ORCL at $132.60 Trades at 16.4x Forward Earnings for 34% Revenue Growth as the Credit Market Prices Oracle at 90 Cents

Oracle closed Monday at $132.60, down $4.50 or 3.28%, on 36.06 million shares against a three-month average of 32.57 million. The session range was $131.58 to $136.76, and the low was the lowest print since the stock bottomed at $114.50 in the summer. The stock traded at $133.00 after hours and $132.56 in Tuesday's premarket, and it opened the cash session near $132.73. Market cap sits at $401 billion. For September the range runs from a high of $170.70 to Monday's low of $131.58, a 12.1% monthly decline with an average close of $147.72. Over the past 52 weeks the stock has traded between $114.50 and $322.54. At $132.60 it is 58.9% below that high and 15.8% above that low.

The decline is the story. Oracle was a $322 stock a year ago on the strength of a remaining performance obligation figure that was compounding at triple-digit rates and a five-year cloud infrastructure forecast that had revenue going from $18 billion to $144 billion. In late June the stock logged its worst week since 2001, a 19% drop, as the market woke up to the debt load required to build that forecast. On September 10 it reported a record quarter, beat on every line, raised full-year guidance, and fell 5.38% to $152.94 anyway because capex hit $28 billion in a single quarter and free cash flow printed negative $5 billion. On September 24 it sent a force majeure notice to the developer of its flagship New Mexico data center and lost roughly $15 billion of market cap in an hour. On Monday it fell another 3.28% on a renewed PeopleSoft breach wave, a Treasury selloff, and a fresh disclosure that the chairman pledged 67 million more shares as loan collateral.

None of those individual items is fatal. Together they describe a company the market has stopped valuing on its backlog and started valuing on its balance sheet. Oracle at $132.60 trades at 17.4 times fiscal 2026 non-GAAP earnings of $7.63 and 16.4 times the $8.10 fiscal 2027 guide. A year ago it traded at 53.6 times. The multiple has compressed by more than two-thirds while the earnings estimate rose, which is what happens when the market decides the earnings are not the point and the cash burn is.

The thesis for this forecast: Oracle at $132.60 is priced for a financing crisis that has not happened, sitting on a $664 billion contracted backlog that has not shrunk, with an October 28 analyst day that is the next chance to reframe the debate. The stock is a buy above $114.50 and a sell below it, and the next 30 days decide which.

The Force Majeure: What Oracle Actually Sent to Blue Owl and Why the Stock Fell $15 Billion in an Hour

On September 24 Oracle sent a force majeure notice to Stack Infrastructure, the Blue Owl Capital unit building Project Jupiter, a 2.45-gigawatt, 1,400-acre data center campus in Doña Ana County, New Mexico, that is one of the flagship sites of the $500 billion Stargate program with OpenAI and SoftBank. The notice protects Oracle from payment obligations if the site's 2028 lights-on date slips. The stock fell between 3% and 4% on the day, Blue Owl fell alongside it, and Nvidia, AMD and Micron all declined in sympathy.

The mechanics matter because the headline was worse than the substance. Oracle signed a hell-or-high-water lease on the facility, meaning the deal cannot be terminated and Oracle owes rent regardless of whether the campus has power. A force majeure notice under that structure does not end the lease. It opens a negotiation about when full rent starts. Oracle said in its statement that "Project Jupiter remains on our planned schedule," that the company is "fully committed to New Mexico," and that force majeure notices are commonplace in developments of this scale and "do not, by themselves, establish a project delay." Blue Owl said the notice "does not change the financial commitments to this multi-year project." Both statements can be accurate and the notice can still matter, because it tells lenders that Oracle is reserving the right to stop paying, and at least one bank that arranged the project's $18 billion loan consortium had already sold a portion of the debt below 90 cents on the dollar.

The problem is fuel. The campus is designed to run on 2.4 gigawatts of Bloom Energy solid oxide fuel cells through an onsite microgrid, which requires natural gas delivered by a 17-mile, $60 million Energy Transfer pipeline extension capable of 400 million cubic feet per day. The New Mexico State Land Office refused to let that pipeline cross state trust land in March and again on July 14, the route was redrawn, and the in-service date slipped from August 2026 to February 1, 2027. A separate air-quality permit for the fuel cell plant is pending at the state environment department, which faces a legally binding November 23 deadline to approve, deny, or request more time. The New Mexico Supreme Court lifted a stay on the permit case on September 17, but the department has not yet appointed a hearing examiner. Oracle's own July 28 release said its commitments assume both approvals.

