To The Moon

The market is worried about $Oracle(ORCL)$  a risky, debt-fueled AI gamble that could strain cash flow and margins for years.

But here's what the headlines are missing.

Cast your mind back to Micron, Sandisk, Western Digital, Seagate. Everyone called them reckless. Too much debt. Too much capacity. "The cycle will destroy them."

Then scarcity hit. And they became cash machines.

We are watching the exact same playbook unfold in real time with GPU compute.

Right now in early 2026:

Spot GPU instances on AWS going for $14/hr per GPU

H100s renewing at the SAME price as 3 years ago — because buyers have zero leverage

Neoclouds have stopped selling single nodes. The supply is that tight.

Nvidia's own CFO just confirmed supply stays "very tight" into 2027

This is not a normal market. This is a scarcity supercycle.

And Oracle? Oracle is the landlord.

They've already locked in $523 BILLION in contracted future revenue. That's not a forecast. That's signed contracts. Cloud revenue growing 44% year over year. 70%+ of all revenue is recurring and sticky.

Yes the debt is real. Yes the capex is brutal — $48B this year alone.

But here's what the memory cycle taught us: once the infrastructure is built and depreciated, every dollar of revenue that comes in hits with devastating margin efficiency. The cash cow doesn't appear during the build. It appears AFTER.

Oracle also has something Micron and Seagate never had — 30 years of enterprise database lock-in sitting underneath the GPU story. Customers don't rip out Oracle databases. Ever. That's a moat the memory companies could only dream of.

Now look at this EPS chart. Really look at it. 👇

📈 May 2016: $2.11 📈 May 2021: $4.67 📈 May 2025: $5.69 📈 May 2026 (E): $7.45 📈 May 2027 (E): $10.73

That's a 408% total change since 2016. 14.5% CAGR.

But here's the number that should make you stop scrolling.

From 2026 to 2027 alone — EPS jumps 44% in a single year.

That's not gradual compounding. That's a step change. That's what happens when years of brutal infrastructure spending suddenly flips into operating leverage. The costs are already sunk. The revenue is just beginning to pour through.

And 2027? That's not the finish line. That's the launchpad.

Once the debt cycle peaks and free cash flow turns positive, the capital that was going into building data centres starts flowing back to shareholders. The GPU scarcity that's been filling Oracle's backlog for 3 years starts printing money at margins Micron investors would recognise instantly.

This thing could go parabolic beyond 2027 — and the market hasn't priced it yet.

Now here's where it gets really interesting. Run the numbers. 👇

Oracle is trading at $146 today.

Using the 2027 EPS estimate of $7.97:

Forward P/E Price Target

12x (bear) $95.64

15x (conservative) $119.55

20x (fair value) $159.40

25x (premium) $199.25

At a completely reasonable 20x forward multiple — which is below what the market typically awards high-growth cloud companies — Oracle is already undervalued at $146. At 25x, you're looking at a 36% gain from here on earnings alone, before you even price in the scarcity supercycle tailwind.

And if the cash cow thesis plays out the way Micron did? The market re-rates the multiple upward just as EPS is accelerating. That's the double engine that creates life-changing returns.

BUT — and this is the critical question nobody is asking:

⏳ HOW LONG DOES THE SCARCITY LAST?

This is everything. Oracle's debt danger window is 2026-2028. Their free cash flow is deeply negative RIGHT NOW. They need the scarcity to hold long enough for that $523B backlog to convert into cash.

Here's what the data says:

→ Supply chain experts: stabilization around 2027 at the earliest → New fabs in US, Europe, Asia don't come online until 2027-2028 → TrendForce calls AI's reallocation of memory capacity permanent → Even when supply eases, baseline pricing resets higher — not lower

That's the Micron moment hiding in plain sight. The shortage doesn't need to last forever. It just needs to outlast Oracle's debt crunch by enough margin to flip the cash flow.

18 to 30 months of sustained scarcity = Oracle becomes a generational cash cow.

Scarcity breaks early = Oracle is dangerously exposed.

Right now? Every data point says the window holds. And that EPS curve agrees.

The bears are looking at the debt. The bulls are looking at the backlog. The smart money is asking how long — and the answer keeps pointing to 2027.

Here's the setup in plain English:

🔴 Market sees: Debt. Layoffs. Negative cash flow. Pain. 🟢 Smart money sees: $523B backlog. 44% EPS jump. GPU scarcity with no end in sight. 30 years of unbreakable enterprise lock-in. A parabolic setup beyond 2027 that nobody has priced in.

Micron shareholders who held through the pain made life-changing returns. This is the same movie. Different sector. Bigger numbers. Higher stakes.

DYOR. Not financial advice. But history rhymes louder than most people listen. 👇

At $146 today with EPS projected to jump 44% in a single year and potentially go parabolic beyond 2027 — is Oracle the most undervalued AI infrastructure play nobody is talking about?

@TigerObserver  @TigerPM  @Tiger_comments  @TigerStars  @Daily_Discussion