Six weeks ago, Oracle (ORCL -4.21%) management guided for about $8.05 in non-GAAP (adjusted) earnings per share this fiscal year. As of this writing, the stock trades around $117 — less than 15 times that figure, after setting a new 52-week low of $114.75 on Friday. A multiple like that is usually reserved for mature software companies whose growth is ending, not for a business that just guided for revenue growth of about 34%.
However, the analysts covering the software and cloud computing giant haven’t followed the stock’s price all the way down. The average price target on Oracle sits at about $248, more than double the current share price of about $117.
To be clear, an average price target isn’t an investment case, and I wouldn’t buy any stock because of one. But a gap this wide is worth understanding. Each side of it is pricing a different answer to the same question: Will Oracle’s enormous backlog of AI (artificial intelligence) contracts convert into cash before the cost of building for it damages the company?
Image source: The Motley Fool.
What the market has stopped paying for
The selling has been relentless. Oracle shares have fallen about 66% from their high of $345.72, and the pressure traces back to spending.
Oracle’s capital expenditures reached $55.7 billion in fiscal 2026 (the year ended May 31, 2026), most of it going into data centers for its cloud infrastructure business. Operating cash flow rose 54% to a record $32 billion, and the build-out consumed all of it. Free cash flow for the year came in at a negative $23.7 billion.
The bill has started arriving in other forms, too. S&P Global Ratings cut Oracle’s credit rating to BBB- earlier this month, one notch above junk status, citing the cost of the build-out. And Oracle has said it expects to raise $40 billion through debt and equity financing this fiscal year, including a $20 billion share sale that will dilute existing shareholders.
When a company is burning more than $20 billion of cash a year, the market stops valuing its earnings and starts scrutinizing its balance sheet. A forward multiple below 15 says the market is worried about more than the durability of growth — but also the risks presented by a deteriorating balance sheet.
What the analysts are still counting
The other side of the argument is the business itself, which keeps performing. Fiscal 2026 revenue rose 17% year over year to $67.4 billion, and growth roughly doubled over the course of the year, with fiscal fourth-quarter revenue up 21%. Even more striking, Oracle’s cloud infrastructure revenue (the business that rents computing capacity to AI customers) grew 77% for the full year and 93% year over year in fiscal Q4, reaching $5.8 billion for the quarter.
Profits kept up, too. Fiscal 2026 earnings per share came in at $5.83 under generally accepted accounting principles (GAAP), up 34%, though one-time gains on the Ampere chip-business sale and Bloom Energy warrants did much of that lifting — excluding them, adjusted earnings per share rose 13%.
Then there’s the backlog. Remaining performance obligations (Oracle’s signed contract value that hasn’t yet become revenue) finished fiscal 2026 at $638 billion after growing $85 billion in the final quarter. The prepaid and customer-supplied hardware portions of Oracle’s large AI contracts now total about $75 billion — customers paying for their graphics processing units (GPUs) up front or supplying the chips themselves, which moves part of the spending burden onto the customers.
If most of that backlog converts on schedule, the math behind a $248 target isn’t hard to follow. Management’s forecast calls for about $90 billion of revenue this fiscal year, or growth of about 34%. Growth like that, at less than 15 times guided earnings, is exactly what the covering analysts are pointing at.
Today’s Change
(-4.21%) $-5.05
Current Price
$114.99
Key Data Points
Market Cap
Day’s Range
$114.75 – $123.08
52wk Range
$114.75 – $345.72
Volume
46.4M
Avg Vol
30.1M
Gross Margin
63.34%
Dividend Yield
1.74%
So whose number is closer to the truth? Probably neither, fully. The market is treating guidance backed by signed contracts as if it were speculative. And the analysts are extrapolating a backlog whose single largest customer is itself an unprofitable AI company. It’s also worth remembering that a price target costs its publisher nothing.
My own answer is that I don’t need to pick a side yet. The stock is arguably cheap against guidance, but the cash burn is enormous, and the next few quarters will show whether free cash flow is finding a floor while the backlog converts into revenue. That evidence, not the distance to a price target, is what could get me to buy.
Until it shows up, I’m staying on the sidelines. And I’d suggest investors who do buy this dip keep the position small. After all, shares have been beaten down for a reason.
