Why AI Companies Are Paying More to Borrow
AI-linked companies still have access to bond markets. Investors are demanding more yield to lend to them.
Spreads on AI-related corporate bonds have moved towards 115 basis points, compared with roughly 78 basis points across the broader corporate market.
The difference appears while technology companies continue spending heavily on data centres, chips and power infrastructure. Bond investors are pricing the bill.
Strong companies still face a supply problem
Large technology groups can service more debt than weaker borrowers. Credit quality solves only part of the calculation. Hyperscalers are expected to issue roughly $420 billion of debt next year, around 60 percent more than in 2026.
Every new bond competes for space in institutional portfolios. An asset manager already holding several technology issuers needs a reason to add more exposure to the same investment cycle. A higher spread provides one.
Other sectors face less concentration. Pharmaceutical or insurance issuers borrowing at the same time give bond investors another source of yield without adding more AI infrastructure exposure.
Project debt carries another layer of risk
Oracle’s Project Jupiter shows the difference between lending to a company and financing infrastructure tied more closely to a specific project.
Around $18 billion of loans associated with the New Mexico data-centre project have traded at roughly 89 to 91 cents on the dollar. Investors are assessing Oracle’s borrowing, the project itself, power requirements and local opposition. A normal corporate bond relies on cash flows across the wider company. Project-related debt concentrates more risk around one development and the economics behind it. More AI construction means more debt sitting between those two structures.
Watch debt against the cash produced by AI
Higher spreads do not require investors to expect defaults. Lenders charge more when issuers borrow heavily before new assets produce enough cash to cover the investment cycle.
Three conditions put more pressure on pricing: issuance keeps rising, AI revenue arrives more slowly than expected and government-bond yields remain high. Stronger cash generation changes the equation.
For traders, compare new borrowing with additional cash generated from AI products and infrastructure. Debt growing much faster than cash flow gives bond investors a reason to keep demanding more yield.