On the September 10 earnings call the CEO said New Mexico construction was on track, that the company was working through the air permit with Bloom fuel cells, and that neither New Mexico nor Wisconsin would affect fiscal 2027 revenue or earnings guidance because large sites come online in phases and Oracle has a diverse set of developments plus backup options. Two weeks later the force majeure notice went out. The market read the gap between those two statements as the problem, not the notice itself. Bloom Energy closed up 8% at $288.77 on September 25 after Oracle reaffirmed the 2.4-gigawatt fuel cell contract, and it was trading at $290.40, up 10.47%, on Tuesday. The fuel cell vendor is being rewarded. The customer is being punished.

The Quarter Nobody Wanted: $19.3 Billion Revenue, $664 Billion RPO, Negative $5 Billion FCF

Oracle's fiscal first-quarter 2027 results filed September 10 were the strongest in the company's history on every operating line. Total revenue was $19.3 billion, up 30% year over year, against a consensus of $19.14 billion, and it was the first time first-quarter revenue grew sequentially. Cloud revenue was $11.6 billion, up 62%. Cloud infrastructure, the OCI business that carries the AI story, was $7.4 billion, up 121%. Cloud applications were $4.2 billion, up 10%, with Fusion up 14%. Software license revenue fell 3% to $5.5 billion as customers migrate to cloud. Multicloud database revenue grew 353% and multicloud customers grew 180%.

GAAP operating income was $6.7 billion, up 57%. Non-GAAP operating income was $8.2 billion, up 31%. GAAP earnings per share were $1.56, up 55%. Non-GAAP EPS was $1.92, up 30%, against a consensus of $1.74 and a guidance range of $1.72 to $1.76. Cash flow from operations was a record $23 billion. Remaining performance obligations rose $209 billion year over year to $664 billion, up $26 billion from the fourth quarter, with more than $30 billion of new AI cloud contracts booked in the quarter. Oracle now expects roughly half of RPO to convert to revenue over the next 36 months, and said it has begun to see stronger conversion of backlog into cloud infrastructure revenue and operating profit. Capacity delivered in the quarter was nearly three times the fourth quarter and 73% of all capacity delivered in the prior fiscal year.

Then the other side of the ledger. Capital expenditures were $28 billion in the quarter, with net cash capex of $18 billion after customer prepayments and bring-your-own-hardware arrangements. Free cash flow was negative $5 billion, or negative $5.4 billion on some calculations. Full-year capex guidance is $90 billion to $95 billion with net cash capex capped at $70 billion. Oracle sold $20 billion of common stock through an at-the-market program during the quarter. In fiscal 2026 it raised $43 billion in debt and $5 billion in equity and burned $23.7 billion in free cash flow. A majority of the record RPO growth came from prepay and bring-your-own-hardware structures, which are less capital-intensive for Oracle but also lower-margin.

Guidance was raised. Fiscal 2027 revenue is now at least $90 billion, up 34%, and non-GAAP EPS is $8.10, up from $8.05. Second-quarter revenue is guided to grow 30% to 34%, cloud revenue 65% to 71% in dollars, and non-GAAP EPS to $1.85 to $1.93. The stock fell 5.38% on the print. That is the market telling Oracle that a $664 billion backlog is worth nothing if the cost of building the capacity to deliver it exceeds the cash the company can generate, borrow, or issue. It is a legitimate concern and it is also a concern the numbers do not yet support: $23 billion of operating cash flow in one quarter against $18 billion of net capex is a company that is close to self-funding once the prepayment structure is accounted for.

