Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Oracle’s fiscal third-quarter 2025 results showed why CEO Safra Catz called AI a “motivator” for cloud adoption: customers could gain access to AI-enabled applications, databases and infrastructure by modernizing systems that had remained on premises. The quarter also showed the limits of that thesis. Cloud demand and contracted future work grew quickly, but Oracle said data-center capacity, power and components constrained how fast it could deliver.
Oracle reported the results on March 10, 2025, for the quarter ended February 28. Its cloud infrastructure growth and rising GPU use offered evidence of strong demand, while its $130 billion remaining performance obligations (RPO) figure represented future contracted work—not revenue already earned.
What Catz meant by AI as a “motivator”
On Oracle’s earnings call, an analyst asked whether AI functionality was becoming a tipping point for customers that had postponed moving workloads from on-premises systems to the cloud. Catz answered, “Yes, I mean this is the motivator.” She was describing AI-enabled systems as a practical reason to modernize: they could handle more routine work, save time, help reduce operating costs and draw more insight from company data.
The point was about what AI could prompt customers to adopt—not a claim that every Oracle customer was moving because of AI alone. Nor did Oracle report a separate, comprehensive AI-revenue line. Catz’s argument was that AI features in applications and database services could make a broader cloud or software purchase more compelling. The earnings-call transcript provides the exchange and management’s explanation.
#1 Best Overall
- New
- Mint Condition
- Dispatch same day for order received before 12 noon
- Guaranteed packaging
- No quibbles returns
Oracle’s fiscal Q3 2025 scorecard
The reported figures point to a company growing much faster in cloud infrastructure than in its overall business. Growth rates below are year over year; USD means reported U.S. dollar growth, while constant currency adjusts for exchange-rate movements.
| Metric | Fiscal Q3 2025 | Year-over-year change |
|---|---|---|
| Total revenue | $14.1 billion | +6% USD; +8% constant currency |
| Total cloud revenue (IaaS and SaaS) | $6.2 billion | +23% USD; +25% constant currency |
| Cloud infrastructure revenue | $2.7 billion | +49% USD; +51% constant currency |
| Cloud applications revenue | $3.6 billion | +9% USD; +10% constant currency |
| Fusion Cloud ERP revenue | $0.9 billion | +16% USD; +18% constant currency |
| NetSuite Cloud ERP revenue | $0.9 billion | +16% USD; +17% constant currency |
| Remaining performance obligations (RPO) | $130 billion | +62% USD; +63% constant currency |
| Non-GAAP earnings per share | $1.47 | +4% USD; +7% constant currency |
| GAAP earnings per share | $1.02 | +20% USD; +25% constant currency |
These figures are from Oracle’s SEC-filed earnings release, which is also available through Oracle Investor Relations. The cloud infrastructure figure is not synonymous with AI revenue: it includes infrastructure services, and Oracle did not disclose a complete AI-only revenue measure.
Where AI demand showed up: OCI and GPU use
Oracle Cloud Infrastructure (OCI) was the clearest reported growth engine associated with AI demand. Management said OCI revenue rose 51% in the quarter, OCI consumption revenue rose 57%, and GPU consumption revenue was nearly 3.5 times its year-earlier level. Those are management-reported usage and revenue measures; they do not establish what share of OCI revenue came from AI workloads.
Oracle also pointed to the scale of its build-out. In the earnings release, Executive Chairman and CTO Larry Ellison said the company had crossed 101 cloud regions and that GPU consumption for AI training had grown 244% over the previous 12 months. He cited plans for a 64,000-GPU liquid-cooled NVIDIA GB200 cluster and a multibillion-dollar contract with AMD for a cluster of 30,000 MI355X GPUs. These are company statements about its footprint, usage and planned infrastructure, not independent assessments of performance or profitability. The release and call discussion are in Oracle’s filing and the call transcript.
Multicloud lets Oracle meet customers where they run workloads
Oracle’s strategy is not limited to persuading customers to move everything to OCI. Through its Database@ services, Oracle makes database technology available in or alongside other hyperscalers’ environments. Management discussed services associated with Microsoft Azure, Google Cloud and Amazon Web Services, as well as OCI and dedicated deployment models such as Cloud@Customer.
- Database@ services: Oracle database capabilities deployed in connection with a hyperscaler environment, allowing customers to use Oracle databases alongside that provider’s services.
- OCI: Oracle’s own public-cloud infrastructure, where customers can run infrastructure, database and other workloads.
- Cloud adoption: A customer can modernize an Oracle database or application without making OCI its sole or primary cloud provider.
Oracle said multicloud database revenue from Microsoft, Google and Amazon rose 92% in the last three months, without specifying a comparable segment revenue total in the cited release. Call coverage also described 18 live regions for Database@ services and 40 more planned. The regional figures were management commentary, not a statement that all planned sites were already operating. MarketBeat’s call display and the transcript report those details.
