To The Moon

$Microsoft(MSFT)$  

$Alphabet(GOOG)$  

$Amazon.com(AMZN)$  

$Intel(INTC)$  

Over roughly 48 hours, the market just ran one of the best experiments we have had on the AI boom.

Billions upon billions are being spent.

Cloud demand is exploding.

Data centres are being built.

Memory is being consumed.

Yet the stocks reacted completely differently.

Why?

Because Wall Street has changed the question.

In 2024 and 2025, investors asked:

“Who is spending on AI?”

In 2026, the question is becoming:

“WHO IS ACTUALLY GETTING PAID?”

And after studying these five reports together, I think the answer is becoming surprisingly clear.

🥇 MICROSOFT: THE CLEANEST AI CONVERSION

Microsoft delivered:

$92.3B revenue, +19%

$45.9B operating income, +21%

$4.89 adjusted EPS

And Azure revenue reached roughly:

$33.2B, +48% YoY

This is why Microsoft remains the benchmark.

The company is spending enormous amounts on infrastructure.

But the spending is translating into enormous cloud revenue and operating profit.

That is the equation investors want:

CAPEX ↑

↓

CAPACITY ↑

↓

AI USAGE ↑

↓

REVENUE ↑

↓

PROFIT ↑

Microsoft is showing that the machine works.

But there is an important catch.

Microsoft's problem is increasingly:

EXPECTATIONS.

When everyone already knows you are the best student in class, scoring 95 instead of 93 does not necessarily make the stock explode.

That is why a great company can produce a great quarter and still generate an ordinary stock reaction.

🔴 ALPHABET: CLOUD WON. CASH FLOW LOST.

Alphabet may actually be the more fascinating report.

Google Cloud revenue exploded to roughly:

$30.7B

That is more than 100% YoY growth.

Amazing.

But Alphabet's overall revenue of approximately $127.6B missed expectations slightly.

Advertising disappointed.

And the number I dislike most:

FREE CASH FLOW: NEGATIVE $9.1B

Capex reached approximately:

$51B in ONE quarter.

And management expects 2027 capex to rise substantially again.

This is the exact tension I have been writing about.

Google can absolutely win in AI.

But shareholders need more than proof that Gemini is popular and Cloud demand is huge.

They need proof that:

AI REVENUE GROWTH > AI CAPITAL INTENSITY

Cloud doubling is fantastic.

But if cash generation deteriorates while spending keeps accelerating, the market will eventually ask what return it is receiving on every new dollar invested.

That is why I would not call Alphabet's report an AI failure.

I would call it:

An AI monetisation success with a cash-flow warning attached.

🟡 AMAZON: THE SAME PROBLEM, DIFFERENT NUMBERS

Then Amazon reported.

And suddenly the pattern became even clearer.

AWS growth accelerated to roughly:

39%

That is excellent.

The cloud demand story is alive.

But investors focused on something else:

CAPEX.

Amazon is now looking at roughly $220B of infrastructure spending.

That is an extraordinary number.

Again:

DEMAND = 🟢

CLOUD GROWTH = 🟢

AI OPPORTUNITY = 🟢

CAPITAL INTENSITY = 🔴

The market is no longer willing to reward the first three while ignoring the fourth.

This may be the single biggest change in the AI trade.

🚨 THE HYPERSCALER PARADOX

Now put Microsoft, Alphabet and Amazon together.

All three are experiencing extraordinary AI demand.

All three are investing enormous sums.

Yet investors are becoming nervous.

Why?

Because hyperscalers have to:

PAY FIRST.

They buy GPUs.

They buy memory.

They buy networking.

They buy storage.

They secure power.

They build data centres.

They depreciate those assets.

And only then do they monetise the capacity.

Which led me to a different question.

WHO GETS PAID FIRST?

And suddenly one earnings report looks very different.

🥇 SANDISK MAY HAVE GIVEN US THE MOST IMPORTANT REPORT

SanDisk reported:

$11.0B revenue

up 23% sequentially.

Gross margin:

85.0%

Adjusted free cash flow:

$5.7B

But look at this:

DATACENTER REVENUE: $4.8B

That was up:

61% QUARTER ON QUARTER.

Read that again.

While hyperscalers debate whether hundreds of billions in AI capex will eventually generate sufficient returns, SanDisk is already selling them the infrastructure required to build it.

And management guided the next quarter to:

$12.2B-$12.6B revenue

with roughly:

84.5%-85.5% non-GAAP gross margin.

This is why I keep coming back to the AI picks-and-shovels trade.

SanDisk does not need to prove that an AI chatbot can eventually monetise.

It needs AI infrastructure spending to continue.

Microsoft, Google, Amazon, Oracle and others are effectively telling us:

IT IS CONTINUING.

🧠 THIS CHANGES HOW I VIEW MU TOO

SanDisk's report matters enormously for Micron.

Remember my memory thesis:

HBM

↓

DRAM

↓

ENTERPRISE SSD

↓

NAND

If only HBM were booming, I would worry that the AI memory cycle was narrow.

But SanDisk's datacenter revenue growing 61% sequentially tells me the demand is spreading deeper into storage.

That is exactly what I wanted to see.

It suggests we may increasingly be moving from:

HBM SUPERCYCLE

toward:

BROADER AI MEMORY SUPERCYCLE.

And that keeps MU extremely interesting to me.