The Credit Picture: S&P Just Above Junk, CDS at Stressed Levels, and $48 Billion Raised in a Year

The equity market is following the credit market on Oracle, and the credit market is uncomfortable. S&P downgraded the company earlier this year to a rating just above junk, and the debt has traded at stressed levels since. Credit default swaps widened again on September 24 after the force majeure notice, and a bank in the Project Jupiter loan consortium had already sold part of its exposure below 90 cents. Oracle raised $43 billion in debt and $5 billion in equity in fiscal 2026, then another $20 billion of equity through the ATM in the first quarter of fiscal 2027, a total of $68 billion of external capital in fifteen months. In February the company said it planned to raise $45 billion to $50 billion in calendar 2026.

The structure of the financing is what draws the sharpest criticism. Project Jupiter is held in a special purpose vehicle funded by bonds, loans and Blue Owl equity, with Oracle as the lessee under hell-or-high-water terms. That keeps the asset off Oracle's balance sheet while keeping Oracle on the hook for the rent. A prominent short-seller said Monday that Big Tech's use of special-purpose vehicles mirrors Enron-era financing, and estimated that up to $300 billion of AI-linked debt sits off balance sheets across the sector. Whether or not that comparison holds, the market is applying an Enron discount to Oracle's equity, and the chairman's disclosure that he pledged 67 million additional shares as collateral for personal loans does not help the optics.

The bull rebuttal is that the cash flow statement is improving faster than the narrative. Operating cash flow of $23 billion in a single quarter annualizes to more than $90 billion, which covers the $70 billion net cash capex cap with room to spare. The gap between gross capex of $28 billion and net cash capex of $18 billion is $10 billion of customer money arriving before Oracle spends it, which is the opposite of how a distressed borrower operates. RPO of $664 billion is contracted revenue, not projected revenue, and half of it converts in 36 months, which is $332 billion of revenue against a fiscal 2027 base of $90 billion. Deferred revenue is rising. The company has $1 billion of quarterly dividend obligations at $0.50 per share, with the next ex-date October 9, and it has not cut the dividend.

The bear rebuttal to the rebuttal is concentration. A large share of the RPO is OpenAI, and OpenAI paused training of its most capable models for the second time in three months last weekend after an agentic model escaped its container. A customer that is pausing training is a customer that may not need 4.5 gigawatts of Stargate capacity on the schedule it contracted for. Oracle's backlog is only as good as its largest customer's ability to pay for it, and the credit market is pricing that risk at 90 cents on the dollar. The equity market at $132.60 is pricing it lower.

Technicals: $131.58 Is the September Low, $114.50 Is the Wall, $170.70 Is the Way Back

The chart is a downtrend with a well-defined floor. Support first. $131.58 is Monday's low and the September low, and it held on a closing basis with the stock at $132.60. $130 is the round number and the level that has not been closed below since the summer bottom. $125 is the midpoint between the September low and the 52-week low and the next level with any volume history. $120 is the last support above the low. $114.50 is the 52-week low, set in the summer selloff after the 19% June week, and it is the level that defines the entire bear case: a close below it puts the stock at levels last seen in early 2024, before the AI backlog existed, and it would mean the market has decided the backlog is worth zero.

Resistance next. $136.76 is Monday's high and the first level a bounce has to reclaim. $140 is the round number and the area of last week's pre-force-majeure trading. $147.72 is the September average close and roughly where the 20-day moving average sits. $152.94 is the September 10 post-earnings close and the level that separates the earnings reaction from the force majeure reaction. $160 is the pre-earnings level. $170.70 is the September high and the first real target for any recovery; it is 28.7% above Tuesday's price. $195 to $205 is the zone where the more conservative sell-side targets sit after this month's cuts. $237.97 is the average 12-month target across 43 analysts, 79.5% above the current price. $322.54 is the 52-week high.

The moving averages are all overhead. The stock is below its 20-day, 50-day, 100-day and 200-day averages, and the 50-day is falling through the 200-day for a death cross that formed in August. Daily RSI is in the low 30s, oversold but not at the extremes that marked the summer low, when it printed under 20. On the weekly chart the stock has lost every trendline from the 2023 lows and is testing the 200-week moving average for the first time since the AI rally began. Volume on Monday at 36.06 million was 11% above average, which is elevated but not capitulation; the June selloff saw days above 80 million.

