Oracle’s AI expansion has become a credit story as much as a technology story. S&P Global has downgraded Oracle’s rating from "BBB" to "BBB-", leaving it one notch above junk status, and it points to OpenAI as a "key credit risk" behind that move.
The issue is not simply that Oracle is investing in AI. It is that the scale, timing, and concentration of those commitments create a sharper financial risk if demand does not arrive as planned or if OpenAI’s position weakens.
Why S&P Global cut Oracle’s rating
According to S&P, Oracle’s AI business is using far more cash than previously expected. Capital spending is now projected to reach $95 billion by 2027, compared with an earlier estimate of $60 billion.
That increase matters because the revenue tied to those investments is not expected to arrive immediately. The source describes a gap between large near-term spending and revenue that will not materialize for years. For a credit rating agency, that timing can change the risk profile quickly.
The downgrade from "BBB" to "BBB-" does not say Oracle has failed. It says the margin of safety has narrowed. A rating one notch above junk status signals that the company still holds an investment-grade rating, but with less room for shocks.
In plain terms, Oracle is committing large amounts of capital now for an AI buildout whose payoff depends on future demand. S&P Global’s concern is that one major customer sits at the center of that risk.
OpenAI is central to the concern
OpenAI accounts for roughly half of Oracle’s $638 billion in contractual obligations. That concentration is the heart of the credit issue. If a large share of expected demand is tied to one company, Oracle’s infrastructure plans become more exposed to that company’s future.
S&P Global views OpenAI as a "key credit risk" because Oracle could be left with massive data center capacity it could not fill if OpenAI collapsed. The risk is not described as a current collapse. It is a scenario in which Oracle’s spending and capacity commitments become difficult to absorb if that customer can no longer support the expected demand.
That is why the credit analysis focuses on more than headline AI growth. A company can have a strong position in a fast-growing market and still face balance-sheet pressure if it spends ahead of revenue, especially when future revenue depends heavily on a concentrated set of obligations.
- Capital spending: projected at $95 billion by 2027, up from $60 billion.
- Contractual obligations: OpenAI represents roughly half of Oracle’s $638 billion total.
- Credit rating: S&P Global cut Oracle from "BBB" to "BBB-".
- Main risk: unused data center capacity if OpenAI collapsed.
Why Oracle differs from larger cloud rivals
S&P says Oracle is in a tougher position than AWS, Google, and Microsoft. The reason given is that those companies have internal workloads that could absorb excess capacity, along with deeper financial reserves.
That comparison is important because cloud infrastructure risk is not only about building data centers. It is also about what happens if expected external demand falls short. A company with major internal workloads has more ways to use capacity that customers do not take up.
Oracle, as described in the source, faces a more concentrated challenge. If OpenAI collapsed and Oracle had capacity it could not fill, the company would have less protection than competitors with broader internal demand and larger reserves.
Still, the source also makes clear that this is not only an Oracle problem. Even the balance sheets of AWS, Google, and Microsoft would take a serious hit if OpenAI went under. That point shows how much financial weight the AI infrastructure boom has placed on expectations around OpenAI.
Doubts are appearing beyond Oracle
The source points to signs of concern elsewhere in the market. SoftBank reportedly had to reduce a loan backed by OpenAI shares from $10 billion to $6 billion because lenders struggled to value the privately held company.
That detail does not prove OpenAI is weak. But it does show that valuation uncertainty is affecting financing tied to the company. When lenders struggle to value a privately held company, deals connected to that value can become harder to complete at the original size.
OpenAI has also pushed its IPO back to 2027. In the context of Oracle’s credit risk, that delay matters because public markets can provide a clearer valuation reference and a different funding path. Without that near-term public-market benchmark, uncertainty around OpenAI remains more difficult for lenders and counterparties to resolve.
What this means for the AI infrastructure race
The downgrade highlights a basic tension in the current AI buildout: infrastructure must be financed before the full revenue stream arrives. Companies that build data center capacity are making large commitments based on future usage, and those commitments can become risky when demand is concentrated.
For Oracle, S&P Global’s decision frames AI growth as a credit tradeoff. The opportunity is large enough to drive major capital spending, but the financial exposure is large enough to affect the company’s rating.
The broader implication is that AI infrastructure is no longer being judged only by technical ambition or customer announcements. It is being judged by funding needs, balance-sheet resilience, customer concentration, and the timing of future revenue.
OpenAI remains central to that equation in the source’s account. Its role in Oracle’s contractual obligations, its importance to cloud capacity planning, and the reported difficulty of valuing shares tied to lending all point to the same issue: AI demand may be powerful, but the financial structure around it is under closer scrutiny.