NEW YORK — Microsoft has signed a $9.7 billion, five-year agreement with IREN Limited, an Australian data center and cloud infrastructure company, to secure large-scale AI cloud capacity powered by NVIDIA GB300 GPUs.
The deal, announced on November 3, 2025, grants Microsoft access to high-performance GPU infrastructure hosted at IREN’s 750MW Childress, Texas campus, which is being developed to support up to 200MW of critical IT load across four phases (Horizon 1–4).
Under the agreement:
IREN will provide Microsoft with GPU cloud services over five years.
The contract includes a 20% prepayment by Microsoft.
IREN will purchase approximately $5.8 billion worth of GPUs and ancillary equipment from Dell Technologies.
The infrastructure will be deployed in phases through 2026, with liquid-cooled data centers supporting AI workloads.
This strategic move allows Microsoft to scale its AI offerings rapidly without the delays and capital intensity of building new data centers from scratch. It also reflects the company’s urgency to meet surging demand for AI compute, driven by its partnerships with OpenAI and other enterprise AI initiatives.
IREN, listed on NASDAQ, is positioning itself as a key player in the global AI infrastructure market. The deal with Microsoft significantly boosts its profile and underscores the growing importance of third-party GPU cloud providers in the AI ecosystem.
