Oracle Market Analysis 2026 – $638B Backlog

Oracle Market Analysis

Oracle’s Backlog Just Hit a Record $638 Billion — Its Stock Is Down 64% Anyway, and a Credit Downgrade Explains Why

Key Questions

Oracle’s backlog grew 363% to a record $638 billion — so why is the stock down 64% from its high

Worth resolving this apparent contradiction directly, since both numbers come from the same earnings report and neither is a misprint. Oracle closed fiscal 2026 (the year ended May 31, 2026) with $638 billion in Remaining Performance Obligations (RPO — contracted future revenue not yet recognized), up 363% from $138 billion a year earlier, after adding $85 billion in the fourth quarter alone. Full-year revenue reached a record $67.4 billion, up 17%, and non-GAAP earnings per share rose 27% to $7.63. Here’s the reconciliation: building the data-center capacity behind that backlog cost $55.7 billion in capital expenditures for the year (up 162% and roughly 83% of revenue), which turned a genuinely strong $32 billion in operating cash flow into a negative $23.7 billion free-cash-flow position. So what does this mean in practice: Oracle’s stock isn’t falling because demand is fake — the $638 billion figure is a real, signed number — it’s falling because the market is pricing the cost, financing, and concentration risk of actually building out the capacity to deliver on it, not the headline backlog itself. Shares have fallen roughly 64% since closing at an all-time high near $325 in September 2025, a decline that briefly took Oracle’s market capitalization from above $875 billion to about $345 billion.

How much of that $638 billion depends on one customer, and can that customer pay for it

Worth separating the backlog’s scale from its concentration, since this is the specific point Oracle’s own credit rating agency flagged. A five-year cloud-computing agreement with OpenAI, reported by the Wall Street Journal at roughly $300 billion starting in 2027, accounts for an estimated 47% to more than 50% of Oracle’s total RPO, according to Bank of America and other analysts (Oracle itself does not disclose the customer split). Once fully ramped, that single contract implies close to $60 billion a year in obligations — more than OpenAI’s entire current business generates. Here’s the client financial health check this deserves: OpenAI is not a publicly traded, self-funding, profitable customer the way Meta or Microsoft would be. It is a private company whose reported revenue run-rate has ranged from roughly $10 billion to $25 billion depending on source and timing (Reuters cited a ~$25 billion figure in early 2026; a Financial Times-confirmed leak put actual 2025 sales at $13 billion against a $21 billion loss), and whose CEO Sam Altman has told investors the company does not expect profitability before 2029, with cumulative losses projected near $44 billion before then. Its spending is underwritten by outside investors, including Microsoft, SoftBank, Nvidia, and Amazon, rather than its own operating cash flow. What this means in practice: Oracle’s own fiscal 2026 annual report explicitly warns that some large backlog customers “may be highly leveraged and subject to their own operating and regulatory risks,” creating a real risk of non-payment — language that is analytically relevant to exactly this kind of exposure rather than mere boilerplate.

Oracle just got its credit rating cut and plans to raise another $40 billion — what happens at the September 14 earnings report

Worth being precise about the financing mechanics rather than a general “Oracle is taking on debt” framing. S&P Global Ratings downgraded Oracle’s long-term issuer credit rating from BBB to BBB- on July 9, 2026, explicitly citing OpenAI-related concentration risk as a factor — a rare instance of a rating agency naming a specific counterparty as the reason for a downgrade of an investment-grade technology company. Total debt now approaches $130 billion, following an $18 billion bond sale in September 2025 and a further $20-25 billion, eight-part bond offering in February 2026, together with equity-linked issuance, as part of a roughly $45-50 billion 2026 funding plan. Oracle has flagged plans to raise an additional $40 billion in fiscal 2027 to keep funding the buildout. So what does this mean in practice: equity markets initially shrugged off the July 9 downgrade — the stock barely moved that day — while credit markets have been considerably more sensitive, with Oracle’s bonds trading 60-80 basis points wider than issuance at points this year. Oracle reports its first fiscal 2027 quarter on Wednesday, September 14, 2026, after market close, against its own guidance of 27-29% revenue growth and non-GAAP EPS of $1.72-$1.76 for the quarter. That report is the next concrete test of whether the RPO conversion Oracle has guided to — roughly 12% of the backlog into revenue over the next 12 months — is tracking, or slipping.

This week at a glance:

• Stock price: ↑ $119.90 (July 27 close), up 4.27% that session, rebounding off a 52-week low set July 24
• 52-week range: ↓↓ $114.50-$345.72, down roughly 64% from the September 2025 high
• RPO backlog: ↑↑ $638 billion, up 363% year-over-year, of which roughly half is tied to a single OpenAI contract
• Credit rating: ↓ S&P cut to BBB- from BBB on July 9, citing OpenAI concentration explicitly
• Free cash flow: ↓↓ negative $23.7 billion for fiscal 2026, on $55.7 billion of capital expenditure
• Next earnings: ➖ Wednesday, September 14, 2026, after market close (Q1 fiscal 2027)

