Oracle's gamble?
Let's talk about Oracle's report again – and this time, some surprising insights.
The key figure in Oracle's report was, of course, RPO, which jumped 15% in the fourth quarter.
Following the surprising surprise last quarter, RPO (future revenues from existing contracts) grew by an additional $68 billion in the second quarter to $523 billion (a 438% increase year-over-year).
For comparison, this is approximately 8 times Oracle's annual revenue.
This provides tremendous visibility for the business, but there's a catch: this figure assumes customers can pay and Oracle can deliver.
Oracle Cloud Infrastructure (OCI) revenue growth accelerated to 68% year-over-year (up from 55% in the first quarter and 49% in the fourth quarter).
This data confirms that Oracle is capturing significant demand for AI training, limited only by the speed at which it can deploy GPUs and power.
Earnings per share slightly exceeded expectations ($2.26 versus approximately $1.63 expected) – but note – this figure included a $2.7 billion gain from the sale of holdings in chipmaker Ampere (buried under 'Other' in the top right corner of the revenue flow chart below).
Without this one-time gain, earnings would have actually been a significant miss, which likely contributed to market skepticism.
What else is particularly important?
Cash flow:
Operating cash flow year-to-date increased 10% to $22.3 billion.
But free cash flow recently was negative $13.2 billion and remains under immense pressure due to the increase in capital expenditures.
And there's also the capital expenditure shock:
Management raised its capital expenditure guidance for fiscal year 2026 to $50 billion.
This is a $15 billion increase from previous guidance – Oracle is now investing in infrastructure at a scale that rivals leading hyperscale providers (Amazon, Microsoft, Google).
And all of this ultimately converges on the balance sheet:
Oracle's net debt is $105 billion (approximately 4x net leverage based on $26 billion EBITDA in the comparable period).
What does all this actually tell us?
The order backlog is indeed real, but timing is everything:
The $523 billion RPO is an astronomical figure.
And Project Stargate (OpenAI's massive deal) and others will not convert to cash until data centers are built.
On the other hand, continuous RPO growth indicates that demand is not slowing.
Larry Ellison essentially says, 'Demand is so assured, we will spend whatever it takes.'
Management hinted at potential relief valves, such as leasing chips or allowing customers to bring their own chips, to reduce the capital burden.
This leads us to the following conclusion that confuses investors:
On the one hand, if the AI cycle reverses, Oracle will be left with massive debt and empty server racks.
But on the other hand, if it continues, they could become the fourth hyperscaler.
The most bullish sign is that OCI is accelerating to 68% growth.
This is essentially the core thesis.
Oracle is growing its cloud infrastructure faster than AWS, Azure, or Google Cloud (from a smaller base), which shows its GPU Supercluster architecture is gaining traction.
Revenue declined slightly because an unfinished data center cannot be billed.
The bottleneck has shifted from sales to construction.
Management cited power availability and construction timelines as factors affecting growth.
The stock's decline reflects a sober look at the bottom line – combined with a debt-fueled spending spree, Wall Street demands short-term performance over long-term promises.
The questions investors are asking and for which there are no clear answers:
1. Can Oracle finance $50 billion in capital expenditures without jeopardizing its dividend or credit rating?
Management avoided precisely answering how much capital it needs to raise, so debt offerings or even equity dilution should be closely monitored.
2. How quickly will the $523 billion RPO convert to OCI revenue?
We need to see IaaS revenue accelerate towards a quarterly run rate of $5-6 billion to justify the expenditure.
It's also important to remember that while IaaS is growing, the traditional business (licenses/support/SaaS) remains flat or declining.
Oracle needs the old business to continue printing cash to fund the new business. If the old business erodes too quickly, the calculations become more difficult.
In summary, Oracle is taking a gamble, leveraging itself with massive debt, based on immense demand, assuming it can indeed build all the projects it plans and that customers can actually pay for them. If successful, Oracle is poised to become one of the largest companies in the world, joining the club with companies like Amazon, Meta, and Google.