Data status: September 8, 2026. Oracle will report fiscal first-quarter 2027 results after the market closes on September 10. Few large-cap earnings reports carry a stranger combination of momentum and financial stress.
The operating story looks spectacular. Oracle ended fiscal 2026 with $67.4 billion of revenue, $34.0 billion of cloud revenue and $18.1 billion of cloud infrastructure revenue. In the fourth quarter alone, Oracle Cloud Infrastructure, or OCI, grew 93% year over year. Remaining performance obligations reached $638 billion, up 363% from a year earlier. Management guided fiscal Q1 2027 revenue growth to 27%–29% and cloud growth to 58%–64% in U.S. dollars.
The financing story looks almost inverted. Oracle spent $55.7 billion on capital expenditures in fiscal 2026, generated $32.0 billion of operating cash flow and therefore reported negative free cash flow of $23.7 billion. The company raised $43 billion of debt and $5 billion of equity financing during the year, and management expects roughly another $40 billion of debt and equity financing in fiscal 2027.
This is why Oracle stock earnings on September 10 matter more than a simple beat or miss. Investors are no longer asking whether Oracle has AI demand. The backlog has already answered that. The question is whether Oracle can convert that demand into revenue and cash fast enough to justify one of the most aggressive infrastructure buildouts in corporate history.
The earnings setup in one sentence
Oracle is trying to turn a software company into an AI infrastructure utility without giving up the economics that made the software company attractive in the first place.
That sentence captures the entire tension. Traditional Oracle generated recurring software support cash flows, enjoyed high margins and needed comparatively modest physical capital. The new Oracle is leasing land, contracting power, buying or hosting GPUs and building data centers at a scale more familiar to industrial infrastructure than enterprise software.
If utilization stays high and customers honor enormous multi-year contracts, the capital intensity can create a durable moat. If AI demand, pricing or customer concentration disappoints, Oracle will have financed a very expensive asset base before the economics were fully proven.
What Oracle actually reported in fiscal 2026
Oracle’s June results were not a normal software quarter. Fourth-quarter revenue reached $19.2 billion, up 21%. Cloud revenue reached $9.9 billion, up 47%. OCI revenue was $5.8 billion, up 93%, while SaaS revenue grew a much slower 10%.
For the full year, revenue rose 17% to $67.4 billion. Cloud revenue rose 39% to $34.0 billion. OCI revenue increased 77% to $18.1 billion. GAAP operating income was $20.6 billion and GAAP net income available to common shareholders reached $17.0 billion.
The growth mix matters because OCI is becoming the center of the valuation debate. Oracle is not merely selling more database licenses into the same enterprise installed base. It is competing for hyperscale AI workloads against companies such as Microsoft and Amazon, both of which have much larger cloud franchises and enormous balance sheets.
That comparison is useful. Our Microsoft stock analysis shows what the AI infrastructure cycle looks like when a hyperscaler can fund capex from a giant existing cash machine. Oracle is attempting a similar physical expansion from a smaller financial base.
Source: Oracle Q4 and FY2026 earnings release. Growth rates are reported in U.S. dollars.
The $638 billion backlog is both the bull case and the source of confusion
Oracle’s remaining performance obligations ended fiscal 2026 at $638 billion. The figure increased by $85 billion in the fourth quarter alone and by 363% year over year.
That number is extraordinary relative to current annual revenue. It tells investors that Oracle has signed commitments whose nominal value stretches far into the future.
But backlog is not revenue. Revenue is not profit. Profit is not free cash flow.
The conversion path matters. A contract can be signed today, require data-center construction over several years and only begin producing material revenue when capacity is installed and accepted. Some contracts include customer prepayments or customer-supplied GPUs, which improve Oracle’s financing economics. Others still require Oracle to secure power, buildings, networking and supporting equipment before the contract becomes a productive asset.
Oracle disclosed that prepaid and customer-supplied hardware portions of large AI contracts totaled about $75 billion. That is a critical detail. It means customers are helping finance the very GPUs that create future cloud revenue, reducing the amount Oracle must fund from its own balance sheet.
The structure resembles parts of the financing model we see at specialized AI infrastructure providers. Our Applied Digital analysis makes the same distinction: a signed contract reduces demand risk, but the investor still has to ask who funds the physical asset and what return remains for equity holders after financing.
The Q1 guidance sets a very high bar
Oracle’s official fiscal Q1 2027 guidance calls for total revenue growth of 27%–29% and total cloud revenue growth of 58%–64% in U.S. dollars. Non-GAAP EPS is expected between $1.72 and $1.76.
Those numbers are important for two reasons.
