IREN Orders 50,000 Nvidia GPUs, Projects $3.7B AI Cloud Revenue
IREN agreed to purchase 50,000 Nvidia B300 GPUs, expanding its GPU fleet to 150,000 units across sites in British Columbia and Texas. The company forecasts this capacity could drive annualized AI cloud revenue above $3.7 billion by end-2026, funded partly through a $6 billion ATM equity program raising dilution concerns.
1. Nvidia GPU Purchase
IREN signed purchase agreements for 50,000 Nvidia B300 GPUs, taking its installed and contracted fleet to 150,000 units. The GPUs will be deployed in stages at air-cooled data centers in Mackenzie, British Columbia, and Childress, Texas.
2. Deployment Plan and Capacity
The company plans to roll out the GPUs during the second half of 2026, leveraging existing site infrastructure for rapid deployment. The additional capacity aims to support AI training and inference workloads, meeting growing customer demand.
3. Revenue Projection
IREN projects that a 150,000 GPU footprint could generate more than $3.7 billion in annualized AI cloud revenue by end-2026. This projection is based on internal assumptions and illustrates the potential scale of the expanded compute infrastructure.
4. Financing Strategy and Dilution
To fund incremental capital spending of about $3.5 billion tied to the GPU orders, IREN launched a $6 billion at-the-market equity program alongside other financing sources. Critics have raised concerns over dilution after the company issued shares totaling roughly $1 billion.