Treasury Yields Surge, Raising Costs for AI Infrastructure Development
- Treasury yields are at their highest since 2007, increasing borrowing costs for companies.
- JPMorgan predicts $4.1 trillion in AI-related debt issuance by 2030.
- SoftBank raised $11.1 billion in junk bonds, prioritizing capital for AI projects.
As Treasury yields climb to their highest levels since 2007, businesses that rely on debt are bracing for increased borrowing costs. This development suggests that the already high expenses associated with AI infrastructure buildout will escalate further. In June, JPMorgan Chase estimated that $4.1 trillion in AI-related debt will be issued through 2030, as companies race to enhance capacity to meet soaring demand for AI services.
Currently, the 10-year Treasury yield is around 5.17%, a rise of approximately one percentage point since January. As a result, companies will need to offer more attractive returns to draw in investors, particularly in the AI sector.
Despite these pressures, the market is not in panic. Shares of CoreWeave, a prominent neocloud provider, emerged resilient, gaining almost 8% in the past week. Conversely, Oracle, which has heavily relied on debt for its AI expansion, suffered a downturn, falling 7% this week and approximately 30% this year.
On the investment front, Japan’s SoftBank, a key capital provider for AI initiatives, raised $11.1 billion from a junk bond sale earlier this week, with yield rates reaching up to 9.75% for the 7-year tranche. Mark Malek, chief investment officer at Siebert Financial, noted that firms like SoftBank are often “price insensitive,” meaning they need substantial capital influx to stay competitive in the burgeoning AI landscape.
Leading AI model developers such as OpenAI and Anthropic are central to current market dynamics, each valued close to $1 trillion in private markets. To support the infrastructure for their advanced models, major tech firms like Amazon, Google, Meta, and Microsoft have committed to extensive capital expenditures, with expectations for increases by 2027. However, these tech giants benefit from investment-grade credit ratings, allowing them to acquire cheaper capital than smaller firms, which may face significant challenges ahead.
Challenges Ahead for Neocloud Financing
A senior private credit investor, speaking anonymously, indicated that financing for neocloud deals would become increasingly difficult due to diminished capacity to absorb increased costs. Riley Thompson, a vice president at Mitsubishi HC Capital America, reported that lenders are becoming more selective about the projects they will finance, regardless of higher interest rates.
CoreWeave cautioned in its recent SEC filings about the impact of rising rates on its operations. An increase of every 100 basis points (1 percentage point) could lead to a $30 million rise in its interest expenses based on floating rate debt. Furthermore, Oracle’s recent stock drop followed a Bloomberg report that revealed it issued a “force majeure” notice concerning its New Mexico data center project in response to rising costs. Oracle maintains that Project Jupiter remains on schedule, slated for completion in 2028.
The current environment isn’t solely defined by rising interest rates. Prior to recent increases, leading executives from OpenAI and Anthropic voiced concerns regarding the pace of AI development, advocating for a slowdown due to fears of advanced models operating beyond human control.
Furthermore, a backlash against AI data centers plays a significant role as the mid-term elections approach. A recent NBC News poll found that 69% of respondents opposed the construction of such facilities in their locales. In response, Texas Governor Greg Abbott has temporarily halted all data center-related environmental permits amid a competitive reelection campaign.
Regardless, demand for AI services is surging, as evidenced by Meta’s Muse personal assistant application, which exceeded 2.5 million downloads within its first two weeks, surpassing ChatGPT in popularity on the Apple App Store. Evercore’s Mark Mahaney forecasts that Muse could attract 100 million users within the next 6 to 12 months.
Financial experts like Andrew Giudici from KBRA believe that despite rising rates impacting future deals, overall borrower demand won’t significantly diminish. He anticipates continued substantial issuance in the capital markets. Haim Zaltzman, vice chair at Latham & Watkins, agrees, stating that while costs may rise, the demand structure makes it more manageable for companies to absorb these expenses. Borrowers committed to significant contracts, such as those with OpenAI and Anthropic, remain eager to secure financing, even at higher costs.
News Courtesy of CNBC
