Breaking: Microsoft has secured a significant expansion of its AI cloud infrastructure with a new $9.7 billion, five-year agreement with Australian data center provider IREN. The deal, announced Monday, will provide Microsoft with access to compute capacity powered by Nvidia‘s GB300 GPUs.
The infrastructure will be deployed in phases thru 2026 at IREN’s facility in Childress, Texas, and is planned to support 750 megawatts of capacity. IREN is also investing in its own infrastructure, purchasing GPUs and equipment from Dell for approximately $5.8 billion.
This agreement underscores Microsoft’s aggressive strategy to bolster its AI capabilities amid surging customer demand. the company recently launched its first production cluster featuring Nvidia’s GB300 NVL72 systems for Azure, designed to optimize performance for reasoning models, agentic AI systems, and multi-modal generative AI.
Microsoft also signed a deal last month with Nscale for roughly 200,000 Nvidia GB300 GPUs to be deployed across data centers in Europe and the U.S.
IREN, like competitor CoreWeave, transitioned from a bitcoin-mining operation to focus on the more lucrative AI workload market. IREN CEO Daniel Roberts anticipates the Microsoft deal will represent approximately 10% of the company’s total capacity and generate around $1.94 billion in annualized revenue, according to a Bloomberg report.