This approach can lower the friction of adopting Oracle database services for organizations already committed to Azure, AWS or Google Cloud. It also means Oracle may share parts of the customer relationship and economics with those providers. Oracle did not disclose enough information in the cited materials to compare margins by multicloud channel, so the financial trade-off cannot be quantified from these results.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Applications and databases give the AI thesis a customer-facing form
AI Data Platform and Oracle Database
Oracle introduced an AI Data Platform intended to connect models from providers including OpenAI, xAI and Meta with data held in Oracle databases. The proposition is that organizations can apply AI to enterprise data while keeping it within an environment they control. That describes the product strategy, not proof of broad adoption or independently measured customer results. Management’s product discussion appears in the earnings-call transcript.
Fusion Cloud Applications
Oracle said it was embedding AI agents in Fusion applications for finance, supply chain, human capital management and other workflows. The commercial logic is to improve the usefulness of the application suite through built-in assistance and automation, rather than necessarily selling each agent as a standalone product. Catz described potential productivity and cost benefits, but Oracle did not provide a broad, independently audited measure of customer savings from those agents.
NetSuite Cloud ERP
NetSuite Cloud ERP revenue increased 16% in U.S. dollars and 17% in constant currency. That growth shows strength in an applications business as well as in large infrastructure deployments; it does not by itself establish how much of the increase was caused by AI.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.RPO showed demand visibility, not revenue already recognized
Oracle said it signed more than $48 billion in sales contracts during the quarter, taking RPO above $130 billion. RPO is contracted future performance that has not yet been recognized as revenue. It can include commitments spanning multiple years, with revenue recognized as Oracle delivers the contracted services over time.
The distinction matters because delivery depends on capacity. Oracle said converting RPO into revenue partly depended on bringing data centers online and securing available power. A fast-growing backlog is evidence of demand and gives Oracle visibility into contracted work, but it also raises the execution bar: the company must build and operate the infrastructure needed to serve it. Oracle’s release said the cited Q3 RPO did not include anticipated business related to Stargate; it should not be counted as Q3 booked revenue. See the SEC-filed release and call transcript.
Capacity and capital spending were the central execution risks
Oracle expected fiscal 2025 capital expenditures to reach approximately $16 billion, more than double the prior year, as it expanded data-center and GPU capacity. Management said data-center capacity was expected to double during calendar 2025, available power capacity to double during calendar 2025 and to triple by the end of the following fiscal year. It also cited component delays and said demand exceeded supply. These were forward-looking management expectations as of the March 2025 call, not completed capacity additions.
The underlying risk is a conversion bottleneck: customers may sign commitments and consume available GPU capacity, but Oracle must secure components, power, facilities and operating capability to provide the contracted service. Until capacity is available and services are delivered, demand does not translate automatically into recognized revenue. Higher investment can enable growth, but the reported results did not establish the eventual return on that spending.
What Oracle forecast for Q4 and beyond
For fiscal Q4 2025, management guided to total revenue growth of approximately 9% to 11%, excluding foreign exchange. It expected total cloud revenue growth of approximately 24% to 28% in constant currency and 25% to 27% in U.S. dollars, and non-GAAP EPS of approximately $1.61 to $1.65 in U.S. dollars. The EPS outlook included a negative effect of more than $0.03 from losses on an investment in another company; the assumed tax rate was 19%, subject to one-time tax fluctuations. These were forecasts at the time, not subsequent reported results. Details are in the call transcript and MarketBeat’s transcript display.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Oracle also said its confidence in a $66 billion total-revenue target for fiscal 2026 had strengthened and that it had raised its fiscal 2027 growth expectation to 20%. Those statements, reported in CRN’s coverage of the call, were management forecasts rather than results already achieved.
What executives and investors should watch
- Revenue conversion: Whether RPO growth is followed by reported cloud revenue as contracted services are delivered.
- OCI and GPU consumption: Whether infrastructure growth and usage remain strong, while remembering that neither measure isolates all AI revenue.
- Application growth: Whether Fusion and NetSuite continue expanding alongside OCI, indicating that the cloud story is not only about GPU clusters.
- Capacity delivery: Whether Oracle meets its stated build-out expectations for data centers and power, and resolves component constraints.
- Capital efficiency: Whether sharply higher investment produces durable revenue and returns; the Q3 release alone did not answer that question.
- Customer evidence: Whether named customer outcomes substantiate claims about AI-driven productivity, cost savings and migration decisions.
Oracle’s Q3 results supported the view that AI was helping make cloud modernization more urgent for some customers. They also showed why that demand thesis cannot be assessed from bookings or GPU growth alone: sustained growth depends on converting commitments into delivered services while funding and building enough capacity.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
