DigitalOcean secures $725 million equipment facility for AI cloud capacity
DigitalOcean lined up bank equipment financing for GPUs and servers. The credit lesson is the same one funders face in every asset file: demand only helps when asset control, revenue timing and repayment evidence connect.
DigitalOcean said Sept. 10 it entered a $725 million equipment finance facility to buy GPU, CPU and other infrastructure for its AI native cloud platform. The facility matures Sept. 10, 2030, and includes an accordion option of up to $300 million, subject to lender commitments and other conditions. The useful read for small business finance is not the AI label. It is the shape of the credit. DigitalOcean is matching equipment cash outflows against expected customer revenue, while a bank group takes collateral support around the hardware that creates that capacity. The same discipline applies in MCA, RBF, factoring, equipment finance and ABL. A growth claim becomes financeable only when the file ties spend, asset control, revenue timing and repayment evidence together. The lender group also matters. The company said MUFG Bank served as sole administrative agent and collateral agent. MUFG, Axos Bank, BMO Bank and Wells Fargo acted as joint lead arrangers and joint bookrunners, with PNC Bank as document agent. MonitorDaily carried the same facility terms on Sept. 11 for the equipment finance audience. DigitalOcean described the transaction as proactive financing for 2027 and 2028 demand. Management said the company has low leverage and healthy adjusted free cash flow margins, but the underwriting question sits below that corporate language. Lenders funding compute capacity have to care about what equipment is bought, how fast it ages, where it sits, who can use it, what customer contracts support the spend and how quickly demand can turn into collected cash. That is why this is closer to equipment and asset based finance than a simple tech growth headline. GPUs and servers can be financed, but they are not static collateral like a building. Their value depends on model demand, utilization, power, cooling, maintenance, resale market depth and the speed of replacement cycles. A funder cannot stop at a purchase order and a demand forecast. The file needs equipment schedules, serial level records, lease or data center rights, insurance, service obligations, vendor terms, customer concentration and cash application. It also needs a plan for what happens when a rack is delayed, reallocated, impaired or no longer economic to run. The operator read AI infrastructure credit is a cleaner version of a common small business problem. The borrower can have real demand and still fail the file test if the asset list, revenue path and repayment record do not connect. ABF Journal’s Sept. 14 middle market debt weekly placed the DigitalOcean facility next to retail, steel and automotive ABL transactions. Its point was that collateral secured capital remains available, even as floating rate costs may rise if the Federal Reserve moves again. For borrowers, that makes structure matter. A cheaper spread can help, but only if collateral reporting and coverage can survive a more expensive rate path. For a broker or funder, the DigitalOcean deal is a reminder to separate the financed asset from the market story around it. In a restaurant file, card receipts need processor history and chargeback context. In a packaging or food company file, inventory needs aging, purchase orders and deduction history. In a compute infrastructure file, equipment needs location, control, utilization and contract support. Different assets, same test: can the lender prove the repayment source without leaning on the borrower’s momentum story? Opinion Equipment finance only works when the equipment file stays current The risk is not that lenders are financing servers. The risk is treating server demand as repayment evidence by itself. The better file starts with the asset schedule, then follows control, utilization, customer cash and amendment history until the lender can explain what happens in a weak month. Plutus tracks funder behavior and credit file discipline at plutusai.net.