Hive Digital Technologies’ Buzz High Performance Computing business has secured customer agreements valued at up to $30 million.

The contracts span two-year fixed terms and are for AI cloud capacity operated by the company.

The deals are related to Buzz’s upcoming GPU deployment at its Canada West location in Manitoba. The first phase of capacity is expected to go online sometime this quarter of 2026, and will consist of 504 liquid-cooled Dell server-based GPUs.

The first phase is expected to generate $15m in annual recurring revenue once operational. Once Buzz has fully deployed its cluster, this is expected to increase again.

Frank Holmes, executive chairman of Hive, said: “We are entering 2026 with strong momentum in our HPC and GPU cloud business. Hive has built a track record as one of the longest-standing publicly traded crypto Tier-I data center operators, performing through multiple market cycles while protecting cash flow and balance sheet strength. Now, with Buzz, we are leveraging that foundation to build a high-growth AI cloud platform spanning Canada, Sweden, and Paraguay.”

Holmes added that the company is using Tier-III standard facilities for its GPU clusters, and rather than constructing its own facilities, Buzz is partnering with other data center providers.

“Where Hive owns land and buildings and operates its Tier-I facilities, we are pursuing selective Tier-III conversions and colocation strategies for HPC. This showcases our vertically integrated model and diversified revenue streams from both HPC colocation and GPU AI cloud services, reinforcing Hive’s dual-engine strategy of hashrate services and high-performance computing.”

Hive’s CEO, Aydin Kilic, added that the company is aiming to achieve $140 million in ARR over the next year, and ultimately plans to deploy 2,000 GPUs at the Canada West site.

Buzz HPC and Bell Canada previously revealed they would partner on Nvidia-based AI infrastructure in Canada, including a planned 5MW deployment at a Bell Canada data center in Manitoba.

Hive was founded in 2017 and originally functioned as a Bitcoin mining company. It has since expanded into the AI cloud compute space and is building and operating data centers for both purposes across Canada, Sweden, Paraguay, and the US. It is anticipated to have 540MW of capacity by the end of the year, including 400MW in Paraguay and 140MW in Canada and Sweden.

According to a previous release, the company aimed to have more than 11,000 GPUs running by the end of 2026.