Key Facts

Price & Valuation

• Current price: $119.90 (July 27, 2026 close); market capitalization approximately $345 billion, down from a peak above $875 billion last September
• 52-week range: $114.50-$345.72 — the low set July 24, 2026, the high set in September 2025 when shares closed at an all-time high near $325
• Trailing P/E: roughly 19-21x on trailing GAAP earnings (data providers differ slightly); roughly 15.7x measured against fiscal 2026’s non-GAAP EPS of $7.63
• Forward P/E: approximately 14.9x against Oracle’s own fiscal 2027 non-GAAP EPS guidance of $8.05 — meaningfully below the trailing figure, but only cheap if that guidance holds
• Dividend: $0.50 per share quarterly (roughly 1.6-1.7% yield); ex-dividend date July 10, 2026
• Trading volume: approximately 37 million shares, in line with recent averages

Fiscal 2026 Results (Reported June 2026, Year Ended May 31, 2026)

• Total revenue: $67.4 billion, up 17% year-over-year, a company record
• Q4 revenue: $19.2 billion, up 21%; Q4 cloud revenue $9.9 billion, up 47%; Q4 Cloud Infrastructure (IaaS) $5.8 billion, up 93%; Cloud Applications (SaaS) $4.1 billion, up 10%
• Non-GAAP EPS: $7.63 for the full year, up 27%; Q4 non-GAAP EPS $2.11, up 24%
• Operating cash flow: $32 billion, up 54% year-over-year
• Free cash flow: negative $23.7 billion, driven by capital expenditures of $55.7 billion (up 162%, roughly 83% of revenue)
• Remaining Performance Obligations (RPO): $638 billion, up 363% year-over-year, after growing $85 billion in Q4 alone (from $553 billion to $638 billion)
• Gross margin declined roughly 5 points year-over-year as the infrastructure buildout scaled

Fiscal 2027 Guidance (Oracle’s Own, Issued With Q4 Results)

• Full-year revenue: approximately $90 billion, up 34% in constant currency
• Full-year non-GAAP EPS: $8.05, up 18%
• Q1 fiscal 2027 (quarter ended August 31, 2026): revenue growth guided at 27-29%; cloud revenue growth 58-64%; non-GAAP EPS $1.72-$1.76
• Long-term outlook: management has guided to a 31% revenue CAGR and 28% non-GAAP EPS CAGR from fiscal 2026 through fiscal 2030
• Next earnings: Wednesday, September 14, 2026, after market close

Backlog & Customer Concentration

• RPO conversion guidance (Oracle’s own risk disclosure): approximately 12% converts to revenue over the next 12 months, a further 34% over the following 24 months
• OpenAI’s five-year cloud-computing contract, reported at roughly $300 billion starting in 2027, is estimated to represent 47% to more than 50% of Oracle’s total RPO (Bank of America, other analyst estimates; Oracle does not disclose the customer split)
• Once fully ramped, the OpenAI contract implies close to $60 billion a year in obligations — more than OpenAI’s entire current revenue
• OpenAI’s reported financials: revenue run-rate estimates range from roughly $10 billion to $25 billion depending on source and timing; a Financial Times-confirmed leak put 2025 sales at $13 billion against a $21 billion loss; CEO Sam Altman has said the company does not expect profitability before 2029, with cumulative losses projected near $44 billion before then
• OpenAI is privately held, not publicly traded, and its buildout spending is backed by outside investors including Microsoft, SoftBank, Nvidia, and Amazon

Financing & Credit

• S&P Global Ratings downgraded Oracle’s long-term issuer credit rating from BBB to BBB- on July 9, 2026, explicitly citing OpenAI-related concentration risk
• Total debt approaching $130 billion (more than $122 billion long-term)
• Bond issuance: $18 billion (six-part, September 2025), followed by $20-25 billion (eight-part, 3-40 year tenors, February 2026), as part of a roughly $45-50 billion 2026 funding plan split about evenly between debt and equity-linked issuance
• Oracle has flagged plans to raise an additional $40 billion (debt and equity combined) in fiscal 2027

Recent Contract Wins & Other Context

• A five-year U.S. Department of War (Navy) enterprise software IDIQ contract, awarded July 23, 2026, with a $3.31 billion base value and options that could lift the total to approximately $7 billion — a diversification data point outside the AI-cloud concentration story
• Separately, Oracle co-founder and Chairman Larry Ellison has personally guaranteed roughly $40.4 billion tied to a media deal involving his son, which a number of U.S. states are now suing to block — a personal financial matter distinct from Oracle’s own balance sheet, though worth noting given Ellison’s central role at the company

Oracle’s setup heading into September is a genuine test of which side of this ledger investors weigh more heavily: a record, real backlog and the fastest cloud infrastructure growth rate among the major providers, or a credit downgrade, a widening cash deficit, and a customer concentration risk large enough that Oracle’s own regulatory filings warn about it directly. Both sides are backed by specific, dated facts rather than speculation.

The live chart below reflects current ORCL share price action in real time.

