The early-October street pair is not where Friday’s close points. Oracle closed Friday, 16 October 2026, at $128.23, and the last 90 daily log-returns, carried across the 52 New York sessions from 19 October through 31 December, put the bull at $143.70, the base at $97.56 and the bear at $75.36.
The missing fact is the drift. A bull near $200 and a bear at $110 were street targets tied to an early-October close, not a 52-session path from Friday. The drift-only base, $97.56, already finishes the year $12.44 under that $110. The bull, $143.70, is only $15.47, or 12.06 percent, above Friday, because the mean log-return on those 90 sessions is negative. Stifel’s $200 and Alexander Haissl’s $110 are still on a ratings table. They are not this distribution. Even a no-drift version of the same 1.5-deviation upside, which is a sensitivity and not the bull, prints at $188.87 and still misses $200.
Friday was also outside an ordinary band. The log-return from Thursday’s $134.47 was -0.04752, or 1.33 of the sample standard deviation of 0.03580. The one-deviation band Thursday already implied ran from $129.74 to $139.37. Friday closed through the bottom of it. The largest day of the week was still Wednesday, not Friday.
Key facts
- Friday 16 October 2026 official close $128.23, down $6.24, or 4.64 percent, from Thursday’s $134.47. Open $133.49, high $133.80, low $127.36, volume 36,109,050. Source: Nasdaq historical prices, file saved 19 October 2026. A premarket last sale of $129.666 at 7:12 a.m. ET is not a close and is not a chart bar.
- From the 9 October close of $141.40 the week is down $13.17, or 9.31 percent. Wednesday 14 October closed at $130.55. Thursday closed at $134.47. Same Nasdaq file, saved 19 October 2026.
- Nasdaq’s summary, saved that morning, prints a previous close of $128.23, a one-year target of $225.00 and a market cap of $393,082,627,986. The file’s high close is $313.00 on 16 October 2025. The low close is $114.99 on 24 July 2026. This note uses the rows, not the summary screen of $304.28 / $114.50.
- Quarter ended 31 August 2026: revenue $19.3 billion, up 30 percent. Cloud revenue $11.6 billion, up 62 percent. Cloud infrastructure $7.4 billion, up 121 percent. Remaining performance obligations $664 billion. Operating cash flow $23 billion. Free cash flow negative $5 billion. Source: Oracle’s 10 September 2026 release.
- The same release guides fiscal 2027 revenue to at least $90 billion and non-GAAP earnings to $8.10. Second-quarter revenue growth is guided at 30 percent to 34 percent, cloud growth at 65 percent to 71 percent in dollars, and non-GAAP earnings at $1.85 to $1.93. A $0.50 dividend is paid 23 October 2026.
- The Form 10-Q tags operating cash flow at $23,103 million and capital-expenditure payments at $28,499 million. The difference is negative $5,396 million. The at-the-market program issued about 141 million shares for net proceeds of $19.9 billion.
- StockAnalysis, fetched 19 October: Stifel $200 Buy, 25 September. Yorkville Ives $155 Hold, 7 October. Rothschild and Co Redburn $110 Sell, 21 September. Bernstein $325 Buy, 25 September. RBC Hold, 12 October, with no target on the table.
- House path, not a street target: N = 52, mu = -0.0052569765, sigma = 0.0358013530. Bull $143.70, base $97.56, bear $75.36. The rounded pair is not within $1 of 200 and 110, 205 and 98, or 245 and 95, so the horizon stays 31 December.
What just happened, and why the obvious reading is wrong
The obvious reading is that Friday’s 4.64 percent drop was the OpenAI revenue story breaking Oracle again. Two dates in the file refuse that reading. The revenue report hit the tape on Thursday 8 October, when the close fell from $143.56 to $135.69, a decline of 5.48 percent. Friday 9 October closed at $141.40, up 4.21 percent, and put back part of that drop. That bounce was already a separate note. It was not a reversal, and it is not this week’s close.
The week from that $141.40 print to Friday is down $13.17, or 9.31 percent. Wednesday’s log-return, the natural log of $130.55 divided by $143.20, was -0.09249, or 2.58 sample deviations. Friday’s -0.04752 was 1.33. Thursday’s rise to $134.47 is the point against a five-session collapse. The closes used here are the Nasdaq rows saved on 19 October, before the cash open.
