
Key Points
- Oracle delivered a strong set of results, beating both revenue and EPS expectations. The stock declined by approximately 5% in after-hours trading, weighed down by higher-than-expected capital expenditure and management's announcement of plans to secure an additional US$40 billion in financing during FY2027.
- While operational execution remains strong, Oracle will need to continue delivering robust earnings growth and converting its substantial backlog into recognised revenue over the coming quarters.
- Therefore, we remain Neutral on the stock.
Oracle delivered a strong set of results, beating both revenue and EPS expectations. Cloud Infrastructure (IaaS) revenue growth accelerated to 93% year-on-year, while Remaining Performance Obligations (RPO) surged to US$638 billion, representing record levels by virtually every measure. Despite the impressive results, the stock declined by approximately 5% in after-hours trading, weighed down by higher-than-expected capital expenditure and management's announcement of plans to secure an additional US$40 billion in financing during FY2027.
AI as the key growth engine
Cloud Infrastructure (IaaS) revenue reached an impressive US$5.8 billion, growing 93% year-on-year in US dollar terms, driven by robust demand for AI training and inference workloads. This was the standout metric of the quarter and exceeded market expectations.
At the same time, Remaining Performance Obligations (RPO)—representing contracted revenue that has yet to be recognised—reached US$638 billion, up 363% year-on-year and significantly above analyst expectations of US$595.7 billion. Management indicated that approximately 12% of the current RPO balance is expected to be recognised over the next 12 months, with a further 34% anticipated between months 13 and 36. Importantly, management noted that both percentages are expected to increase over time. This provides an unusually high degree of revenue visibility and is a key factor supporting the credibility of Oracle's US$90 billion revenue target for FY2027.
While Oracle does not disclose operating margins for its cloud infrastructure business, management stated that cloud margins are expected to expand in the coming years. This helps explain why the company reaffirmed its US$90 billion FY2027 revenue target while simultaneously raising its non-GAAP EPS guidance to US$8.05, ahead of consensus expectations. That said, management has not provided sufficient detail regarding the drivers of future margin expansion, making it difficult to assess the relative contribution from operational efficiencies versus scale benefits..
Another notable disclosure was Oracle's announcement where prepaid customer commitments and customer-supplied hardware associated with its large AI contracts now total approximately US$75 billion. While this represents only around 12% of the company's total RPO balance, we believe it may represent a potential competitive advantage relative to peers if customer demand remains sufficiently strong.
As highlighted in our previous article, this structure allows Oracle to reduce a portion of the capital expenditure burden typically associated with large-scale AI infrastructure deployments. However, the trade-off is that Oracle may forgo a portion of the long-term economics of these contracts, resulting in lower margins over time compared with a fully self-funded infrastructure model.
Table 1: Key financial metrics

Source: Oracle Quarterly Report, Analyst estimates, Data as of 11 June 2026.
Table 2: Oracle data centre projects under Stargate project

Source: Oracle Quarterly Report, Analyst estimates, Data as of 11 June 2026.
Tale of two halves – Cloud Applications
Despite the significant beat in cloud infrastructure results, one of the key areas that disappointed investors was the Cloud Applications (SaaS) segment. Cloud Applications revenue came in at US$4.13 billion versus consensus expectations of US$4.17 billion—a small miss in absolute terms, but a meaningful disappointment in the context of elevated investor expectations and Oracle's broader AI-driven narrative.
The result attracted attention because recent results from other software names, including Snowflake, had reinforced the view that investors may have overstated the disruptive impact of AI on SaaS companies, particularly for vendors operating in systems of record (SoRs), which were thought to be relatively resilient. Oracle's results did not clearly reinforce that narrative.
Management highlighted strong adoption of AI-related offerings and noted that Oracle's cloud database business grew 29% year-on-year, while multi-cloud revenue surged 404% year-on-year. The company has also begun introducing agent-based pricing models (see appendix). However, the strength in AI demand has not yet translated into a meaningful acceleration in reported Cloud Applications growth. One possible explanation is that a significant portion of current AI-related revenue maybe coming from Oracle's existing on-premise customer base, while migrations from on-premise to cloud may be partially cannibalising revenue within the same segment.
