ORCL Oracle Corporation
A software cash cow levered into AI infrastructure: the backlog is real, but returns on that capital are unproven.
The story
Oracle is a mature, high-margin database and applications franchise using its cash flows and a lot of new debt to become a hyperscale AI infrastructure provider (OCI). The contracted backlog is real and gives revenue visibility. But the business is turning from a capital-light software company into a capital-heavy compute landlord: TTM capex of $75.7B is about 7x D&A and above revenue. Oracle is effectively starting a second, lower-return life cycle on top of a cash cow.
I kept the drivers from a month ago and only nudged year-1 growth from 20% to 22% (not a material move), because OCI backlog is converting into revenue. The TTM margin of 37.4% still flatters the future, since GPU-rental mix and depreciation from the capex surge should pull margins back toward 30%. Sales-to-capital stays at 0.75 on the assumption that today's 0.43 improves as capacity comes online, though it is still far below software norms. Beta of 1.3 and a 6% failure probability reflect $98B of net debt and counterparty concentration among AI labs.
Value drivers
| Revenue growth (Y1) | 22.0% |
| Terminal growth | 3.5% |
| Forecast horizon | 10y |
| Target operating margin | 30.0% |
| Years to target margin | 6 |
| Sales-to-capital | 0.75 |
| Beta | 1.30 |
| Failure probability | 6.0% |
| Cost of capital (WACC) | 9.9% |
| Terminal WACC | 8.9% |
Valuation bridge
| PV of explicit FCFF | 126.24B |
| PV of terminal value | 347.12B |
| Equity value | 347.43B |
| ÷ shares → per share | $114.61 |
Projected free cash flow to the firm
| Yr | Revenue | Growth | Margin | NOPAT | Reinvest | FCFF | PV |
|---|---|---|---|---|---|---|---|
| 1 | 87.57B | 22.0% | 36.2% | 27.67B | 21.05B | 6.61B | 6.02B |
| 2 | 105.03B | 19.9% | 34.9% | 32.06B | 23.29B | 8.77B | 7.26B |
| 3 | 123.82B | 17.9% | 33.7% | 36.46B | 25.05B | 11.41B | 8.59B |
| 4 | 143.43B | 15.8% | 32.5% | 40.69B | 26.14B | 14.55B | 9.97B |
| 5 | 163.19B | 13.8% | 31.2% | 44.54B | 26.35B | 18.19B | 11.35B |
| 6 | 182.32B | 11.7% | 30.0% | 47.79B | 25.51B | 22.29B | 12.65B |
| 7 | 199.94B | 9.7% | 30.0% | 52.41B | 23.50B | 28.92B | 14.93B |
| 8 | 215.16B | 7.6% | 30.0% | 56.40B | 20.29B | 36.11B | 16.97B |
| 9 | 227.11B | 5.6% | 30.0% | 59.54B | 15.94B | 43.60B | 18.64B |
| 10 | 235.06B | 3.5% | 30.0% | 61.62B | 10.60B | 51.02B | 19.85B |
Key risks
- Customer concentration: a few AI labs with weak balance sheets underpin much of the backlog, so contract renegotiation or default would strand capacity
- Capex outruns returns: GPU depreciation and financing costs compress margins below 30% and ROIC below the cost of capital
- Rising leverage and credit-rating pressure raise the cost of debt and limit flexibility
Catalysts
- Quarterly evidence that the backlog is converting into OCI revenue at stable gross margins
- A capex peak followed by free cash flow turning positive, with deleveraging