A one-deviation move is exp(0.0358013530) minus 1, or 3.64 percent. Thursday’s band was $129.74 to $139.37, and Friday closed $1.51 under the floor. The next band, from Friday, is $123.72 to $132.90. The premarket last sale of $129.666 sits inside it and is not drawn.
Who is exposed
Equity holders take the first loss. The Nasdaq summary market cap is $393,082,627,986. Divided by Friday’s $128.23, that is about 3.065 billion shares, so Friday’s $6.24 decline is about $19.13 billion of market value. That is arithmetic, not a second official cap. The 10-Q says the quarter’s equity sale was about 141 million shares for net proceeds of $19.9 billion.
Customers sit in the cloud line. Infrastructure revenue of $7.4 billion was 38.3 percent of the $19.3 billion quarter. Software revenue was $5.5 billion, down 3 percent. The desk record of the quarter is separate from this path. No October page fetched here revises the release.
Lenders and one director sit beside the equity. Project Jupiter is about $18 billion of construction loans, quoted at 89 to 91 cents on 18 September, as the debt note carries from Reuters and the Financial Times. The same note reads the 10-Q fair value of $125.0 billion of borrowings at $105.7 billion on 31 August, about 84.6 cents. Director Stephen H. Rusckowski’s 25,000 shares, at a weighted average of $139.352, cost $3,483,800. At Friday’s close they are worth $3,205,750, down $278,050, or 7.98 percent.
What the tape and the primary documents show
No new revenue figure arrived with Friday’s close. The tape, the 10 September release and the 10-Q answer a narrower question than a broken cloud franchise.
| Item | Obvious reading | What the record shows |
|---|---|---|
| The week | Friday was the OpenAI break | 9 October $141.40 to 16 October $128.23, down $13.17, or 9.31 percent. The revenue-report session was 8 October, at $135.69 |
| Friday | Ordinary noise | $128.23, down 4.64 percent. Log-return -0.04752, or 1.33 sample deviations, through the $129.74 floor of Thursday’s band |
| Wednesday | The same as Friday | $130.55, down 8.83 percent from $143.20. Log-return -0.09249, or 2.58 sample deviations |
| Thursday | The slide never paused | $134.47, up 3.00 percent from $130.55. One of the five sessions rose |
| Year to date | A new bear market this week | 2 January close $195.71 to $128.23, down $67.48, or 34.48 percent. The file high close is $313.00 on 16 October 2025 |
| The backlog | Demand vanished on Friday | Remaining performance obligations $664 billion at 31 August, up $209 billion year over year. No October filing fetched here cuts it |
| Capex | A new spending guide this week | Hilary Maxson, 10 September: $90 billion to $95 billion for the year, net cash capex not more than $70 billion. No revision in the pages fetched 19 October |
| Street targets | The title bull and bear | Stifel $200, Haissl $110, Ives $155, Bernstein $325, Nasdaq screen $225. This bull is $143.70 and this bear is $75.36 |
The cash gap is in the filing, not in Friday’s headline. Operating cash flow of $23,103 million minus capital-expenditure payments of $28,499 million is negative $5,396 million. The release calls free cash flow negative $5 billion. Friday did not replace the 10 September revenue guide of at least $90 billion.
What Oracle and named officials have said
The last full statements from the operating executives are on the 10 September call, not this week. Clay Magouyrk, co-chief executive, said, as Benzinga quoted the call on 11 September: “Anyone that’s been in the business of doing construction or large-scale infrastructure development, if their plan relies on 100% achievement of every one of their deliverables, we have a term for that, it’s called a bad plan.” He was answering a question about data-center delays in New Mexico, not a share price. The Motley Fool transcript of the same call carries the line in speech-to-text form.
Hilary Maxson, chief financial officer, put the spending number on that call. The transcript has her saying: “We continue to anticipate $90 billion to $95 billion in CapEx for the full year with not more than $70 billion in net cash CapEx.” She also said “the vast majority of those new contracts were via prepay or bring your own hardware or a similar mechanic so will not require incremental capital from Oracle.” Data Center Dynamics quotes Magouyrk: “While there clearly are capital expenditures, it does not require Oracle to go out and find additional cash to do it.”
