Microsoft has entered into a $9.7 billion agreement with data center operator IREN to secure access to Nvidia’s AI chips, a move aimed at overcoming the ongoing global shortage of high-performance computing resources. The five-year deal highlights Microsoft’s strategy to expand artificial intelligence infrastructure rapidly without the heavy capital investment of building new data centers.
Under the agreement, IREN will provide Microsoft with access to Nvidia GB300 GPUs for AI workloads. Microsoft will invest about $5.8 billion in infrastructure that IREN will operate using Dell and Nvidia hardware. Dell Technologies will supply IREN with Nvidia GPUs and related systems as part of the arrangement.
IREN plans to deploy the GPUs in phases through 2026 at its 750-megawatt Childress, Texas, facility. The site will include new liquid-cooled data centers capable of supporting around 200 megawatts of critical IT load. IREN operates multiple data centers across North America with a total power capacity of 2,910 megawatts.
Microsoft’s prepayment of around 20 percent of the total contract value will help finance IREN’s $5.8 billion deal with Dell. However, Microsoft retains the right to terminate the agreement if key delivery milestones are not achieved, Reuters reports.
Daniel Roberts, Co-Founder and Co-CEO of IREN, said the partnership underscores IREN’s position as a vertically integrated AI cloud provider. “This agreement validates IREN’s strength as a trusted provider of AI cloud services and opens access to new customer segments among global hyperscalers,” Roberts said.
Jonathan Tinter, President of Business Development and Ventures at Microsoft, added that the partnership strengthens Microsoft’s AI infrastructure expansion strategy. “IREN’s expertise in building and operating a fully integrated AI cloud, combined with their secured power capacity, makes them a strategic partner,” Tinter said.
This collaboration follows Microsoft’s recent $17.4 billion deal with AI infrastructure company Nebius and similar partnerships with CoreWeave and Lambda. The agreements reflect Microsoft’s growing reliance on specialized “neocloud” providers to meet soaring demand for computing power to support services such as ChatGPT and Copilot.
Rajani Baburajan
