SNDK’s AI Boom Is Finite, and the Clock Is Ticking

After a brutal 35% wipeout earlier this week, Sandisk (SNDK) stock bounced back more than 3% in premarket trading Wednesday. But anyone tempted to call it a comeback should read the fine print. Morningstar just put a $1,000 price target on the shares — a number that sits below where the stock closed Tuesday — and slapped it with a “Very High” uncertainty rating. That’s not exactly a ringing endorsement for a name that’s still up nearly 350% this year.

The note from senior equity analyst William Kerwin is a masterclass in sober analysis. He acknowledges that AI infrastructure spending is fueling an extraordinary run for NAND flash memory. SanDisk’s revenue could compound at 27% annually through fiscal 2030. Gross margins might exceed 80% in 2027. (That’s right — a memory chip company flirting with software-like margins.) But here’s the thing: every boom has a shelf life, and this one already has an expiration date written on it.

The Finite Nature of a Commodity Boom

Kerwin doesn’t mince words. “Exponential pricing growth is tremendous, but finite,” he wrote, and the reason is uncomfortably simple. NAND flash memory remains a commodity business — the kind where one manufacturer’s chips are largely indistinguishable from another’s. In that world, pricing power doesn’t come from brand moats or proprietary wizardry. It comes from the brutal arithmetic of supply and demand.

Right now, demand is white-hot because AI data centers are guzzling storage like never before. At the same time, new fabrication plants take years to build. That mismatch has sent NAND prices soaring and lifted the entire industry. SanDisk is riding that wave beautifully. But as Kerwin points out, these gains are “cyclical” rather than “structural.” When supply eventually catches up — and it always does — the floor can vanish fast.

How AI Is Inflating Margins (and Why That Can’t Last)

Let’s sit with those margin projections for a moment. An 80% gross margin is the kind of number you’d associate with a platform business — think a cloud software company with zero cost of goods sold. For a hardware manufacturer in a cutthroat commodity market, it’s almost surreal. Morningstar expects tight supply to keep those elevated margins alive through fiscal 2027, which gives SanDisk a multi-year runway of astonishing profitability.

But the model hinges on one fragile assumption: that buyers keep paying up without a viable alternative. That’s where things get dicey. NAND chips follow brutal boom-and-bust cycles. When prices spike too high, data center operators and device makers start scrambling for substitutes or delaying purchases. Eventually, capacity comes online, inventories swell, and the pricing umbrella collapses. Morningstar’s timeline puts the peak right around 2028 — just two years after those peak margins arrive.

When the Cycle Peaks: 2028 and the Coming Hangover

What happens in 2028? New manufacturing capacity. A lot of it. SanDisk isn’t the only one racing to capitalize on the AI gold rush. Competitors, both established and emerging, are pouring billions into fab expansions. Morningstar’s analysis points to a sharp downturn in 2029 as that supply surge hits, dragging NAND prices back down to earth. For investors who’ve watched the stock multiply, the risk is watching those paper gains evaporate in a matter of quarters.

Think of it as a Wile E. Coyote moment. The stock runs off the cliff, legs still churning, and only looks down when the support is gone. By the time the cycle turns, the narrative will pivot from “AI-driven scarce resource” to “overbuilt commodity glut.” That’s not a prediction of doom so much as a recognition of how memory markets have worked for decades.

The Long-Term Risks Nobody’s Talking About

Beyond the cycle itself, Kerwin flagged a pair of structural headwinds that make SanDisk’s moat — well, nonexistent. On one side, you’ve got Chinese state-backed competitor YMTC aggressively expanding its NAND footprint. YMTC isn’t just chasing market share; it’s backed by a government willing to subsidize capacity in ways that could permanently depress industry pricing.

On the other, NAND manufacturing is ferociously capital-intensive. Building and equipping a leading-edge fab can cost well north of $20 billion. When demand softens, those fixed costs don’t disappear. They just slice into margins with a cleaver. Together, these forces mean SanDisk is unlikely to earn returns above its cost of capital over a full cycle — the classic definition of a business without an economic moat.

Here’s the shorter list of long-term risks, distilled:

• YMTC’s capacity expansion — state-funded overproduction could flood the market.
• Capital intensity — when downturns hit, high fixed costs crush profitability.
• Commodity status — no real differentiation means pricing power evaporates the moment supply normalizes.

What Retail Investors Are Saying Now

On Stocktwits, the mood has swung from bearish to “extremely bullish” in a single day, with message volume more than doubling. Some traders see the dip below $1,000 as inevitable, a self-fulfilling prophecy that might trigger panic selling. Others are convinced the stock will rally hard in the near term. There’s a palpable sense of whiplash — and honestly, that’s perfectly rational given the numbers.

A 35% drop over a few sessions is jarring even for a stock that’s been on a multi-month tear. But if Morningstar’s cycle framework holds, this kind of volatility isn’t a bug — it’s a feature. The debate over whether $1,000 will hold isn’t just a technical squabble. It’s a proxy for the bigger question: Is the AI memory boom already priced in, or does the real mania still lie ahead?

For now, the party rages on. Revenue is surging, margins are fat, and AI demand shows no sign of letting up. But the analysts who’ve studied these cycles for years are already marking their calendars for the comedown. The AI supercycle isn’t a myth — it’s just on a timer. And the louder the music gets, the fewer people seem to notice the fuse burning at the edge of the dance floor.