Oracle’s $18 Billion AI Loan Stalls as Banks Cut Prices
TREE NEWS reports: A roughly $18 billion syndicated bank loan backing Oracle’s Project Jupiter data center in New Mexico has run into distribution trouble. Banks are now quoting the debt at 89 to 91 cents on the dollar, implying a market value of about $16 billion to $16.4 billion — a steep discount to face value for a facility that was only arranged late last year to kick off construction. Oracle shares fell 2% overnight, bucking a broader rally in AI names and underperforming the sector noticeably in recent sessions.
The project sits inside Stargate, the $500 billion AI infrastructure initiative led by OpenAI, SoftBank and Oracle, aimed at building compute capacity across the United States at unprecedented scale. The loan’s stalled syndication is an early, concrete signal that the market’s appetite for financing the AI buildout may be thinner than the headlines suggest.
Why the Discount Matters
When a loan trades below par, it usually means lenders doubt they can offload it at the original terms. Banks that had planned to sell the debt to a wider pool of investors have instead been forced to keep more Oracle-related exposure on their own balance sheets than intended. That ties up capital and raises the cost of the next deal.
The credit backdrop has been deteriorating. S&P downgraded Oracle’s corporate rating in July, leaving it just one notch above junk. Persistent borrowing to fund AI infrastructure is weighing on the company’s credit profile, and the market is beginning to price that in.
Pressure Is Not Isolated to Oracle
- OpenAI disclosed in an internal July presentation that it expects cumulative negative free cash flow of $278 billion between 2026 and 2030. Revenue is projected to climb from $36 billion in 2026 to $350 billion in 2030, but compute spending is expected to outpace that growth substantially.
- Anthropic has pushed its IPO back to November from an expected October listing, with leaders at several of the largest AI firms calling for a slower pace of development cited among the reasons.
- Capital expenditure guidance from the five largest U.S. cloud companies exceeds $750 billion for 2026 combined and is projected to top $1.1 trillion in 2027. Cumulative AI capex could reach $5.5 trillion by 2030.
JPMorgan estimates the investment-grade corporate credit market will need to supply more than $2.1 trillion of data center financing over the next five years, with high-yield bonds and leveraged loans adding roughly $350 billion. Oracle’s discounted, stalled loan lands right at the start of that wave.
Market Implications
For equities, the read-through is a widening gap between AI infrastructure builders and the rest of the sector. Companies carrying heavy debt loads to fund data centers — Oracle chief among them — face higher funding costs and more skeptical equity investors. That could pressure valuations even as AI demand remains strong.
In credit markets, a loan clearing below par sets a reference point for future AI-related issuance. If investors demand wider spreads and deeper discounts, the marginal cost of the buildout rises, potentially slowing project timelines. Investment-grade spreads, which have been tight, could see modest widening in data-center-linked names.
For rates and macro, the scale of AI borrowing adds to corporate supply in credit markets, which can nudge yields higher at the margin. For crypto, the story is indirect but relevant: risk appetite for long-duration, capital-intensive tech themes is a sentiment input for digital assets, and any repricing of AI enthusiasm tends to spill into speculative corners.
Commodities tied to data center construction — copper, power equipment, uranium for nuclear-powered campuses — remain structurally supported, but a financing slowdown would delay demand rather than eliminate it.
Key Takeaways for Investors
- Watch the secondary market pricing of Oracle’s loan: a move toward 85 cents would signal deeper stress.
- Track credit rating actions on AI infrastructure borrowers; Oracle sits one notch above junk.
- Distinguish between AI demand (still strong) and AI financing capacity (showing strain).
- Expect volatility in AI-linked equities where debt-funded capex is central to the growth story.
- Monitor upcoming AI-related bond and loan deals as a real-time gauge of investor appetite.
