Oracle Puts Its Credit Rating Ahead of Growth as the AI Bill Comes Due

Oracle has made an unusual admission for a company in the middle of an artificial intelligence land grab: keeping its investment-grade rating matters more than hitting the next growth milestone. The company told CNBC that preserving that status is a priority, a statement that reframes the debate over its sprawling AI buildout from order-book bravado to balance-sheet arithmetic.

The remark lands against a tougher backdrop than the headline growth figures suggest. S&P downgraded Oracle to BBB? in July, leaving it on the lowest rung of investment grade. That makes the reaffirmed priority something more than reassuring language — Oracle has named the yardstick by which its expansion will now be judged.

Debt, Leases and the Line S&P Has Drawn

FactSet data put Oracle's long-term debt at $118 billion, alongside $39 billion in capital and lease obligations. The distinction matters: anyone weighing the company's financial load has to look past conventional loans and bonds. S&P has spelled out what could trigger a further cut — sustained leverage above four times, or the absence of a credible path to positive free operating cash flow by fiscal 2029. A more negative view of the AI strategy or the sector could also weigh in.

Consensus forecasts point to negative free cash flow of $23.7 billion this year and $36 billion in 2027. Those are projections rather than reported results, but they show why financing cannot be treated as a footnote. Oracle plans capital spending of $90–95 billion for the fiscal year, while management expects at least $90 billion in revenue for fiscal 2027.

Demand Is Not the Weak Link

Operationally, Oracle has plenty to support the buildout. Revenue reached $19.345 billion in the first quarter of fiscal 2027, up 29.6%, with adjusted earnings per share of $1.92. Cloud infrastructure revenue climbed 121% in the same quarter, and the order backlog stood at $664 billion.

Those are powerful demand signals — but not yet spendable cash. Contractual commitments can underpin future utilization; they do not automatically pay for the investment required today. Between signing a contract and collecting revenue sits the task of building capacity and carrying its financing.

A Power Deal Still on the Drawing Board

The same tension runs through Oracle's energy arrangements for Project Jupiter. On October 13 the company proposed a service agreement with El Paso Electric carrying power outlays of at least $100 million over ten years. The arrangement is proposed, not signed.

A separate piece of the plan would shield other electricity customers from the cost impact. Oracle floated credits of up to $55 million a year, potentially granted for as long as three years. Both the amount and the duration are ceilings, not committed payments, and this element targets the economic spillover of energy supply rather than data-center capacity itself.

Local approval is no formality. According to media reports, members of the county commission questioned the proposal, judging that it could require an adjustment to the project-related Industrial Revenue Bonds. That is a possible need for change, not a decision already taken.

Oracle also puts a number on the project's expected economic contribution. An updated analysis published October 13 projects an effect of more than $13 billion for New Mexico, spanning the construction phase and the first 17 years of full operation. At full capacity, Project Jupiter is expected to support more than 3,100 jobs in the county. Both figures describe expectations, not results already achieved — and a regional economic effect is neither company revenue nor profit, so it offers investors no direct read on returns on deployed capital.

What Shareholders Should Actually Watch

The proposed customer credits and the projected regional benefit sit alongside the financing questions, not in place of them. Reuters reported on Thursday about concerns over borrowing costs, heavy indebtedness and whether the investments will pay off — the same conflict the rating priority exposes.

The most convincing promise, then, is not the largest possible buildout but one whose funding holds up even if projects or receipts slip. Growth still counts. It now has to come with a traceable financing plan and a route to positive cash flow. Oracle's statement deserves attention precisely because it sounds less spectacular than a fresh AI contract: it sets a testable claim that expansion must not crowd out creditworthiness.