Bitcoin Miners Face Fragmented Borrowing Costs for AI Expansion

Securing a high-profile AI customer no longer guarantees favorable debt terms. Operators must still fund the complex physical requirements of these projects, including specialized computing hardware, facility construction, and massive power infrastructure. Lenders are now prioritizing the quality of the counterparty and the specific route to project completion before extending credit.

IREN serves as a primary example of this disparity. The company secured a $3.65 billion GPU financing package at a 6% rate for a Microsoft cloud contract, yet a separate $2.4 billion arrangement for its Mackenzie expansion carries a 9% fixed rate. Analysts emphasize that these figures are not directly comparable, as each package covers different exposures and project milestones. Beyond interest rates, companies like Cipher Digital are integrating equity components into their debt, such as providing Google with warrants for 5.4% of its equity to backstop lease obligations for the Barber Lake project.

Institutional capital is also flowing through convertible debt, with IREN raising $3 billion via senior notes at a 1% coupon. However, the scale of these ambitions remains massive. A June analysis from BlocksBridge estimated that public miners require approximately $50 billion to finalize their planned AI infrastructure. With capital expenditures reaching over $5 billion in the first half of the year against relatively modest AI-related revenue, the sector faces a significant funding gap that is forcing operators to balance dilution risks against high-interest borrowing.