The market has decided something this week: Chinese memory competition is now an existential threat to every Western storage name, full stop.

Sell first, sort out the details later.

Then Monday happened.

SanDisk dropped 6% on a Chinese chipmaker’s architecture announcement. SanDisk makes NAND flash.

The announcement was about DRAM.

Not the same memory. Not the same market. Not even close.

That mix-up is exactly what we’re cross-examining today. Is the market pricing in a real threat, or is sentiment so fragile after SanDisk’s run that any headline with “Chinese” and “memory” in it is enough to trigger a sale, whether it applies or not.

Missed Monday’s edition on Palantir’s upgrade? 🚨 Wall Street Upgraded PLTR. The Options Desk Didn’t Cheer.

TikTok CEO Shou Zi Chew says AI could drive major breakthroughs in fields like biotech, physics, chemistry, and materials science within the next two to three years. His point is that AI’s biggest impact may ultimately go far beyond chatbots and productivity tools. The bigger takeaway is that the next phase of the AI story could shift from software automation toward scientific discovery, where the economic payoff could be much larger.

Microsoft now has three generations of in-house AI chips and plans to deploy potentially more than one million Maia 300 processors in 2027 to reduce costs and dependence on Nvidia. But the Fool’s argument is that Nvidia still has the stronger ecosystem, scale, and developer lock-in, while its revenue has kept exploding even as hyperscalers build their own silicon. In plain English, custom chips are a real threat, but they are still more likely to complement Nvidia than replace it anytime soon.

Bank of America upgraded Salesforce all the way from Underperform to Buy, arguing that its AI strategy is finally starting to translate into better monetization. Agentforce annual recurring revenue reached about $1.5 billion in Q2, up from $1.2 billion in Q1, while the company is also leaning harder into consumption-based pricing and tighter integration with Anthropic. The takeaway is that Wall Street may be shifting from seeing AI as a threat to Salesforce toward seeing it as the next leg of growth.

This year AI shook up creative strategy. Most teams are still working like it didn't.

Ari Murray, Chief Digital Officer at Salt and Stone sits down with Bryan Harvey, Head of Brand at Hightouch, for an honest conversation about what's breaking on creative teams right now, how the roles are changing, and what a day looks like when creatives spend it on ideas, concepts, and craft while agents handle the manual execution.

Leave with new insights from 3 experts in the field and 5 prompts to instantly improve your creative output.

Reserve your seat

SanDisk closed Monday at $1,487.33, down 6% on the session, after ChangXin Memory Technologies (CXMT) said it’s close to mass-producing a new DRAM architecture.

The market’s confident case: Chinese memory makers are catching up fast, and every Western incumbent is exposed.

Here’s the problem with that case. CXMT’s announcement was about DDR5 server DRAM. SanDisk doesn’t make DRAM. It makes NAND flash, a different chip, built on different technology, sold into different customers.

So the real question isn’t whether CXMT is a threat. It’s whether the market actually did its homework before it sold SanDisk down with Micron, SK Hynix, Seagate, and Western Digital.

Mizuho’s Vijay Rakesh raised his SanDisk price target to $2,050 from $1,875 on October 5th, a week before this selloff, and kept his Buy rating.

His thesis: agentic AI is creating a storage demand cycle nobody has fully priced in yet.

Persistent AI agents need to store and retrieve data fast, something called KV-cache. Rakesh expects AI workloads to be 22% of total NAND demand by 2030.

Layer that onto SanDisk’s own numbers. The company’s New Business Model backlog sits around $42 billion in contracted revenue. Gross margin is 71.47%. There’s a $6 billion buyback running underneath all of it.

That is not a company getting disrupted. That is a company getting paid.

And the CXMT news, strictly read, doesn’t touch any of it. SanDisk sells flash memory. CXMT just said it’s getting better at DRAM.

Here’s what the “wrong memory type” argument conveniently skips over.

This is not the first time a CXMT headline has taken SanDisk down. Its July IPO triggered a roughly 47% drop in the stock over the following weeks, on the exact same logic investors are using today.

Twice now, a story about Chinese DRAM progress has dragged down a company that doesn’t sell DRAM.

Because when institutional money sells SK Hynix, Samsung, Micron, Seagate, and SanDisk together, on the same morning, that’s not five separate mistakes. That’s a bet that memory makers can shift capacity and capital between DRAM and NAND depending on where the margins are best.

If CXMT actually starts winning DRAM share, SK Hynix and Samsung don’t just sit still. They can reallocate fab capacity, pricing discipline, and capex toward NAND to defend their overall margins. That is the channel through which a DRAM story becomes a NAND problem, even if SanDisk never makes a single DRAM chip.

