Updated 9 October 2026. Oracle (NYSE: ORCL) closed at $135.69 on Thursday 8 October, down 5.48 percent, after the Financial Times reported that OpenAI’s annualised revenue was approaching $50 billion at the end of September, roughly $20 billion below the $68-70 billion figure that circulated in late September (StockAnalysis, FT). The stock traded as low as $134.66 on volume of 42.4 million shares, more than twice the previous day’s. Verdict: the $50 billion number is a correction of a figure OpenAI never published, not a miss, but Oracle is the most OpenAI-exposed large cap on the market and the same day brought a report of new chip-financing talks and a reminder that long-term debt now exceeds $160 billion. The bear case is the Street’s low target of $110; the base case is Dan Ives’s new $155 Hold; the bull case is Stifel’s $200.
Key facts
- Share price: $135.69 close on 8 October (-5.48%), from $143.56 on 7 October. Day range $134.66-$142.78 (StockAnalysis). Oracle is down nearly 30 percent in 2026 (Yahoo Finance).
- The OpenAI number: the FT reported on 8 October that OpenAI told investors annualised revenue was approaching $50 billion at end-September. Figures of $68-70 billion had been widely reported late in the month.
- The explanation: a person familiar with the documents told CNBC the higher figure did not come from OpenAI and probably arose from comparisons with Anthropic, whose reported number includes gross revenue booked through cloud partners; OpenAI’s figure is net revenue.
- Market reaction: Nvidia fell 2.9 percent, Broadcom nearly 4 percent, Microsoft more than 1 percent and CoreWeave about 8 percent, with the Nasdaq down 1.25 percent to 27,193 (CNBC, CNN).
- Financing talks: the Wall Street Journal reported the same day that Oracle is in talks with Apollo and Goldman Sachs on a structure in which outside investors fund a separate company that buys chips and leases them to Oracle, with a deal targeted as soon as this year (as reported by Yahoo Finance).
- Balance sheet: long-term debt has nearly doubled in two years to more than $160 billion, making Oracle the fifth-largest borrower in the US corporate bond market; capital spending was $28.5 billion in the June-August quarter against $2.3 billion two years earlier (Bloomberg, via Yahoo Finance).
- Backlog: remaining performance obligations of $664 billion at the end of August. The five-year, $300 billion OpenAI cloud agreement reported in September 2025 accounts for roughly 45 percent of it.
- Analysts: 43 analysts tracked by S&P Global average $237.97 with a range of $110 to $400; Yorkville Ives & Co initiated at Hold with a $155 target on 7 October; Stifel $200 and Bernstein $325, both maintained 25 September (StockAnalysis).
What the FT reported, and what it does not mean
OpenAI’s own figure, as reported by the FT, is that annualised revenue was approaching $50 billion at the end of September. That is a large and fast-growing number. The problem is the comparison: for two weeks the market had been working with $68-70 billion, which several outlets, including the FT, had reported. A $20 billion gap between what investors thought and what the company told them is what moved the AI complex on 8 October.
The gap is a definitional one, according to the explanation given to CNBC. The higher number appears to have been built to compare OpenAI with Anthropic, whose headline revenue includes gross sales made through cloud providers. OpenAI’s $50 billion is net. Nothing in the reporting says OpenAI lost sales, cut prices or missed a target. FinanceFeeds is therefore not describing this as a miss; it is a correction of the denominator the market was using to value OpenAI’s suppliers.
For Oracle that distinction is cold comfort. The company’s $664 billion backlog is the largest in software, and a contract reported at $300 billion over five years with OpenAI is close to half of it. If investors recalibrate OpenAI’s ability to pay from a $70 billion run-rate to a $50 billion one, every dollar of that backlog is discounted a little harder. That is the whole trade in one sentence.
The debt is the real story
The same day’s second headline was financing. Oracle is reportedly in talks with Apollo and Goldman Sachs on an off-balance-sheet structure: investors fund a special-purpose company that buys the chips, and Oracle leases them back over time instead of borrowing to buy them outright. Broadcom and SpaceX were reported to be pursuing similar deals. The attraction for Oracle is obvious. Bloomberg puts its long-term debt above $160 billion, almost double the level of two years ago, and capital spending in the latest quarter was $28.5 billion. In February the company said it expected to raise $45-50 billion of gross proceeds across equity and debt during calendar 2026.
FinanceFeeds has covered the arithmetic before: the first-quarter results showed free cash flow turning negative as capex ran ahead of operating cash, and the backlog-and-debt analysis explained why the bond market would not price the new paper at par. A chip-leasing vehicle does not change the amount of capital the build-out consumes; it changes who holds the risk and at what price.
Why the stock fell 5.5 percent and not 15
Three reasons. First, Oracle had already de-rated: the shares are down nearly 30 percent this year and were trading at $138 on 1 October before the Tencent compute deal lifted them briefly, as FinanceFeeds reported on 2 October. Second, the OpenAI figure is a revision of a rumour rather than a disclosure by a customer that it will spend less. Third, the backlog is contractual; the question is collection risk over five years, not demand today.
The analyst sheet reflects that split. The consensus target of $237.97 is 75 percent above the current price, but the dispersion is extreme: Bernstein at $325, Stifel at $200, and a new Hold initiation from Dan Ives at Yorkville Ives & Co at $155 on 7 October, the day before the FT report. The Street low is $110.