Capex Tab Unnerves the Market
Capital expenditure for the latest quarter reached US$15.9 billion, bringing total FY2026 capital expenditure to US$55.7 billion, above prior guidance of US$50 billion.
What drew the greatest investor scrutiny was management's outlook for FY2027. Oracle guided to approximately US$70 billion of net cash capex in FY2027 and indicated plans to raise roughly US$40 billion through debt and equity financing, including the previously announced US$20 billion at-the-market equity issuance. Management also introduced a new framing for capital intensity: net cash capex excludes the US$20–25 billion of customer advance payments expected to be received in FY2027. On a gross basis, total capex could therefore approach roughly US$90 billion.
This guidance came after Oracle reported negative free cash flow (FCF) of US$23.7 billion for FY2026. With even higher capex outlays expected in FY2027, we estimate that FCF could deteriorate further, potentially reaching approximately negative US$42.4 billion, which helps explain the market's negative reaction.
That said, we also expect very strong top-line growth, particularly in Cloud Infrastructure (IaaS), which should improve EBITDA generation and reduce the company's debt-to-EBITDA ratio relative to the scale of its infrastructure investments. Nevertheless, because the capex programme is expected to be financed through a combination of debt and equity issuance, some degree of shareholder dilution appears likely.
Figure 1: FCF and Total Debt to EBITDA

Stronger growth; but underlying risks remain
Oracle has demonstrated exceptionally strong cloud infrastructure growth on a percentage basis, outpacing larger hyperscaler peers. However, this growth has been driven by management's highly aggressive expansion strategy. Compared with other hyperscalers that are also investing heavily to capture AI-related demand, Oracle's financial position and balance sheet remain among the weakest within the peer group.
Despite ongoing improvements in scale and utilisation, Oracle's cost of goods sold (COGS) continues to grow faster than revenue, resulting in pressure on gross margins. Although gross margins have improved over the past three quarters, the underlying trend suggests that the company has yet to fully achieve the operating leverage typically associated with large-scale cloud infrastructure businesses.
Another key concern is customer concentration risk, particularly with OpenAI. Based on the latest estimates, OpenAI may account for a significant portion of Oracle's contracted backlog. While this highlights Oracle's strong positioning within the AI infrastructure ecosystem, it also creates a meaningful dependency on a single customer.
OpenAI reportedly remains unprofitable despite its rapid growth. Should a customer of this scale encounter financial difficulties or become unable to fulfil its contractual commitments, a significant portion of Oracle's backlog could ultimately fail to convert into recognised revenue. While we view such an outcome as unlikely at present, the concentration risk remains elevated and warrants close monitoring as Oracle continues to scale its AI infrastructure business.
Figure 2: COGS outpacing revenue growth

Strong Quarter; But More Is Needed, Remain Neutral on Oracle
Oracle is executing at a pace not seen in more than 15 years. The strength of its cloud infrastructure business and the unprecedented US$638 billion backlog make the long-term bullish case increasingly difficult to ignore. However, the near-term concerns remain largely financial in nature. Oracle still needs to demonstrate that its current strategy—characterised by negative free cash flow, rising debt levels and a US$20 billion equity issuance—is capable of generating sufficient long-term returns to justify the risks being assumed by investors.
Similar to the previous quarter, this was not the type of result that fully addresses the key concerns that have historically made investors cautious on the stock. While operational execution remains strong, Oracle will need to continue delivering robust earnings growth and converting its substantial backlog into recognised revenue over the coming quarters.
That said, we are encouraged by several developments. A more constructive stance would likely require evidence of stronger cost discipline. We are seeing early signs, where the company has been able to lower operating expenses—including line items such as sales and marketing which improves net income margin. We also expect to see significant improvements after a larger bulk of RPO being recognised in the coming year.
Nonetheless, at current valuation levels, we believe the investment case still requires a high degree of execution consistency relative to the quality of the balance sheet, the capital intensity of the business and the concentration of its demand base. While the growth opportunity remains substantial, investors are effectively being asked to underwrite significant operational, financing and customer concentration risks.
