Oracle just posted some of the best headline numbers in its history, and the stock still couldn’t hold its footing. Shares snapped a brutal nine-session slide only after investors had already wiped out roughly a quarter of the company’s market value in a matter of weeks. The trigger wasn’t disappointing revenue. It was what sat beneath the revenue line.
OCI grew 93% and the backlog hit $638B, but cash flow fell to negative $24 billion.
Oracle Cloud Infrastructure (OCI) revenue jumped 93% in the fiscal fourth quarter, and the company’s remaining performance obligations (RPO), essentially its contracted order book, surged 363% year-over-year to $638 billion. At the same time, free cash flow for fiscal 2026 collapsed to negative $23.7 billion. Two record-setting stories, one company, one earnings report. That contradiction is exactly why Wall Street couldn’t agree on how to price the stock.
What Triggered the Selloff Despite a Record Quarter
On paper, Oracle’s fiscal Q4 2026 results looked like a win. Revenue reached $19.2 billion, up 21% year-over-year, and non-GAAP earnings per share of $2.11 beat consensus estimates of $1.97. Full-year revenue climbed 17% to $67.4 billion, and cloud revenue grew 39% to $34 billion.
Yet the stock fell roughly 11% the day after the report, and it kept sliding for several sessions afterward, eventually shedding more than half its value from its 52-week high before stabilizing. A few factors combined to spook investors:
- Guidance on future capital spending. CFO Hilary Maxson guided for approximately $70 billion in net capital expenditures for fiscal 2027, well above the $67.66 billion analysts had modeled.
- A credit downgrade. S&P lowered Oracle’s rating to BBB-, just one notch above junk status, citing rising leverage and the capital intensity of the AI buildout.
- Margin warnings. Maxson also flagged that gross margins would “step down” as new data centers ramp up before they generate meaningful revenue.
- Financing plans. Oracle confirmed it would raise close to $40 billion in fresh debt and equity in fiscal 2027, on top of $43 billion in debt and $5 billion in equity already raised in fiscal 2026.
In short, investors weren’t punishing Oracle’s sales. They were repricing the risk of how that growth gets funded.
The Bill Behind the Backlog
The number that unsettled the market most wasn’t in the headline it was in the cash flow statement. Oracle’s capital expenditures reached $55.7 billion in fiscal 2026, a 162% jump from $21.2 billion the year before. That spending is almost entirely tied to building and leasing data center capacity for AI workloads.
| Metric | FY2025 | FY2026 | Change |
|---|---|---|---|
| Capital expenditures | $21.2B | $55.7B | +162% |
| Free cash flow | -$0.4B | -$23.7B | Sharp decline |
| Operating cash flow | $20.8B | $32.0B | +54% |
| RPO (backlog) | $138B | $638B | +363% |
| Long-term debt | Under $100B | Over $100B | Rising |
| Restructuring/severance costs | $374M | $1.8B | +381% |
To close the funding gap, Oracle has leaned hard on debt markets rather than existing cash reserves. That approach works as long as lenders and bondholders remain confident the backlog converts into real revenue on schedule. It’s a bet, not a guarantee, and it’s the reason rating agencies are watching Oracle’s leverage as closely as its growth.
See More: ASML Earnings Wednesday: EUV Bookings Will Show Whether the AI Chip Boom Is Sustainable
Record Earnings, Negative Cash Flow: Why Both Can Be True Simultaneously
This is the part of Oracle’s story that trips up casual readers of earnings headlines. Net income, operating income, and revenue can all set records in the same period that free cash flow turns sharply negative. Here’s why:
- Accounting profit isn’t cash. GAAP net income reached $17 billion for the year, up 36%, but that figure includes non-cash items and revenue that’s been recognized on paper before cash physically arrives.
- Depreciation is rising fast. Depreciation on new data center hardware nearly doubled to $7.62 billion, a cost that hits the income statement gradually even though the cash for the equipment goes out upfront.
- Capex sits outside operating cash flow. Operating cash flow actually grew 54% to a record $32 billion. Free cash flow subtracts capital expenditures from that figure, and Oracle’s $55.7 billion in capex simply outpaced the cash the business generated.
- Prepayments help, but don’t erase the gap. Oracle noted that $75 billion of its backlog involves customers prepaying for GPUs or supplying their own hardware, which reduces the capital Oracle itself needs to raise. It softens the burn; it doesn’t eliminate it.
Put simply: Oracle is profitable on paper and cash-negative in practice because it is spending faster than it collects, betting that today’s construction becomes tomorrow’s recurring cloud revenue.
The OpenAI Dependency That Defines Oracle’s Bull and Bear Cases
A large share of Oracle’s backlog traces back to a single customer relationship. Bank of America analysts estimate that more than half of Oracle’s $638 billion RPO is linked to OpenAI, reportedly anchored by a multi-year contract worth roughly $300 billion that doesn’t begin generating revenue until 2027.