Micron sits across HBM and DRAM, while SanDisk gives us evidence from the NAND and enterprise-storage side.

Different products.

Same underlying customer:

The AI data centre.

🔴 THEN THERE IS INTEL

Intel's numbers initially looked impressive.

Revenue:

$16.5B, +21%

DCAI revenue:

$7.3B, +78%

GAAP gross margin:

43%

Those are not bad numbers.

Yet the stock was crushed.

Why?

Because Intel entered earnings with something the other four did not:

A massive turnaround premium already embedded in the stock.

This is exactly why I was cautious about holding a large INTC position around the previous elevated levels.

A company can improve dramatically.

But if the stock has already priced in an even faster improvement:

Good fundamentals can still produce bad price action.

Intel therefore teaches us another AI lesson.

Microsoft shows:

execution matters.

Alphabet and Amazon show:

capital efficiency matters.

SanDisk shows:

scarcity economics matter.

Intel shows:

VALUATION MATTERS.

🔥 THE AI MONEY FLOW IS BECOMING CLEARER

Think about where one dollar of hyperscaler AI capex goes.

MICROSOFT / GOOGLE / AMAZON

↓

DATA CENTRE

↓

GPU

↓

MEMORY

↓

NETWORKING

↓

STORAGE

↓

POWER

The hyperscaler spends the dollar.

The infrastructure supplier receives it.

That does not mean infrastructure suppliers are automatically better investments.

Eventually supply catches up.

Pricing normalises.

Cycles turn.

But during the scarcity phase?

I want to know:

Who has pricing power before capacity catches demand?

Right now, memory and storage continue giving us very interesting answers.

🚦 MY AI EARNINGS SCORECARD

If I rank what these earnings taught me:

🥇 SNDK: SCARCITY WINNER

The surprise standout.

Datacenter +61% sequentially.

Margins around 85%.

FCF already substantial.

🥈 MSFT: MONETISATION WINNER

The cleanest hyperscaler equation.

Azure +48%.

Huge operating profit.

AI investment is already translating into revenue.

🥉 AMZN: DEMAND WINNER, CAPEX QUESTION

AWS acceleration is impressive.

But $220B of infrastructure spending raises the hurdle dramatically.

4️⃣ GOOGL: CLOUD WINNER, CASH-FLOW LOSER

Cloud growth is extraordinary.

But negative FCF and another capex step-up demand scrutiny.

5️⃣ INTC: FUNDAMENTALS IMPROVING, EXPECTATIONS RESETTING

DCAI growth is real.

But the stock had already priced an enormous turnaround.

This is why buying price matters.

🔵 SO WHOSE AI SPEND PAYS OFF FIRST?

I think the question itself needs changing.

The hyperscalers may eventually generate enormous returns from AI.

Microsoft is already proving much of that thesis.

But during the infrastructure buildout:

THE SUPPLIERS GET PAID BEFORE THE BUILDERS PROVE THEIR ROI.

That is the insight I am trading.

And it is why my hierarchy currently looks like this:

1. MEMORY / STORAGE

MU + SNDK

2. GPU / COMPUTE

NVDA and the wider accelerator ecosystem

3. NETWORKING / OPTICS

The next bottleneck I am watching closely

4. QUALITY HYPERSCALERS

MSFT first, then selective opportunities when valuation resets

The hyperscalers are writing the cheques.

The suppliers are cashing them.

🔮 WHAT WOULD CHANGE MY VIEW?

Three things.

🔴 1. AI CAPEX CUTS

If Microsoft, Google, Amazon, Oracle and Meta begin cutting infrastructure plans simultaneously, my supplier thesis weakens.

🔴 2. MEMORY SUPPLY CATCHES DEMAND

If HBM, DRAM and NAND pricing starts falling materially while inventories rise, I become much more cautious.

🔴 3. HYPERSCALER CASH FLOW INFLECTS HARDER

If Cloud and AI revenue suddenly accelerate faster than capex, the balance shifts.

Then hyperscalers themselves may offer the superior risk/reward.

Until then, I continue to like companies getting paid during the buildout.

MY CALL

These earnings did not tell me AI is a bubble.

Quite the opposite.

They showed extraordinary real demand.

Microsoft's Azure growth.

Google Cloud's acceleration.

AWS growth.

Intel's DCAI demand.

SanDisk's datacenter explosion.

The AI infrastructure cycle is very real.

But the market is moving into the next phase.

It is no longer enough to say:

“WE ARE SPENDING BILLIONS ON AI.”

Investors now want:

“SHOW ME WHAT THOSE BILLIONS EARN.”

And that creates a fascinating asymmetry.

The companies spending the money must prove ROI.

The companies supplying the bottlenecks simply need the spending to continue.

For now, that is why I remain fascinated by:

MU + SNDK + NVDA + THE AI INFRASTRUCTURE CHAIN.

The next stage of the AI trade may not be about finding who spends the most.

It may be about finding who gets paid first.

I am not a financial advisor. Trade wisely, Comrades!

After this earnings week, would you rather own the AI spenders, MSFT/GOOGL/AMZN, or the suppliers, NVDA/MU/SNDK?

If you find this post interesting, give it wings! 🪽 Repost and share the insights.

Do consider ‘Follow me’ and get firsthand read of my daily new post. Thank you. @TigerPM  @TigerStars  @Daily_Discussion  @TigerEvents  @Tiger_SG