Data center developer IREN has entered into a large-scale, multi-year agreement with Microsoft to provide cloud infrastructure for artificial intelligence workloads valued at approximately $9.7 billion.
Under the terms of the partnership, Microsoft will receive access to NVIDIA GB300 GPUs over a five-year period, backed by a 20% upfront payment.
In addition, IREN has signed a separate procurement deal with Dell Technologies worth around $5.8 billion to supply GPUs and supporting hardware. The equipment will be deployed progressively through 2026 at IREN’s 750 MW campus in Childress, Texas, where new data centers with liquid-cooling systems are being constructed to support up to 200 MW of high-density AI compute (Horizon 1–4).
The project will be financed using existing funds, customer prepayments, operating income, and additional financing mechanisms.
IREN co-founder and co-CEO Daniel Roberts said the agreement underscores the scalability of the company’s vertically integrated AI infrastructure platform:
“This collaboration highlights our ability to deliver reliable, large-scale AI compute. It not only strengthens IREN’s role as a trusted cloud provider but also opens the door to new hyperscale clients.”
He added that expanding GPU clusters remains a core strategic priority across IREN’s 3 GW power portfolio in North America.
Microsoft’s Corporate VP for Business Development and Ventures, Jonathan Tinter, commented:
“Together with IREN, we are building next-generation AI infrastructure for our customers. Their expertise across data center development, GPU integration and energy supply makes them a strategic partner for Microsoft.”
In October 2025, IREN secured $875 million through debt financing after its public listing, aimed at scaling both energy and mining assets.
Meanwhile, Microsoft is facing a class-action lawsuit in the U.S., where plaintiffs claim the company intentionally constrained OpenAI’s compute capacity to inflate ChatGPT pricing over a three-year span. According to the lawsuit, pricing dropped by nearly 80% following OpenAI’s switch to Google infrastructure.