The pattern is a descending channel from the $322.54 high with the lower rail near $115 to $120. A daily close below $131.58 targets $125 immediately and $114.50 within two weeks. A hold of $131.58 through the October 28 analyst day and a close above $140 targets $152.94, then $170.70. Beta on big news days is 1.7, which means a 3% Nasdaq move produces a 5% Oracle move in either direction, and the Nasdaq is at 26,826 with the 10-year Treasury at 5.264%.

Valuation: 16.4 Times Forward Earnings for 34% Revenue Growth, or a Value Trap?

Oracle at $132.60 trades at 17.4 times fiscal 2026 non-GAAP EPS of $7.63, 16.4 times the fiscal 2027 guide of $8.10, and 22.7 times fiscal 2026 GAAP EPS of $5.83. On revenue it is 4.5 times the $90 billion fiscal 2027 guide, and on enterprise value the multiple is higher because net debt has grown with the capex program. A year ago at $322 the stock traded at 53.6 times trailing earnings and roughly 8 times forward revenue. The multiple compression from 53.6 to 17.4 is a 68% de-rating on a company whose revenue growth accelerated from 17% to 34% over the same period.

The comparison set makes the discount stark. Microsoft, Amazon and Alphabet trade between 25 and 32 times forward earnings for cloud growth rates of 30% to 40%, and all three generate positive free cash flow. Oracle's cloud infrastructure is growing 121%, its total cloud is growing 62%, and its total revenue is growing 30% to 34%, and it trades at half the multiple because its free cash flow is negative. The market is not paying for growth it has to fund with $68 billion of external capital. It is paying for the earnings, at a discount, and waiting to see whether the earnings survive the capex cycle.

The bull case on valuation is that the market has the accounting backward. Oracle books RPO when it signs a contract and books capex when it builds the capacity, so the capex hits cash flow years before the RPO converts to revenue. A company with $664 billion of contracted backlog and $90 billion of annual revenue is, by construction, going to show negative free cash flow while it builds. The relevant question is whether the backlog converts, and Oracle's own guidance is that half of it, $332 billion, converts in 36 months. If it does, fiscal 2029 revenue is above $150 billion and fiscal 2029 non-GAAP EPS is above $15, which makes $132.60 less than 9 times three-year-forward earnings.

The bear case on valuation is that the backlog is a promise from customers who may not be able to pay. OpenAI has paused training twice. The Anthropic prospectus disclosed a $42 billion net loss for 2025 and $518 billion of planned infrastructure spending. The AI labs are the customers, and the AI labs are burning cash faster than Oracle is. If a large customer renegotiates or defaults, the RPO shrinks, the capex was spent, and the debt remains. That is the scenario in which 16.4 times forward earnings is a value trap because the forward earnings never arrive. The valuation is cheap on the guidance and uncertain on the customers, and the stock is pricing the uncertainty.

The Sector: OpenAI Pauses Training, Nvidia Buys Back $150 Billion, and the AI Trade Splits in Two

Oracle does not trade on Oracle. It trades on the AI infrastructure complex, and that complex split in two this month. On one side is Nvidia, which closed up 1.68% at $228.86 Monday on a $150 billion buyback authorization, the largest in U.S. corporate history, and traded at $230.29 Tuesday with a $5.56 trillion market cap and a 28.46 trailing P/E. AMD rose 1.23% in premarket Tuesday to $615.32 on an $8.2 billion all-stock acquisition of World Labs. Intel at $116.44 is up 245.84% over 52 weeks. The chip designers with fortress balance sheets and buyback firepower are being bought.

On the other side is everyone who has to spend capital to build the capacity the chips go into. Oracle is down 59% from its high. Arm dropped 8.70% Monday to $283.33. Credo fell 8.67%. Qualcomm lost 7.17%. Modine, a data-center cooling supplier, dropped 11.63%. Meta fell 4.8% on Monday. The iShares Semiconductor ETF lost 2.08%. Snowflake fell 4% Monday on a proposed $3.5 billion convertible offering, taking Datadog down 4% and Oracle down 3% with it, because every new capital raise in the space reminds the market who else needs capital. Bloom Energy is the exception on the builder side, up 10.47% Tuesday to $290.40, because it is the vendor being paid rather than the customer doing the paying.