ORCL stock scenarios: bear, base and bull
All percentages are measured from the 8 October close of $135.69.
| Scenario | ORCL level | What has to happen | Anchor |
|---|---|---|---|
| Bear | ~$110 (-18.9%) | OpenAI’s funding or spending plans are scaled back, the chip-leasing deal prices at a visibly high cost, and the December quarter shows RPO growth slowing while capex stays near $30 billion. | Lowest Street target (S&P Global, via StockAnalysis) |
| Base | ~$155 (+14.2%) | The Apollo-Goldman structure closes on reasonable terms, OpenAI’s revenue keeps compounding from the $50 billion base, and fiscal Q2 results in December meet the 30-34 percent revenue-growth guide. | Yorkville Ives & Co Hold initiation, 7 October |
| Bull | ~$200 (+47.4%) | Cloud revenue growth holds in the 60-70 percent range, free cash flow turns up as customer prepayments offset capex, and the OpenAI run-rate story is forgotten by the next print. | Stifel Buy target, 25 September |
The bear case is a genuine downside: $110 sits 19 percent below the current price and close to the levels Oracle traded at in the spring. Even so, note that the base case is only a Hold-rated target; the market’s own view, expressed by the average, remains far more optimistic than the price.
What to watch next
- OpenAI’s own numbers. Any direct statement from OpenAI on run-rate revenue or on its compute commitments would reset the debate; so would an updated FT figure.
- The chip-leasing deal. Size, lease rate and whether it is consolidated. A deal “as soon as this year” means the terms could be public before Oracle reports.
- Fiscal Q2 results. Oracle has not confirmed the date; its last two second-quarter reports came in the second week of December. Guidance was for revenue growth of 30-34 percent and cloud growth of 65-71 percent.
- Bond spreads. With long-term debt above $160 billion, Oracle’s credit is now a market signal in its own right; wider spreads on the 2026 paper would lead the equity lower.
Quick take: ORCL at $135.69 is being marked down for a number its biggest customer never actually published. The real exposure is not whether OpenAI makes $50 billion or $70 billion this year but whether it can keep funding a $300 billion commitment while Oracle borrows to build the capacity. The December quarter and the terms of the chip-leasing vehicle will answer that better than any revenue leak.
FAQ
Why did Oracle stock fall on 8 October 2026?
The Financial Times reported that OpenAI’s annualised revenue was approaching $50 billion at end-September, about $20 billion below figures that circulated in late September. Oracle, whose $664 billion backlog includes a reported $300 billion OpenAI contract, fell 5.48 percent to $135.69. A Wall Street Journal report on chip-financing talks with Apollo and Goldman Sachs added to the pressure.
Did OpenAI miss its revenue target?
No reporting says so. According to a source cited by CNBC, the $68-70 billion figure did not come from OpenAI and probably reflected a gross-revenue comparison with Anthropic. OpenAI’s $50 billion is a net figure. The correction changed the market’s assumption, not OpenAI’s business.
How exposed is Oracle to OpenAI?
The five-year, $300 billion cloud agreement reported in September 2025 represents roughly 45 percent of Oracle’s $664 billion of remaining performance obligations. OpenAI is one of its largest cloud customers.
What is the analyst price target for ORCL?
StockAnalysis shows 43 analysts with an average target of $237.97, a low of $110 and a high of $400 as of 8 October 2026. Recent calls include Yorkville Ives & Co at $155 (Hold, 7 October), Stifel at $200 and Bernstein at $325 (both Buy, 25 September).
How much debt does Oracle have?
Bloomberg puts long-term debt above $160 billion, nearly double the level two years ago, making Oracle the fifth-largest borrower in the US corporate bond market. The company said in February it planned to raise $45-50 billion of gross proceeds in calendar 2026.
When does Oracle report next?
Oracle has not announced the date for its fiscal second quarter, which runs September to November. Its two previous second-quarter reports came in the second week of December; the company typically confirms the date about a week in advance.
Is $110 a realistic bear case for Oracle?
It is the lowest published 12-month target and sits about 19 percent below the current price. Reaching it would require OpenAI’s spending plans to be cut or the chip-financing deal to price at a cost that signals credit stress, together with slowing backlog growth in the December quarter.
Related coverage
- Oracle stock and the $7 billion Tencent AI compute deal
- Oracle stock price prediction: $205 bull case, $98 bear case
- Oracle stock after Q1: cloud revenue up 62%, RPO $664 billion
- Nebius, CoreWeave and Nvidia: GPU prices up 21 percent
Sources: Financial Times, 8 October 2026 (OpenAI annualised revenue approaching $50 billion); CNBC, 8 October 2026 (explanation of the gap, Nvidia, CoreWeave and index moves); CNN, 8 October 2026 (Broadcom, Microsoft, Intel moves); Wall Street Journal via Yahoo Finance, 8-9 October 2026 (Apollo and Goldman Sachs chip-financing talks); Bloomberg via Yahoo Finance (long-term debt, capital spending, bond-market rank, year-to-date performance); Oracle Corporation Q1 FY2027 results, 10 September 2026 (RPO, capex, guidance); Oracle Corporation financing plan announcement, February 2026; StockAnalysis (daily prices 1-8 October 2026, analyst ratings and targets as of 8 October 2026).
This article is for information only and is not investment advice. Share prices move continuously and the figures above were accurate at the time of writing. Nothing here is a recommendation to buy or sell any security. Do your own research and consider your own circumstances before investing.
