Oracle Stock: $664B Backlog, Negative $5B Free Cash Flow
Oracle has $664 billion of future contracted revenue and still posted negative $5 billion in free cash flow. That is not an automatic buy signal. It is a capital-intensive bet.
Oracle has $664 billion of future contracted revenue and still posted negative $5 billion in free cash flow.
That is not an automatic buy signal. It is a capital-intensive bet.
The number everyone will repeat — and the one they should fear
On September 10, Oracle reported a monster first quarter for fiscal 2027. Revenue rose 30% to $19.3 billion. Cloud revenue rose 62% to $11.6 billion. Its cloud-infrastructure business — the bit powering the AI gold rush — grew 121% to $7.4 billion. Oracle also booked more than $30 billion in new AI cloud contracts, taking its remaining performance obligations, or RPO, to $664 billion. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
That $664 billion figure is the headline. It should be. It is enormous.
But RPO is not cash in the bank. It is future revenue contracted with customers that Oracle still has to deliver, support, finance and turn into real profit. Anyone who has built a business knows the difference between a signed contract and money safely sitting in your account. The gap between the two is where projects blow up, customers renegotiate and margins disappear.
Oracle’s own numbers tell both sides of the story. Operating cash flow hit a record roughly $23 billion in the quarter, up 184%. Yet free cash flow was negative roughly $5 billion because the company is spending aggressively to build data centres and install the capacity those contracts require. It brought 850 megawatts of additional data-centre capacity online and delivered more than 300,000 GPUs to customers during the quarter. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
That is not a software company collecting lovely subscription cheques from a high-margin asset-light product. This is a company pouring concrete, buying very expensive hardware and racing to build infrastructure before competitors do.
It may work brilliantly. It may also be a bloody expensive way to learn that demand is not the same as durable returns.
Oracle is no longer the boring database company
For years, Oracle was the dependable old bloke of enterprise tech: databases, finance systems, large corporate customers and a business model that printed cash without much drama.
Now Larry Ellison’s company is trying to become a serious AI-infrastructure supplier beside Amazon Web Services, Microsoft Azure and Google Cloud. Its latest quarter shows it has real momentum, not just an AI sticker slapped on a legacy business. Cloud infrastructure growth of 121% is hard evidence. The $664 billion RPO balance is hard evidence too. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
But the legacy business is telling you something as well. Oracle’s software revenue fell 3% to $5.5 billion in the quarter as customers continued moving from on-premise products to cloud services. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
That matters because the old Oracle was built around reliably profitable software. The new Oracle is asking investors to accept heavier capital spending, negative free cash flow and more execution risk in exchange for faster growth.
That is a fair trade only if management converts that contracted demand into cash at attractive economics. Not revenue. Not backlog. Cash.
This is where plenty of investors get seduced by big numbers. A backlog is exciting because it makes the future look pre-sold. But a backlog can also become a capital obligation. You have promised capacity. Now you need land, power, chips, networking, staff and financing — before your customer has paid enough to make the effort worthwhile.
The $90 billion target is impressive. It is not proof.
Oracle lifted its fiscal 2027 outlook to at least $90 billion in revenue and $8.10 in non-GAAP earnings per share. It expects second-quarter revenue growth of 30% to 34%, with cloud revenue growth of 65% to 71%. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
Those are massive forecasts. Put the $664 billion RPO figure beside the $90 billion annual revenue target and the backlog is about 7.4 times that target. That is why the market is paying attention. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
But here is the bit I would write on the whiteboard if I owned Oracle shares: What does it cost to earn each dollar of that backlog?
Reuters reported capital spending of $28.5 billion in the quarter, versus $8.5 billion in the same quarter a year earlier. Oracle’s free cash flow remained negative even as operating cash flow surged. ([investing.com](https://www.investing.com/news/stock-market-news/oracle-beats-fourthquarter-revenue-estimates-4736107))
There is no shame in spending heavily when the opportunity is real. Amazon did it. Plenty of great businesses did. But spending is not a moat. Spending is merely the entry fee.
