Oracle reported $664 billion in remaining performance obligations, or RPO, at the end of its fiscal first quarter of 2027.
The number is enormous. It is also easy to read too much into it.
RPO is not $664 billion of current revenue, and it is certainly not $664 billion of cash waiting to arrive. It is contracted revenue that Oracle expects to recognize as it delivers services over time. The real question is therefore no longer whether Oracle can sign huge AI contracts.
It is this:
How much of that backlog can Oracle convert into revenue and, eventually, cash without having to spend an equally extraordinary amount of capital first?
Q1 gave us a much better answer than the backlog headline alone. Oracle Cloud Infrastructure revenue rose 121% year over year to $7.4 billion, which shows that the demand is beginning to turn into actual business. At the same time, Oracle spent $28.5 billion on capital expenditures and still reported negative free cash flow.
That tension—not the size of RPO by itself—is where I think the Oracle story now gets interesting.
Table of Contents
$664 Billion Is a Pipeline, Not Next Year’s Revenue
Oracle finished Q1 FY2027 with RPO of $664 billion, up $209 billion from a year earlier. The company said it signed more than $30 billion of additional AI cloud contracts during the quarter. Oracle Q1 FY2027 results
RPO represents contracted revenue that has not yet been recognized. Some of it will become revenue relatively soon; much of it will not.
That timing matters.
At the end of fiscal 2026, when Oracle’s RPO stood at $638 billion, the company expected to recognize only about 12% over the following 12 months. Another 34% was expected in months 13 through 36, another 34% in months 37 through 60, and the remainder later. Oracle FY2026 filing
Those percentages apply to the May 31, 2026 balance, not automatically to the new $664 billion Q1 figure. The mix could have changed as Oracle signed new contracts.
Still, the disclosure tells us something important: this is a multi-year infrastructure pipeline, not a near-term revenue forecast.
| Q1 FY2027 metric | Result |
|---|---|
| Total revenue | $19.3B |
| Cloud revenue | $11.6B |
| OCI infrastructure revenue | $7.4B |
| OCI revenue growth | +121% YoY |
| Operating cash flow | $23.1B |
| Capital expenditures | $28.5B |
| Free cash flow | -$5.4B |
| RPO | $664B |
Source: Oracle Q1 FY2027 results.
This is why comparing $664 billion directly with Oracle’s current annual revenue can be misleading. The economic value of that backlog depends on when the capacity becomes available, how quickly customers consume it, what margins Oracle earns and how much capital has to be committed before the revenue appears.
The contract value is real. The timing and economics are the part still being tested.
Q1 Proved Revenue Conversion—Not Yet Capital Efficiency
There is one major reason to take Oracle’s backlog more seriously after Q1: the infrastructure business is now growing fast enough to show up clearly in reported revenue.
OCI revenue reached $7.4 billion, up 121% year over year. Total cloud revenue climbed 62% to $11.6 billion, while total company revenue rose 30% to $19.3 billion. Oracle also said it added 850 MW of data-center capacity during the quarter and had delivered more than 300,000 GPUs to AI cloud customers since the end of Q4.
So the skeptical argument that Oracle merely has a giant pile of distant contracts is getting harder to make.
The business is scaling.
But scaling it is expensive.
Oracle spent $28.499 billion on capital expenditures in one quarter, compared with $8.502 billion a year earlier. Free cash flow came in at negative $5.396 billion.
That combination captures Oracle’s current challenge rather neatly: demand can be excellent and the business can still consume cash because the physical infrastructure has to exist before much of the contracted revenue can be recognized.
There is also an important detail in that $23.1 billion operating cash flow figure.
Oracle recorded $11.363 billion of increased deferred revenue from customer prepayments with a significant financing component during Q1. In other words, a meaningful part of the quarter’s cash inflow arrived because customers were helping fund infrastructure before Oracle delivered all of the corresponding service.
That is not a problem. In fact, it may be one of Oracle’s biggest advantages.
But it does mean I would not look at $23.1 billion of operating cash flow and conclude that the underlying infrastructure business suddenly became self-funding. The prepayment structure is part of the story, not something to strip out and forget.
The Customers Are Helping Pay for the Buildout
This is where Oracle’s AI contracts become more interesting than the headline backlog number.
At the end of FY2026, Oracle said the prepaid and customer-supplied hardware portions of its large AI contracts totaled about $75 billion. In some large contracts, customers prepaid Oracle so it could purchase GPUs. In others, the customers bought the GPUs themselves and supplied them to Oracle.
That shifts a significant part of the infrastructure funding burden away from Oracle.
Q1 made the effect visible in the cash-flow statement. Oracle reported $28.499 billion of CapEx, but its supplemental measure of net cash outlay for capital expenditures was $17.966 billion after accounting for customer prepayments and capital-expenditure-related financing cash flows.
That distinction deserves more attention than another record RPO headline.
If Oracle had to finance every GPU, server and data-center expansion entirely from its own balance sheet, a backlog of this size could become a financing problem almost as quickly as it became a growth opportunity.
Customer funding changes that equation.
It does not make the infrastructure free. Oracle still has to deliver the capacity, operate the facilities and earn an acceptable return from those assets. But it reduces the amount of Oracle capital that has to go into the ground before the customer begins consuming compute.
For a business scaling this quickly, that is a meaningful difference.
Prepayments Help With Funding. They Do Not Guarantee Good Economics.
