Oracle After the Fall: What the Stock's Halving Tells You About Its New Business

Oracle's stock today sits near $156, roughly half of its September 2025 peak above $345. In one year the company went from a $900 billion market cap to $450 billion, and the decline accelerated into the worst week since August 2001, during the dot-com bust. The story that produced the selloff is straightforward: Oracle bet the company on AI infrastructure, spent more to build data centers than it earned, and piled on enough debt to push S&P to downgrade its credit rating to one notch above junk status.

The question most investors should ask before earnings after market close today isn't whether Oracle is growing. The growth is real and rapid. The question is whether this new business can ever finance itself, and whether the stock's 55% drop has already done the work of pricing that risk.

The numbers behind the growth

Oracle's fiscal year 2026 results, reported in June, told a tale of two companies. Total revenue was $67.4 billion, up 17%. Cloud revenue alone reached $34 billion, up 39%. The growth was concentrated in cloud infrastructure, which jumped 77% to $18.1 billion for the year and 93% in the final quarter to $5.8 billion. A remaining performance obligations backlog — essentially signed contracts not yet delivered — hit a record $638 billion, up 363% from a year earlier.

Oracle also guided for $90 billion in fiscal 2027 revenue, a 34% jump, and non-GAAP earnings per share of $8.05.

These numbers alone would be enough to keep any tech stock trading near its highs. But the growth came at a cost that changed the character of the business entirely.

When growth stops producing cash

Oracle spent $55.7 billion on capital expenditures in fiscal 2026, a 162% increase from the prior year. Operating cash flow, meanwhile, was $32 billion. The result: negative free cash flow of $23.7 billion.

That is the number that matters most to understanding what happened to this stock.

Oracle has operated for decades as a cash machine. Software and database licensing, with its high margins and recurring renewals, produced free cash flow that exceeded net income. The company used that cash for dividends, buybacks, and acquisitions. In fiscal 2025, free cash flow margin was still roughly 21%. In fiscal 2026, it was negative 35%.

The shift happened because building AI data centers is not a software business. It's a heavy capital business. Oracle is buying GPUs, leasing land, securing power, and opening data center facilities in Michigan, New Mexico, Texas, and Wisconsin. S&P projects the company's free cash flow deficit could widen to negative $42 billion in fiscal 2027.

Compare this to what Oracle's rivals have managed. Amazon, Microsoft, and Google all invested heavily in AI infrastructure during the same period. Their distinction was that revenue growth outpaced spending enough that free cash flow stayed positive. Oracle is the only major cloud provider that went negative.

The debt and the concentration

Oracle's total debt stands at roughly $219 billion against $43 billion in equity, for a debt-to-equity ratio above 3 to 1. The company raised $43 billion in debt and $5 billion in equity during fiscal 2026, and plans to raise another $40 billion in fiscal 2027, including a $20 billion share sale.

S&P downgraded Oracle's credit rating from BBB to BBB- on July 9, citing the scale of capital spending, the projected free cash flow deficit, and one specific risk that deserves attention: OpenAI.

Approximately half of Oracle's $638 billion backlog is tied to OpenAI, which committed to spending more than $300 billion on Oracle's infrastructure over five years. OpenAI posted a $20.9 billion operating loss in 2025 and has not yet turned profitable. If that single customer slows its spending, delays its IPO, or renegotiates terms, Oracle faces idle data centers that continue costing billions to maintain.

The credit market has already reacted. Oracle's five-year credit default swap spread doubled to 102 basis points, implying a 1.7% annual default probability. Moody's placed a negative outlook on Oracle's Baa2 rating months before S&P acted. Barclays downgraded Oracle's debt to underweight, warning of a potential funding crunch.

What the multiple says about the reset

Today's valuation tells you what the market has decided about these risks. Oracle trades at roughly 26.5 times trailing earnings and about 36 times forward earnings. That sounds expensive until you look at the trajectory. The stock was trading near 50 times earnings at its peak. The multiple compression tracks the business deterioration.

A PEG ratio of 0.80 suggests the trailing multiple may be modest relative to revenue growth. But PEG ratios built on trailing earnings and current revenue growth don't capture the forward reality: margins are compressing, free cash flow is negative, and the capital intensity of the next two years could be even higher. The $90 billion revenue target for fiscal 2027 requires Oracle to sustain roughly 34% growth while burning through $70 billion to $95 billion in capital expenditure.

This is the bridge between business quality and stock quality. The growth is there. The backlog is massive. But the economics of the business have shifted from high-margin software to capital-intensive infrastructure with lower margins. Gross margin fell from roughly 71% in early fiscal 2025 to 65.8% by year end. Oracle has been forced into lower-margin work renting GPUs to AI customers, which is offsetting the higher margins of its traditional software business.

What today's earnings might change

Oracle is set to report fiscal Q1 2027 results after market close today. Analysts expect roughly $19.1 billion in revenue, about 18% growth, and non-GAAP EPS near $1.74. Management guided for total revenue growth of 27% to 29% and cloud revenue growth of 58% to 64%.

Three things will determine whether today moves the needle:

Cloud infrastructure growth pace. OCI grew 93% in the prior quarter. If growth slows sharply, it suggests the AI demand Oracle has been counting on is less durable than the backlog implies. If it holds near or above 90%, it confirms the demand side of the equation.

Capital expenditure update. The market will want to know whether the spending trajectory is accelerating, flattening, or decelerating. Oracle has roughly $75 billion in customer prepayments and customer-supplied hardware that reduce its funding burden. If management signals that capex is becoming more self-financing through these arrangements, it would be meaningful.

OpenAI and concentration commentary. Any guidance or qualitative update on the OpenAI relationship, additional customer diversification, or deployment timelines would address the single biggest credit risk hanging over the stock.

Where this leaves the investment case

Oracle is not the same company it was eighteen months ago. It was a cash-generative software and database business trading at a premium for its moat. It is now an AI infrastructure builder with the growth of a tech company and the balance sheet of an industrial expansion. The market has priced that transition aggressively, cutting the stock in half.

The bear case is structurally sound. The free cash flow deficit is real. The debt load is unprecedented for a software company. The OpenAI concentration is a genuine counterparty risk. A credit downgrade below investment grade would force institutional funds to sell, creating a secondary pressure on the stock.

The counterpoint is that the valuation reset has been sharp enough to absorb a substantial amount of that risk. A $450 billion market cap for a company growing revenue at 17% to 34% with a $638 billion backlog is not a frothy multiple. The average analyst target near $242 implies roughly 55% upside, though those targets predate the full extent of the capex shock and may be slow to adjust.

The right frame for this stock is not whether Oracle is a good company. The growth numbers prove it is. The frame is whether you can tolerate a company that requires billions in external financing each year, trades one notch above junk status on its credit rating, and has half its future revenue tied to an unprofitable customer — all while hoping the AI infrastructure builds fast enough to turn the economics around.

That is a risk an investor can accept at $156 if they believe in the AI buildout and Oracle's execution. It is a risk that is hard to accept if you believe the spending curve will extend beyond the revenue ramp, or that OpenAI's commitments will not materialize on the timeline Oracle is counting on. Today's earnings and guidance will narrow that uncertainty. They will not resolve it entirely. The resolution requires seeing whether Oracle's free cash flow deficit narrows over the next two quarters, not whether one quarter's cloud growth number is high.