Oracle spent $28.5 billion on capex in one quarter, and at $115.16 the price is about fair for that bet, so the verdict is hold

In the three months to August, Oracle generated a record $23.1 billion of operating cash flow, almost half of it prepayments from customers, and still had $5.4 billion less cash from operations than it spent on data centers. Capital spending was $28.5 billion in a single quarter. To pay for it, Oracle sold $20 billion of new shares, and its 10-Q lists $288 billion of data center leases that have been signed but not yet started. The stock, which traded above $300 in September 2025, closed at $115.16 on November 5, down 6.9% on the day.

At $115.16, Oracle trades 4.7% below my Base value of $120.86. The Bear value is $42.71, 62.9% below the price, and the Bull value, built on the company's own fiscal 2030 target, is $240.94, more than double. I assign no probabilities to the cases. The verdict is Hold, with a buy range of $59.59 to $82.91 for new money.

Valuation route: a mature software business attached to a fast-growing, capital-heavy AI cloud. The primary method is a ten-year free cash flow model that counts stock-based pay, depreciation and Oracle's own capital spending in full; the cross-check is a forward multiple of enterprise value to fiscal 2029 operating profit after stock pay. Both use the same debt and share count. Oracle's fiscal year ends on May 31, so fiscal 2027 runs from June 2026 to May 2027. Values are present fair-value estimates, not 12-month price targets.

What Oracle is

For forty years Oracle sold databases and business software, collected maintenance fees and earned some of the highest margins in technology. That business still exists: license and support revenue was $5.5 billion in the quarter, down 3%, and cloud applications such as Fusion and NetSuite were $4.2 billion, up 10%.

The growth is now Oracle Cloud Infrastructure (OCI), rented computing capacity for training and running AI models. It grew 121% to $7.39 billion and was 38% of revenue. That business looks nothing like software. Oracle leases the buildings for 15 to 19 years, buys or finances the chips, and rents the capacity to a few very large AI customers under multi-year contracts. Revenue is contracted, but so are the costs.

The latest results

Contracted backlog, the remaining performance obligations (RPO), reached $664 billion, up $209 billion in a year, with more than $30 billion of new AI contracts in the quarter. Only 13% of it, about $86 billion, is expected to turn into revenue in the next twelve months; half within three years. Oracle guides fiscal 2027 revenue to at least $90 billion and adjusted earnings of $8.10 a share, and for the second quarter it expects revenue growth of 30% to 34% and cloud growth of 65% to 71%.

In June management said fiscal 2027 capital spending would be up to $95 billion: about $70 billion of its own and $20 to $25 billion that customers are expected to repay. In fiscal 2026 it spent about $55.7 billion against a $50 billion target. Interest expense in the quarter was $1.43 billion, up 55%.

Which numbers to trust

• Operating cash flow is flattered by prepayments. $11.4 billion of the quarter's $23.1 billion came from customer prepayments, which Oracle itself says carry a significant financing component. Deferred revenue jumped to $30.8 billion. I treat the part above Oracle's normal level of about $12 billion, roughly $18.8 billion (my estimate), like debt.
• Depreciation is the new big cost. Adjusted operating income includes depreciation of the data center equipment, and it will rise fast. I let it go from about 17% of revenue this year to 23% in fiscal 2029 before easing.
• Stock-based pay is a cost. About $1.1 billion a quarter, 5% to 6% of revenue. I deduct it.
• Leases sit in the costs, not the debt. Rent for data centers runs through operating costs, so I do not add the $34.6 billion of operating leases or the $288 billion of future leases to debt. They show up as lower margins instead. Finance leases of $9.2 billion are counted as debt.

Debt, shares and funding

On August 31 Oracle owed $125.3 billion of notes and borrowings plus $9.2 billion of finance leases and held $37.1 billion of cash and securities: net debt of about $97 billion. In the quarter it repaid $4.2 billion of debt, paid $1.6 billion of dividends and raised $19.9 billion by selling shares through an at-the-market program. For the full year management plans to raise about $40 billion of debt and equity, including that sale. There were 3.02 billion shares on September 7.

Larry Ellison, the chairman and chief technology officer, is the largest shareholder. That keeps the strategy consistent, and it also means the decision to bet the balance sheet on AI capacity is his to make.

