Dossier · SNDK · Dormant
SNDK · Sandisk Corporation · Stock research
Last analysed ·
Current thesis
Memory supercycle unwinding: SNDK is -40% from its June peak and -25% in the week to 2026-07-17, driven by a Chinese memory IPO filing (7/16) and Kimi K3's cheap-compute shock (7/17). The 8.1x forward multiple is peak-earnings cheap, and a ~2026-07-30 Q4 print into broken structure is the binary. Stand aside until it bases.
Invalidation trigger
A weekly close below $1000 breaks the July capitulation floor and confirms full cycle rollover rather than a shakeout; reinforced if the ~2026-07-30 Q4 FY26 print guides NAND/SSD ASPs lower or the Chinese memory IPO prices.
Thesis status
Open commitment catalyst in 11dscored if the trigger above fires How this is scored →Latest analysis and events for SNDK —
As of 2026-07-19, orbyd's latest analysis for Sandisk Corporation (SNDK): Memory supercycle unwinding: SNDK is -40% from its June peak and -25% in the week to 2026-07-17, driven by a Chinese memory IPO filing (7/16) and Kimi K3's cheap-compute shock (7/17). The 8.1x forward multiple is peak-earnings cheap, and a ~2026-07-30 Q4 print into broken structure is the binary. Stand aside until it bases.
Invalidation trigger: A weekly close below $1000 breaks the July capitulation floor and confirms full cycle rollover rather than a shakeout; reinforced if the ~2026-07-30 Q4 FY26 print guides NAND/SSD ASPs lower or the Chinese memory IPO prices.
Next dated event on file: — catalyst in 11d.
Current Thesis
The AI-memory supercycle that carried SanDisk roughly 6,000% off its post-spin base has entered its unwind phase. The stock is down about 40% from its June peak and 25% in the week ending 2026-07-17 alone, and the damage is now cycle-wide: Micron, Seagate and Kioxia all broke together, technology fell 5.9% on the week (2026-07-17), and the trigger was exogenous Moonshot AI's Kimi K3 launch (2026-07-17) revived the cheap-Chinese-compute scare that hits memory demand assumptions first. What was a valuation debate in June is now a structural one: a new Chinese memory IPO filing (2026-07-16) put dated supply on the calendar, and US lawmakers urging a ban on Chinese memory chips (FT, 2026-07-16) confirms the threat is real enough to legislate against. The forward multiple compressing to 8.1x (2026-07-17) is not the bull case memory always looks cheapest on peak earnings, which is exactly what a Q4 FY26 print into a broken tape (~2026-07-30) is set up to reveal. This is the higher-beta derivative of Micron in a theme that has flipped from accelerating to saturated. The setup is a stand-aside until it bases.
Bullish and bearish views on Sandisk Corporation
The model's bull view on Sandisk Corporation (SNDK), in brief: Policy tailwind against the main bear, 2026-07-16: US lawmakers pressed the administration to ban Chinese memory chips (FT). The bear view: The drawdown is not a dip, 2026-07-16: down ~40% from the June peak and ~25% in a single week. Both cases follow in full.
Bull Case
- Policy tailwind against the main bear, 2026-07-16: US lawmakers pressed the administration to ban Chinese memory chips (FT). If enacted, the CXMT/YMTC supply threat that broke the tape gets legislated out of the US market a direct margin defense for SanDisk, Micron and Western Digital.
- Multiple has genuinely reset, 2026-07-17: SanDisk at 8.1x forward and Micron at 6.2x are now among the cheapest names in the Nasdaq 100 after the July selloff. Downside from a growth-multiple derate is largely spent even if the cycle argument is not.
- Flow is fading the fade, 2026-07-18: DRAM ETF inflows jumped despite the Micron/SanDisk/Seagate/Kioxia slump allocators are adding memory-cycle exposure into the drawdown rather than liquidating it.
- Supply deficit still asserted by a lead supplier, 2026-07-10: SK Hynix's chief told Bloomberg the memory shortage should run into the next decade. No supplier has walked back the deficit framing since.
- Options desks are engaged, 2026-07-16/17: SanDisk appeared repeatedly in Benzinga's information-technology whale-activity screens through the worst of the drawdown large-lot positioning is being built, not just unwound.
Bear Case
- The drawdown is not a dip, 2026-07-16: down ~40% from the June peak and ~25% in a single week. A move of that size in five sessions is liquidation, and liquidation rarely ends on the first washout.
- Chinese supply is now dated, not theoretical, 2026-07-16: a new Chinese memory IPO filing was the named driver of the 2026-07-16 Western Digital and SanDisk declines. Citi Wealth first flagged the CXMT/YMTC profit squeeze on 2026-07-08; the market is now pricing it.
