Gross margin went from 26% to 85% in four quarters and management referenced 2023 unprompted while showing zero hedging language.
Thesis: SNDK converted a quarterly spot-price memory franchise into multi-year contracted revenue at ~80% gross margin with eight customers who have no near-term BiCS8-scale alternative. The structural shift is real and the cash flow confirms it. The risk is that guarantee coverage sits at 17.6% of the $93.9B commitment with releases back-loaded to years 3-5, and management is building inventory against an allocation narrative while consumer segments contract 32% QoQ. Whether the NBM floor holds in year 3 depends on hyperscaler capex decisions SNDK does not control.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Investor Day next week where management will detail HBF progress, BiCS10 economics, and potentially new NBM signings. Q1 print of $10.3B+ at or above 83% GM would confirm the contract base holds pricing.
Key Risk: Guarantee coverage of 17.6% ($16.5B against $93.9B) with releases timed to the end of agreements, combined with 50%+ of FY27 bits locked to eight hyperscalers whose capex SNDK does not control. If inference architecture shifts or capex normalizes in year 3, the released guarantees offer no cushion and the non-NBM half faces spot pricing with no floor.
The Tell: When CJ Muse asked about margin durability, David voluntarily referenced 2023: 'In 2023, we were having very different earnings calls than we're having now.' Unprompted admission of prior cycle risk while showing zero hedging on forward guidance. He then immediately pivoted to 'we want to get this boom and bust out of it,' revealing awareness that the boom they are printing is the same pattern that preceded the bust.
Friction Level: MODERATE_FRICTION — Both sides agree on the NBM numbers ($93.9B floor, $16.5B guarantees, 17.6% coverage). Bull sees Backlog Fortress with contracted floors at 80% GM. Bear sees thin guarantee coverage released over time on 5-year commitments, with pricing ceilings capping upside on the committed half. Same facts, opposite read on whether the floor holds.
2026Q4 · Aug 5, 2026
3 of 5 readings agreed3 HOLD · 2 LONGRead the full call
84.6% Margin From 26.4% In One Year, Guarantees Cover 17.6% Of Backlog
SNDK converted a quarterly spot-price memory franchise into multi-year contracted revenue at ~80% gross margin with eight customers who have no near-term BiCS8-scale alternative. The structural shift is real and the cash flow confirms it. The risk is that guarantee coverage sits at 17.6% of the $93.9B commitment with releases back-loaded to years 3-5, and management is building inventory against an allocation narrative while consumer segments contract 32% QoQ. Whether the NBM floor holds in year 3 depends on hyperscaler capex decisions SNDK does not control.
When CJ Muse asked about margin durability, David voluntarily referenced 2023: 'In 2023, we were having very different earnings calls than we're having now.' Unprompted admission of prior cycle risk while showing zero hedging on forward guidance. He then immediately pivoted to 'we want to get this boom and bust out of it,' revealing awareness that the boom they are printing is the same pattern that preceded the bust.
Revenue printed $8.97B against guided $7.75-8.25B, beating the high end by $714M (~9%). EPS of $39.25 against $30-33 range beat top by 19%. Two-thirds of the $1.1B sequential jump came from pricing. Buy-side was modeling the guided range.
Two-thirds of sequential revenue growth came from price, not volume. Gross margin expanded 5 consecutive quarters from 22.7% to 84.6%. Q1 guide of 83-85% assumes continued pricing on top of current levels. Eight customers with no near-term BiCS8-scale alternative.
Ten NBMs signed with eight customers totaling $93.9B at floor pricing. RPO of $59.8B growing to $91.1B including post-quarter deals. $16.5B in customer financial guarantees. Four-year weighted average duration with five-year deals. Customers returning to increase volume commitments one quarter later.
The $93.9B NBM figure is floor-pricing backed by only $16.5B in guarantees (17.6% coverage). Management states guarantees are released 'towards the end of the agreement,' meaning years 3-5 carry the thinnest protection. Variable pricing component has a ceiling as well as a floor, capping upside on more than half of FY27 bits.
Management says demand exceeds supply and bits stay on allocation beyond CY2027 while simultaneously raising inventory days and reducing sellable bits to mid-teens for FY27. The inventory build absorbs the same excess supply the allocation narrative claims does not exist.
Gross margin went from 26.4% to 84.6% in four quarters. Zero hedging language across the entire call. Management's own reference to 2023 ('very different earnings calls') acknowledged the previous downcycle produced 26% gross margins within 18 months, yet no downside scenario is presented.
Repurchased $4.5B in one quarter, authorized additional $14B (total remaining authorization $15.5B). CapEx at 6.3% of revenue guiding to ~6% full year while delivering mid-teens bit growth through nodal transitions. Cash flow from operations of $7.1B with $5.0B adjusted FCF.
Mid-teens bit growth from nodal transitions caps supply, and reduced sellable bits to mid-teens for FY27 means SNDK cannot accelerate even if demand surprises to the upside. Management presents this as intentional NBM fulfillment strategy, but it removes optionality.
Entire demand thesis anchored to inference scaling. Management admits 'customers are driving the requirements' and the architecture is 'evolving very quickly,' meaning SNDK's bit-mix forecast is hostage to decisions they do not control.
Beat above high end on revenue, margin, and EPS simultaneously. Q1 guide of $10.3-10.8B represents 15-21% sequential growth. GM guide at 83-85% with management acknowledging 'the high end of the guide is 85%' and characterizing component cost assumptions as prudent.
+85
CONFIDENT"Zero hedging across the entire call. David uses 'thrilled,' 'extraordinarily optimistic,' 'incredible,' 'enormous' without qualification. Luis gives specific margin and revenue numbers without deflection. Every analyst question met with direct, detailed responses."
Management provided specific NBM counts, guarantee amounts, RPO figures, book-to-bill context, end-market revenue breakdowns, and detailed guidance with component cost acknowledgment. Nearly every question answered with numbers rather than narrative.
Investor Day next week where management will detail HBF progress, BiCS10 economics, and potentially new NBM signings. Q1 print of $10.3B+ at or above 83% GM would confirm the contract base holds pricing.
Guarantee coverage of 17.6% ($16.5B against $93.9B) with releases timed to the end of agreements, combined with 50%+ of FY27 bits locked to eight hyperscalers whose capex SNDK does not control. If inference architecture shifts or capex normalizes in year 3, the released guarantees offer no cushion and the non-NBM half faces spot pricing with no floor.
The verdict is public. The case behind it is Pro: the forensic reading of guidance and language, the bull and bear cases, the red flags, and the play. It opens to everyone on Nov 3, 2026.