To The Moon

$Micron Technology(MU)$  

Micron has fallen from almost $1,000 to around $896 pre-market.

Again.

And I can already feel the frustration among MU investors.

SK Hynix posts extraordinary numbers.

Memory remains scarce.

AI data-centre demand remains enormous.

SanDisk reports explosive datacentre growth.

Micron itself is producing extraordinary margins.

Yet MU falls.

So naturally investors are asking:

WHAT IS WRONG WITH MICRON?

My answer?

Less than the stock price suggests.

But something HAS changed.

Wall Street has moved one level higher in the AI supply chain.

The question is no longer:

“Is memory demand strong?”

We already know it is.

The new question is:

“CAN THE COMPANIES BUYING ALL THIS MEMORY KEEP SPENDING?”

And that explains almost everything happening to MU right now.

🔴 FIRST, THIS IS NOT JUST A MICRON SELLOFF

This distinction is critical.

Micron is down.

SanDisk is down.

SK Hynix is down.

The semiconductor complex is down.

That tells me the market is not suddenly discovering a Micron-specific disaster.

It is repricing:

AI INFRASTRUCTURE RISK.

Alphabet shocked investors with massive AI spending and negative free cash flow.

Amazon's enormous infrastructure bill frightened investors.

Intel warned about cash burn and weak guidance.

Samsung produced record earnings, yet still disappointed a market expecting perfection.

Put these together and Wall Street has suddenly asked:

“Wait.”

“If hyperscalers have to spend hundreds of billions to build AI…”

“…what happens if shareholders force them to slow down?”

And immediately the market sells the suppliers.

NVDA.

MU.

SNDK.

Networking.

Semiconductors.

That is the transmission mechanism.

🧠 THE IRONY IS THAT MEMORY DEMAND HAS NOT COLLAPSED

This is the strange part.

Look at Micron's current business outlook.

Revenue guidance:

$61.5B ± $1.5B

Non-GAAP gross margin:

~86.25%

Non-GAAP EPS:

$38.15 ± $1

Those are not recession numbers.

Those are not memory-downcycle numbers.

Those are extraordinary numbers.

HBM4 is progressing.

Server memory demand remains strong.

AI infrastructure continues consuming enormous quantities of DRAM.

And the wider memory industry continues describing supply as tight.

So we have a fascinating contradiction:

MEMORY FUNDAMENTALS:

🟢

MEMORY STOCKS:

🔴

Whenever that happens, I ask:

DID THE BUSINESS CHANGE, OR DID THE PRICE PEOPLE ARE WILLING TO PAY FOR THE BUSINESS CHANGE?

Right now, I think it is primarily the second.

🔥 THE MARKET HAS MOVED FROM SCARCITY TO ROI

This is the most important shift.

For much of the memory rally, the equation was:

AI DEMAND ↑

↓

HBM SHORTAGE

↓

DRAM SHORTAGE

↓

PRICING POWER ↑

↓

MU EARNINGS ↑

↓

MU STOCK ↑

Simple.

But now Wall Street has inserted another variable.

HYPERSCALER CASH FLOW.

The equation is becoming:

AI DEMAND

↓

HYPERSCALER CAPEX

↓

MEMORY ORDERS

↓

MEMORY PRICING

↓

MU PROFITS

The market is attacking the second link.

Not the fourth.

Not yet.

That distinction matters enormously.

🚨 THIS IS WHY ALPHABET AND AMAZON MATTER TO MICRON

You might own MU and think:

“Why should I care whether Google free cash flow is negative?”

Because Google is one of the companies writing the cheque.

Same with Microsoft.

Amazon.

Meta.

Oracle.

The hyperscalers are upstream of Micron's earnings.

Remember the AI money flow I wrote about:

MICROSOFT / GOOGLE / AMAZON / META

↓

DATA CENTRES

↓

GPU

↓

HBM

↓

DRAM

↓

STORAGE

The hyperscalers spend.

Micron gets paid.

So when investors become frightened about the spender's return on capital, they immediately question how long the supplier's supercycle can last.

That is what MU is pricing.

🔥 BUT HERE IS WHERE I DISAGREE WITH THE SELL-OFF

There is a major difference between:

AI CAPEX BECOMING LESS PROFITABLE

and

AI CAPEX ACTUALLY BEING CUT.

We have evidence of the first.

I do NOT yet see convincing evidence of the second.

That is crucial.

Alphabet may be spending enormous amounts.

Amazon may be spending enormous amounts.

Investors may hate the spending.

But from Micron's perspective:

THE CHEQUES ARE STILL BEING WRITTEN.

And while those cheques continue:

HBM gets purchased.

Server DRAM gets purchased.

