HIVE Digital Technologies shifts capacity from bitcoin mining to AI data centers
Context and chronology
HIVE Digital Technologies has begun an operational pivot: management told markets it will reduce ASIC hashrate at the Boden, Sweden site and redeploy that power and cooling footprint to support liquid-cooled GPU clusters and other high-performance computing (HPC) deployments. The company is advancing a Boden upgrade to Tier‑III HPC standards to host systems built around NVIDIA GB300‑class hardware while accelerating North American capacity through its BUZZ High Performance Computing unit and a commercial partnership with Bell Canada AI Fabric.
Execution plans in Canada are explicit and measurable: HIVE intends to move its local capacity from roughly 4 MW to an aggregated 16.6 MW, including an immediate 5 MW build in British Columbia and an optional expansion of 7.6 MW. That footprint underpins a near-term deployment above 4,000 GPUs with a defined growth path beyond 6,000 GPUs and a targeted contracted annualized revenue run‑rate of approximately $200 million by the fiscal year ending March 31, 2027.
Operationally and commercially this is consistent with a wider sector trend: several public miners and asset owners are converting grid‑connected, power‑dense campuses into colocation and AI compute sites. Comparable industry moves range from Core Scientific’s monetization of bitcoin treasuries to fund conversions, to Hut 8’s hyperscaler‑anchored, multi‑hundred‑megawatt commitments, to MARA and TeraWulf’s partnerships that seek to accelerate tenant sourcing and de‑risk buildouts with developer or hyperscaler partners. Those peers show divergent tactical approaches — treasury monetization, long‑term hyperscaler anchors, or developer JV models — which imply distinct timelines, liquidity profiles and revenue‑risk allocations relative to HIVE’s revenue‑from‑colo push.
Market and technical constraints are material: the sector faces accelerator (GPU) procurement backlogs, OEM channel constraints, packaging and test throughput limits, and local permitting and interconnection windows that can extend conversion timelines from months into quarters. Complementary market signals show Bitcoin network hashrate easing below ~1,000 EH/s and mining difficulty softening (reported declines in recent weeks from roughly 156T toward ~146.5T), which moderates immediate mining economics and helps justify redeploying stranded power into higher‑margin HPC workloads.
Financial signaling was immediate: HIVE’s equity showed a single‑session uptick (intraday movement reported near ~5.6%) on the announcement, reflecting investor recognition of the strategic reorientation. Yet public examples from peers illustrate the tradeoffs: Core Scientific reported near‑term revenue and EPS pressure as hashing rigs were retired and sites retrofitted despite monetizations that bolstered liquidity; Hut 8’s hyperscaler‑backed commitments reduce revenue risk but do not remove construction, interconnection and GPU delivery risk.
Implications for local power and colo markets are asymmetric. If HIVE materially repurposes Boden capacity, northern Sweden could see freed firmed megawatts and potential short‑term easing in local wholesale power tightness for miners; conversely, Canadian corridors where HIVE adds multi‑megawatt liquid‑cooled capacity may tighten colocation availability and widen scarcity premiums for AI‑optimized racks, especially where firmed, low‑carbon energy and fiber interconnects are scarce.
The technical realities are blunt: transforming ASIC campuses to host thousands of GPUs requires more than swapping compute — it demands upgraded power distribution, rack‑level PDUs sized for high per‑rack wattages, liquid‑cooling topology, PCIe and Ethernet/InfiniBand fabrics, and procurement of accelerator inventories that remain constrained. These factors, together with customer onboarding cadence for colocations, will determine when announced capacity converts into contracted revenue and whether the $200M run‑rate target is met on schedule.
Overall, HIVE’s announcement is a concrete example of a broader capital‑intensive reallocation across the sector: it monetizes under‑used, energy‑dense facilities into higher‑margin HPC services while inheriting the schedule, equipment and regulatory risks that have defined other miners’ transitions.
