Oracle's $18 billion AI loan is being sold at a discount, highlighting the pressure on its AI financing chain.

Distribution of a syndicated loan of approximately $18 billion for Oracle's Project Jupiter data center in New Mexico has stalled, with bank offers falling to 89 to 91 cents per cent, corresponding to a loan value of approximately $16 billion to $16.4 billion. Oracle's stock price bucked the trend and fell 2% overnight, significantly underperforming the AI sector as a whole recently.

The loan, provided by a syndicate of banks at the end of last year, was used to launch Project Jupiter. This project is part of the Stargate initiative—a $500 billion AI infrastructure project led by OpenAI, SoftBank, and Oracle, aiming to build large-scale AI computing infrastructure across the United States.

Market concerns are intensifying regarding Oracle's continued borrowing and weakening credit profile. In July, S&P downgraded Oracle's corporate credit rating, bringing it just one notch below junk status. Efforts by banks to sell the loan to a wider investor base have stalled, with banks reportedly forced to maintain an additional level of Oracle-related debt exposure on their balance sheets, according to the Financial Times.

AI giants are facing cash flow pressures.

It's not just Oracle. According to the Financial Times, OpenAI disclosed in an internal presentation in July that it expects cumulative negative free cash flow to reach $278 billion between 2026 and 2030. Although its revenue is projected to grow from $36 billion in 2026 to $350 billion in 2030, the massive investment in computing power will far exceed the scale of its revenue.

According to a Wall Street Journal report on the 18th, Anthropic plans to postpone its IPO until November, after the market had previously expected a record-breaking initial public offering in October. Sources familiar with the matter said one reason for the delay is that executives at some of the largest companies in the AI industry have called for a slowdown in the pace of technological development.

Trillion-dollar financing demands test the credit market

AI investment continues to expand, but funding is not without its limitations. The combined capital expenditure guidance of the five largest US cloud computing companies exceeds $750 billion in 2026 and is expected to surpass $1.1 trillion in 2027; by 2030, the cumulative scale of AI capital expenditure may reach $5.5 trillion.

JPMorgan Chase projects that the investment-grade corporate credit market will provide over $2.1 trillion in financing for data centers over the next five years, with high-yield bonds and leveraged loans contributing approximately $350 billion. The discounting and stagnation in the distribution of Oracle's $18 billion loan occurred at the very beginning of this financing wave.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.