the new five-year GB300 contract, with full payback of GPU capital expenditures in approximately 3 years, leaving two years of free cash flow after payback

Segment performance

1. Bitcoin Mining Segment: Generated $72.1 million in revenue for the quarter, representing 90% of total company revenue. Revenue grew 7% quarter-over-quarter and 77% year-over-year, driven by full-scale operations at the Paraguay facility. The company holds 24 exahash of operational hashing power (25.3 exahash installed), generating approximately $750,000 in daily revenue and $275 million in annualized revenue (ARR) as of the quarter end. 2. High Performance Compute (HPC) / AI Segment: Generated $7.1 million in revenue for the quarter, representing 10% of total company revenue. This translates to a $28 million annualized run rate exiting Q1, and $35 million ARR as of the call date (mid August). The segment has $180 million in total active and contracted ARR for GPU cloud business, including a newly signed five-year $350 million total contract value (TCV) deal for 2088 GB300 GPUs to be deployed in Q4 2026.

Risks & headwinds

- Ongoing tax dispute with Swedish regulators: the company has recorded an $84.7 million non-cash provision for a potential VAT liability following an adverse court of appeal ruling, and the company continues to pursue all available legal avenues to contest the ruling - High inherent volatility in Hive's stock price, which is 3x more volatile than Bitcoin and 4x more volatile than Nvidia, driven by macro sentiment shifts in both Bitcoin and AI sectors - Contagion risk from the unwinding of the Japanese carry trade: the recent sell-off in AI and crypto equities was driven by this deleveraging, and further rising interest rates in Japan could create continued downward pressure on global risk assets - Bitcoin mining sector headwinds: the company has faced declining Bitcoin prices and increasing network difficulty, though it has maintained profitability via scale economies and low-cost power in Paraguay - Execution risk for large data center development projects: substation construction, permitting, and financing for new greenfield sites can experience delays that push back revenue generation

Analyst Q&A

• Q: What are the expected returns and financing for the new 5-year GB300 GPU deal, and what is the next milestone for the Paraguay power buildout?

  A: Total CapEx for the GB300 cluster is $185 million, with expected 75-80% EBITDA margins. The investment will pay for itself in approximately 3 years, leaving 2 years of free cash flow for a ~1.6x total return before owning the GPUs outright. The $245 million in convertible note proceeds raised this quarter covers the required down payment, with vendor financing covering the rest. The Paraguay substation is expected to be energized by the end of 2026, with regular public updates on construction progress.

• Q: How close is the Bowdoin co-location definitive agreement, and what are current demand and pricing trends for GPU and co-location services?

  A: Negotiations are well advanced, with a draft agreement already exchanged, and the company expects to announce a definitive deal in September. Demand is very strong across all segments of the market: frontier labs want the latest cutting-edge hardware, while other users prioritize lower-cost older generation GPUs for inference. Pricing has been rising, with recent rental rates far above levels seen 6-9 months prior, and demand exceeds the company's current 200 million ARR target, with room for substantial further growth.

• Q: How does management balance HPC co-location opportunities versus owned GPU cluster opportunities across the company's power portfolio?

  A: The company plans to deploy 10,500 GPUs across existing Bell Canada partnership facilities first, after which it has ~400 megawatts of available utility load in Canada for further development. The company will evaluate the highest return option for each site: GPU clusters have very strong near-term demand with attractive economics, while large co-location leases can generate very large long-term contracted revenue. The company continues to source new land and power to meet sustained unmet demand for AI compute capacity.

• Q: What is the timing and funding plan for the GTA Gigafactory project, and will a decision on cloud vs co-location need to be made upfront?

  A: The project has a completed basis of design, and is still on track for energization at the end of 2027 with compute going live in early 2028. The company will likely pursue pre-signed off-take agreements before securing construction financing (most likely via investment grade corporate bonds). The project can support a mix of both co-location and owned GPU cloud capacity, and a final mix decision does not need to be made during initial groundbreaking.