That concentration cuts both ways:
- The bull case: OpenAI’s compute demand is enormous and growing, and locking in a contract of that scale gives Oracle years of visible, high-value cloud revenue that few competitors could match.
- The bear case: OpenAI is not yet profitable, and it depends on continued outside funding to meet its own obligations. If OpenAI’s growth slows or its financing tightens, Oracle’s largest revenue driver becomes its biggest vulnerability.
Analysts note that only a fraction of the total backlog roughly 12%, according to Oracle’s own filings is expected to convert to revenue within the next fiscal year. The rest depends on multi-year execution, data center construction timelines, power availability, and customer renewals holding up as planned.
How Oracle Cloud Infrastructure Actually Works
For readers less familiar with the technical side, OCI is Oracle’s cloud computing platform, competing directly with Amazon Web Services, Microsoft Azure, and Google Cloud. It provides:
- Compute capacity, including the GPU clusters needed to train and run large AI models
- Storage and networking infrastructure for enterprise workloads
- Database services, an area where Oracle has decades of enterprise relationships to draw on
- AI infrastructure specifically built for large-scale model training, which is the segment driving Oracle’s 93% OCI growth
Oracle’s pitch to AI companies is capacity and price. Because Oracle entered the hyperscale cloud race later than its rivals, it has been aggressive about signing large multi-year contracts with AI labs that need guaranteed GPU access, even if that means carrying heavier debt to build the data centers those contracts require.
21,000 Jobs Cut, $1.8 Billion in Severance
While Oracle poured money into AI infrastructure, it simultaneously shrank its workforce. According to Oracle’s annual regulatory filing, the company’s global headcount fell to 141,000 employees as of May 31, 2026, down from 162,000 a year earlier a reduction of about 21,000 people, or nearly 13% of the workforce.
Key details from the filing:
- Severance and restructuring costs hit $1.8 billion, up sharply from $374 million the prior year
- Oracle directly stated that “the adoption and deployment of AI technologies across our operations have resulted, and may continue to result, in reductions to our workforce”
- The company cited additional factors beyond AI, including management changes, performance-based cuts, strategic shifts, and acquisitions
- Oracle warned that further workforce reductions could follow as AI adoption continues internally
The cuts illustrate a broader pattern across the tech sector in 2026, where companies like Meta, Microsoft, and Amazon have paired heavy AI capital spending with workforce reductions elsewhere in the business.
What Analysts See From Here
Wall Street remains split, but not evenly negative. Several analysts, including those at Bank of America, maintain buy ratings on Oracle, pointing to the scale and quality of the backlog as justification for the spending. Average price targets from bullish analysts imply significant upside from current trading levels, with the expectation that payoff arrives after fiscal 2027 once new data centers reach full utilization.
The more cautious view centers on:
- Execution risk across dozens of simultaneous data center builds
- Power availability and construction delays that could push revenue recognition further out
- Oracle’s rising debt load and the BBB- credit rating, which raises future borrowing costs
- Customer concentration risk tied to OpenAI’s own financial trajectory
Neither side disputes that Oracle’s cloud business is growing quickly. The disagreement is about timing, funding cost, and whether the backlog converts to cash before the debt becomes a real constraint.
Conclusion
Oracle’s fiscal 2026 results capture the defining tension of the current AI infrastructure boom: massive contracted demand colliding with the enormous upfront cost of building the capacity to serve it. Record revenue, record backlog, and record operating cash flow are all real. So is a $23.7 billion free cash flow deficit, over $100 billion in debt, and a workforce that shrank by 21,000 people in the same year. Investors aren’t questioning whether AI demand for Oracle’s cloud is genuine. They’re questioning whether the company can fund the buildout fast enough, and profitably enough, before the bill comes due.
Frequently Asked Questions
Why is Oracle stock falling in 2026 despite reporting record revenue?
Investors are focused on Oracle’s negative $23.7 billion free cash flow, rising debt, and heavy AI capital spending rather than its record revenue and backlog figures.
What is the difference between Oracle’s “record earnings” and its negative free cash flow?
Earnings reflect accounting profit, including non-cash items, while free cash flow subtracts real capital spending from actual cash generated, and Oracle’s $55.7 billion in capex outpaced its cash inflows.
Did Oracle actually cut 21,000 jobs because of AI?
Oracle’s SEC filing confirms AI adoption contributed to workforce reductions, alongside management changes, performance issues, and strategic shifts, cutting headcount from 162,000 to 141,000.
What should investors watch to determine if Oracle’s AI bet pays off?
Watch backlog-to-revenue conversion rates, data center construction timelines, Oracle’s debt costs, and the financial health of major customers like OpenAI.
I’m a professional blog writer with over 10 years of experience in crafting engaging and original content, particularly in the blessings and quotes niche. My focus is on delivering high-quality, reader-friendly, and well-researched content that both inspires and adds meaningful value to readers.
Through blessingpress.info, I consistently create and share unique content aimed at building trust, strengthening authority, and connecting with a loyal and growing audience.