First, the guidance already assumes a major acceleration. A company the size of Oracle growing total revenue near 30% would be exceptional. Investors therefore need to distinguish a strong result from a strong result relative to expectations. A 25% revenue increase would look impressive in isolation and still be a disappointment if the company guided to 27%–29%.
Second, cloud growth remains the proof point. Oracle’s old software base is durable, but the current investment case depends on OCI becoming a much larger percentage of the company. If OCI growth decelerates materially while capex remains elevated, the market will have to reconsider the return profile of the infrastructure buildout.
The most important number may be revenue recognized from the backlog
I would pay special attention to management commentary on the timing of RPO conversion.
Backlog growth can keep the narrative alive for years. Revenue conversion determines whether the economics are arriving on schedule. If Oracle adds another enormous block of RPO but delays capacity delivery, investors may eventually discount the backlog more heavily. If revenue growth accelerates into the high 20s while RPO remains strong, the contracts begin to look less like a promise and more like an operating system.
For Q1, I would want answers to four questions:
- How much new OCI capacity came online during the quarter?
- Did large AI contracts begin recognizing revenue on schedule?
- Are customer prepayments increasing fast enough to reduce Oracle-funded capex?
- Is the backlog broadening across customers, or becoming more concentrated?
Why Oracle’s negative free cash flow is not automatically a red flag
Oracle generated $32.0 billion of operating cash flow in fiscal 2026. That is real cash produced by the business. Free cash flow turned negative because capital expenditures rose to $55.7 billion from $21.2 billion a year earlier.
This difference matters.
A company burning cash because customers are leaving is very different from a company consuming cash because contracted demand requires new physical capacity. Oracle’s operating cash flow increased 54% in fiscal 2026. The underlying business is not collapsing.
But growth capex is not free. Investors sometimes make the opposite mistake and assume every dollar of AI capex should be excluded because it is “for growth.” A data center must still earn a return. GPUs depreciate. Power contracts can become expensive. Buildings and networking consume capital long before the full revenue stream appears.
The correct question is not whether the capex is maintenance or growth. The correct question is whether the future incremental cash flow generated by that capex exceeds the cost of the capital used to fund it.
The financing bill is now part of the product
Oracle issued about $42.7 billion of senior notes net of issuance costs during fiscal 2026 and raised $5.0 billion through mandatory convertible preferred stock. Management has said it expects roughly $40 billion of additional debt and equity financing in fiscal 2027, including a previously announced $20 billion at-the-market equity program.
This changes how I think about Oracle’s AI cloud economics.
The old software model could be analyzed mainly through revenue growth, margins and retention. The new infrastructure model requires a fourth variable: cost of capital.
If a data-center project earns a high return on invested capital, debt can amplify equity returns. If the project earns only modest returns, higher interest expense and dilution can absorb the benefit. That is why investors should not celebrate backlog without asking how much equity is needed to deliver it.
This issue is even more visible at pure-play AI infrastructure companies. Our CoreWeave deep dive examines what happens when the financing stack becomes almost inseparable from the product economics.
Customer concentration is the hidden variable
Large AI infrastructure contracts can make quarterly growth spectacular while also increasing concentration risk.
Oracle has benefited from very large AI commitments, including well-publicized relationships with frontier model developers. The credit quality and strategic importance of those customers matter. So does the percentage of Oracle’s future capacity tied to a small number of counterparties.
A giant contract with a high-quality customer can be safer than thousands of small speculative customers. But it can also create bargaining power, pricing pressure and renegotiation risk if too much capacity is built around one relationship.
I would therefore listen carefully for customer diversification, contract duration, prepayment structures and any changes in deployment timing.
The software business still matters more than the AI headlines suggest
Oracle is not starting from zero. It has a deeply embedded database franchise, enterprise applications, NetSuite, Fusion, Cerner and a massive installed base.
That legacy distribution is one reason the cloud transition is credible. Existing database customers can move workloads into OCI without treating Oracle as a completely new vendor. Oracle can bundle database economics, applications and infrastructure in ways a smaller neocloud cannot.
The risk is that investors begin valuing the company entirely through the AI infrastructure lens and forget that software growth outside OCI is slower. In Q4, SaaS growth was only 10%. A valuation that assumes hyperscale growth across the entire company would be too generous.
What would count as a strong Q1 result?
For me, a strong quarter would need more than beating EPS by a few cents.
- Total revenue growth at or above the 27%–29% guide. This would confirm that the backlog is converting quickly enough to change the consolidated company.
- Cloud growth near the top of the 58%–64% range. The faster OCI scales, the more credible the infrastructure thesis becomes.
- RPO remains very strong without deteriorating contract quality. Growth should not come from weaker financing terms or unsustainably concentrated customers.
- Capex remains high but increasingly matched by prepayments and operating cash flow. The best outcome is not lower capex by itself. It is better funding efficiency.
