IREN closes $3.65B GPU loan to fund $9.7B Microsoft AI deal

• IREN secures $3.65B investment-grade financing, covering 96% of its GPU purchase obligations in a $9.7B Microsoft AI cloud contract
• Facility is the first U.S. private placement of GPU-backed, investment-grade debt, signaling a shift in AI infrastructure financing

On June 1, 2026 (UTC), GlobeNewswire reported that IREN Limited (NASDAQ: IREN) closed a $3.65 billion investment-grade debt facility to fund NVIDIA GPU hardware for a multiyear AI cloud services contract with Microsoft. The deal marks the first deployment of GPU-backed, investment-grade debt in the U.S. private placement market. Fitch rated the facility A, and DBRS rated it A(low). Collateral combines both NVIDIA GPUs and cash flows from Microsoft’s contract, providing relatively low risk and high credit quality for institutional lenders.

The facility features two tranches: $2.1 billion in private placement bonds at SOFR+2.13% (fixed), and $1.55 billion in delayed draw term loans at SOFR+2.25% (floating, hedged). The blended loan cost is 6.0%. Microsoft’s prepayments lower the average cost of all financing to 3.31%. Altogether, these funds cover 96% of IREN’s $5.81 billion GPU purchasing obligations under the $9.7 billion, five-year Microsoft agreement. The funding structure significantly reduces risk for IREN’s AI cloud expansion and frees capital for further growth.

This financing sets a new standard by attaching investment-grade credit ratings to GPU-backed debt, collateralizing both hardware and contract cash flows, and allowing for near-total hardware financing. The access to institutional capital and risk reduction marks a departure from typical venture capital or costly private credit in the AI infrastructure space. Goldman Sachs and J.P. Morgan structured and sold the deal, highlighting increased institutional involvement in this sector.

IREN’s market cap stands at $23.62 billion. The company maintains strong liquidity and targets 480MW of AI cloud capacity by the end of 2026, leveraging its owned power and land assets for further expansion. The transaction’s structure provides scalable, relatively low-cost capital access for AI data center operators, reflecting a broader trend toward asset-backed, investment-grade financing as demand for AI infrastructure surges.