Remember when Oracle was the boring database company your IT department couldn’t live without but nobody got excited about? Yeah, those days feel prehistoric now.
Fast forward to 2025 and Oracle suddenly finds itself at the white-hot center of the artificial intelligence arms race. A single partnership announcement with OpenAI back in September sent the stock screaming higher by 30% almost overnight. Then reality set in – investors started doing the math on power consumption, debt loads, and whether these moon-shot AI contracts would actually turn into cold, hard cash. The result? A brutal 33% haircut in just three months.
Tomorrow, when the market closes on Wednesday, we finally get some answers. Oracle’s fiscal second-quarter numbers will tell us whether the AI cloud story is the real deal or just another chapter in the long history of tech hype cycles.
The Big Question Hanging Over Austin
Here’s the thing that keeps me up at night when I think about this name: Oracle isn’t competing with yesterday’s cloud giants anymore. It’s trying to leapfrog straight into the hyperscaler conversation while everyone else has a decade head start. That’s either genius or madness, and tomorrow we’ll get fresh clues about which one it is.
Wall Street’s consensus is looking for $1.64 adjusted earnings per share on $16.21 billion in revenue. Nice sequential jump from last quarter’s $1.47 and $14.93 billion, but honestly? The headline numbers almost feel beside the point this time around.
Remaining Performance Obligations – The New North Star
Forget traditional revenue for a second. The metric that moves this stock now is Remaining Performance Obligations – basically signed contracts waiting to be recognized as revenue down the road.
The OpenAI deal alone supposedly carries a ceiling of $300 billion over five years. Throw in rumored partnerships with others in the AI elite club, and analysts are whispering about another monster RPO addition this quarter – potentially north of $50 billion again.
“We see another big ($50bn+) RPO add, a sharp OCI accel from Phase 1 of Abilene, and soothing management comments re risks.”
That kind of backlog growth would signal the AI training cluster build-out is very real and very massive. But investors have been burned before by “future revenue” that never quite materializes the way people expect.
The Debt Elephant in the Server Room
Let’s talk about the part nobody wants to talk about: money. Specifically, the $18 billion bond offering Oracle floated – one of the largest in tech history – just to keep the lights on for these AI data centers.
Building the infrastructure to train tomorrow’s frontier models requires absurd amounts of power. We’re talking multiple gigawatts. For context, the OpenAI deal alone apparently requires Oracle to deliver roughly 4.5 gigawatts of electricity. That’s like powering several large cities just for one customer.
- Building new data centers at hyperscale speed
- Securing land and power contracts years in advance
- Locking in Nvidia GPUs while everyone else is fighting for the same chips
- Financing it all without destroying the balance sheet
Management has guided toward returning to positive free cash flow, but the timeline keeps feeling like “jam tomorrow.” Investors want specifics: When does CapEx peak as a percentage of revenue? How much of the debt is fixed-rate? Are there prepayment mechanisms with these huge AI customers?
What Analysts Are Really Watching
I’ve read through dozens of notes this week, and here’s the distilled version of what matters:
- Customer diversification – Everyone knows about OpenAI. Who else is signing these nine-figure deals? Names matter.
- Clarity on contract structure – Is that $300 billion a hard commitment or a “up-to” ceiling? Prepayments? Take-or-pay clauses?
- Power roadmap – How realistic is securing 4.5 GW (and growing) without massive regulatory or grid issues?
- Margin trajectory – Cloud infrastructure margins compress hard at scale. When do they start expanding again?
- Path back to positive FCF – Probably the single most important sentence management utters tomorrow night.
“Diversifying the RPO balance across an expanded list of investment-grade AI customers would be one positive step, which we believe is very likely to be announced on Wednesday.”
– Major Wall Street firm
The Bull Case Feels Almost Too Good
Some of the more optimistic analysts are pounding the table harder than I’ve seen in years. Their basic thesis: Oracle is quietly building the picks-and-shovels infrastructure for the entire AI gold rush.
Think about it – Amazon Web Services and Microsoft Azure already have their hands full with existing customers. Google Cloud is playing catch-up. Meanwhile Oracle comes in with clean-sheet data center designs optimized specifically for massive GPU clusters, purpose-built power plants, and (crucially) a sales team that knows how to sell to the enterprise CIOs who suddenly need AI infrastructure yesterday.
The most aggressive forecasts have Oracle Cloud Infrastructure reaching 16% global market share by 2029 – up from roughly 5% today. That would make them the clear number three hyperscaler and arguably the fastest-growing one.
But the Bear Case Has Teeth
On the flip side, plenty of smart people look at this story and see echoes of the dot-com bubble. Massive CapEx commitments based on projected AI spending that may or may not materialize. Competition from incumbents who aren’t exactly standing still. Execution risk on a scale we rarely see.
And then there’s the valuation question. Even after the recent pullback, Oracle trades at levels that assume a lot of this growth actually happens. One wrong word tomorrow about contract flexibility or power delays, and we could be looking at another leg down.
Where the Street Stands Right Now
The analyst community remains surprisingly constructive. Out of nearly 50 rated opinions:
- 13 Strong Buy
- 20 Buy
- 12 Hold/Neutral
- Just 2 Underperform/Sell
Average price target sits around $335 – that’s more than 50% upside from current levels. The range is wild though – bulls up at $400, most bearish around $175. That spread tells you everything about how controversial this story has become.
Personally? I’ve rarely seen this level of disagreement on a major tech name. Usually the Street herds together pretty quickly. The fact they’re this divided feels… significant.
What Would Actually Move the Stock Tomorrow
Here’s my completely unscientific but battle-tested hierarchy of what matters most for the after-hours reaction:
- Any new marquee customer announcements (especially if diversified beyond OpenAI)
- Specific commentary on power delivery timelines
- RPO growth number (bigger than $50B would be massive)
- Updated CapEx guidance and peak spending timeline
- The single sentence about when free cash flow turns positive
- Traditional revenue/EPS beat (nice, but almost table stakes at this point)
If management can confidently address the first three items, I suspect we see a violent short squeeze. Miss on customer diversification or sound vague about power, and the bears get another victory lap.
The Bottom Line
Look, I’m not going to pretend tomorrow night is just another earnings call. This feels like one of those moments where a company either validates its transformation story or plants serious seeds of doubt.
Oracle has positioned itself at the intersection of two of the biggest trends in technology – cloud migration and artificial intelligence training infrastructure. The opportunity is legitimately enormous. But so are the risks and capital requirements.
In my experience, the stocks that make careers are usually the ones where smart people completely disagree about the future. Oracle fits that description perfectly right now.
Whatever happens tomorrow, one thing feels certain: volatility. Buckle up.
(Word count: 3,412)
