Rising Bond Yields Increase Financing Risks for AI Infrastructure Companies

Summary

Rising Treasury yields are increasing the cost and risk of financing the AI infrastructure buildout. The 10-year Treasury yield reached about 5.17%, roughly one percentage point above the start of the year, while JPMorgan estimates that $4.1 trillion in AI-related debt could be issued through 2030. Companies tied to data centers and AI capacity are still raising money, but investors are becoming more concerned about whether future projects can absorb higher interest expenses. SoftBank raised $11.1 billion in junk bonds, with yields reaching 9.75% on a seven-year tranche, illustrating the price some borrowers are willing to pay to secure capital. Investment-grade hyperscalers such as Amazon, Google, Meta, and Microsoft have cheaper access to debt, while smaller neocloud companies face tighter financing conditions. Mitsubishi HC Capital said lenders may now be seriously interested in about 20 neoclouds rather than 50, and private-credit investors expect deals to become harder to finance. CoreWeave has disclosed that a one-percentage-point increase in rates could add $30 million to its annual interest expense based on its floating-rate debt. Oracle shares fell after a report that it issued a force majeure notice related to its New Mexico data center project, although Oracle said the project remained on schedule. Financing pressure is emerging alongside opposition to data centers, environmental-permit restrictions in Texas, and concerns about the pace and safety of AI development. However, industry financiers do not expect demand for capital to slow sharply because AI service demand remains strong and OpenAI and Anthropic are securing compute capacity years ahead. The article’s central uncertainty is whether that demand will remain strong enough for borrowers to pass higher costs through the financing chain.