Bitcoin mining firm HIVE Digital Technologies is raising $75 million through a private placement of 0% exchangeable senior notes due in 2031. The capital will be used to fund a strategic pivot from pure-play Bitcoin mining toward artificial intelligence (AI) cloud services and high-performance computing (HPC).
The notes will be issued by the company's wholly owned subsidiary, HIVE Bermuda 2026 Ltd., to qualified institutional investors. The deal includes a 13-day option for an additional $15 million, potentially bringing the total offering to $90 million. HIVE stated the net proceeds will be allocated for general corporate purposes and capital investment, specifically for purchasing graphics processing units (GPUs) and expanding its data center infrastructure.
To protect existing shareholders from potential dilution if the notes are exchanged for equity, HIVE "intends to fund capped call transactions using cash on hand." Part of the offering's proceeds may also be used to reimburse these hedging costs.
Concurrently, HIVE has received conditional approval to uplist its common shares from the TSX Venture Exchange to the Toronto Stock Exchange (TSX). The transition is expected around April 30, 2026, pending the fulfillment of standard listing requirements by June 30, 2026. The company's shares closed at $2.47 on Nasdaq on the day of the announcement.
This financing move follows what HIVE called a record fiscal third quarter, ended December 31, 2025, where it reported revenue of $93.1 million—a 219% year-over-year increase. However, the company also posted a net loss of $91.3 million, driven by accelerated depreciation from its Paraguay expansion and non-cash revaluation adjustments.
The capital raise accelerates HIVE's ongoing business transformation. In March, the company announced plans to progressively phase down ASIC-based Bitcoin mining at its Boden facility in Sweden, converting it into a Tier-III HPC data center. It has already launched its first GPU cluster in Asunción, Paraguay, where its BUZZ AI Cloud platform is processing large language model training workloads.
