Oracle's Earnings Beat. The Stock Still Tells a Different Story.

Oracle just delivered a quarter that, by nearly every metric, beat what Wall Street expected. Revenue grew 30% to $19.3 billion. Cloud infrastructure revenue more than doubled, up 121%. Adjusted earnings per share cleared the consensus line of $1.74. Remaining performance obligations hit $664 billion. The stock jumped 7% in after-hours trading.

But then the stock fell 22% this year. It's down roughly 55% from its $900 billion market-cap peak last September. The company now sits at a $433 billion market value.

These two facts belong to the same story. Oracle's earnings beat. Investors are still worried.

The bull and bear camps both read the same quarterly report. They agree on the 30% growth, the $28.5 billion in capital spending, and the $20 billion stock offering that diluted shareholders in a single quarter. They are fighting over what those numbers mean for the economics of a company that used to sell software and now builds data centers.

The shared record

Reporting date: September 10, 2026. Period ended August 31, 2026. Stock price: roughly $150 (trading down 1.7% on September 11). Market cap: $433 billion. Total debt: approximately $125 billion. Cash: $36.4 billion.

Revenue: $19.3 billion, up 30% year-over-year. Cloud infrastructure: $7.4 billion, up 121%. Cloud applications: $4.2 billion, up 10%. Licensing and legacy software: $5.5 billion, down 3% — customers are migrating off premises, as Oracle has wanted for years.

Non-GAAP EPS: $1.92, up 30% — beat the consensus estimate of $1.74. Operating cash flow: $23 billion, up 184%. Capital expenditures: $28.5 billion, up from $8.5 billion a year ago. Free cash flow: negative $5.4 billion.

Capital raised this quarter: $20 billion in common stock through an at-the-market offering. Gross margins fell roughly 5 percentage points year-over-year as the business mix shifted toward lower-margin hardware. Oracle's infrastructure utilization reached 97.9%.

Full-year FY2027 guidance: at least $90 billion in revenue and $8.10 non-GAAP EPS. Full-year capital expenditure guide: $90 billion to $95 billion — meaning Oracle will spend roughly one dollar of capital for every dollar of revenue it earns this year.

Round 1: Is the growth real, and does it persist?

The bull's strongest case starts with the pipeline. Oracle's remaining performance obligations — contracted but not yet recognized revenue — grew to $664 billion, up $209 billion from a year ago. The company added more than $30 billion in new AI cloud contracts this quarter alone and delivered over 300,000 GPUs to customers. That is almost triple the GPU capacity delivered the previous quarter. Infrastructure utilization hit 97.9%, meaning almost every server rack is occupied. Demand isn't theoretical.

Here's the fact that embarrasses the bull: Oracle has never converted a pipeline this large before. A $664 billion backlog with a $90 billion capex target is untested territory. Even Oracle's own management acknowledges that construction timelines, energy permits, and grid connections can delay delivery. The bull has to trust that execution scales as smoothly as the contracts signed.

The bear's strongest case on this variable is simpler: the revenue is real, but the old comparison set is gone. Oracle grew 93% in cloud infrastructure last quarter and 121% this quarter. Those are compounding from a smaller base — $3.3 billion in infrastructure revenue a year ago. Even modest deceleration to 80% or 60% would still produce impressive dollar growth. The bear isn't arguing the growth stops; the bear is arguing it can't sustain triple-digit rates indefinitely, and the price assumed it could.

The bull answers: Q1 revenue grew sequentially for the first time, breaking Oracle's historical pattern of a lighter first quarter. Growth is accelerating, not decelerating. Q2 guidance of 30% to 34% revenue growth implies momentum continues.

The bull wins this round. The growth is verified, the pipeline is deep, and utilization is near capacity. The bear's deceleration argument is reasonable but speculative at this stage.

Round 2: Will the economics survive the transition?

This is where the fight gets real. Oracle was historically a software company with gross margins above 80% and capital expenditures well below 5% of revenue. That business still exists — licensing declined just 3% — but the center of gravity is shifting. Cloud infrastructure now represents roughly 38% of total revenue, up from about 22% a year ago. The business is becoming more hardware, less software.

Oracle's management told analysts that AI data center steady-state margins target 30% to 40%, and that gross margin pressure from the current ramp is expected to flatten over the next couple of years. GPU renewals are closing at a 20% premium to prior contracts, and management asserts pricing offsets cost increases.

The bear's best point: Oracle's non-GAAP operating margins expanded by only 35 basis points to 42.1%, despite 30% revenue growth. A company growing at 30% should show meaningful operating leverage. The fact that it doesn't tells you the new business — infrastructure — carries different economics than the old one — software. Even at $23 billion of operating cash flow, Oracle couldn't cover its own capex. The $28.5 billion spent this quarter was 147% of revenue.