Oracle previously said it could spend as much as $95 billion in fiscal 2027, while expecting some customer reimbursements. Reuters reported in June that the company expected up to $70 billion of its own capital spending, plus another $20 billion to $25 billion it expected to be repaid for. ([investing.com](https://www.investing.com/news/stock-market-news/oracle-beats-fourthquarter-revenue-estimates-4736107))
That is not a trivial detail buried in the footnotes. It is the investment case.
The overlooked risk is not AI demand. It is financing discipline.
Most of the commentary around AI has been embarrassingly simplistic: AI is big, chips are needed, data centres are needed, therefore buy everything with a server rack.
No. That is how people confuse an industrial boom with an easy investing boom.
The important question is who earns the best return after funding the boom. The company selling the scarce inputs can do very well. The company underwriting billions of dollars of capacity for demanding customers might do very well too — or it might wear the cost while someone else captures the upside.
Oracle completed a $20 billion sale of common stock during the quarter through its at-the-market equity programme. Management said its new contract structure does not create an incremental need to raise capital beyond existing plans. ([prnewswire.com](https://www.prnewswire.com/news-releases/oracle-announces-q1-results-driven-by-triple-digit-growth-in-cloud-infrastructure-revenues-302875728.html))
Read that sentence carefully. It is not a disaster. It is not a reason to panic. But it is a reminder that this growth story needs funding.
A business can be genuinely excellent and still be a rotten investment if you pay too much, if margins get squeezed, or if fresh capital keeps being required before the promised cash arrives. Founders learn this the hard way. Investors often learn it later, after a few pretty presentations and a nasty share-price chart.
Oracle’s adjusted earnings beat expectations and its shares rose 6.9% after hours following the report. Yet the same reporting noted investor scrutiny around negative free cash flow, backlog conversion and the funding required for the buildout. ([investing.com](https://www.investing.com/news/earnings/oracle-raises-full-year-profit-guidance-stock-rises-after-hours-4896781))
Good. That scrutiny is healthy. Markets need fewer cheerleaders and more adults asking where the cash is going.
The contrarian angle: Oracle may be more useful as a test than a trade
I am not saying Oracle is doomed. That would be lazy.
I am saying Oracle is now one of the cleanest tests of whether the AI infrastructure boom creates broad shareholder wealth or simply causes companies to spend absurd amounts of money chasing strategic relevance.
If Oracle turns $664 billion of contracted work into strong free cash flow, sensible margins and modest funding needs, it will have pulled off a serious reinvention. It will also validate a good deal of the wider AI-capex thesis.
If revenue races ahead while free cash flow stays negative and fresh capital becomes routine, the lesson will be equally useful: demand can be real while returns are ordinary.
That is why I would watch Oracle’s next few quarters for four things, in this order:
1. Free cash flow: Is it moving toward positive territory as revenue expands? 2. Capital spending: Is growth becoming less capital-hungry, or is the bill climbing faster than the business? 3. Cloud margins: More cloud revenue is good only if the economics improve after depreciation, power and hardware costs. 4. Backlog conversion: Does the RPO balance become recognised revenue and cash on a timetable that makes sense?
Notice what is not on that list: breathless commentary about AI agents, keynote demos, or a one-day share-price move.
What this means for you
If you are a long-term saver with broad index funds, do not let one Oracle earnings report turn you into a data-centre trader. Your advantage is time, diversification and the discipline to keep buying productive assets through noise. Don’t throw that away because a $664 billion number gave you FOMO.
If you own individual shares, size Oracle as an execution-heavy infrastructure bet, not as the old stable software company. Treat the backlog as evidence of demand — not as guaranteed profit. Review the cash-flow statement before you celebrate the revenue line.
And if you run a business, steal the real lesson. Never brag about pipeline, bookings or contracted revenue without being brutally clear on three things: the cash needed to deliver it, the gross margin after delivery, and the time it takes to get paid.
I have seen businesses die while sitting on terrific sales numbers. Big contracts can make you feel rich long before they make you rich.
Oracle’s $664 billion backlog might become one of the great corporate transformations of this decade. But the next dollar of free cash flow will tell us more than the next hundred billion dollars of hype.