From a risk-reward perspective, we continue to prefer larger hyperscale cloud providers with stronger balance sheets, more diversified customer bases and proven free cash flow generation. Our preferred names remain Alphabet Inc., Microsoft Corporation and Amazon.com, Inc..
Overall, Oracle's AI-driven growth story remains compelling, but we believe additional evidence of sustainable margin expansion, stronger free cash flow generation and improved balance sheet discipline will be required before adopting a more constructive stance on the shares.
Table 3: Valuation
|
2025 |
2026 |
2027 Est |
2028 Est |
|
|
Revenue, Adj |
57,399.00 |
67,357.0 |
89,205.0 |
130,976.0 |
|
Growth %, YoY |
8.4 |
17.3 |
32.4 |
46.9 |
|
EPS, Adj |
4.43 |
5.94 |
6.28 |
9.01 |
|
Growth %, YoY |
13.98 |
33.2 |
5.7 |
43.4 |
|
P/E |
37.4 |
40.55 |
22.84 |
16.75 |
|
Fair P/E |
19 |
|||
|
Upside Potential |
13.83% |
|||
|
Price |
209 |
|||
|
Source: Bloomberg Finance L.P., iFAST compilations. Data as of 17 June 2026. |
||||
Extending my thoughts for the rest of the hyperscalers
When talking about customer diversification, the management mentioned how’s the situation of the renting business is doing – “Our infrastructure is fundamentally multi-tenant by design, enabling us to dynamically reallocate resources across customers. In the fourth quarter, 35,000 GPUs from 59 distinct customers reached renewal, with 49% of those customers renewing 92% of their GPUs. However, this does not mean 8% of GPUs were idle—most of these GPUs were quickly resold to other customers within the same quarter. Our global GPU utilization rate stands at 97.5%.”
This means that the GPUs that weren't renewed were almost immediately snapped up by other customers. Demand is strong enough that there's essentially a queue of buyers waiting.
This suggests AI infrastructure demand remains robust. However, the degree to which this experience is shared across other hyperscalers may vary depending on customer mix and capacity availability.
Appendix (Oracle’s Agent-Based Pricing Model)
Layer 1 — Included at no extra charge (baseline)
Core AI improvements — including embedded AI and standard agentic capabilities — are included within existing application subscriptions. Customers subscribing to Fusion financials, HCM, supply chain, or other applications receive these enhancements as part of the normal service evolution, much like database upgrades.
There's a baseline allocation included in a standard Fusion subscription — 20,000 AI Units per month — giving customers a chance to explore what is possible before making bigger commercial decisions.
Layer 2 — AI Units (consumption pricing)
Fusion AI runs on a consumption model. Every AI action — answering a question, generating a document, processing an approval — consumes AI Units. One AI Unit costs one cent.
Oracle landed on a single consumption measure — the AI Unit — as one model across Fusion, rather than a patchwork of different charging rules depending on which product area or type of agent is being used.
When you need more than the 20,000 monthly baseline, you buy token bundles. Oracle began a limited rollout of token bundles in Q4 FY2026, with 33 customers including Aon Services Corporation and Liberty Energy pre-purchasing tokens.
Layer 3 — Fusion Agentic Applications (premium subscription)
Beyond the standard AI Unit model, Oracle has introduced a separate subscription for Fusion Agentic Applications — which includes 30 million AI Units per year in addition to the 20,000 monthly units in the base Fusion subscription, the right to publish agentic applications to production environments, and access to the Agentic Applications Builder, a low-code environment that allows users to describe a desired outcome in natural language and have the platform construct the agent workflow. Oracle currently offers 22 pre-built agentic applications across ERP, SCM, HCM, and CX.
Layer 4 — Outcome-based pricing (the strategic shift)
This is the most forward-looking piece and what Oracle is actively expanding. Oracle is expanding outcome-based pricing across its applications portfolio — the model already exists in areas such as construction, hospitality, and healthcare, and is now being extended across the broader applications suite including Fusion. Examples include interview agents priced by the number of candidates screened, and hospitality upsell agents priced by a percentage of end-consumer upsell transactions.
Declaration:
For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) and the analyst who produced this report hold a NIL position in the abovementioned securities.
This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report – including all investment theses, ratings, price targets and conclusions – has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.
