IREN Expands AI Cloud Capacity to 150,000 GPUs

The growing demand for AI infrastructure

As AI workloads continue to scale, the availability and timing of compute capacity is becoming increasingly important. Larger models, longer training runs, and sustained inference demand are placing growing pressure on infrastructure, making speed and certainty of deployment critical. 

IREN has entered into purchase agreements for over 50,000 NVIDIA B300 GPUs, expanding its total fleet to 150,000 GPUs. 

NVIDIA B300 GPU deployment phases

This expansion builds on existing deployments and will be delivered in phases through the second half of 2026 across IREN’s air-cooled data centers in Mackenzie, British Columbia and Childress, Texas. 

Time-to-compute is an increasingly important factor in today’s AI cloud market. Early procurement enables IREN to bring cloud capacity online with greater speed and certainty, ensuring deployment timelines remain on track as customer agreements progress. 

At full scale, the 150,000 GPU fleet is expected to support AI Cloud annualized run-rate revenue of over $3.7bn by the end of 2026. Existing data centers at Canal Flats and Childress provide capacity to support additional GPUs over time. 

$9.3B in secured funding

Over the past eight months, IREN has secured $9.3bn of funding across customer prepayments, convertible notes, GPU leasing, and GPU financing. 

IREN expects to leverage these and other capital sources to finance approximately $3.5bn of additional capex for these orders expected in H2 2026, including GPUs, servers, storage, networking, labor and ancillary equipment. 

Payment terms are structured on a post-shipment basis, enhancing working capital efficiency. Hardware procurement will continue to be sequenced in line with commercial milestones and capital availability. IREN has also established an at-the-market equity program as part of its broader capital management framework. 

Daniel Roberts, Co-Founder and Co-CEO of IREN, commented: 

“Scaling to 150,000 GPUs positions IREN among the largest AI cloud infrastructure providers globally and underscores the strength of our vertically integrated platform. In a supply-constrained environment, early hardware procurement reduces time-to-compute and increases execution certainty as we scale.”  

Delivering AI Cloud capacity for training and inference 

IREN’s expansion to 150,000 GPUs builds on its approach to delivering large-scale infrastructure for AI training and inference. Learn more about IREN Cloud™.