- Demand-side shock from China AI, 2026-07-17: Moonshot AI's Kimi K3 unveiling extended the semiconductor rout. Cheaper frontier compute undercuts the hyperscaler capex arithmetic the entire memory bid rests on.
- Earnings into a broken tape, ~2026-07-30: memory equities historically screen cheapest immediately before estimates break (noted 2026-07-16). A Q4 FY26 print that guides NAND/SSD ASPs lower converts an 8x multiple into a 14x multiple without the stock moving.
- Sector correlation removes the idiosyncratic case, 2026-07-15/16: SanDisk sank 13% on 2026-07-15 alongside the broader memory complex; the name trades as beta to the group, so any long here is a leveraged bet on Micron rather than a company-specific view.
- Macro cover is gone, 2026-07-17: June inflation cooled to 3.5% and tech still fell 5.9%, with IBM posting its worst day on record. The AI trade broke on a friendly print that is de-rating, not macro noise.
Setup & Price Structure
The 50-day moving average was lost on 2026-07-08 and never reclaimed; the July decline has since taken out the pre-June base entirely. Distribution started 2026-06-26, one day after the Micron-driven rip, and has run without a meaningful counter-trend rally in three weeks. The correlated DRAM ETF has been in a >20% bear market since 2026-07-12 and has not repaired. Nothing about this structure offers a defined-risk entry: there is no higher low, no volume-dry pullback, and no reclaim of a broken level to trade against. A base needs a weekly close that holds, a retest that does not undercut it, and a first higher low none of which exist yet. Buying a 40% drawdown four sessions before an earnings print with no support reference is the classic averaging-into-broken-structure error, and the "cheapest Nasdaq 100 stock" headlines are the sentiment cover that usually accompanies it.
Catalyst Calendar (next 30 days)
- ~2026-07-30 (est.) Q4 FY26 results. The binary. NAND/SSD ASP guidance and any commentary on Chinese capacity determine whether 8.1x forward is real or an artifact of peak estimates.
- ~2026-08-01 to 08-15 (est.) post-print sell-side revisions. Watch whether the $1750–3250 target dispersion collapses downward; capitulation from the high-end targets is typically what marks the low.
Elapsed catalysts
- Ongoing, July–August 2026 Chinese memory IPO roadshow and pricing following the 2026-07-16 filing. Pricing dates are the cleanest scheduled bear catalyst for the incumbent margin structure. _(passed 3d ago)_
- Undated, active US legislative action on a Chinese memory chip ban (FT, 2026-07-16). Any formal rule-making or executive action is the single largest upside surprise available. _(passed 3d ago)_
What Would Change Our Mind
A weekly close that reclaims the broken 50-day moving average, with the DRAM ETF exiting its bear-market drawdown at the same time, would mark the cycle scare as a shakeout rather than a top. A Q4 FY26 print around 2026-07-30 that guides NAND ASPs flat-to-higher and explicitly addresses Chinese capacity would remove the estimate risk the multiple is signalling. Formal US restriction on Chinese memory imports would break the primary bear leg outright. Absent those, price needs to build a base with a defined higher low before any structural view improves a low-conviction stance on fresh entries either way.
Correlation Notes
Micron, Western Digital, SK Hynix, Samsung, Kioxia, Seagate and SanDisk trade as one risk unit; the 2026-07-15 through 2026-07-18 sessions moved them together with no dispersion. Any memory exposure is a single concentrated AI-memory bet, not diversification. Micron remains the cheaper (6.2x vs 8.1x forward), more liquid anchor for the identical cycle the incremental case for owning the higher-beta derivative instead is hard to construct at these levels. Secondary correlations run to the DRAM ETF as the theme's flow proxy and to Nasdaq 100 semiconductor beta, which was the transmission channel for both the 2026-07-16 and 2026-07-17 declines.
Notes
- c
- o
- r
- r
- u
- p
- t
- e
- d
- n
- o
- t
- e
- s
- b
- l
- o
- b
- 8
- 4
- M
- B
- b
- a
- c
- k
- s
- l
- a
- s
- h
- c
- a
- s
- c
- a
- d
- e
- w
- i
- p
- e
- d
- 2
- 0
- 2
- 6
- 0
- 5
- 1
- 4
- s
- e
- e
- b
- a
- k
- Next SNDK earnings (Q4 FY26) est. late July / early Aug 2026 outside the 30d window; re-check blackout proximity as it approaches.
- China memory IPOs (CXMT/YMTC-type) are the cleanest structural bear catalyst watch for pricing/roadshow dates (flagged 2026-06-03).
- Memory complex (MU/WDC/SK Hynix/Samsung/Kioxia) trades as one risk unit treat any memory long as a single concentrated AI-memory bet, not diversification.
- Theme maturing→saturated signals: top-searched-ticker status (2026-06-02) + bubble-burst headlines (2026-06-06) + viral DRAM ETF -20% (2026-06-07). Require a pullback that holds the 20-EMA and bases before any fresh entry.
