SanDisk Corporation (SNDK)
Verdict history · SNDK
Period return = price change from verdict date to next verdict date (or present for Active). Prices are from published frontmatter — a stated fact, not a live feed. Full track record →
The case against outweighs the case for at today's price.
NASDAQ · Semiconductors / NAND Memory · 2026-07-01 · analysis, not advice
The panel's take
Verdict: SELL · Conviction: MEDIUM · Last price: USD 2,060.04 (weekly close; traded to ~$2,238 intraday on 2026-07-01 on a broker upgrade)
SanDisk is the panel's clearest cyclical-peak warning — and, unusually, the caution comes despite genuinely spectacular numbers. The NAND memory shortage is real: fiscal Q3 revenue jumped 251% year-over-year to $5.95B, non-GAAP gross margin exploded to ~78% (from ~51% the prior quarter), EPS hit $23.41 against a $14 estimate, the balance sheet is now debt-free with a $42B multi-year contracted backlog, and the stock is the S&P 500's best performer, up roughly 4,700% over twelve months. The problem is what you're being asked to pay for it. Five of six frameworks — value, growth, valuation, forensic, and fundamental — land bearish, and their objection is identical: this is a commodity with a 40-year boom-bust history, its ~78% gross margin is a transient shortage rent (mid-cycle is 25-35%, the FY23 trough was 7%), and the market is capitalizing peak-cycle earnings as if they're permanent. Normalized through the cycle, independent fair-value estimates cluster from roughly $400 to $750 — 65-75% below the price — and the analyst average target ($1,751) actually sits below the current quote, with insiders net sellers and short interest at record highs into an all-time high. Only the momentum lens stays constructive, and even it is "actively trimming." The demand is genuine and the shortage may run into 2028, which is why this is a Medium-conviction call rather than a high one — but paying peak multiples on peak earnings in commodity memory is, historically, the single most reliable way to lose money in this industry. The synthesis is SELL.
Key levels
Key levels · SNDK
USD · as of 2026-07-01R1 · 2,280 · +10.7% from current
All-time-high shelf
S1 · 1,850 · −10.2% from current
Prior breakout shelf
S3 · 1,200 · −41.7% from current
Deeper parabolic-unwind zone
Analyst consensus target 1,751 USD · range 1,000–3,000
Key support & resistance and analyst consensus — educational analysis, not advice. These are not entry or exit prices. Trading involves risk of loss.
Key resistance starts at the ~$2,280 all-time-high shelf (the 52-week high is ~$2,354), then round-number projections at ~$2,400 and ~$2,500 — a parabolic advance has little overhead structure, so these are projection, not memory. Key support sits at ~$1,850-1,900 (the prior breakout shelf during the vertical leg), then ~$1,600 (a ~38% retracement with thin volume beneath it), with a deeper parabolic-unwind zone near ~$1,100-1,200 (~61.8% retracement — where blow-off moves often revert). The analyst consensus target spans an extraordinarily wide USD 1,000 (a "bubble" fair-value call) to 3,000 (a fresh Street high), averaging near USD 1,751 — notably below the current price, a sign consensus is chasing the move rather than leading it; the panel's own through-cycle fair values ($400-750) sit lower still.
What legendary investors think
We ran SanDisk past a panel of six legendary investors' frameworks.
The panel · 6 investors
5 bearish · 1 bullishMCThe Moat CompounderValueBearishHigh conviction
"A cheap multiple on peak, unsustainable earnings isn't cheap — it's the classic value trap." NAND is a commodity; SanDisk controls its price "no more than a wheat farmer controls the wheat price." Up 4,700% in a year isn't the business getting 47x better, it's the market pricing a cycle top as forever. No moat, lousy through-cycle returns — outside the circle on cycle timing.
GOThe GARP OperatorGrowth (GARP)BearishHigh conviction
"A cycle peak dressed up as a supercycle." The P/E *lies* at both ends — trailing ~68x actually understates the extension because peak pricing inflates earnings. A 55-point gross-margin jump in a year on a commodity is the tell. The fortress balance sheet survives the downturn, "but it doesn't protect you from paying 68x peak earnings." Sell when the P/E looks cheap and everyone says "new paradigm."
DSThe Disruptive-Innovation SeekerGrowthBearishMedium conviction
The AI-storage demand is real (datacenter NAND growing high-60% in exabytes), "but this is not our kind of disruptive platform." NAND is a mature, commoditized S-curve; the margin surge is a supply-discipline pricing event, not a proprietary moat. "We'd rather own the AI compute and data layer where IP compounds — not the interchangeable commodity that gets squeezed when the next capacity wave lands."
IMThe Intrinsic-Value ModelerValuationBearishHigh conviction
A through-cycle DCF (normalized revenue ~$10-12B, mid-cycle margin ~13%, 10-11% WACC) yields fair value ~$550-750 — 65-75% downside. "7% gross margin in FY23, 80% in FY26 — that amplitude *is* the story, and the market is pricing the top of it as the new floor." Even the bull case (structurally higher mid-cycle margins) leaves it ~half overvalued.
FSThe Forensic SkepticContrarianBearishMed-High conviction
"The lowest-P/E-at-the-top trap I've seen in every commodity cycle." Q3's +251% YoY / +97% QoQ is a peak comp that never repeats; capacity is coming (Kioxia/SanDisk capex +41%, Micron +63%, with a 12-18 month lag before the glut); insiders are selling into strength; the $42B fixed-price backlog "becomes a liability, not an asset, when spot rolls over." Through-cycle value ~$400-700.
MOThe Macro OpportunistMacroBullishMedium conviction
"That's not a stock chart, that's a commodity-squeeze chart" — weekly RSI pinned above 99, price 4x above its 200-day, negative momentum divergence at the high. The shortage is genuine (NAND contract prices +33-38% QoQ) so he won't short it, "but the Fed isn't easing and this parabola is financed by scarcity and index flows, not liquidity." Trim into strength; new money is a trading position sized for a 30-40% air pocket.