On the Jupiter loans, the debt note says Oracle, Santander and Jefferies did not respond to Reuters. On the 8 October revenue report, the Friday note says OpenAI declined to comment, as CNN reported. The $50 billion annualized figure is not an Oracle filing, and neither is the earlier $68 billion or $70 billion figure.
The director’s purchase is a filing, not a comment. The 2 October note reads the Form 4: 25,000 shares, code P, on 29 September, at a weighted average of $139.352, held by The Stephen H. Rusckowski Living Trust, with fills from $138.96 to $139.74. The form does not mention Tencent. A Sunday post states the same purchase at a rounded $139.35 and says that at $128.23 the insider is underwater. A later post restates the $3.48 million buy next to the $664 billion backlog. The Mizuho target in that post is not on the StockAnalysis table saved here, so it is not used.
The credit and the unconfirmed lease
Two constraints sit on the guide, and Friday resolved neither. The debt note says S&P cut Oracle to BBB- on 9 July 2026, one notch above junk, with a stable outlook, tied to data-center spending and backlog concentration. A further notch would change which mandates can own the paper. The 89 to 91 cent loan quote, from 18 September, has not been refreshed in a page fetched today. It is a credit price, not a share-price target.
The second constraint is a contract the company has not named. The Financial Times reported on 1 October, as the Tencent note carries it, a five-year lease of about $7 billion for roughly 100,000 advanced AI chips in Southeast Asia, with about 30 percent paid upfront. Oracle does not mention Tencent in the 10 September release, the call, or the 10-Q, and Tencent has not confirmed the lease in any statement that note could tie to the report. Seven billion dollars is 1.05 percent of the $664 billion obligation. It is not an input to the bull or the bear.
The investor-relations FAQ opened on 19 October prints no date for the next earnings release. This note does not pick 9 December or 14 December. The $0.50 dividend on 23 October is the only company cash date still on the 10 September release.
Bull, base, and bear through 31 December
These three prices are the distribution, not analyst targets. A log-return is the natural log of one official close divided by the prior close. Mu is the average of the last 90. Sigma is the sample standard deviation, divided by n minus 1. The prior close for the first return is 9 June 2026, at $205.81. The last return is Friday’s. Mu is -0.0052569765. Sigma is 0.0358013530. N is 52, the sessions from Monday 19 October through Thursday 31 December, skipping weekends and only Thursday 26 November and Friday 25 December. Friday 27 November and Thursday 24 December still count. The base is the spot times exp(N times mu). The bull adds 1.5 deviations. The bear subtracts one. The bull still prints above the spot, so the drift stays in the formula.
In steps: 52 times -0.0052569765 is -0.2733627789. The square root of 52 is 7.21110255. Sigma times that square root is 0.2581672282. The bull’s exponent is -0.2733627789 plus 1.5 times 0.2581672282, or 0.1138880634. The bear’s exponent is -0.2733627789 minus 0.2581672282, or -0.5315300071. Applied to $128.23, exp(0.1138880634) is 1.12062668, exp(-0.2733627789) is 0.76081673, and exp(-0.5315300071) is 0.58770509. Rounded to the cent, the bull is $143.70, the base is $97.56 and the bear is $75.36. The bull is $15.47, or 12.06 percent, above the close. The base is $30.67, or 23.92 percent, below it. The bear is $52.87, or 41.23 percent, below it.
| Case | Price | Versus $128.23 | What has to be true |
|---|---|---|---|
| Bull | $143.70 | Up 12.06 percent | The negative drift holds, and the path still finishes 1.5 deviations above it. The Q2 cloud guide is not cut |
| Base | $97.56 | Down 23.92 percent | Only the drift. No new guide is required. This already sits $12.44 under the $110 street low |
| Bear | $75.36 | Down 41.23 percent | Drift minus one deviation. A miss of the 30 percent revenue-growth floor is the fundamental version |
Street targets are twelve-month opinions. Stifel’s $200 is $56.30 above this bull. Haissl’s $110 is $12.44 above this base. A no-drift 1.5-deviation upside is $188.87, still $11.13 short of $200, and it is a sensitivity, not the bull. The rounded pair is not within $1 of 200 and 110, 205 and 98, or 245 and 95. N stays 52.