- Next SNDK earnings (Q4 FY26) est. late July / early Aug 2026 at/beyond the 30d edge; re-check blackout proximity as it approaches.
- Memory complex (MU/WDC/SK Hynix/Samsung/Kioxia/STX) trades as one risk unit treat any memory long as a single concentrated AI-memory bet, not diversification.
- Saturation signals stacked: peak retail-search status, target dispersion $1750-3250, viral DRAM ETF -20% (2026-06-07), first distribution day 6/26 after the Micron-driven 16% rip. Require a pullback that holds the 20-EMA and bases before any fresh entry.
- MU is the cheaper, more liquid anchor for the same cycle (cheapest multiple of MU/WDC/SNDK per 2026-06-25 comparison); SNDK's incremental edge over MU at current price is hard to locate.
- Q4 FY26 earnings est. ~2026-07-30 (late July / early Aug) now INSIDE the 30d window; treat as an approaching binary and re-check exact date + blackout proximity.
- China memory IPOs (CXMT/YMTC-type) escalated from background risk to a named, dated bear catalyst (Citi Wealth 2026-07-08) watch for roadshow/pricing dates as the cleanest way to break the shortage-into-2028/next-decade frame.
- Structure broke: SNDK closed below its 50-DMA on 2026-07-08; require a reclaim + higher-low base that holds on volume before any fresh entry do not buy the first pullback into a chip rout.
- Memory complex (MU/WDC/SK Hynix/Samsung/Kioxia/STX) trades as one risk unit any memory long is a single concentrated AI-memory bet, not diversification.
- MU remains the cheaper, more liquid anchor for the same cycle; SNDK's incremental edge over MU at current levels is hard to locate.
- Saturation signals stacked: peak retail-search status, target dispersion $1750–$3250, DRAM ETF still >20% off peak (2026-07-12), first distribution day 2026-06-26, H2 'selling winners, buying losers' rotation (2026-07-09).
- Q4 FY26 earnings est. ~2026-07-30 INSIDE the 30d window; confirm exact date and blackout proximity before any entry.
- Chinese memory IPO filed 2026-07-16 roadshow/pricing dates are the cleanest dated bear catalyst; track them.
- US lawmakers urging a ban on Chinese memory chips (FT, 2026-07-16) is the single largest undated upside catalyst; any formal rulemaking flips the structural bear.
- Memory complex (MU/WDC/SK Hynix/Samsung/Kioxia/STX/SNDK) trades as one risk unit any memory long is a single concentrated AI-memory bet, not diversification.
- MU remains the cheaper (6.2x vs 8.1x fwd), more liquid anchor for the same cycle; SNDK's incremental edge is hard to locate.
- Theme status flipped ACCELERATING → SATURATED/rolling: 50-DMA lost 2026-07-08, DRAM ETF >20% bear market since 2026-07-12, distribution from 2026-06-26.
- Memory equities screen cheapest right before estimates break treat a single-digit forward multiple here as an estimate-risk signal, not a valuation floor.
- Require a weekly close reclaiming the 50-DMA plus a first higher low before treating any bounce as a base.
Related · shared themes
SIMO
Silicon Motion Technology Corporation
NAND-controller toll-booth on the steepest flash shortage in ~15 years: TrendForce H1 2026 contract +>100% cumulative, H2 still rising with no capacity adds. Q1 record +105% YoY, sell-side chasing to $400/$450. Fundamental leg ACCELERATING, but the tape is digesting a 4x into insider selling the 2026-07-29 Q2 print is the next binary.
AMD
Advanced Micro Devices, Inc.
MI450/Helios rack-scale plus EPYC agentic-CPU narrative drove ATHs near $561.80 (7/1), but the July target cluster ($600 BNP, $620 BofA, $725 high) is arriving into a falling tape 20-day MA lost 7/8, chip-led Nasdaq selloff 7/17, sentiment into Fear. Sell-side catch-up after the move, with the 8/4 print as the binary.
HPE
Hewlett Packard Enterprise Company
AI-server and enterprise-networking pivot is real at the theme level, but HPE's own chart broke: the record +25% Q2 FY26 blowoff (2026-06-02, ATH $64.25) fully round-tripped, the ~$48 pre-print shelf failed (last observed close $43.71 on 6/26), and nothing is dated until the Q3 print in early September. A fresh buy here catches the breakdown, not a base.
MU
Micron Technology, Inc.
HBM scarcity intact SK Hynix sees shortage into the next decade, Qualcomm/Hyundai auto deals signed 2026-07-16, US push to ban Chinese memory but the tape broke: SOXX −18.6% in July, worst month since 2008, and the $854.50 shelf failed. At 6.2x forward the setup is cycle-timing, not momentum; no higher low yet.
See also · stocks to watch