Rising US Treasury bond yields are increasing the cost of capital for AI infrastructure projects, pressuring project economics.

Rising government bond yields in the United States could increase borrowing costs for companies financing AI infrastructure construction. The yield on 10-year US Treasury bonds has reached approximately 5.17%, about one percentage point higher than at the start of the year, according to CNBC.

JPMorgan Chase estimated in June that AI-related debt obligations could reach $4.1 trillion by 2030. Data center operators and other AI infrastructure participants are expanding capacity to meet growing demand.

Higher government bond yields force corporate borrowers to offer investors more competitive returns. SoftBank recently raised $11.1 billion through high-yield bonds, with the seven-year tranche yielding 9.75%.

CoreWeave, a company that relies heavily on debt financing, reported in its quarterly filing that a 100-basis-point increase in interest rates could raise its interest expenses by approximately $30 million due to its outstanding floating-rate debt. CoreWeave shares rose nearly 8% over the week, while Oracle shares declined 7% during the same period and approximately 30% year-to-date.

Lenders are becoming increasingly selective about financing deals in the neocloud segment, even when borrowers offer higher rates. Riley Thompson, vice president at Mitsubishi HC Capital America, said the market is genuinely interested in only about 20 such companies rather than 50.

Oracle shares declined after Bloomberg reported that Oracle invoked a force majeure notice on Project Jupiter, its data center campus in New Mexico, seeking to defer payment if the facility does not become operational as planned in 2028. Oracle stated that the project remains on schedule.

Major technology companies—Amazon, Google, Meta, and Microsoft—maintain investment-grade credit ratings, giving them cheaper access to capital. Andrew Giudici of rating agency KBRA believes higher rates may affect future deals but does not expect a significant reduction in borrowing demand for AI infrastructure.