IREN closes US$3.65 billion GPU financing facility tied to Microsoft AI cloud deal, indicating push toward a transmission-ready 800MW data center campus

IREN Limited, the data-center and infrastructure company, has closed a US$3.65 billion investment-grade GPU financing facility that it says is linked to its Microsoft AI cloud contract. The financing is designed to support the deployment of graphics processing units, or GPUs, which are specialized computer chips used to run training and inference workloads for artificial intelligence.

The move positions IREN to scale up compute capacity at a pace that depends not only on chip supply, but also on the power required to operate large data-center campuses. In the same timeframe, the company outlined advanced plans for a data center site with 800MW of capacity that is described as “transmission-ready,” meaning the power delivery and grid connection work are expected to support large load growth.

A “GPU financing facility” generally refers to structured funding that helps a company pay for GPU-related buildouts, often including the supporting technology and infrastructure needed to house and operate accelerators at scale. In IREN’s case, the financing was tied to its Microsoft AI cloud commercial relationship, suggesting the company is treating AI-linked demand as a central driver of new capacity rather than a side line of traditional colocation.

The announcement comes as hyperscalers and enterprise customers increasingly rely on accelerated computing, which tends to concentrate both capex and power requirements into new infrastructure phases. For data-center operators, the financing terms and the ability to secure power at large scale can be as consequential as the compute itself, particularly when projects require multi-year build cycles and high initial investment.

Microsoft, for its part, has been expanding its cloud and AI footprint across markets through its Azure platform and related services. While the reported closure of IREN’s facility points to a Microsoft-linked demand pipeline, the company has not, in the information available here, detailed the contract economics, the duration of any arrangement, or whether the GPU capacity is dedicated, flexible, or tied to performance milestones.

In sector terms, the data-center industry is moving from “capacity-first” planning to “power-and-compute synchronized” planning. Investors typically look for developers that can line up power access, grid readiness, and financing before committing to full-scale builds. IREN’s description of a transmission-ready 800MW campus, paired with a large investment-grade GPU financing package, aligns with that broader industry approach.

What remains unclear is the specific structure of the Microsoft-backed linkage. The disclosed information does not spell out which Microsoft programs the financing supports, the share of future GPU buildouts covered by the facility, or how much of the 800MW plan is already under contract versus in active development. It also does not provide details on timing, expected capital deployment schedules, or any revenue or margin estimates tied to the GPU deployments.

Going forward, the key things to watch are project milestones for the 800MW campus, additional disclosure on how power interconnection and data hall construction progress, and whether IREN provides further operational metrics that translate GPU capacity planning into measurable utilization or contracted demand under its Microsoft arrangements. Any update on the financing draw schedule and related cost assumptions would also be material for understanding how quickly the company can convert commitments into revenue.