APLD
CRWV
The customer isn’t just paying rent — it’s lowering the bill
Applied Digital just handed investors a nice little plot twist: the real prize from its mysterious hyperscaler customer may not be the lease revenue. It may be the fact that the customer’s shiny investment-grade balance sheet is helping the company borrow money more cheaply.
That matters because AI infrastructure is basically a capital-eating machine. You don’t just build one data center and call it a day — you keep pouring concrete, power, and cash into the thing like it’s a very expensive Lego set.
The financing glow-up
Management said its latest debt financing came in with a coupon around 7%, down from roughly 9.25% previously. That 225-basis-point haircut is the kind of thing that quietly changes the math on a project, especially when you’re building campuses at scale.
A few key takeaways:
- Applied Digital is leaning harder into leases with investment-grade hyperscalers instead of younger AI model companies.
- CoreWeave gets a shoutout as a customer that helped shape the financing model.
- The company says future campuses leased to similar blue-chip customers should keep financing costs attractive.
Why investors should care
If you’re building AI data centers, capital costs can make or break returns. A lower interest rate doesn’t just save money on one loan — it can improve the economics of every new campus you finance after that.
Applied Digital says it already has about 1.41 gigawatts of contracted capacity and another 1.7 gigawatts it’s still marketing. Translation: this isn’t a one-off win. If the company keeps getting cheaper debt because its tenants have fortress-like balance sheets, that advantage could compound fast.
Big picture: in AI infrastructure, the winner may not just be the company with the most GPUs. It may be the one that can fund the whole party without getting crushed by interest expense.