The fact that embarrasses the bear: management said these projects are "self-funding" once ramped, with 100% free cash flow conversion to post-tax EBITDA. Oracle also demonstrated pricing power with 20% premium renewals and 97.9% utilization. If the infrastructure fills and renews at that premium, the margin question resolves itself. The bear has to assume the utilization and renewal numbers degrade.

This round is contested. The bull has pricing evidence and utilization data. The bear has the actual quarterly margin result showing barely any operating leverage. Both sides are right about different parts of the picture: the demand is there, but the economics are structurally different from what Oracle delivered for decades.

Round 3: Can Oracle fund the buildout without destroying shareholder value?

Capital expenditure hit $28.5 billion in a single quarter. Full-year guidance is $90 billion to $95 billion. Operating cash flow of $23 billion covered $23 billion of spending — but a significant portion came from customer prepayments, which Oracle disclosed were $11.3 billion, more than double the prior year's full-year total. That's not cash generation; that's borrowing from customers.

Oracle then sold $20 billion in common stock to bridge the gap. Diluted share count rose 3.1% year-over-year. For a company that previously used excess cash to repurchase shares, this is a regime change. Total debt now sits at $125 billion.

The bull's strongest defense: Oracle has $40 billion in RPO that doesn't require additional Oracle capex — customer prepayments and "bring your own hardware" models. The company co-CEO Clay Magouyrk told analysts capital doesn't all need to flow from Oracle's balance sheet. The equity offering was a strategic choice to avoid loading the balance sheet with even more debt, not a sign of desperation. The $20 billion ATM offering was structured so Oracle controls the timing and pace of share sales.

The bear's strongest counter: Oracle previously told investors that customer prepayments would cover the capital gap. The $20 billion stock issuance proves that assumption was wrong. The company now needs external equity funding at scale. That $20 billion at current prices adds a meaningful number of shares, and those shares dilute every future earnings dollar. The forward P/E of roughly 35x non-GAAP earnings is priced on a share count that may keep growing.

The bear wins this round. Whether you call it strategic or reactive, Oracle needed to sell $20 billion of equity in one quarter because its cash generation couldn't cover its own spending. The bull's "self-funding" timeline remains unspecified. The share count is now a live variable, not a historical footnote.

What the price demands

Oracle trades at $150, which translates to roughly 35x forward non-GAAP earnings on the $8.10 full-year guide. The trailing P/E sits at 23x. The price-to-sales ratio is about 6x on a $67.4 billion revenue run-rate, heading toward $90 billion.

A 35x forward multiple on a company spending $90 billion in capex means the market is pricing in two simultaneous assumptions: first, that the revenue growth continues at roughly 30% plus, and second, that the infrastructure business eventually produces durable cash returns after the capex cycle normalizes. The multiple isn't wrong for a high-growth story. It's unforgiving of execution mistakes.

Now make both stories pay rent. The bull story works if Oracle hits $90 billion in revenue, maintains operating margins near 40%, and free cash flow turns positive by FY2029 when the capex cycle moderates. That path supports the current valuation — and potentially more. The bear story works if growth slows below 25%, margins compress to the low 30s as hardware weight increases, and dilution from continued equity raises erodes per-share returns. On that path, 35x forward earnings is expensive for a business that isn't generating the cash to justify the multiple.

The current price — down 22% year-to-date and 55% from its peak — already incorporates a meaningful dose of the bear's concerns. The question isn't whether the bear's risks exist. They do. The question is whether they are large enough to warrant the discount.

The ruling

The business case goes to the bull. Oracle has verified demand, near-full utilization, pricing power on renewals, and a $664 billion contracted pipeline that dwarfs what the company was doing two years ago. The infrastructure growth is real, not projected.

The stock call goes to the bear at this price. Not because the growth story is broken, but because 35x forward earnings on a $90 billion capex year with negative free cash flow and a growing share count demands a near-perfect execution path. The margin transition from software economics to infrastructure economics hasn't been proven yet — it's been assumed. The equity financing hasn't been proven manageable — it's been attempted. And the "self-funding" ramp hasn't been dated — it's been described.

Oracle isn't a failing company. It's a rapidly growing company that has fundamentally changed what kind of business it runs, and the price hasn't fully adjusted to the different economics. The stock fell because investors realized that $90 billion in annual spending produces different cash returns than the $3 billion in spending Oracle did five years ago, even when the revenue is growing faster than ever.

The ruling flips if Oracle reports positive free cash flow for two consecutive quarters before the end of calendar year 2028, combined with non-GAAP operating margins stabilizing above 40%. That would prove the infrastructure model converts to cash returns without requiring perpetual equity financing. Until then, the bear carries the edge — not because the business is weak, but because the price requires certainty that the company hasn't yet earned.

The strongest reason the bull could still win: Oracle delivers the $90 billion revenue guide, capex moderates in FY2028 as the buildout completes, and the share growth from the ATM offering stops. That scenario would separate the stock from its current discount within 12 months. The evidence events that test it start next quarter, when Oracle reports Q2 FY2027 and reveals whether capex intensity begins to bend toward revenue growth.