Strengthening its global AI infrastructure, Microsoft signed a $9.7 billion partnership with IREN to boost data centre and cloud capabilities.
Microsoft Corp. has signed a five-year, $9.7 billion agreement with Australian data center operator IREN Ltd. to purchase AI cloud capacity, making Microsoft the company’s largest customer. The deal underscores Microsoft’s aggressive push to expand its AI infrastructure as demand for cloud computing and AI services surges worldwide.
Under the agreement, Microsoft will gain access to Nvidia accelerator systems built on the GB300 architecture for AI workloads at IREN’s Texas facilities. The deal includes a 20% prepayment and is expected to generate about $1.94 billion in annualized revenue once fully implemented, according to IREN Chief Executive Officer Daniel Roberts.
To support the agreement, IREN will spend $5.8 billion purchasing GPUs and related equipment from Dell Technologies. The installation of Nvidia’s GB300 systems will roll out in phases through next year, utilizing roughly 10% of IREN’s total capacity. The company operates data centers with a combined potential of nearly 2.75 gigawatts in Texas, including its Sweetwater Hub near Abilene.
Roberts said IREN has been in talks with several major hyperscalers as AI demand accelerates. Following the announcement, IREN shares surged over 28% in pre-market U.S. trading, while Dell gained about 4%.
IREN joins a growing group of “neocloud” providers including CoreWeave and Crusoe that began as crypto miners and have since pivoted to supplying high-performance computing power for AI giants like Microsoft, Meta, and OpenAI.
