The latest quarterly and annual financial results from Oracle were released after the U.S. market closed on June 10th. Driven by exceptionally strong demand for AI computing power and the company's multi-cloud database strategy, the scale of its order backlog, revenue growth, and future capital expenditure plans have reached astonishing levels.
At the end of the fourth fiscal quarter, Oracle's remaining performance obligations (RPO) reached a record $638 billion, representing a net increase of $85 billion for the quarter and a staggering year-over-year growth of 363%. This provides an exceptionally high degree of certainty for the company's revenue growth in the coming years.
Consequently, management has issued high guidance for fiscal year 2027, projecting revenue growth of 34%. Based on fiscal 2026 revenue exceeding $67 billion, achieving this 34% growth would bring fiscal 2027 revenue close to the $90 billion mark.
However, concerns about data center costs are intensifying. Oracle's capital expenditures for fiscal 2027 are projected to be approximately $70 billion, and the company plans to double its financing for the new fiscal year to $40 billion. Management emphasized that they do not anticipate taking on additional debt within the 2026 calendar year.
New CFO Hilary Maxson sought to ease market concerns. She stated that once infrastructure projects reach a steady state, the return on invested capital at the project level could reach the high 20% range. Management also reiterated the long-term potential for the infrastructure business to achieve 30% to 40% operating margins.
AI Enthusiasm Takes Concrete Form
Within the earnings report, the metric that most shocked the market was the remaining performance obligations (RPO).
At the end of Q4, Oracle's RPO hit a record $638 billion, a net quarterly increase of $85 billion and a year-over-year surge of 363%. This provides a very high level of visibility for future revenue growth.
Where did these astronomical order additions come from? CEO Clay Magouyrk provided the answer on the call: "This quarter we signed $67 billion in AI infrastructure contracts, the majority of which were 'Bring Your Own Hardware' (BYOH) or 'prepaid' models. The cumulative total for these two contract types has now reached $75 billion."
Facing the surge in computing demand, Clay could not hide his excitement, stating plainly: "AI infrastructure makes the existing cloud infrastructure market look small. Everything we see indicates this market is worth trillions of dollars annually."
He even used data to dispel market concerns about an "AI computing glut": "Our global GPU utilization is 97.5%. In Q4, 35,000 GPUs from 59 different customers were up for renewal; 49% of those customers renewed 92% of the GPUs. But that doesn't mean the other 8% sat idle; the vast majority of those returned GPUs were immediately sold to other customers within the same quarter."
The "Good Problem" of High Capital Expenditure
The surge in orders necessitates a high-intensity capital expansion. CFO Hilary Maxson indicated that net cash outflows for capital expenditures in fiscal 2027 are expected to be around $70 billion.
To support this massive capital expenditure plan, the company expects to raise approximately $40 billion in fiscal 2027 through debt and equity financing, including a previously announced $20 billion at-the-market equity offering.
Hilary explicitly stated there are "no plans to add additional debt financing in calendar year 2026," aiming to protect the company's investment-grade credit rating.
This large-scale, asset-heavy expansion plan sparked short-term market concerns after hours. However, from management's perspective, this is a necessary cost to lock in future revenue.
When asked about the high capital expenditure, Clay candidly admitted: "My job is to figure out how to spend money faster, so that sometimes I can get the increased revenue."
He explained that due to supply chain and component price increases (like memory, SSDs), Oracle signs fixed-price contracts with customers when costs are certain; when costs are uncertain, a floating mechanism is used to ensure "Oracle is not in a position where its margins are declining."
To alleviate investor concerns about the asset-heavy model compressing profits, CFO Hilary laid out the economics: "We believe the returns on the infrastructure business (CPU and GPU business) are very strong. Roughly speaking, when large projects reach steady-state revenue, the return on invested capital can be in the high 20% range. And in models like 'Bring Your Own Hardware,' that return is even higher."
AI's Full-Scale Implementation
Beyond the underlying computing power, Oracle's AI monetization in software and databases is also accelerating.
Co-CEO Mike Sicilia noted: "Our customers have moved past the experimentation phase with AI. They are ready to implement enterprise-grade, full agentic solutions to help run their businesses."
To lower the barrier for customer AI adoption and provide a fast return on investment, Oracle introduced a disruptive business model—"pay-for-outcomes" pricing. Mike explained: "For example, an interview agent is priced based on the number of candidates screened, or a hotel upsell agent is priced based on the percentage of end-consumer upsell transactions closed."
Additionally, Oracle launched "Token bundles," allowing customers to access more advanced AI inference capabilities by pre-purchasing tokens.
On the database front, the AI data strategy is driving exponential growth in multi-cloud business.
In Q4, Oracle Cloud database revenue grew 29%, with multi-cloud revenue skyrocketing 404% year-over-year and bookings up 325%, making it one of the company's fastest-growing business lines.
Looking ahead, Oracle expects total revenue growth to accelerate to 27%-29% in fiscal Q1 2027, with cloud revenue growth jumping further to 57%-64%.
Concurrently, the company maintained its full-year revenue target of $90 billion and raised its non-GAAP EPS guidance to $8.05 (representing approximately 18% growth year-over-year after excluding one-time gains), demonstrating management's absolute confidence in converting the massive order backlog into actual profits.
Comments