- No material increase in financing stress. Investors need evidence that debt and dilution remain tools rather than symptoms.
What could disappoint even if revenue beats?
Oracle could report a good headline quarter and still produce a negative stock reaction.
One risk is that OCI growth beats while capex guidance rises faster. Another is that RPO grows but revenue conversion slips. A third is that management announces more debt or equity issuance than investors expected. A fourth is margin pressure as data-center depreciation and infrastructure costs flow through the income statement.
In a capital-intensive growth cycle, the market eventually stops rewarding revenue in isolation. It asks whether each additional dollar of revenue is creating enough economic profit.
Valuation: Oracle is cheaper than the narrative sounds, but the cash-flow math is harder
Oracle’s share price has fallen sharply from its prior peak, which makes the stock look less expensive on conventional earnings multiples than many AI infrastructure names. But a simple P/E ratio can understate the risk because reported earnings do not immediately reflect the cash required to build future cloud capacity.
I would use a two-layer framework.
Layer one: operating earnings. If revenue grows near 30% and operating margins remain resilient, Oracle can support strong earnings growth even while the cloud mix changes.
Layer two: capital return. Investors should then subtract the economic cost of the capital program. That means tracking free cash flow, debt, share count, interest expense and the timing of revenue from newly built capacity.
A useful way to avoid false precision is a reverse-valuation question: what level of sustainable OCI growth and free-cash-flow recovery would be required for the current valuation to look attractive five years from now? Our DCF valuation guide explains why this expectations-based approach is often more useful than pretending one target price is objectively correct.
Three scenarios after Q1
Bear case: OCI growth begins slowing faster than expected, capacity delivery slips, capex stays above $50 billion annually and debt plus equity issuance continue. Oracle still grows, but the economics resemble an expensive infrastructure build rather than a high-return software transition.
Base case: Oracle broadly meets Q1 guidance, cloud growth remains around 60%, the $638 billion backlog converts steadily and customer prepayments continue to offset hardware funding. Free cash flow stays negative near term but begins moving toward normalization as operating cash flow catches up.
Bull case: OCI growth stays well above 60% for longer than expected, RPO expands while customer concentration improves, financing needs peak and operating cash flow rises fast enough to make the infrastructure program self-funding. In that world, Oracle becomes a genuine hyperscale AI cloud platform rather than a software company renting a temporary AI boom.
What I would watch on the earnings call
- OCI revenue growth and capacity commentary.
- RPO growth and the share tied to large AI contracts.
- Customer prepayments and customer-supplied GPU structures.
- Fiscal 2027 capex guidance.
- Debt and equity financing needs.
- Operating cash-flow growth.
- Cloud gross margin and depreciation.
- NetSuite, Fusion and Cerner growth outside OCI.
- Any change to fiscal-year revenue expectations.
My view before September 10
Oracle has one of the most unusual risk/reward profiles among mega-cap software companies.
The bull case is not speculative demand. The contracts exist. OCI growth is already extraordinary. The backlog is already enormous. The company is clearly becoming more important to the global AI infrastructure market.
The uncertainty is whether Oracle can finance and deliver that growth without transferring too much of the economics to creditors, equipment suppliers and newly issued shareholders.
That is why I would not frame the September 10 earnings report as a simple question of whether Oracle beats consensus. The more important test is whether the company shows evidence that the capital cycle is becoming more efficient as the revenue cycle accelerates.
Oracle does not need to prove that AI demand exists. It needs to prove that the $638 billion backlog can become attractive per-share cash flow.
If that conversion begins happening faster than the financing bill grows, the stock can look materially more compelling. If capital needs continue rising faster than the cash return, investors may discover that an extraordinary backlog can coexist with an ordinary equity return.
Oracle stock earnings FAQ
When does Oracle report Q1 FY2027 earnings?
Oracle is scheduled to report fiscal Q1 2027 results on Thursday, September 10, 2026, after the U.S. market closes.
What is Oracle guiding for Q1 revenue growth?
Oracle guided total revenue growth to 27%–29% in U.S. dollars and total cloud revenue growth to 58%–64%.
How large is Oracle’s backlog?
Remaining performance obligations ended fiscal 2026 at approximately $638 billion.
Why is Oracle free cash flow negative?
Fiscal 2026 operating cash flow was $32.0 billion, but capital expenditures reached $55.7 billion as Oracle rapidly expanded data-center capacity, producing negative free cash flow of about $23.7 billion.
What is the biggest risk for Oracle stock?
The biggest risk is that infrastructure spending, debt and dilution grow faster than the cash economics generated by OCI and the AI backlog.
Sources
This article is independent financial analysis for educational purposes and does not constitute investment advice.
