IREN Secures 85% of AI Cloud ARR as Compute Demand Outpaces Capacity

IREN ($IREN) said it has already secured roughly 85% of its targeted year-end 'AI cloud' annual recurring revenue (ARR), underscoring how demand for compute continues to outstrip the pace at which new infrastructure can be brought online—a gap that is increasingly shaping competition across the data-center and digital-asset infrastructure ecosystem.

The company’s shares closed at $36.80 on Friday in U.S. trading, down 3.8% on the day. The stock was off 0.7% on the week, though recent price action has taken a back seat to investor focus on IREN’s capacity expansion timeline and the durability of contracted revenue tied to hyperscale customers.

According to the company, the majority of its planned AI cloud ARR for the end of this year is already at the contract stage. While IREN did not detail every counterparty in the update, the progress suggests that customer commitments are being made well ahead of delivery—an important signal in a market where AI model training and inference workloads have tightened availability for high-performance GPUs and power-dense facilities.

IREN is pursuing a phased buildout designed to scale its data-center footprint to 480 megawatts (MW) by 2026 and 1.2 gigawatts (GW) by 2027. The roadmap reflects a broader industry shift: operators that can secure power, land, transmission, and hardware financing are moving fastest to capture 'liquidity inflow' from multi-year AI infrastructure deals.

Beyond North America, IREN said it has signed a transmission connection agreement to support an 800MW data-center campus in southern Australia, targeting operations in 2028. The project highlights the strategic value of regions with supportive grid access and developable sites, as AI deployment pushes the sector toward larger, campus-style builds that can accommodate next-generation power and cooling requirements.

To support delivery on a high-profile AI cloud agreement with Microsoft ($MSFT), IREN also announced it has secured $3.65 billion in investment-grade GPU financing. The structure is notable in a capital-intensive environment where debt markets have been selective, and where equipment financing tied directly to revenue-generating assets can lower overall funding friction for large-scale rollouts.

Market participants generally view investment-grade terms as a sign that lenders are underwriting not just hardware collateral, but the perceived stability of contracted cash flows. The association with Microsoft is likely to be watched closely as an indicator of customer quality and execution credibility—factors that can influence follow-on demand from other enterprise and hyperscale buyers.

While IREN’s stock moved lower into the close, the company’s update points to a fundamental narrative that remains constructive for the AI infrastructure trade: long-dated customer commitments are arriving faster than capacity, and firms able to translate power access and financing into deliverable megawatts stand to strengthen their position as global suppliers of AI compute.