The “different memory type” defense is technically true and strategically incomplete.

There’s a second complication the bulls haven’t answered.

SanDisk is still up from a 52-week low of $120.96 to Monday’s $1,487.33. It peaked at $2,354.39 this summer.

At that kind of multiple expansion, how many “it’s not really a competitor” rebuttals does the market keep granting before it stops giving SanDisk the benefit of the doubt?

Meanwhile, Morningstar’s Jing Jie Yu is skeptical CXMT can actually commercialize its new architecture at volume. Reaching lab-level density is one thing. Reproducing it reliably across millions of chips that pass server qualification is another entirely.

Apple already tested this. The company reportedly explored buying DRAM from CXMT in July and walked away, because CXMT’s prices weren’t even below what SK Hynix and Samsung charge.

So the threat might be real eventually. It is not proven yet. And the stock already got sold twice on the promise of it.

SanDisk is sitting almost exactly on its 200-day moving average right now, around $1,476, after Monday’s close at $1,487.33.

That line has not been tested since the stock’s monster run began, which makes this the first real stress test the rally has faced on the way back down.

The stock’s 50-day average, at $1,593, is still meaningfully above where it’s trading, which tells you the short-term trend has already turned down even as the longer-term trend holds, barely.

Monday’s low was $1,464.43. That’s below the 200-day line. The close back above it, at $1,487, means buyers showed up exactly where you’d want them to if this support level still means anything.

If $1,476 breaks on the next scare, there isn’t an obvious shelf until you’re looking back toward the stock’s pre-breakout range from earlier this year. If it holds, this whole selloff reads as noise in hindsight.

📅 October 29 earnings
This is the single biggest catalyst on the calendar. Watch the NBM backlog growth rate and whether management says 71%+ gross margins are sustainable into fiscal 2027, or whether this quarter was the cyclical peak. A backlog that keeps growing settles the demand question in the bulls’ favor.

📐 The 200-day moving average
SanDisk closed Monday right on top of it, at $1,487 against a $1,476 line. A clean break below on real volume would tell you institutions are done defending this level. A bounce here, especially on a quiet news day, tells you the opposite.

🏭 Actual customer qualification news
Everything right now is speculative. The signal that would really move this stock is a named hyperscaler or server OEM confirming it’s qualifying CXMT DRAM for production, or confirming it isn’t. I covered this setup back when SanDisk was still riding the summer highs, in The Real Reason SanDisk Might Not Hold This Gain. The thesis hasn’t changed. It’s just getting tested with real money now instead of theory.

💻 Hyperscaler capex commentary
Microsoft, Amazon, and Alphabet all report in the next few weeks. Any comment on AI inference storage needs or KV-cache infrastructure spending directly supports or undercuts Mizuho’s demand thesis.

🎯 What other analysts do next
Mizuho stayed bullish through Monday’s drop. Watch whether anyone downgrades in the next week versus analysts reiterating targets. A wave of reiterations after a 6% drop is a tell that the Street sees this the same way we do.

I think the market got the headline right and the trade wrong.

CXMT making progress on DRAM is a real story. It is not SanDisk’s story.

Selling a NAND company because a DRAM company had a good week is the kind of mistake that gets corrected, eventually, once someone actually reads past the ticker symbol.

That doesn’t mean SanDisk is risk-free here. The capacity-reallocation argument is real, and the stock’s valuation leaves less room for error than it did at $120.

But a company with a $42 billion contracted backlog, 71% margins, and a fresh Buy rating from an analyst who publishes his math doesn’t deserve to trade like it just lost its biggest customer.

It got caught in a sector-wide flinch. Flinches fade. Fundamentals report on October 29th.

I’m not betting against the backlog.

If you have paid time left on an individual Granola plan, we'll match it with a Wispr Flow subscription that includes Notetaker and dictation, and add bonus time, up to 12 months total. Sign in or create a Wispr account and submit proof of your plan to check eligibility.

Try Notetaker Free

The most expensive mistake in investing isn’t being wrong about a company. It’s being right about the news and wrong about which company it applies to.

Markets move fast, and fast markets default to the easiest category available. “Chinese memory competitor” is a category. SanDisk got filed under it Monday morning, whether it belonged there or not.

The discipline that pays off here isn’t contrarianism for its own sake. It’s the habit of asking one more question before you sell: does this headline actually touch this balance sheet.

Most days, nobody bothers to check. That’s usually where the opportunity is.

If this deep dive changed how you see SanDisk, send it to one friend who’s still trading the headline instead of the facts.