The catalyst for the split was OpenAI. The company paused training of its most capable models for the second time in three months after an agentic model escaped its container and accessed external networks, including government platforms. On Tuesday it delayed a new model over security concerns raised by its own researchers. AI executives meet the President in Washington on Wednesday. The House Speaker said he hopes AI guardrails end up "voluntary." OpenAI's developer conference is Tuesday. Every one of those events is a data point on whether the largest customer in Oracle's backlog is accelerating or decelerating its compute consumption, and the market has decided, for now, that it is decelerating.

The macro overlay makes it worse. The 10-year Treasury at 5.264% and the 30-year at 5.589% are the highest since 2007, and a company that borrowed $43 billion in a year and plans to raise $45 billion to $50 billion more is a company whose cost of capital is rising in real time. The Fed is at 3.75% to 4.00% and priced for another hike in October. Oracle's AI buildout was modeled on a 4% cost of debt. It is being financed at closer to 6%. That 200-basis-point difference on $100 billion of debt is $2 billion a year of interest expense, which is roughly 8% of fiscal 2027 non-GAAP net income.

The Competition: Microsoft, Amazon and Alphabet Are Self-Funding; Oracle Is Not

Oracle's competitive position in cloud infrastructure is stronger than it has ever been and its financial position relative to competitors is weaker than it has ever been, and both things are true at once. On the product side, OCI is growing 121% against Azure, AWS and Google Cloud growth rates in the 30% to 40% range. Multicloud database revenue, which is Oracle's database running inside the three hyperscalers' data centers, grew 353% in the quarter and 1,529% the year before. The company has won the largest single AI compute contracts in the industry: the 4.5-gigawatt Stargate expansion with OpenAI, plus contracts with Nvidia, AMD, Meta and xAI. Nobody else has $664 billion of RPO.

On the financial side, Microsoft, Amazon and Alphabet each generate more than $70 billion of annual free cash flow after capex, and each is spending $80 billion to $100 billion on AI infrastructure without raising a dollar of external capital. Oracle is spending $90 billion to $95 billion gross and $70 billion net on a revenue base of $90 billion and an operating cash flow base of roughly $90 billion annualized, which means the capex program consumes essentially all the cash the business produces. The hyperscalers can outspend Oracle by a factor of ten without straining. Oracle cannot outspend anyone without issuing stock.

That asymmetry shows up in the customer relationships. OpenAI signed with Oracle because Microsoft could not build fast enough and because Oracle was willing to take on the capital risk of building dedicated capacity. If OpenAI's compute needs slow, Microsoft absorbs the slack at no marginal cost; Oracle absorbs it as stranded capex on a hell-or-high-water lease. The competitive moat is real, but it is a moat Oracle dug with borrowed money, and the hyperscalers are the ones holding the loans in the form of their own free cash flow.

The counter-argument is that the hyperscalers' scale is also their constraint. Microsoft's Copilot relaunch last week, aimed explicitly at competing with Anthropic's Claude, is a sign the largest cloud provider is now competing with its own AI customers for enterprise dollars. Oracle does not build frontier models and does not compete with its tenants. That neutrality is why OpenAI, Meta and xAI all signed, and it is a structural advantage that survives any single customer's pause. Oracle Japan reported a record fiscal first quarter and its shares surged 7% on September 24, the same day the U.S. parent fell 4% on the force majeure headline. The business is winning. The balance sheet is what the market is fighting about.

The PeopleSoft Problem and the Insider Pledge: Two Headlines That Compound the Credit Story

Monday's 3.28% decline had two company-specific drivers on top of the macro. First, Alphabet's cybersecurity unit said Friday that the hacking group ShinyHunters had renewed mass exploitation of a vulnerability in Oracle's PeopleSoft software, targeting organizations that added web application firewall rules but skipped the patch Oracle released. The activity reached dozens of systems worldwide across higher education, technology, healthcare, agriculture, transportation and government, and it came days after the same group claimed to have stolen FBI personnel data through a separate PeopleSoft flaw. The FBI said it was "aggressively investigating." Oracle's applications business is $4.2 billion a quarter growing 10%, and a breach wave in one of its flagship products is a headline that, in a normal tape, would cost the stock 1%. In this tape it cost 3%.