Primary valuation: ten years of free cash flow

The Base takes fiscal 2027 at the $90 billion guidance, then $118 billion, $145 billion and $170 billion by fiscal 2030, against the company's October 2025 target of $225 billion. After that growth slows from 12% to 4%. Operating margin after stock pay falls from 36% to 27% as low-margin AI infrastructure becomes the bulk of revenue and depreciation catches up. Oracle's own capital spending falls from 78% of revenue this year to 21% in the long run, a bit above depreciation. I discount at 9.5% a year with 3% growth after fiscal 2036.

Free cash flow turns positive in fiscal 2029. The present value of the flows to 2036 is $127.0 billion and of everything after that $365.1 billion, so the terminal value is 74.2% of the total. Enterprise value is $492.1 billion. Less $116.2 billion of net debt, finance leases and excess prepayments, that leaves $375.9 billion, or $124.30 a share.

The discount rate moves this a lot, which is normal for a company whose cash comes late:

Secondary valuation: a forward multiple

The cross-check applies 13 times to fiscal 2029 operating profit after stock pay of $41.3 billion, as a value at the end of fiscal 2028, discounted to today. That gives an enterprise value of $466.0 billion and $115.69 a share.

I use operating profit, not EBITDA, because depreciation and capital spending are now the biggest numbers in Oracle's accounts, and an EBITDA multiple would hide them. Once capex settles near depreciation, about 91% of after-tax operating profit becomes free cash flow, so 13 times is a free cash flow yield of about 5.8%, consistent with 3.7% growth a year forever at a 9.5% cost of capital. I did not build a dated peer comparison for this analysis.

The two methods are 7.2% apart and both rest on the same revenue path, so their agreement is not an independent confirmation.

History and earnings check

Oracle traded above $300 in September 2025, when the backlog first jumped by hundreds of billions. Since then the market has moved from celebrating the contracts to counting the cost of delivering them. On adjusted earnings the stock looks cheap: $115.16 is 14.2 times the $8.10 guided for fiscal 2027, and the company's October 2025 target of $21 for fiscal 2030 would put it near 5.5 times. Those earnings figures exclude stock pay and say nothing about capital spending, which is why I do not lean on them. The growth-adjusted price to earnings ratio is not meaningful during a build-out of this size.

Valuation triangulation

• Cash flow model (weight 60%): $124.30.
• Forward multiple (weight 40%): $115.69.
• Base value: $120.86, inside the method range of $115.69 to $124.30.
• The cash flow model gets more weight because it follows the capital spending and depreciation year by year.

What is the market already pricing in?

The market is close to my Base. It assumes Oracle gets to about $164 billion of revenue in fiscal 2030, roughly 73% of its target, with margins near mine. It is not paying for the target and it is not pricing in a collapse either. If the Base comes true, the stock returns about 9.8% a year, a little above the 9.5% cost of capital.

Where the return would come from: if the market keeps paying 7.6 times forward EBITDA, the Base gives about 14.8% a year to May 2029 including dividends, all of it from EBITDA growth, because net debt keeps rising until free cash flow turns positive. In the Bear, the same multiple gives -3.1% a year.

Bear, base and bull scenarios

• Bear: one or more large AI customers cannot fund what they signed, backlog converts slowly, and Oracle is left with leases and equipment it cannot fill at good prices. Revenue reaches only $128 billion in fiscal 2030 and margins settle at 22%. Cash flow model $40.38, multiple $46.20.
• Base: as described above. Cash flow model $124.30, multiple $115.69.
• Bull: the fiscal 2030 target of $225 billion is met, margins hold at 32% and capex falls to 20% of revenue. Cash flow model $260.72, multiple $211.28.

Each case blends the two methods 60/40. The discount rate is 9.5% in all three.

How the buy price is set

Confidence is C, which carries a margin of safety of 42.5%. Applied to the Base that gives $69.49. But one point of discount rate moves the Base by $17.22 to $23.34, more than 15% of the value, so a single buy price would look more precise than it is. Using the Base at 10.5% and 8.5%, the buy range is $59.59 to $82.91, and I treat the lower end as the maximum buy price.

Other single-input effects: three points of operating margin move the Base by $21.10, 10% more or less revenue from fiscal 2028 by $16.34, two points of capex every year by $11.74, two turns of multiple by $9.48, and each extra $10 billion of net debt by -$3.31. Without my prepayment adjustment the Base would be $127.07.

For reference on the other side: above about $125 the expected return under my Base falls below the 9.5% cost of capital. That is where I would review the position, not a target.

