Hive's Buzz, Dell enter deal for AI infrastructure

Executive Summary

• Hive Digital Technologies’ subsidiary, Buzz High Performance Computing, has signed an agreement with Dell Technologies to deploy a 63-node cluster of liquid-cooled servers for its AI cloud expansion.
• Buzz has executed operating agreements for the first phase of deployment within the Bell AI fabric data centre, building on existing partnerships to expand its sovereign AI cloud capabilities.
• The company projects significant growth in GPU capacity and revenue, targeting over 11,000 total GPUs and $120 million in additional annual run-rate revenue by the end of 2026.

Key Details

• Dell Agreement: Buzz signed an agreement to deploy a 63-node cluster of liquid-cooled Dell PowerEdge XE9680L servers.
• Hardware Specifications: The deployment includes 504 latest-generation GPUs optimized for AI compute, delivered through Dell integrated rack scalable systems (IR5000 racks).
• Deployment Location: Delivery and installation will occur at the Bell AI fabric data centre.
• Bell Partnership: Operating agreements have been executed for the commencement of the first phase of the Buzz and Bell AI fabric deployment.
• Rankings: Buzz recently achieved bronze status in the semi-analysis ClusterMax rankings, surpassing publicly traded peers in the AI cloud sector.
• Future Capacity (2026):
  • Buzz is procuring long-lead components for sites in Canada and Sweden, with go-live targeted for the second half of calendar 2026.
  • Each Canadian and Swedish site, alongside the Bell AI fabric facility, will support 2,000 additional latest-generation GPUs.
  • Total target fleet: Over 6,000 latest-generation GPUs for AI cloud by the end of 2026.
• Financial Projections:
  • The projected expansion is expected to generate an additional $120 million in annual run-rate revenue.
  • Operating margin is estimated at 80% after electrical and data centre costs.
  • This increase is in addition to the company's current $20 million Annual Recurring Revenue (ARR).
• Total GPU Count: Combined with over 5,000 GPUs already in operation, the Buzz AI cloud will reach a total of over 11,000 GPUs by the end of 2026.
• Grand Falls Facility: The Hive 70-megawatt Tier I data centre in Grand Falls, N.B., sits on 32.5 acres of land and is positioned for upgrade to a Tier 3+ hyperscaler co-location facility, capable of supporting over 25,000 latest-generation GPUs based on reference architecture.

Notable Quotes

• No direct quotes from executives were included in the provided text.