Oracle’s post-earnings reset looks overdone because the market is obsessing over the cost of the AI buildout while skipping the clearest proof that demand is already real: remaining performance obligations reached $638 billion, up 363% year over year. That is not a hope trade. It is contracted business, and management said most of the jump came from large-scale AI deals tied to training and inferencing workloads. At $131.66, the stock is being treated like an AI spender with shaky demand when the numbers say Oracle is becoming an AI capacity seller with revenue already spoken for.
The backlog is the whole story here. Oracle added $85 billion of RPO sequentially and tied most of that increase to large AI contracts, which matters more than any one quarter’s free cash flow swing because it shows customers are committing in size before the revenue is recognized. Even better, $75 billion of those large AI contracts are tied to customer prepayments and customer-supplied hardware, which undercuts the simplest bear argument that Oracle must shoulder the entire funding burden alone.
The income statement is already moving in the right direction. Revenue grew 17.4% year over year, EPS grew 33.2%, and net income grew 37.3%, while Oracle still posted a 30.8% operating margin and a 25.4% net margin. That combination is why the TickerSpark Score gives Oracle a 95 for Profitability and a 95 for Growth. This is not a low-quality turnaround being priced for perfection; it is a highly profitable software and infrastructure business accelerating into AI demand.
The forward setup also looks stronger than the tape suggests. Oracle guided Q1 FY2027 revenue growth of 27% to 29% and cloud revenue growth of 57% to 64%, numbers that do not fit the market’s current mood around the stock. Valuation helps too: ORCL trades at 22.2 times trailing earnings with a PEG of 0.67, while Adobe sits at 9.68 times earnings but only 10.5% revenue growth, and Synopsys trades at 31.86 times earnings with 15.1% growth. Oracle is not the cheapest stock in software, but for a company growing revenue 17.4% and sitting on a $638 billion contracted backlog, it does not look expensive enough to justify this kind of panic.
The real risk is not fake demand. It is that the economics get uglier before they get cleaner. FY2026 capex ran about $55.66 billion, above the prior $50 billion target, FY2027 capex could reach $90 billion to $95 billion, and the free cash flow deficit widened to $23.7 billion from just $394 million a year earlier. Oracle also raised $43 billion of debt and $5 billion of equity in FY2026 and expects about $40 billion more in FY2027, which is exactly why the TickerSpark Score only gives it a 48 for Financial Health.
That is a legitimate objection, and the weak chart backs up the stress. ORCL is down 32.6% year to date versus technology up 27.1%, the RSI is 27.25, and the shares are sitting near the 52-week low of $131.35. Insider activity does not help the optics either, with 9 recent sells totaling $61.30 million and no buys. Even so, those numbers argue for volatility, not for a broken thesis. If demand were cracking, Oracle would not be posting a 363% jump in RPO and guiding cloud growth as high as 64%.
That leaves Oracle looking like a classic contrarian AI infrastructure trade: messy funding optics, ugly momentum, and much stronger underlying demand than the stock action implies. We would rather own ORCL here than ADBE because Oracle has the more powerful growth reacceleration and the more obvious AI monetization path already embedded in signed contracts.
What matters now is conversion. The next earnings cycle in September needs to show that backlog is turning into revenue fast enough to justify the capex ramp, and any update that narrows the FY2027 spending range or reinforces the customer-funded nature of new AI contracts would strengthen the case. The trigger that would change our mind is not another ugly headline about spending; it would be evidence that RPO growth stalls while leverage keeps climbing. Until that happens, the selloff looks like the market pricing the pain of the buildout while underpricing the value of the demand already locked in.