Read Our Expert Analysis
Create an account or login for free to unlock our expert analysis and key takeaways for this development.
By continuing, you agree to receive marketing communications and our weekly newsletter. You can opt-out at any time.
Recommended for you
Big Tech’s AI Spending Supercharges Bitcoin Miners’ Pivot to Cloud and HPC
Aggressive AI procurement by Meta, Microsoft and other hyperscalers is expanding demand for dense compute beyond traditional data centers, creating a fast-growing commercial outlet for bitcoin miners that retooled sites for GPUs and HPC. Early megawatt-scale contracts (including a reported 300 MW deal) and visible company-level moves — set against a backdrop of falling bitcoin hashrate and ongoing chip and permitting constraints — validate the strategy but leave miners exposed to accelerator supply, local permitting, and power-delivery risks.
Nebius boosts GPU and data‑center spending to lock in AI capacity
Nebius sharply increased quarterly capital spending to buy AI processors and expand its global data‑center footprint, pushing secured electrical capacity above 2 GW and raising its year‑end target to more than 3 GW. The build‑out — including a planned 240 MW, GPU‑dense campus in Béthune, France — widens near‑term losses but is aimed at underpinning a multibillion‑dollar annualized revenue run‑rate by the end of 2026.
MARA partners with Starwood to repurpose mining sites into AI data centers
MARA:US is converting mining campuses into cloud and AI-ready data centers, launching with an immediate 1 GW of capacity and upside to 2.5 GW . The announcement sent MARA stock up roughly 15% in after-hours trading and signals a new pathway to monetize stranded energy assets.
Bitdeer liquidates Bitcoin treasury, unveils $300M convertible debt to fund AI and data‑center push
Bitdeer sold its corporate Bitcoin stash, reducing its treasury to 0 BTC after liquidating 1,132.9 BTC . The miner simultaneously filed a $300M convertible debt package (plus a $45M option), sparking a steep share selloff and signaling a reallocation toward AI and data‑center capacity; peers have pursued varied, sometimes less‑aggressive paths (e.g., Cango converted 4,451 BTC to USDT while preserving mining and repurposing campuses for modular GPU clusters).
Hut 8 Accelerates AI Data‑Center Pivot with $7B Google‑Backed Lease
Hut 8 reported a hefty FY2025 loss driven by digital‑asset writedowns while signing a 15‑year, $7B agreement for 245 MW of AI IT capacity underwritten by Google — a deal that shifts the company from spot crypto exposure to contracted AI hosting. The transaction sits alongside broader market moves (private‑credit for greenfield builds, hyperscaler strategic stakes, and miners repurposing grid sites) and highlights divergent financing and execution risk profiles across the emerging AI‑compute supply chain.
Global AI datacenter boom risks oversupply and wasted capacity
Rapid expansion of GPU‑heavy datacenter capacity for generative AI is outpacing measurable production demand and colliding with local permitting, financing and grid constraints. Absent tighter demand validation, better utilization mechanisms and coordinated grid planning, the sector faces lower returns, schedule risk and heightened public pushback.
US Communities Push Back on Power-Hungry AI Hubs, Echoing Bitcoin Mining Conflicts
Communities in multiple US states are increasingly resisting large, power-intensive AI data center projects, raising questions about long-term grid strain and local costs. Industry tracking shows roughly $64 billion in US data center developments have been delayed or blocked, prompting tech firms to adopt new cost-sharing and community engagement tactics.
Australian AI infrastructure firm wins $10B financing to accelerate data‑center buildout
Firmus Technologies closed a $10 billion private‑credit facility led by Blackstone‑backed vehicles and Coatue to underwrite a rapid roll‑out of AI‑optimized campuses in Australia. The debt package targets deployment of Nvidia accelerators and up to 1.6 gigawatts of aggregate IT power by 2028, embedding the project in a wider global wave of specialized, high‑power data‑center financing.