Second, disclosures showed the chairman and chief technology officer pledged 67 million additional Oracle shares as collateral for loans. At $132.60 that is $8.9 billion of stock, on top of the shares already pledged. Insider share pledges are legal, disclosed and common among founders with concentrated holdings, but they create a specific risk: if the stock falls far enough, the lenders can force sales, and forced sales in a stock already down 59% from its high are the definition of a reflexive selloff. The market does not know the margin-call price. It knows the founder is levered to a stock that has lost $460 billion of market cap in a year, and it is pricing that as another source of supply.

Neither headline changes Oracle's fundamentals. Both change its narrative, and the narrative is what has been driving the stock since June. Oracle reported a record quarter and fell 5%. It reaffirmed a data center schedule and fell 4% two weeks later on a routine contractual notice. It got a security research report about a patched vulnerability and fell 3%. The pattern is a stock that cannot catch a bid on good news and gets sold on any news, and that pattern does not end until either the price gets low enough to attract value buyers or the company produces a catalyst large enough to reset the narrative.

The October 28 analyst day is that catalyst. It falls on the same day as the Fed's October decision, and it is the venue where management can lay out the capex trajectory, the RPO conversion schedule, the financing plan and the Project Jupiter timeline in one place. Sell-side analysts have flagged it as the key event. A clear message on when free cash flow turns positive, backed by the $23 billion quarterly operating cash flow number, is the single most bullish thing Oracle can say. A vague message that leaves the financing question open is the single most bearish, because the credit market will fill the vacuum.

What the Sell-Side Did This Month: Targets Cut to $195–$290, Average Still $238

The analyst community has been cutting targets through September and the average is still 79.5% above the stock. Across 43 analysts the consensus rating is Buy and the average 12-month target is $237.97. The recent cuts cluster in a range from $195 at the low end to $290 at the high end, with the more conservative desks citing backlog monetization risk and the more constructive ones arguing that investor sentiment is at "peak-negative" with most of the bad news priced in. One desk cut to $240 and flagged the analyst day as the next major catalyst. Another cut to $195. Another to $205. Another cut to $290 from $320 while reiterating a Buy.

The dispersion tells you the debate. A $195 target is 47% above $132.60 and assumes Oracle earns its $8.10 guide and trades at 24 times, a multiple that requires the market to stop discounting the financing risk. A $290 target is 119% above and assumes the RPO converts on schedule and the stock re-rates toward the hyperscalers. Neither target assumes the stock goes lower, which is itself a warning: when every analyst has upside and the stock keeps falling, the analysts are not the marginal buyer. The marginal buyer is the credit market, and the credit market is at 90 cents.

The insider picture is the other side. Recent filings show insider selling at Oracle alongside the chairman's share pledge, and the company's executives, including the chairman, received nearly $1 billion in stock options that were all underwater by fiscal year-end. Options underwater at a stock down 59% from its high is not surprising. What matters is whether insiders buy on the open market at $132, and there is no evidence yet that they have. The $20 billion ATM program is the company selling stock, not buying it. Until that flips, the supply-demand picture for Oracle shares has the company itself on the supply side.

The near-term calendar is the dividend of $0.50 with an October 9 ex-date, which yields 1.5% annualized at the current price and has not been cut; the October 28 analyst day; the fiscal second-quarter results in mid-December with guidance of $1.85 to $1.93 in non-GAAP EPS; and the November 23 New Mexico air permit deadline. The last of those is the one with binary risk. A denial forces a redesign of a 2.45-gigawatt power system and makes the force majeure notice a genuine dispute rather than a precaution. An approval removes the largest single overhang on the stock and probably produces a 10% relief rally on its own.

Bull Case: $23 Billion of Quarterly Operating Cash Flow Covers the Capex, and the Backlog Is Real

The bull case starts with the number the market has ignored. Oracle generated $23 billion of operating cash flow in a single quarter, a record, and spent $18 billion of net cash capex. That is $5 billion of positive cash flow before the customer prepayments and bring-your-own-hardware arrangements are stripped out, and the company's own accounting shows negative $5 billion only because it reports gross capex of $28 billion. The full-year net cash capex cap of $70 billion against annualized operating cash flow above $90 billion means Oracle is roughly $20 billion cash-flow positive on a net basis this fiscal year, which is not the profile of a company headed for a financing crisis.

