IREN unlocks $363M from AI servers already earning revenue
What's the deal? IREN has closed a $363 million equipment financing deal arranged through Wingspire Equipment FinanceDealroom has a profile for this one. Try Dealroom →, backed by a six-lender group led by Eldridge Capital Management and Silver Point CapitalDealroom has a profile for this one. Try Dealroom →. The facility refinances revenue-generating AI compute infrastructure — GPU-based servers and networking systems — already deployed in the company's high-density data centres.
Why it matters: The operative word is refinancing. Unlike a planned site or a new customer order, the deal converts existing, income-producing equipment into cash IREN can redeploy for growth. It is secured mainly by the financed hardware and backed by multiyear customer contracts.
What's the endgame? IREN is scaling its AI build-out fast. It has reported $4 billion in contracted annualised revenue tied to 2026 capacity, but plans $25–30 billion in fiscal 2027 capital spending — making it a prominent name among AI-adjacent infrastructure operators.
The expansion: On October 21, IREN formed IREN Europe and secured civil-works approval for a planned AI data-centre campus in Badajoz, Spain. The roughly 300MW site, with grid-connected power and a construction permit both secured, targets first operations in 2028 — a future goal, not running capacity.
What could go wrong? The refinancing is a positive funding signal, not proof the broader build-out is fully financed. IREN has disclosed GPU financing at rates of 6% to 9%, and lenders are scrutinising AI infrastructure risk as fresh borrowing grows costly.
The signal: IREN shares rose 7% on news that Anthropic filed for an IPO, lifting AI-adjacent stocks. But sentiment and financing are different stories: one prices the sector's mood, the other marks a concrete step by the company. The deal shows deployed AI capacity is starting to generate financing value — a sharper read than betting on the next AI winner.
Read more: newscase.com
Image credit: Idaho National Laboratory