Dividend

Oracle pays $0.50 a quarter, $2.00 a year, a yield of 1.7%. With free cash flow negative, the dividend is currently paid from borrowing and share sales. It is not at risk while the capital markets stay open, but it is not covered by cash flow until fiscal 2029 in my Base.

Reasons to own Oracle

• Revenue growth of 30% and cloud infrastructure growth of 121% in the latest quarter, with fiscal 2027 revenue guided to at least $90 billion.
• A contracted backlog of $664 billion, half of it expected within three years.
• Customers prepay and repay part of the capital spending, which lowers Oracle's own funding need.
• The database and applications business remains one of the stickiest in software.
• After a fall of about two thirds from the September 2025 high, the price is close to my Base.

What could go wrong?

• Customer concentration. Much of the backlog comes from a few AI companies whose ability to pay depends on their own fundraising.
• Commitments. Up to $95 billion of capital spending this year and $288 billion of leases that start between now and fiscal 2029 for 15 to 19 years.
• Leverage. $125.3 billion of borrowings, interest up 55%, and share sales to fill the gap.
• Margins. AI infrastructure earns far less than software, and depreciation is only starting.
• Technology. Chips age quickly; if the useful life is shorter than the depreciation schedule, profits are overstated.

Management execution

• Revenue and cloud growth have beaten guidance, and fiscal 2027 revenue guidance was raised to at least $90 billion.
• The backlog has grown by $209 billion in a year, with over 300,000 GPUs and 850 megawatts delivered since May.
• Fiscal 2026 capex came in at about $55.7 billion against a $50 billion target, and the fiscal 2027 plan rose to up to $95 billion.
• The company is funding the build with debt, leases, customer money and $20 billion of new shares while keeping the dividend.
• Adjusted earnings exclude stock pay and say little about the cash cost of the build.

On my ten-point management check (two points each), Oracle scores 8 of 18 scored points; the buyback point does not apply. Growth is delivered, the cash discipline is not yet visible.

Stock Analyza scorecard

The valuation score is 35 of 100.

The balance sheet is the weak spot: net debt near $100 billion, leases on top and negative free cash flow. Business quality and valuation are shown separately and are not averaged.

Confidence score

The grade measures how well the model is identified, not the chance of a loss.

What I would watch from here

The second quarter report in mid-December has to show revenue growth of 30% to 34% and cloud growth of 65% to 71%, capital spending in line with the $95 billion annual plan, backlog still rising and a clear funding plan for the rest of the year.

The three most important thesis breakers

1. An AI customer that cannot pay for contracted capacity, which leads toward the Bear because the leases and equipment stay.
2. A fiscal 2028 capex plan above this year's with no increase in the revenue outlook.
3. The fiscal 2030 target of $225 billion reconfirmed while capex falls, which leads toward the Bull.

Stress test and independent check

I rebuilt the cash flow model, the multiple, the buy range and the price ladder in a second, separately written script, and the results match to the cent. That checks the arithmetic, not the forecasts.

The weakest point is the capex path: the Base needs Oracle's own spending to fall from $70 billion to about $50 billion within five years while revenue more than doubles. The strongest bull argument is the contracted backlog, $664 billion with prepayments and customer funding attached. The strongest bear argument is that the contracts are only as good as the few customers behind them, while Oracle's leases run for up to 19 years.

Final verdict: hold at $115.16

Oracle has turned itself into one of the largest builders of AI computing capacity, financed partly by its customers, partly by its lenders and partly by its shareholders. At $115.16 the market prices that bet at about what my Base says it is worth. That is a fair price for a wide range of outcomes, not a margin of safety.

Verdict: HOLD at $115.16, watchlist for new money. Base $120.86, 4.9% above the price, Bear $42.71, Bull $240.94, Confidence C (16.25/25). Required margin of safety 42.5%, buy range $59.59 to $82.91, maximum buy price $59.59.

New money waits for the buy range. For an existing holder the original cost is irrelevant; the price is close to fair value, and the next two reports will show whether capital spending starts to bend while revenue keeps up. The backlog says Oracle sold the capacity. The cash flow has to show that the buyers can pay for it.

Sources

Primary sources are Oracle's first quarter fiscal 2027 results release (September 10, 2026) and its Form 10-Q for the quarter ended August 31, 2026. All filings are on SEC EDGAR.

The fiscal 2027 capex plan and fiscal 2026 capex come from Outlook Business (June 11, 2026), and the fiscal 2030 targets from TipRanks (October 2025). The reference market price is $115.16, the November 5, 2026 close.