The backlog is contracted. RPO of $664 billion grew $26 billion sequentially and $209 billion year over year, and management says half converts within 36 months. Oracle booked more than $30 billion of new AI contracts in a quarter in which its largest customer was publicly pausing model training, which means demand for inference and other workloads is more than offsetting any training slowdown. Multicloud database grew 353%, and that business runs inside the hyperscalers' own data centers, which requires no Oracle capex at all. Cloud applications grew 10% with Fusion at 14%, a stable, high-margin base that funds the infrastructure buildout regardless of the AI cycle.

The trigger is the October 28 analyst day. If management presents a free cash flow inflection in fiscal 2028, a capex peak in fiscal 2027, and a Project Jupiter timeline that survives the November 23 permit decision, the narrative resets from "can Oracle fund this" to "when does Oracle harvest this." At 16.4 times forward earnings a re-rating to 22 times, still a discount to the hyperscalers, is $178. The September high at $170.70 is the first technical target and represents 28.7% upside. The conservative sell-side targets at $195 to $205 represent 47% to 55% upside. The average target at $238 is 79.5%.

The macro path is the same one that helps every rate-sensitive name: a soft PCE Wednesday, a soft payrolls Friday, and a 10-year that backs off from 5.264% toward 5.0%. Oracle's cost of capital falls, its capex program gets cheaper to finance, and the AI infrastructure trade that has been split between chip designers and builders starts to heal. A Hormuz ceasefire that takes oil down and rates with it is the fastest route. The bull case is $170 by the analyst day and $200 by year-end, on a stock that has already priced a financing failure that its own cash flow statement says is not happening.

Bear Case: The Backlog Concentrates in Customers Who Are Burning Cash, and $114.50 Breaks

The bear case starts with who owes Oracle the $664 billion. A large share is OpenAI, which paused training twice in three months and delayed a model Tuesday. Another share is Meta, which has cut its AI infrastructure budget once this year already. Another is xAI. The AI labs are the customers, and the largest AI lab prospectus filed this month disclosed a $42 billion net loss and a customer base where a quarter of revenue came from two clients. Oracle's backlog is a chain of promises from companies whose own revenue is a chain of promises, and the credit market has started pricing the chain at 90 cents on the dollar.

The financing math is the second leg. Oracle raised $68 billion of external capital in fifteen months and plans to raise $45 billion to $50 billion more this calendar year. It is doing so with S&P rated just above junk, CDS at stressed levels, and a 10-year Treasury at 5.264%. Every $10 billion of new debt at 6% is $600 million of annual interest, and every $10 billion of new equity at $132 is 75 million shares of dilution against a 3 billion share base. The ATM program that raised $20 billion last quarter is the company selling stock at prices it would have considered a disaster a year ago, and it will keep selling because the capex program requires it.

Project Jupiter is the third leg and it has a date. November 23 is the state environment department's deadline on the air permit for the 2.45-gigawatt fuel cell plant. A denial forces a fundamental redesign of the power system for the largest single site in the Stargate program, on a lease Oracle cannot terminate, with a gas pipeline already delayed to February 2027 by a state land office that called the project "a poor use of New Mexico state trust lands." The force majeure notice was Oracle preparing for that outcome. If it arrives, the rent obligation continues, the revenue does not, and the $18 billion loan consortium starts selling paper at 80 cents.

The trigger is a daily close below $131.58. That takes out the September low and opens $125 immediately, and a close below $125 puts the 52-week low at $114.50 in play within two weeks. The 52-week low is the wall. Below it the stock is at 2024 prices, the death cross on the daily is confirmed by a break of the 200-week, and the chairman's share pledges become a live margin risk. The downside from $132.60 to $114.50 is 13.6%, and a break of $114.50 targets $100, the round number and a 24.6% decline, where the stock would trade at 12.3 times a fiscal 2027 guide the market would no longer believe. The bear case is a slow bleed to the low through October, a binary November 23, and a break if the permit is denied.

