HIVE Digital Technologies Reaches AI Cloud Milestone in Paraguay, Powers Columbia University LLM Research from New York to Asuncion

Executive Summary

• HIVE Digital Technologies announced that its BUZZ AI Cloud platform in Asunción, Paraguay is now live, with Columbia University researchers running GPU‑accelerated LLM training workloads.
• The deployment marks the first Tier‑III GPU cluster under HIVE’s phased strategy to layer AI/HPC infrastructure onto its 300 MW renewable hydroelectric power base in Paraguay and serves as a proof‑of‑concept for future expansion to Yguazú by 2027.
• Management highlighted that the live workload will generate performance data (latency, throughput, cost efficiency) to guide scaling of commercial HPC capacity, supporting HIVE’s roadmap for large‑scale AI compute services in South America.

Key Details

• Location & Facility: BUZZ AI Cloud GPU cluster hosted in a Tier‑III data center operated by Paraguay’s largest telecommunications provider, powered by hydroelectric generation and connected via nationwide fiber backbone.
• First Live Workload: Columbia University research team (New York) conducting LLM pre‑training on models ranging from 0.2 B to >8 B parameters using multi‑GPU distributed training frameworks.
• Performance Claim: Early experiments with the “Muon” algorithm showed ~1.3× greater efficiency versus standard baselines, providing early validation of GPU compute performance in the region.
• Strategic Purpose: Generates real‑world data (latency, throughput, workload management) to inform HIVE’s roadmap for scaling Tier‑III HPC capacity in Paraguay, with additional deployments targeted through 2027.
• Renewable Power Base: HIVE currently operates 300 MW of hydroelectric power in Paraguay, with an additional 100 MW under development, providing low‑cost, reliable electricity for AI/HPC workloads.
• Future Expansion Plans: Planned Tier‑III data center capacity expansion in Yguazú, contingent on customer adoption and the company’s capital position.
• Management Commentary:
• Frank Holmes (Executive Chairman): Emphasized the beta‑test nature of the deployment and its role in validating the AI factory concept built on existing renewable energy assets.
• Aydin Kilic (President & CEO): Stressed the significance of Columbia University’s LLM training as a validation step toward large‑scale HPC capacity by 2027.

Notable Quotes

“This deployment marks our first live GPU compute workload in Asunción and provides the real‑world performance data we need to guide our Tier‑III expansion roadmap.” – Frank Holmes, Executive Chairman

“Having a research team from Columbia University running LLM training workloads on HIVE's BUZZ Cloud infrastructure in Asunción is a powerful validation of what we are building.” – Aydin Kilic, President & CEO