What happens next
Monday 19 October is the next cash session. The exchanges had not opened when these rows were saved. A close between $123.72 and $132.90 is inside one sample deviation of Friday. A close under $123.72 is a fresh one-deviation down day. Friday was already 1.33 deviations, so a quiet Monday would mean the selling slowed, not that Wednesday’s 2.58-deviation drop has been repaired. The premarket print of $129.666 does not decide that.
Thursday 23 October is the dividend payment date, $0.50 a share, to holders of record on 9 October. Four quarters at that rate are $2.00, or 1.56 percent of Friday’s close. The $0.50 itself is 0.39 percent of $128.23, inside the 3.64 percent one-deviation day. The check does not revise the $90 billion revenue guide or the $664 billion obligation.
The next company number in the path is the second-quarter print against revenue growth of 30 percent to 34 percent, cloud growth of 65 percent to 71 percent in dollars, and non-GAAP earnings of $1.85 to $1.93. The FAQ fetched here does not date that print, so the horizon stays 31 December. A miss of the 30 percent revenue floor is the fundamental bear. Inside the band, the base stays $97.56 unless the drift itself changes. The bull at $143.70 still needs the upside tail on a negative mean. The bear at $75.36 still needs one deviation the other way.
Frequently asked questions
Does the stock’s own volatility still point near $200 and $110 by 31 December, or has the distribution moved?
It has moved. From the Friday 16 October close of $128.23, the last 90 daily log-returns and 52 sessions put the bull at $143.70, the base at $97.56 and the bear at $75.36. The base already sits $12.44 under the $110 street low. Stifel’s $200 and Haissl’s $110 remain separate twelve-month targets on the StockAnalysis table fetched 19 October. They are not this path, and a no-drift upside of $188.87 still misses $200.
What did Oracle stock close at on Friday 16 October 2026?
The official close was $128.23, down $6.24, or 4.64 percent, from Thursday’s $134.47. The open was $133.49, the high $133.80, the low $127.36 and the volume 36,109,050 shares, from the Nasdaq historical file saved 19 October. A premarket last sale of $129.666 at 7:12 a.m. ET that morning is not a close and is not on the chart. The week from 9 October is down 9.31 percent.
Why is the year-end bull only $143.70?
Because the drift is negative and the horizon is 52 sessions. Mu on the last 90 log-returns is -0.0052569765. Times 52, that is -0.27336, and the exponential of that is 0.7608, so the base is $97.56. Adding 1.5 deviations lifts the bull only to $143.70, $15.47 above the close. A no-drift version of the same upside is $188.87. That figure is a sensitivity, not the bull in the title.
Did Oracle confirm a $7 billion Tencent lease?
No. The Financial Times report of a five-year lease for about 100,000 chips, as carried in the 2 October FinanceFeeds note, is not in the 10 September release, the call, or the Form 10-Q for the quarter ended 31 August. Seven billion dollars is 1.05 percent of the $664 billion remaining performance obligation. Until Oracle names the contract, it does not change the bull at $143.70 or the bear at $75.36.
What has to be true for the bear at $75.36?
The bear is the drift minus one deviation over 52 sessions, not a new street low. It does not require a fresh downgrade. It does require the recent log-return path to keep going. A print that misses the second-quarter guide, revenue growth of 30 percent to 34 percent, would be the fundamental version of that tail. The guide is from 10 September. Friday did not revise it, and the investor FAQ fetched here does not date the print.
Does the $0.50 dividend hold the stock at $128.23?
No. The board declared $0.50 a share, paid on 23 October 2026 to holders of record on 9 October. Four quarters at that rate are $2.00, or 1.56 percent of Friday’s close. The payment is cash leaving the company, not a new cloud contract. It does not change mu, sigma, or the 52-session count. The year-end bull and bear are still $143.70 and $75.36.
This is not financial advice.