Microsoft is continuing to accelerate in the race to expand its AI cloud computing infrastructure by signing a $9.7 billion deal with Australian company IREN. The five-year contract is designed to help Microsoft secure huge computing resources to serve the global AI needs.
According to the plan, IREN will deploy the Nvidia GB300 GPU system at a facility in Childress, Texas, with a total capacity of up to 750 megawatt, expected to be completed in 2026. IREN will also buy GPUs and equipment from Dell alone for about $5.8 billion.
Previously, Microsoft launched the first production cluster using the Nvidia GB300 NVL72 system on the Azure platform, optimized for reasoning models, generative AI and multi-modal systems.
At the same time, the company also signed with Nscale to provide 200,000 Nvidia GB300 GPUs to data centers in Europe and the US.
IREN, originally a Bitcoin mining company, has shifted to providing GPUs for AI and quickly reaped the benefits.
CEO Daniel Roberts said that working with Microsoft accounts for only about 10% of the company's capacity, but can generate 1.94 billion USD in annual revenue, according to Bloomberg.
