AI infrastructure faces market stress test! OpenAI's related data center plans to issue $38 billion in bonds

The market is about to face a critical moment to test whether investors are willing to inject "trillions of dollars in capital" into AI infrastructure for the next decade.

According to reports obtained by Zhitong Finance APP, Vantage Data Centers is expected to initiate two debt financing deals totaling up to $38 billion as early as next week, aimed at constructing large data centers in Texas and Wisconsin. Once completed, these assets will be leased in their entirety to Oracle (ORCL.US) to provide cloud computing and AI infrastructure capabilities for OpenAI, as part of both parties' participation in the U.S. government-led $500 billion "Stargate" project. Vantage has not yet responded to requests for comment.

This financing comes at a time when credit market sentiment is weakening, with bankruptcies in the subprime auto loan industry, rising overnight funding pressures, and the Federal Reserve's balance sheet reduction leading to a decline in repurchase agreement balances to their lowest level since 2021, all of which have weakened the supply of funds. The strength of demand for this bond issuance may become a key indicator of whether the upward momentum driven by AI capital expenditures in U.S. stocks and the U.S. economy can be sustained.

Last month, the CEO of Brookfield Asset Management estimated that approximately $7 trillion in capital will be needed over the next decade to meet AI demands for factories, computing infrastructure, and energy systems. Recent practices show that investors remain enthusiastic about this sector. Meta (META.US) recently financed about $29 billion for the expansion of its data center in Louisiana; Oracle completed an $18 billion multi-tranche bond issuance in October, with order demand reaching $88 billion.

In the same week that Vantage's financing is expected to close, five of the seven major tech giants will release quarterly earnings and disclose their capital expenditures and AI investment plans for next year. Citibank estimated last month that total capital expenditures for big tech will rise to about $490 billion by 2026. For companies like Meta, Microsoft (MSFT.US), and Apple (AAPL.US) that have ample cash flow, sourcing funds is not a challenge; however, for Oracle, which may face a cash flow shortfall of $16 billion, and for OpenAI, which still has limited revenue, the path is narrower.

Goldman Sachs analysts expect that OpenAI may need to raise about $75 billion through bond issuance or equity offerings next year to fulfill its computing capacity supply agreements with Oracle, NVIDIA (NVDA.US), Broadcom (AVGO.US), and AMD (AMD.US). Last week, Oracle raised its long-term revenue guidance, expecting revenue to reach $225 billion by the end of this century, with a compound growth rate exceeding 30%, but did not provide plans for capital expenditures beyond 2026. Bank of America estimates that Oracle's cumulative capital expenditures over the next seven years may need to reach $206 billion, while total revenue during the same period is estimated to be about $317 billion.

Additionally, media reports revealed that the cloud computing agreement signed between Oracle and OpenAI last month is valued at $300 billion, accounting for approximately 65% of Oracle's $455 billion "performance obligation" backlog, driving the stock to a historical high in September. This means that Oracle's future capital expenditure funding sources will largely depend on whether OpenAI can successfully complete large-scale debt financing in the coming year