A Record Backlog Should Mean a Resolved Demand Debate — Here’s the Honest Complication

The instinctive read on Oracle’s $638 billion RPO figure is that it settles the question of whether AI cloud demand is real. The data says something more specific: the backlog is a genuine, signed number, but Oracle’s own guidance says only about 12% of it converts to recognized revenue in the next 12 months, with the bulk landing between 2027 and 2030. So what does this mean in practice: a $638 billion headline is measuring committed future demand, not current cash flow — and the gap between those two things is exactly what’s costing $55.7 billion a year to build and pushing free cash flow $23.7 billion negative right now. The client concentration point deserves its own weight here, not just a passing mention: because roughly half the backlog sits with one privately held, currently unprofitable counterparty, the conversion percentage isn’t just a timing question — it’s also a credit question. A comparable backlog concentrated among Meta, Microsoft, and Google (all profitable, investment-grade, and self-funding) would carry a fundamentally different risk profile than one nearly half-dependent on a company whose own CEO has told investors not to expect profitability until 2029.

The GPU supply side deserves more than one sentence too: Oracle’s ability to convert backlog into revenue depends on physically securing enough GPU capacity — chiefly from Nvidia — to build the data centers its contracts promise, and this is an industry-wide bottleneck rather than an Oracle-specific execution failure. Every major cloud provider expanding AI infrastructure faces the same chip-allocation constraint; Oracle’s newer entrant status in the hyperscaler tier means it is competing against AWS, Azure, and Google Cloud for the same limited supply, on top of needing the power and physical construction capacity to house it. Oracle’s own annual report goes further, warning it could end up overestimating demand and be unable to repurpose data-center capacity built for one customer if that customer’s needs change — a genuine operational risk sitting alongside the financing risk, not a separate story.

Worth putting Oracle’s growth rate in context against its larger cloud rivals, without overstating the comparison: Oracle’s Cloud Infrastructure (IaaS) revenue grew 93% year-over-year last quarter to $5.8 billion, faster than Google Cloud’s 82% growth to $24.8 billion (Q2 2026) and faster than Azure’s roughly 40% growth reported for the same period. Oracle is genuinely growing faster in percentage terms — but from a base roughly a quarter the size of Google Cloud’s and a fraction of Azure’s, meaning the comparison reflects a smaller company scaling quickly, not yet a change in the competitive pecking order by absolute revenue.

Valuation Context

Oracle’s headline trailing P/E of roughly 19-21x differs materially from the picture produced by non-GAAP and forward earnings, where the dated figures in this article imply multiples in the mid-teens. The lower forward multiple is only informative if backlog converts roughly as scheduled while Oracle manages a large capital-spending program, negative free cash flow, customer concentration and substantial debt without further deterioration in credit quality.

Third-party analyst targets remain far above the dated market price used in this article, but MatrixPro24 treats those targets only as evidence of external expectations and potential model lag after the credit downgrade. They are not a MatrixPro24 price target, expected-return forecast, or recommendation.

Current Market Data

Oracle trades on the NYSE under ticker ORCL. As of the most recent confirmed session, July 27, 2026, shares closed at $119.90, up 4.27% on the day, within a 52-week range of $114.50-$345.72 — down roughly 64% from the September 2025 high — on volume of approximately 37 million shares. Market capitalization stands at approximately $345 billion. Oracle reports Q1 fiscal 2027 earnings on Wednesday, September 14, 2026, after market close. The live chart below reflects current price action.

Sources

• Oracle Investor Relations — Oracle Announces Record Q4 and FY 2026 Results
• EBC Financial Group — Why Is Oracle Stock Falling Despite Its $638B AI Backlog?
• The Motley Fool — Oracle Stock Is Down 64% in 10 Months: Buying Opportunity or Falling Knife?
• Yahoo Finance (The Motley Fool) — Oracle Stock Plummeted, and This Dire AI Warning Might Be Why
• TheStreet — Oracle’s $638 Billion Promise Comes With One Big Catch
• Morningstar — Oracle Corp (ORCL) Stock Quote & Analysis
• TipRanks — Oracle (ORCL) Earnings Date & History
• Benzinga — What’s Going On With Oracle Stock Tuesday?

Oracle’s OpenAI concentration risk connects directly to Microsoft, which discloses that roughly 45% of its own commercial RPO balance likewise comes from OpenAI, and to Nvidia, whose GPU supply constraints directly gate how quickly Oracle can convert its backlog into delivered, revenue-generating capacity.

About MP24 Analyst X

Published by MP24 Analyst X. Read our Editorial and Content Policy to understand our compliance and brand publishing standards.

MP24 Analyst X is the public-facing pseudonym used for MatrixPro24’s research and editorial work across macroeconomics, commodities, cryptocurrencies, equities, and global financial markets. MatrixPro24 research emphasizes evidence, transparency, and structured reasoning over speculation and market hype.

Disclaimer

This analysis is for informational and educational purposes only and does not constitute personalized financial or investment advice or a MatrixPro24 recommendation to buy, sell, or hold a financial instrument. This editorial/methodology review was completed on August 13, 2026. Market-price figures and dated market reactions remain tied to the observation dates stated in the article; use the live TradingView chart for current market pricing. Past performance is not indicative of future results. Conduct your own independent research before making financial decisions.