Microsoft is making another big move in the AI race. The Redmond-based tech giant has signed a $9.7 billion, five-year deal with Australia’s IREN to secure more AI cloud power. Under the agreement, Microsoft will gain access to a powerful computing infrastructure built with Nvidia’s GB300 GPUs.
IREN CEO Daniel Roberts said the Microsoft partnership will use only about 10% of the company’s total capacity, yet it’s projected to bring in roughly $1.94 billion in yearly revenue.
The new systems will roll out in phases through 2026 at IREN’s Childress, Texas, facility, which is being designed to handle up to 750 megawatts of capacity. This partnership marks a major step in Microsoft’s push to expand its AI cloud capabilities and meet the surging global demand for computing power.
IREN is also making a big investment of its own. The company announced plans to spend around $5.8 billion on GPUs and other hardware from Dell to strengthen its computing infrastructure.
The deal follows Microsoft’s recent launch of its first Azure production cluster powered by Nvidia’s GB300 NVL72 systems. According to the company, the new setup is built to handle the next wave of AI applications, from advanced reasoning models to agent-style tools and multimodal generative AI.
Just last month, Microsoft struck another major agreement — this time with Nscale — to secure around 200,000 Nvidia GB300 GPUs. The hardware will power three data centers in Europe and one in the United States, further expanding Microsoft’s global AI infrastructure.
Much like rivals such as CoreWeave, IREN began its journey in bitcoin mining. But the company soon recognized that its powerful GPU infrastructure could be put to better use in the rapidly growing AI industry. The pivot has paid off, with IREN seeing significant gains from its shift toward AI-focused computing.
Also Read: Broken Benchmarks: How Flawed Tests Threaten AI Progress
