The competitor narrative is straightforward: Sandisk trades at a wider multiple than Micron, therefore the gap looks excessive. That comparison only works if you assume both companies face the same earnings volatility and product mix risk. They don't. But that doesn't make Sandisk's current valuation safe - it just shifts the problem from the gap to Sandisk itself.
I'm maintaining a Hold on Sandisk (NASDAQ: SNDK). The stock's operating metrics are outstanding, but the multiple has run far ahead of what a cyclical memory name can sustain - even one with $42 billion backlog and AI-driven NAND demand. At this point, the risk/reward has flipped. The thesis still works. The price no longer does.
What happened
Sandisk spun off from Western Digital in February 2025, starting as a $5.6 billion company trading under SNDK. Fast forward 16 months and the market cap has swollen to roughly $260 billion. The stock is up more than 218% in 2026 alone and has more than quintupled from spinoff. It hit a $200 billion market cap for the first time in early May, then kept climbing.
Micron (NASDAQ: MU), for comparison, just broke through $1 trillion in market cap in late May after surging nearly 20% in a single day. Both stocks are riding the same AI memory infrastructure build-out. The difference is what the market is paying for exposure.
The numbers don't lie - they just don't support this multiple
Sandisk's latest quarter (Q3 FY2026, reported April 30) was spectacular on the surface. Revenue of $5.95 billion - up 97% sequentially and 251% year-over-year, against guidance of $4.4–4.8 billion. Adjusted EPS of $23.41, well above the $14.36 consensus. The company generated $2.96 billion in adjusted free cash flow, a 49.7% margin, and ended the quarter with $3.74 billion in cash. Management highlighted a $42 billion backlog of hyperscaler orders.
Those are not cyclical trough numbers. They're the best kind of cycle peak.
The problem is that the stock now trades at roughly 20 times trailing sales and about 62 times trailing earnings. The forward P/E sits around 19.5x. Even granting that cyclical memory companies deserve higher multiples in an upcycle, this pricing assumes no quarter disappoints, no margin contracts, and no NAND supply expansion for the foreseeable future. For a business whose margins have historically swung from deep losses to extraordinary profits, that's an expensive bet.
Micron, by contrast, trades at a forward P/E of roughly 11x. Its trailing P/E is inflated at around 50x because prior-year earnings were depressed - the classic cycle-bottom denominator effect. But forward, the market is pricing Micron as a company that will earn substantially more going forward, and it's not asking you to pay a premium for that assumption. Micron's Q2 FY2026 revenue of $23.86 billion (beating $20 billion estimates) and EPS of $12.20 (vs. $8.60–9.31 estimates) show it's winning the same AI memory demand battle, just with a diversified DRAM-plus-NAND mix rather than Sandisk's pure NAND play.
Why the gap exists - and why it doesn't justify the premium
The valuation gap between Sandisk and Micron exists for a reason. Sandisk is a pure-play on NAND flash, and hyperscaler AI storage demand is real. The $42 billion backlog gives investors visibility into demand that Micron's DRAM business - with its more fragmented end markets - can't match on a product-by-product basis. Sandisk also delivered software-like free cash flow margins of nearly 50%, which is extraordinary for a memory manufacturer.
But here's what the premium price assumes will stay true: NAND supply stays tight, hyperscaler capex doesn't waver, Sandisk's margin expansion is durable, and the company sustains triple-digit revenue growth quarter after quarter. That's a lot of variables to price at 20x sales. If any one of them softens - and memory cycles always turn - the multiple compression will be severe.
Micron faces the same demand tailwinds but with a different risk profile. DRAM is where the highest-value AI memory dollars flow (HBM for NVIDIA GPUs, for example), and Micron is a top-three DRAM supplier. Its earnings power in this cycle will be enormous. The stock at 11x forward earnings is saying the market is cautious about the cycle turn, not dismissive of the growth. That's a safer place to be.
What would change the rating
I'd upgrade Sandisk to Buy if the stock pulled back to the $160–180 billion market cap range - roughly 30–40% lower - where the backlog visibility and cash flow would start looking like a margin of safety rather than a justification. At that level, the 251% revenue growth and $42 billion order book would be compelling enough to offset cycle risk.
I'd also upgrade if Sandisk's next two earnings reports sustained gross margins above 70% while revenue grew 50%+ sequentially, proving that the current margin profile is structural rather than cyclical. That would justify re-rating the stock as a growth compounder instead of a cycle play. It hasn't done that yet. One quarter of extraordinary results doesn't rewrite a decade of memory cycle history.
The downgrade trigger is already in play for new buyers. If NAND supply catches up to demand - Samsung and SK Hynix are expanding capacity - and Sandisk's gross margins contract from current levels, a stock at 62x trailing earnings will fall hard. Cyclical names punished at cycle peaks don't get second chances on the same narrative.
The takeaway
The Micron comparison is a red herring. The issue isn't that Sandisk trades at a wider multiple than Micron - it's that Sandisk's multiple is too wide for what it is: a cyclical NAND manufacturer, no matter how impressive its current quarter. Micron at 11x forward earnings with the same AI demand tailwinds is the more defensible position.
Sandisk is not broken. The business is firing on all cylinders. But firing on all cylinders at the top of a cycle is exactly when you stop buying.
Rating: Hold. Wait for a pullback, not a catalyst.