SanDisk [SNDK]
Task 6 · built 2026-07-27 · migrated to the Criteria framework 2026-07-29 · framework v1.5.1 Price at build $1,278.23; current price $1,095.78 (scan, 2026-07-28). Research direction: bearish. No position verdict is expressed here or anywhere else in this memo.
Every entry is dated, checkable, and carries an explicit upgrade/downgrade threshold. Hand off to
equity-research:catalyst-calendar and equity-research:thesis-tracker for ongoing maintenance.
| Date | Event | Confirmed by | What to check | Downgrade (→ Short candidate) | Upgrade (→ Avoid / Long review) |
|---|---|---|---|---|---|
| Wed 5 Aug 2026, 1:30pm PT | Q4 FY2026 + FY2026 results | Company press release, 8 July 2026 | Realised Q4 gross margin vs the 78.9–80.9% guide; first FY2027 commentary; reported RPO in the FY2026 10-K vs $41.6bn; bit-shipment growth y/y; Datacenter revenue | FY2027 revenue commentary implying < ~$36bn, OR a Q1 FY2027 GM guide < 70%, OR RPO fails to grow despite two more NBMs signed | Q4 GM > 81% with a Q1 FY2027 guide > 80%, OR RPO > $60bn |
| Thu 13 Aug 2026, 9:00am ET | INVESTOR DAY — the single highest-value event on this calendar | Company press release, 8 July 2026 | The long-term financial model management deferred on the Q3 call ("we'll wrap this all up in a new model"): the long-term gross-margin target, the contracted-share-of-bits trajectory, NBM pricing mechanics, and the capital-return framework | A long-term GM target < 55%, or a refusal to publish one again | A contractually-supported long-term GM target > 65% with the mechanics disclosed → this invalidates the bearish research view outright |
Why the Investor Day is the Catalyst Criteria catalyst. The house-vs-Street disagreement is not about FY2027 execution; it is about the durability of a 78% gross margin. The Investor Day is the event at which management must attach a number to precisely that. It is also, symmetrically, the largest right-tail risk to any negative exposure. (The prior text concluded "which is why the decision is Watchlist and not Short." No decision is emitted; the observation stands, the conclusion is deleted.)
| Expected | Event | Basis | What to check | Threshold |
|---|---|---|---|---|
| ~early Nov 2026 | Q1 FY2027 results | Q1 FY2026 was reported 2025-11-06 | The first quarter that must show whether ASPs fell. Sequential gross margin; sequential ASP direction; contracted share of bits | Sequential GM decline → evidence trigger B1 fires. GM up again → the reversion thesis is a year early |
| ~late Jan 2027 | Q2 FY2027 results | Q2 FY2026 reported 2026-01-29 | Contracted share of FY2027 bits vs the ">one-third, can get above 50%" claim | Contracted share < 40% → the NBM ramp is stalling |
| ~late Apr 2027 | Q3 FY2027 results | Q3 FY2026 reported 2026-04-30 | Whether the 2H-CY2027 rebalancing TrendForce forecasts is visible | Any negative sequential ASP print |
| ~mid-Aug 2027 | Q4/FY2027 results | FY2026 reported 2026-08-05 | RPO run-off vs schedule; first NBM contract anniversaries | RPO below $30bn with no new agreements |
These resolve between prints and are the cheapest way to be early. Cadence and query are stated so the check is executable.
| Item | Cadence | Source | Baseline at 2026-07-27 | Trigger |
|---|---|---|---|---|
| NAND contract-price change, q/q | Monthly / quarterly | TrendForce press centre | +33–38% (1Q26) → +70–75% (2Q26) → +10–15% (3Q26E) | Any negative print → evidence trigger B2 fires. A fourth consecutive print above +20% → thesis is wrong |
| TrendForce supply/demand balance | Quarterly | TrendForce | 2026-07-21: "supply growth to outpace demand in 2027, easing in 2H27" | A retraction or a push-out beyond 2028 → downgrade the mechanism |
| NAND conversion WFE | Quarterly, with LRCX/AMAT/TEL results | Equipment-maker earnings calls | ~$40bn pulled forward, majority before end-CY2027; 2027E WFE $191bn (+28%) | Conversion spend deferred or cut → the supply response is not arriving |
| Kioxia capacity commentary | Quarterly | Kioxia results | Flash Ventures extended to 31 Dec 2034 (10-Q, 29 Jan 2026) | Any greenfield announcement → supply response accelerates |
| Micron read-across | Quarterly | MU results | Gross margin 84.6% (quarter ended 2026-05-28); ρ(SNDK, MU) = +0.78 | A sequential MU gross-margin decline is the highest-value early read on SNDK, and it typically reports first |
| BIS / Federal Register | Monthly | federalregister.gov/api/v1/documents.json, agency industry-and-security-bureau |
Zero NAND-specific actions since 2024-01-01 (verified) | Any NAND-specific export control → re-scope both directions |
| Nanya DRAM arrangement | Quarterly | SNDK 10-Q Notes 11/16; Nanya disclosures | ~$1bn equity purchase settled 2026-04-08; committed annual volumes at variable pricing off Nanya's trailing ASPs | Any disclosure sizing the DRAM cost exposure → this is an unmodelled margin-squeeze vector |
| SanDisk job postings | Monthly | ATS not yet located — Greenhouse, Lever and six Workday tenant/site combinations all returned 404/422 | Not retrieved | Locating the ATS is itself a research task; a hiring mix skewed to go-to-market vs process/engineering would be a counter-signal |
| Date | Item | Source | Relevance |
|---|---|---|---|
| Continuous | $6bn share repurchase, authorised 30 Apr 2026 | 10-Q MD&A | A live right-tail risk to any short; also a value-destruction risk if executed at peak margins |
| Late 2026 | HBF NAND die delivery | Q3 FY26 call | HBF has gone to zero mentions in prepared remarks for two quarters. A shipment announcement would reverse that decay |
| Early–mid 2027 | HBF full system with controller | Q3 FY26 call | The only genuinely new category in the story |
| Q4 FY2026 (now) | Stargate QLC eSSD enters revenue shipment | Q3 FY26 call | Watch the Datacenter line for the QLC contribution |
| Through FY2029 | Kioxia building-depreciation prepayments: $17m (rest of FY26), $119m (FY27), $174m (FY28), $63m (FY29) | 10-Q Note 11 | Committed cash outflow; charged in the model |
| 31 Dec 2034 | Flash Ventures agreements expire (extended 29 Jan 2026 from 2029) | 10-Q Note 11 | Supply security is locked well beyond the forecast horizon |
| ~FY2031 | Last NBM contract expiry (longest term is five years from Q3 FY2026) | Q3 FY26 call | The date the contracted tranche fully rolls off and the model reverts to merchant treatment. Per the regime-change test, this is when the thesis gets retested from scratch |