Microsoft has finalised a massive $9.7 billion, five-year contract with Australian company IREN to expand its artificial intelligence cloud computing capabilities. The agreement will grant Microsoft access to advanced compute infrastructure based at IREN’s facility in Childress, Texas, which is designed to eventually support up to 750 megawatts of capacity.
This infrastructure will feature Nvidia’s latest GB300 graphics processing units, specifically engineered to accelerate AI workloads involving reasoning, agentic systems, and multimodal generative models. The deployment of these resources is scheduled to continue in phases through 2026.
In parallel, IREN is acquiring roughly $5.8 billion worth of GPUs and associated equipment from Dell Technologies to enhance its hardware portfolio. This investment underscores IREN’s quick transformation from a bitcoin-mining operation into a leading provider of high-performance computing power tailored for artificial intelligence applications.
Only last month, Microsoft debuted its initial production cluster using Nvidia’s GB300 NVL72 systems within its Azure cloud platform, emphasising its commitment to delivering scalable computing solutions optimised for complex AI systems.
Around the same time, Microsoft secured a notable deal with Nscale, acquiring approximately 200,000 Nvidia GB300 GPUs distributed across several data centres in the United States and Europe.
IREN’s chief executive officer, Daniel Roberts, revealed to Bloomberg that the Microsoft contract is expected to utilise about one-tenth of IREN’s total compute capacity and generate close to $1.94 billion in annual revenue.