Particle.news
Download on the App Store

HIVE Proposes 320 MW Toronto AI Gigafactory as Investors Move In

Financing, GPU supply and grid approvals will determine whether the campus can shift HIVE to recurring AI hosting revenue.

Overview

  • BUZZ HPC, HIVE’s hosting unit, announced on May 18 that it bought roughly 25 acres in the Greater Toronto Area and plans a 320 MW campus that could hold more than 100,000 GPUs with an estimated CAD $3.5 billion build cost and a target to begin operations in the second half of 2027.
  • Markets reacted strongly to the May 18 announcement with the stock jumping about 34% that day and rising more than 50% from April lows to roughly $4.10 by May 22, prompting analyst upgrades including Cantor Fitzgerald’s higher target and a new Buy from Northland.
  • A Q1 2026 13F filing shows Leopold Aschenbrenner’s Situational Awareness LP opened a position of about 3.4 million HIVE shares, and coverage notes Aschenbrenner’s background as a former OpenAI researcher focused on AI infrastructure investing.
  • HIVE says its GPU cloud is scaling to about 11,000 active units and management is targeting an Annual Recurring Revenue run rate of $225 million from high‑performance computing, with company financial results due June 1 and an earnings call scheduled for June 2 as the next market checkpoints.
  • The plan leans on Ontario’s low‑carbon grid and aims to supply institutional AI demand but faces clear execution risks from the need to secure billions in capital, long GPU supply lines, permits and grid interconnection that will determine if the project relieves predicted compute shortages in 2026–27.