Enterprise storage gets purchased.

GPUs get purchased.

Networking gets purchased.

The suppliers do not need hyperscaler shareholders to LIKE the capex.

They need management teams to KEEP APPROVING IT.

Those are two very different things.

🧠 SANDISK JUST GAVE US AN IMPORTANT CROSS-CHECK

This is why I found SanDisk's earnings so important.

Its datacentre revenue surged roughly:

61% QUARTER ON QUARTER.

That matters for Micron.

Why?

Because my memory thesis has always been:

HBM

↓

DRAM

↓

ENTERPRISE SSD

↓

NAND

If HBM alone were booming while everything else weakened, I would worry.

That could indicate an extremely narrow AI memory cycle.

But when datacentre storage demand is also exploding?

That suggests AI demand is spreading deeper through the infrastructure stack.

So I am not seeing:

HBM BOOM, EVERYTHING ELSE BROKEN.

I am increasingly seeing:

A BROADER AI MEMORY CYCLE.

That is fundamentally bullish for MU.

🔴 BUT I WILL NOT CALL EVERY DROP A BARGAIN

This is where discipline matters.

MU has had an extraordinary run.

The market knows memory is scarce.

The market knows HBM demand is enormous.

The market knows Micron's margins are extraordinary.

These are no longer hidden insights.

And that means the hurdle has risen.

Samsung just demonstrated this perfectly.

It produced extraordinary profits.

The stock still fell.

Why?

EXPECTATIONS.

The memory trade has moved from:

“Show me good numbers.”

to:

“SHOW ME NUMBERS BETTER THAN THE NUMBERS EVERYONE ALREADY EXPECTS.”

That is a much harder game.

🚨 THE REAL BEAR CASE IS NOT TODAY'S PRICE DROP

There are three things that would genuinely worry me.

🔴 1. HYPERSCALERS ACTUALLY CUT CAPEX

Not investors complaining about capex.

Not stocks falling because capex is high.

I mean Microsoft, Google, Meta, Amazon and Oracle materially reducing infrastructure plans.

That would change my thesis.

🔴 2. MEMORY SUPPLY CATCHES DEMAND

Watch contract pricing.

Inventories.

Lead times.

Customer commitments.

If DRAM/HBM availability suddenly loosens while pricing starts falling materially, I become cautious quickly.

🔴 3. CHINA CLOSES THE TECHNOLOGY GAP

CXMT is not something I dismiss.

Its progress in DRAM matters.

But I distinguish:

conventional DRAM competition

from

leading-edge HBM economics.

If Chinese suppliers begin meaningfully threatening high-value server DRAM or advanced HBM economics, the long-term multiple deserves compression.

Today, that is a risk.

It is not yet my base case.

🔥 THE NUMBER THAT STILL MAKES ME BULLISH

Forget today's MU price for a moment.

Look at:

~86% GROSS MARGIN.

That is extraordinary for memory.

Memory historically behaved like a brutal commodity industry.

Supply rose.

Prices collapsed.

Margins disappeared.

Then consolidation happened.

Technology complexity increased.

HBM became difficult to manufacture.

Advanced packaging became scarce.

Customer qualification cycles became longer.

AI customers began securing supply earlier.

The industry structure changed.

That does not mean memory is no longer cyclical.

It absolutely is.

But this cycle is increasingly characterised by:

scarcity

technological complexity

long-term customer commitments

enormous AI demand.

That is why I refuse to call the cycle finished simply because MU dropped 5%.

📍 MY MU PRICE MAP HAS CHANGED

MU around $896 pre-market is now back in a zone I find interesting.

Previously I liked roughly $880-$900.

We are there.

But given how violently sentiment has shifted, I would layer rather than fire everything at once.

🟡 $890-$910: HOLD / SMALL ADD

Current battlefield.

I would not panic here.

But I would also preserve ammunition because momentum remains ugly.

🟢 $850-$875: FIRST PICK

This becomes increasingly attractive to me if memory fundamentals remain intact.

🟢 $800-$835: PREFERRED PICK

Here the risk/reward becomes much more compelling.

The market would be pricing a significant amount of AI-capex fear into MU without necessarily having evidence of a memory downturn.

🟢 $750-$790: FEAR ZONE

I would be interested aggressively only if:

HBM demand remains strong

DRAM pricing remains firm

hyperscaler capex remains intact

Micron guidance remains strong.

🔴 Below $750: THESIS CHECK

At that point I would investigate whether this is more than valuation compression.

I would specifically look for falling memory pricing or cancelled capacity commitments.

Now upside.

🟢 $930-$950: FIRST RECLAIM

This would tell me sellers are beginning to exhaust themselves.

