Oracle just did something Wall Street didn’t expect: it started looking like a real AI contender.
After years of being dismissed as the old database company trying to keep up with AWS and Azure, Oracle’s latest quarter forced investors to pay attention. Q3 FY2026 revenue jumped 22% year over year to $17.2B. Cloud revenue surged 44%. And the eye-popper: infrastructure (IaaS) up 84%. That’s not maintenance-mode growth. That’s a company sprinting.
The headline number, though, is the $553B in remaining performance obligations — up 325% year over year. That’s backlog. Real contracts. Much of it tied to AI workloads and long-term cloud commitments. Translation: customers aren’t just experimenting with Oracle’s AI capacity — they’re signing up in bulk.
And yet, the stock at around $144 doesn’t scream “AI darling.” Why? Because Oracle is spending like one.
The company plans to raise up to $50B to build out data centers and power infrastructure. It has already locked in $30B through bonds and preferreds. Free cash flow has been pressured. Margins are under scrutiny. This is capital-intensive, heavy-lift expansion. Not asset-light SaaS comfort food.
But here’s the thing: this is the cost of competing in AI infrastructure. Nvidia sells the chips. Microsoft and Amazon build capacity. Oracle is betting it can carve out a serious slice by offering cheaper, high-performance cloud infrastructure tied tightly to its database dominance. And it’s working — at least for now.
The skeptics say Oracle is late and overextending. The bulls argue that being “late” doesn’t matter when demand for AI compute is exploding and supply is constrained. Enterprises don’t want just two hyperscaler options. They want negotiating leverage. Oracle is becoming that third call.
What makes this more than hype is execution. EPS is climbing (GAAP up 24% to $1.27). Cloud apps are still growing double digits. Operating cash flow over the trailing 12 months hit $23.5B. This isn’t a startup chasing AI buzzwords. It’s a cash machine redirecting firepower.
The real risk? Debt and overbuild. If AI demand cools or pricing compresses, Oracle will be sitting on very expensive concrete and silicon. But if demand holds — and every major enterprise budget suggests it will — Oracle could look cheap in hindsight.
Wall Street spent a decade underestimating Satya Nadella’s Microsoft. It may be doing something similar with Larry Ellison’s Oracle.
Oracle isn’t trying to win the cloud popularity contest. It’s trying to win the contracts. And right now, it is.
The question isn’t whether Oracle is an AI company. It’s whether investors are ready to treat it like one.
#OracleAI #TechDisruption #AIInfrastructure #CloudComputing #InvestInTech #DataCenters #FutureOfAI #EnterpriseSolutions #MarketTrends #InnovationLeadership