Each view is one investing framework applied to the stock — a perspective, not advice, and identical for every reader. Signals are the panel's own scale, not a recommendation to act.
Where they agree — and where they clash
Common ground
- The numbers are genuinely extraordinary right now: revenue +251%, ~78% gross margin, EPS $23.41, a debt-free balance sheet, and a $42B contracted backlog — this is a real shortage, not a hoax.
- It is nonetheless a commodity cyclical: NAND has a 40-year boom-bust history, and ~78% gross margin is far above the 25-35% mid-cycle norm (7% at the FY23 trough).
- The valuation prices peak economics as durable: ~68x trailing earnings, the richest EV/EBITDA in the NAND peer group (vs Micron ~17x, WDC ~29x), with through-cycle fair values far below the price.
- Positioning is stretched and two-sided: up ~4,700% in a year, insiders net sellers, record-high short interest, and an analyst average target below the current price.
The real debate
- Structural shift or cyclical spike? The demand and the $42B of multi-year fixed-price contracts are real; bulls say AI-storage has re-based the cycle higher, bears say "more visibility is not no cyclicality" — only ~a third of revenue is locked, the rest rides spot pricing.
- Does the momentum keep running? The macro lens says a genuine physical shortage plus index flows can push a parabola further; the value and forensic lenses say a stock 4x above its 200-day with insiders selling is late, not early.
- When does capacity break it? Everyone agrees new NAND capacity ends every cycle; the dispute is timing — the bears see the 2027-28 capex wave already in motion, the bulls see supply capped below demand through 2028.
The question it comes down to: Is SanDisk a structurally re-rated AI-storage winner whose contracted backlog has tamed the old NAND cycle — or the S&P's most extended commodity stock, priced on peak-cycle earnings that a 40-year history says mean-revert hard, with fair value and even the Street's own average target sitting well below the price?
The numbers
| Metric | Value |
|---|---|
| Price / Market cap | USD 2,060.04 (weekly close; ~$2,238 on 2026-07-01) / ~330B |
| P/E (TTM / fwd) | ~65–75x / ~30x (fwd collapses on annualized peak-quarter earnings) |
| Gross margin | ~78% non-GAAP (Q3 FY26; vs ~25-35% mid-cycle, ~7% FY23 trough) |
| EV / EBITDA | ~52x (vs Western Digital ~29x, Micron ~17x) |
| Balance sheet | Cash ~$3.7B, zero long-term debt (term loan extinguished); $42B contracted backlog |
| Revenue (Q3 FY26) | USD 5.95B (+251% YoY, +97% QoQ); Q4 guide $7.75–8.25B |
| YTD / 1-year return | ~+767-857% YTD; ~+4,700% over 12 months (best in the S&P 500) |
Figures as of fiscal Q3 2026 (quarter ended ~April 3, 2026, reported April 30) / late June 2026; sourced from SanDisk IR/SEC filings, Macrotrends, StockAnalysis, TrendForce, TIKR, Benzinga, Morningstar, Bernstein/Citi via news coverage. The ~$2,060 weekly-close price is real (verified across multiple sources) — not a split artifact; the stock traded to ~$2,238 intraday on July 1 on a broker upgrade. Trailing and forward P/E diverge sharply (peak-cycle earnings distortion); a hyper-cyclical commodity is valued through-cycle (normalized margins, mid-cycle revenue) rather than on point-in-time multiples — which is why the panel's fair values sit far below the tape.
The bottom line
SanDisk is the panel's sharpest "great numbers, dangerous price" call. The bull facts are undeniable and, this quarter, historic: a genuine AI-driven NAND shortage drove revenue up 251%, gross margin to ~78%, and EPS to $23.41, all on a now-debt-free balance sheet with $42B of multi-year contracts providing more visibility than any prior memory cycle offered. But five of six frameworks — value, growth, valuation, forensic, and fundamental — reach the same verdict for the same reason: this is a commodity with a 40-year boom-bust record, its ~78% margin is a transient shortage rent against a 25-35% mid-cycle norm, and the market is capitalizing the peak of the amplitude as if it were the floor. Normalized through the cycle, fair value clusters from roughly $400 to $750 — 65-75% below the price — the analyst average target ($1,751) sits below the current quote, insiders are selling into strength, short interest is at record highs, and the one bull (momentum) is already trimming into a stock pinned at RSI 99 and 4x above its 200-day. What would change the call is genuine, durable industry supply discipline that converts most revenue to multi-year fixed pricing and re-bases mid-cycle margins structurally higher — the bull thesis, which the $42B backlog partly supports. What makes it a sell is that, absent that, buyers are paying peak multiples on peak earnings in the most cyclical corner of semiconductors, right as the 2027-28 capacity wave gets underway — and a supercycle, however real, is still a cycle. The magnitude of the demand and the ferocity of the momentum are why conviction is Medium, not high; the direction is not in much doubt.
Verdix's panel is made up of AI archetypes that apply the well-documented, publicly known investment frameworks of famous investors. They are AI agents — not the investors themselves. Verdix is not affiliated with, endorsed by, or authorized by any real individual, and the archetypes do not represent any real person's actual views, holdings, or statements. Every verdict is AI-generated. Meet the panel →
Verdix provides educational equity research and AI-generated multi-perspective analysis. Nothing here constitutes personalized investment advice or a recommendation to buy or sell any security. Verdicts are uniform across all users and do not consider your individual financial situation, risk tolerance, or objectives. Trading and investing involve substantial risk of loss. Past performance does not predict future results. Consult a licensed financial advisor before making investment decisions.