🟢 $980-$1,000: MAJOR BATTLEFIELD

MU needs to prove it can hold four digits.

Repeated rejection here tells us expectations remain a ceiling.

🟢 Above $1,000: CONFIRMATION

If MU reclaims and holds $1,000 while memory fundamentals remain strong:

$1,050-$1,100

becomes my next zone.

Beyond that, earnings revisions need to do the heavy lifting.

🔮 MY NEXT 4-8 WEEK SCENARIOS

🟢 30%: AI CAPEX FEAR FADES

Hyperscalers reiterate spending.

Memory pricing stays strong.

SNDK and SK Hynix continue confirming scarcity.

MU reclaims:

$950

then challenges:

$1,000-$1,050.

The market realises it confused:

bad cash flow for AI buyers

with

bad demand for AI suppliers.

🟡 50%: GREAT FUNDAMENTALS, TIRED STOCK

This is my base case.

MU trades roughly:

$840-$980

Memory fundamentals remain strong.

But investors refuse to expand the multiple until they have more clarity on hyperscaler returns.

This would frustrate bulls.

But it would also allow earnings to catch up with the share price.

I would use weakness selectively.

🔴 20%: CAPEX FEAR BECOMES CAPEX CUTS

This is the scenario that matters.

Hyperscalers begin reducing infrastructure plans.

Memory pricing expectations weaken.

CXMT fears intensify.

Then:

$800

could break.

And MU could test:

$750-$790.

I would not blindly average into that scenario.

I would first determine whether the memory thesis itself had changed.

🧠 HERE IS THE SIGNAL I AM WATCHING MOST

Not Micron's stock price.

Not Reddit sentiment.

Not one scary headline.

Not even one hyperscaler's free cash flow.

I want to see:

MEMORY PRICING + CUSTOMER COMMITMENTS.

Those are closer to the economic truth.

If MU falls while:

HBM stays scarce

DRAM pricing stays strong

customer commitments remain intact

hyperscaler capex continues

then I interpret the decline as:

MULTIPLE COMPRESSION.

Painful?

Absolutely.

But potentially investable.

If MU falls while those fundamentals weaken too?

Then it becomes:

EARNINGS COMPRESSION.

That is far more dangerous.

Never confuse the two.

🔥 THE MOST IMPORTANT DISTINCTION IN THIS SELLOFF

The market currently fears:

THE BUYERS OF AI INFRASTRUCTURE ARE SPENDING TOO MUCH.

That does not automatically mean:

THE SELLERS OF AI INFRASTRUCTURE ARE SELLING TOO LITTLE.

In fact, those can coexist for quite some time.

Google's problem may be that AI infrastructure costs too much.

Micron's opportunity is that:

AI INFRASTRUCTURE COSTS SO MUCH.

Think about that.

The same dollar that hurts Google's free cash flow can become Micron's revenue.

That is why the hyperscaler capex debate has two sides.

For Google shareholders:

“When do we get the return?”

For Micron shareholders:

“Are they still spending?”

Right now, they are.

MY CALL

I don't like the price action.

I do like the underlying memory economics.

And those are not contradictory statements.

MU falling toward $896 tells me sentiment and positioning have deteriorated.

It does NOT yet tell me:

HBM demand collapsed.

DRAM pricing collapsed.

AI data-centre construction stopped.

Micron's margins collapsed.

Hyperscalers cancelled their buildouts.

None of those things has happened.

So I am not abandoning the memory thesis because the stock is red.

But I am upgrading the question I ask.

Previously:

“Is AI creating a memory shortage?”

Clearly yes.

Now:

“HOW LONG CAN THE CUSTOMERS FUND THAT SHORTAGE?”

That is the next phase of the trade.

Until I see actual capex cuts, weakening memory pricing or deteriorating customer commitments, I continue to view MU weakness differently from a broken fundamental thesis.

My sequence:

🟡 $890-$910: HOLD / SMALL ADD

🟢 $850-$875: FIRST PICK

🟢 $800-$835: PREFERRED PICK

🔴 BELOW $750: THESIS CHECK

And on recovery:

🟢 $930-$950: FIRST RECLAIM

🟢 $980-$1,000: BATTLEFIELD

🟢 ABOVE $1,000: CONFIRMATION

Because the biggest mistake investors can make after a violent selloff is asking:

“How much did the stock fall?”

Instead ask:

“WHAT FELL FIRST, THE PRICE OR THE EARNINGS POWER?”

For Micron today?

I see the price falling much faster.

I am watching very carefully to see whether the fundamentals follow.

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

MU investors: does this selloff make you question the memory supercycle, or are you using $800-$900 weakness to accumulate?

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