This is also where I would stop short of calling the financing issue solved.
A customer prepayment helps answer who pays first. It does not answer whether the project ultimately earns enough.
Oracle still faces the economics of running an increasingly infrastructure-heavy business: data centers, GPUs, networking equipment, power, cooling and depreciation. Large long-duration contracts also introduce concentration and execution risk. When individual agreements become enormous, the value of the backlog depends increasingly on a relatively small group of customers continuing to need huge amounts of compute for years.
The shift is visible in Oracle’s own capital plans.
For FY2027, Oracle had said it expected to raise roughly $40 billion through debt and equity, including a $20 billion at-the-market equity program. Oracle completed that entire $20 billion common-stock ATM program during Q1.
That does not mean Oracle is in financial distress. It means the company has moved into a different kind of growth model.
The old Oracle could sell another software license without first constructing a small power-hungry industrial project.
AI infrastructure does not work that way.
Oracle now has to coordinate contract demand with GPUs, buildings, electricity and financing years in advance. Its ability to structure customer-funded capacity may therefore matter almost as much as its ability to win the contracts themselves.
There is one encouraging detail in the latest quarter. Oracle said the more than $30 billion of new AI cloud contracts signed in Q1 would have no incremental impact on its previously announced capital-raising plans because of how those contracts were structured.
That is exactly the kind of evidence I want to see.
More RPO is useful. More RPO without proportionately more Oracle-funded capital is much better.
That broader financing question also connects with how other hyperscalers are funding AI infrastructure. See Why Alphabet Issued Its First Australian Dollar Bond: Inside Google’s AI Financing Playbook.
What I Would Watch From Here
The next quarter does not need another spectacular backlog number to make the Oracle thesis stronger. Oracle already has more contracted AI demand than it can serve immediately.
The useful question is whether the economics improve as that backlog moves through the system.
| Metric | What it tells us |
|---|---|
| OCI revenue growth | Whether contracted demand is actually becoming delivered compute revenue |
| RPO recognition timing | How quickly the backlog can move into reported revenue |
| CapEx | How much infrastructure Oracle must build to support that growth |
| Net cash outlay for CapEx | How much of the build Oracle is effectively funding itself |
| Customer prepayments | How much infrastructure cost customers are willing to finance upfront |
| Debt and equity issuance | Whether growth is requiring progressively more external capital |
| Free cash flow | Whether revenue conversion is ultimately outrunning infrastructure spending |
OCI growth is already answering the first question positively. Revenue is moving.
The harder test is whether Oracle can keep increasing revenue faster than its own capital commitment.
If customer-funded GPU structures continue to scale and Oracle can reuse infrastructure efficiently, the $664 billion backlog becomes a much more valuable asset.
If each new wave of revenue requires another enormous round of Oracle-funded CapEx and external financing, the same backlog begins to look less impressive.
That is why I would rather see better conversion economics than another $100 billion added to RPO.
The $664 Billion Number Is Real. It Just Isn’t the Verdict.
Oracle’s $664 billion backlog is not accounting decoration. It is strong evidence that customers have committed to buying a remarkable amount of cloud capacity.
Q1 also delivered something more important than another contract announcement: OCI revenue rose 121%. The backlog is beginning to turn into actual business.
But I would not judge the success of Oracle’s AI expansion by how high RPO eventually goes.
The number I find more revealing this quarter is the contrast between $28.5 billion of reported CapEx and $17.97 billion of Oracle’s supplemental net cash outlay for CapEx. Customer prepayments are absorbing a meaningful portion of the financing burden.
That is the model Oracle now has to prove it can repeat.
The bullish version is straightforward: customers keep pre-funding hardware, OCI revenue keeps compounding and Oracle turns a huge long-duration contract book into cash-generating infrastructure without constantly leaning harder on its own balance sheet.
The weaker version is just as clear: revenue grows rapidly, but the capital required to deliver it grows almost as fast.
The $664 billion tells us there is demand.
The next few years will tell us what that demand is actually worth to Oracle shareholders.
For another example of why a large headline commitment needs to be separated from its actual financial structure, see Why Is Nvidia Guaranteeing OpenAI’s Data Center? The $105B Structure Explained.
FAQ
Is Oracle’s $664 billion RPO the same as revenue?
No. RPO is contracted revenue that has not yet been recognized. At May 31, 2026, when Oracle had $638 billion of RPO, the company expected about 12% to be recognized during the following 12 months, with most of the balance scheduled further into the future.
Why was Oracle’s free cash flow negative despite rapid cloud growth?
Oracle generated $23.103 billion of operating cash flow in Q1 but spent $28.499 billion on capital expenditures, resulting in negative free cash flow of $5.396 billion. The company is investing heavily in infrastructure to support its rapidly expanding OCI business.
Are Oracle customers helping finance its AI infrastructure?
Yes. Oracle recorded $11.363 billion of increased deferred revenue from customer prepayments with a significant financing component in Q1. The company had also disclosed at the end of FY2026 that prepaid and customer-supplied hardware portions of large AI contracts totaled about $75 billion.
Zillow Paid Redfin $100 Million. Now Redfin Is Coming Back as a Rentals Competitor.
더보기Zillow paid Redfin $100 million for an exclusive multifamily rentals partnership. Now an FTC settlement is bringing Redfin back as an independent competitor just as Zillow’s Rentals business is growing rapidly. Zillow paid Redfin $100 million in
global.kkomccom.com