🚀 SanDisk Just Broke $1,700 — Is This Still a Memory Cycle, or a New AI Storage Era?

SNDK is becoming VERY hard to ignore.

SanDisk has now ripped roughly 35% in just five trading days, pushing from around $1,200 to ~$1,725 today. And this isn’t happening in isolation:

🟢 SNDK: ~$1,725

🟢 MU: ~$1,002

🟢 SK Hynix ADR: ~$172

🟢 WDC: ~$526

The entire memory complex is waking up again — but SNDK is clearly leading the charge. (Barron’s)

And after digging into SanDisk’s Investor Day, I think the market is beginning to price in something much bigger than another NAND cycle.

🔒 The BIG story: 2027–2028 capacity is already being spoken for

SanDisk has signed eight New Business Model agreements, covering approximately 50% of FY27 bit shipments and around two-thirds of FY28.

That’s extremely important.

Traditional memory investing is basically:

Demand ↑ → prices ↑ → everyone adds capacity → oversupply → prices crash 📉

But SanDisk is trying to change that equation:

Long-term contracts → committed demand → disciplined supply → better visibility → stronger margins 💰

And that’s potentially a massive structural improvement.

The company isn’t simply betting that NAND prices stay high.

It’s increasingly selling future capacity before that capacity is even produced.

That’s why I don’t think the “capacity locked through 2028” thesis is hype.

A huge portion of future output already has visibility. 🔐

🤖 AI is making NAND fundamentally more valuable

Here’s the part I think the market may still be underestimating.

AI isn’t only about GPUs.

Every AI workload creates enormous amounts of data — training data, model checkpoints, inference data, enterprise datasets and increasingly real-time workloads.

That means the infrastructure stack needs:

GPU → HBM → DRAM → NAND/storage

And SanDisk sits directly in that storage layer.

The company is also developing HBF (High Bandwidth Flash) for AI inference applications, with samples expected next year.

If HBF successfully moves from sampling → qualification → volume production, SanDisk gets another potential growth engine on top of traditional NAND.

🔥 That’s why the SNDK story is becoming more interesting than simply “NAND prices are rising.”

💰 And then there’s the margin shock

This is perhaps the biggest reason I’m bullish.

SanDisk’s recent gross margin has already reached extraordinary levels, while management is targeting approximately 80% stabilized gross margins over the longer term.

Think about what that means.

If investors previously valued SanDisk as:

“A cyclical NAND manufacturer”

the valuation should be relatively low because earnings can collapse during downturns.

But if SanDisk can actually achieve:

mid/high-teens revenue growth + ~80% gross margins + contracted future capacity

then the market can start valuing it more like a high-quality AI infrastructure supplier.

That is a completely different story.

And today’s rally suggests investors are beginning to believe it.

⚠️ So why am I NOT bearish after +35%?

Because the stock is still more than 25% below its June record high around $2,335.

In other words, today’s $1,700+ price looks insane if you only look at the last five days.

But zoom out:

SNDK hasn’t even reclaimed its previous peak.

That’s why I wouldn’t automatically call this a bubble just because the weekly candle looks ridiculous.

The market is repricing the earnings power.

And there is another important catalyst: U.S. officials are increasingly discouraging American companies from sourcing memory chips from China, which could further benefit U.S.-based suppliers such as SanDisk. (Barron’s)

🏆 SNDK vs MU vs SK Hynix vs WDC

If I had to pick ONE today for the next leg of the memory/AI-storage trade:

🥇 SNDK — my pick

🥈 MU — strongest diversified memory/AI exposure

🥉 SK Hynix — fantastic HBM exposure, but much more dependent on the broader Korean semiconductor cycle

4️⃣ WDC — attractive storage exposure, but I prefer SNDK’s NAND + AI/HBF + contracted-capacity story

Why SNDK?

Because SNDK has the clearest combination of AI demand + NAND pricing + contracted capacity + margin expansion.

MU may have the better overall memory franchise.

SK Hynix may have the better HBM positioning.

But SNDK has the most dramatic potential transformation in its business model.

🎯 The BIG question: BUY at $1,725?

Here’s where I would be disciplined.

I would NOT chase a vertical +35% move.

Not because I’m bearish.

Quite the opposite.

If I’m bullish on SNDK, I want to own it without paying the maximum amount of FOMO premium.

My preferred zones:

🟢 $1,550–1,620: attractive pullback zone

🟢 $1,450–1,550: very attractive

🔥 <$1,450: I’d become seriously interested

Meanwhile, a sustained breakout above $1,750–1,800 with strong volume would tell me momentum is still extremely powerful — but I’d rather wait for a retest than blindly chase.

And remember: SNDK has already shown it can move hundreds of dollars in days.

That’s both the opportunity AND the risk.

🚀 My conclusion

I think the market is making a mistake if it looks at SNDK and says:

“Another memory stock that’s getting overheated.”

The more interesting question is:

What if SanDisk is actually becoming structurally different?

8 NBM agreements.

~50% of FY27 capacity covered.

~⅔ of FY28 covered.

AI inference opportunity through HBF.

~80% long-term gross-margin target.

And a massive secular increase in storage demand from AI.

That’s not the old SanDisk story.

That’s potentially:

AI → exploding data → storage demand → contracted capacity → pricing power → higher margins → higher earnings visibility. 🔥

So yes, I’m bullish on SNDK.

But after today’s move, I wouldn’t chase.

I’d wait for the pullback.

Because if the thesis is truly right, I don’t need to buy the stock at its most euphoric price.

And if SNDK really is entering a new structural AI-storage cycle…

$1,725 may eventually look cheap. 🚀

SNDK > MU > SK Hynix > WDC for my current AI-storage/memory trade.

Not financial advice. Trade at your own risk tigers! [Look back]  [Look back]  [Look back]