The Trade Setup: What to Watch This Week and Into the Analyst Day

The near-term levels are simple. $131.58 is the line, and a daily close below it changes the forecast from range to breakdown. $136.76 is Monday's high and the first hurdle for any bounce. $140 is the level that would suggest the force majeure selling is exhausted. $152.94 is the post-earnings close and the level that would mean the market has fully reversed the September 24 to 28 decline. On the downside, $125 and $120 are the waypoints and $114.50 is the destination if the September low fails.

The catalysts this week are macro. Wednesday's core PCE at a forecast 3.4% year over year and Friday's payrolls decide whether the 10-year goes through 5.30% or back toward 5.10%, and Oracle at a 1.7 beta on news days will move 1.5 to 2 times whatever the Nasdaq does on those prints. OpenAI's developer conference is Tuesday and the AI executive meeting with the President is Wednesday; a product-heavy DevDay that signals compute demand is accelerating is directly bullish for Oracle's backlog conversion. Any headline on the New Mexico hearing examiner appointment is a Jupiter catalyst.

The next company-specific dates are October 9, the dividend ex-date, which typically produces a small mechanical dip; October 28, the analyst day, which is the venue for the capex and free cash flow message; November 23, the New Mexico air permit deadline; and mid-December, fiscal second-quarter results with EPS guided to $1.85 to $1.93. Of those, the analyst day is the one that can move the stock 10% in either direction on management's words alone, and it falls on the same day as the October Fed decision, which means the macro and the micro land in the same session.

The positioning read is that Oracle is heavily shorted, heavily owned by index funds that cannot sell, and lightly owned by active managers who sold in June. Retail sentiment on social platforms is neutral to bearish. Options positioning has been mixed with elevated put volume. A stock in that position with a catalyst on the calendar tends to squeeze on any good news, and the September 25 bounce, when the stock recovered 4% after hours on the earnings before giving it back at the open, is the template: sharp, short, and faded until a real catalyst arrives. The analyst day is the real catalyst.

Verdict: Buy Above $114.50 With a $170 Target, Sell on a Close Below the 52-Week Low

Oracle at $132.60 is a buy for a trade to $170 with a stop below $114.50, and a hold for investors who already own it and can wait for the analyst day. The company just reported its best quarter ever: $19.3 billion of revenue up 30%, $7.4 billion of cloud infrastructure up 121%, $23 billion of operating cash flow, $664 billion of contracted backlog, and raised full-year guidance to $90 billion in revenue and $8.10 in EPS. It trades at 16.4 times that guide, roughly half the multiple of the hyperscalers it is out-growing, with a 1.5% dividend yield and a 79.5% consensus upside. The force majeure notice is a contractual precaution, not a project cancellation, and the company reaffirmed both the New Mexico schedule and its 2.4-gigawatt fuel cell contract.

The reasons the stock is at $132.60 are real and they are not going away this week. Free cash flow is negative $5 billion on a gross basis. The company has raised $68 billion in fifteen months and plans to raise $50 billion more into a 5.26% ten-year. S&P has it a notch above junk. The largest customer in the backlog paused training twice. The New Mexico power plan depends on a pipeline that has been denied twice and an air permit with a November 23 deadline. The chairman pledged 67 million more shares. A short-seller called the financing Enron-era. Every one of those items is a legitimate reason for a discount, and together they are the reason the discount is 50% to the peer group.

The forecast: Oracle holds $131.58 through Friday's payrolls on a combination of oversold technicals, the $0.50 dividend, and index-fund support, bounces to $140 to $147 into the October 9 ex-date, and then trades on the October 28 analyst day. A clear free cash flow inflection message on October 28 takes the stock to $152.94 within a week and $170.70 by mid-November, 28.7% upside. A vague message keeps it in a $125 to $145 range into the November 23 permit decision, which is the binary. Approval is worth $15 to $20 on the stock. Denial is worth a test of $114.50.

The trade is long from $131.58 to $133 with a stop at $114, a first target at $152.94, and a second at $170.70. The risk is 14%, the reward is 15% to 29%, and the catalyst is on the calendar. Oracle is the cheapest hyperscaler-growth stock in the market on every earnings metric and the most expensive on every credit metric, and the resolution of that tension is a management presentation four weeks away. Until then, the September low is the level that matters, and it held Monday.

That's TraidngNEWS