IREN to Triple AI Cloud Capacity with 50,000 NVIDIA GPU Purchase
AI-generated image: synthetic visual, not an actual depiction of events, people, or locations.
IREN confirmed the acquisition of over 50,000 NVIDIA B300 GPUs as part of a broader strategy to scale its AI cloud infrastructure. This purchase is expected to bring the company's total hardware fleet to 150,000 units. Management noted that early procurement is intended to accelerate deployment timelines and provide greater certainty for customer agreements in a market currently facing supply constraints.
The new hardware will be deployed in stages throughout the second half of 2026. These installations are scheduled for the company's existing air-cooled data centers located in Mackenzie, British Columbia, and Childress, Texas. IREN also maintains capacity for future expansion at its Canal Flats facility.
Financially, the company anticipates that the 150,000 GPU fleet will generate an annualized revenue run-rate exceeding $3.7 billion by the end of 2026. This projection is based on internal assumptions regarding hardware utilization, pricing, and contract terms, though the company cautioned that these results are not yet fully contracted.
To support this growth, IREN has raised $9.3 billion over the last eight months through a combination of customer prepayments, convertible notes, and various GPU leasing and financing arrangements. The company expects to allocate approximately $3.5 billion in capital expenditures during the latter half of 2026 to cover the costs of the GPUs, servers, networking, and associated infrastructure. Additionally, IREN has established an at-the-market equity program to further diversify its funding sources.
This article was generated with the support of our AI agent, which has been rigorously trained under the supervision of well-qualified journalists. While we strive for the highest quality in every article, if you find anything amiss, please contact us to let us know.
