$42B RPO, $11B Financial Guarantees, $6B Buyback, Q4 EPS Guide $30-33: The Structural Transformation Is Complete — Maintaining Outperform (High Conviction)
Key Takeaways
- Q3 FY26 print obliterates expectations: revenue $5.95B (+97% QoQ / +251% YoY) vs guide $4.40-4.80B; non-GAAP EPS $23.41 vs guide $12-14 (+80% above mid); non-GAAP GM 78.4% (+2,730bp QoQ) vs guide 65-67%; Q3 FCF $2.955B (49.7% margin). Data center +233% QoQ to $1.467B (+645% YoY).
- 5 multiyear New Business Model (NBM) agreements signed (3 in Q3 + 2 in Q4 to date) with $42 billion minimum contractual revenue (RPO) and $11+ billion in financial guarantees. Over 1/3 of FY27 bits under firm customer commitments — Goeckeler targets >50%. This is the structural transformation of NAND from a quarterly-negotiated commodity to a multi-year-contracted strategic input.
- Q4 FY26 guide: revenue $7.75-8.25B (midpoint $8.0B; +35% above Street); non-GAAP GM 79-81%; non-GAAP EPS $30-33 (midpoint $31.50). Annualized EPS run-rate at midpoint: ~$126. The Q4 guide alone exceeds Street's full-year FY26 estimate.
- $6 billion share repurchase authorization announced (immediate, no expiration). Combined with debt-free balance sheet, $1B Nanya DRAM investment, and $3.735B cash, capital allocation is now structured for sustained shareholder return. Plus Yokaiichi JV extended through 2034 + Stargate begins shipping for revenue Q4 FY26.
- Rating: Maintaining Outperform (High conviction). The transformation thesis we established at Q1 FY26 is now complete. Fair value range raised to $1,400-1,800 (~11-15x FY27 EPS of ~$120 annualized). Post-print level ~$1,200; meaningful upside remains but the multi-quarter setup is now compounding execution rather than incremental margin expansion.
Results vs. Consensus
| Metric | Q3 FY26 Actual | Guidance | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $5.95B | $4.40-4.80B | Beat | +$1.30B / +29% above mid |
| Non-GAAP Gross Margin | 78.4% (+2,730bp QoQ) | 65-67% | Beat | +1,200bp above |
| Non-GAAP Operating Margin | 70.9% (+3,340bp QoQ) | — | — | — |
| Non-GAAP EPS | $23.41 | $12-14 | Beat | +$10.41 / +80% above mid |
| Bit shipments | flat YoY; -high-teens QoQ | -mid-single QoQ | Intentional under-ship | Building inventory for Q4 NBM/Stargate ramp |
| Adjusted FCF | $2.955B (49.7% margin) | — | — | — |
| Cash | $3.735B | — | — | Debt-free ($650M TLB paid off) |
| Data Center revenue | $1.467B (+233% QoQ / +645% YoY) | — | — | 25% of revenue (was 14%) |
Segment Performance
| Segment | Revenue | QoQ | Notable |
|---|---|---|---|
| Edge | $3.163B | +118% | Pricing flow-through across PC + premium smartphones |
| Consumer | $820M | -10% | Historical Q3 seasonality (-12-14% norm) |
| Data Center | $1.467B | +233% | TLC enterprise SSD driving; Stargate starts Q4 |
| Total | $5.95B | +97% | +251% YoY; record quarter |
The NBM Disclosure — The Most Important Strategic Change in SanDisk History
The single most consequential disclosure on the call was the New Business Model (NBM) framework — 5 multiyear partnership agreements signed (3 in Q3, 2 in Q4 to date), $42 billion minimum contractual revenue from the 3 Q3 contracts alone (disclosed as RPO in 10-Q), and $11+ billion in financial guarantees across the 5 agreements.
| NBM Element | Disclosure | Implication |
|---|---|---|
| Number of NBMs signed | 5 (3 in Q3 FY26 + 2 in Q4 FY26 to date) | Industry-wide LTA adoption confirmed |
| RPO from 3 Q3 contracts | $42 billion minimum revenue | First time NAND industry has disclosed RPO at this magnitude |
| Financial guarantees across 5 agreements | >$11 billion | Downside protection for SanDisk if customer commitments fail |
| Prepayments on balance sheet | $400M (Q3 BS) | Cash up-front; reduces working capital requirements |
| Longest contract length | 5 years | Multi-year pricing visibility |
| Bit coverage | >1/3 of FY27 bits | Goeckeler target: >50% |
| Pricing structure | Fixed + variable elements | Mutual upside-capture |
| Volume structure | Quarterly commitments; volumes increase over contract life | Structural growth embedded |
The NBM structure fundamentally transforms SanDisk from a quarterly-negotiated commodity NAND vendor into a multi-year-contracted strategic input supplier. The $42B RPO alone is roughly equivalent to 2 years of SanDisk's current revenue run rate — meaning customers have committed to buy roughly $42B of SanDisk product through CY2027-2031 (depending on contract durations). This is unprecedented for the NAND industry.
Key Topics & Management Commentary
Overall Management Tone: The most forward-leaning prepared remarks Goeckeler has delivered. The framing — "a fundamental evolution of our business centered on deeper customer alignment, enhanced visibility, and long-term value creation" — explicitly stakes the position that SanDisk has structurally transformed. The disclosures supporting that framing (NBMs, RPO, financial guarantees, Yokaiichi extension, Nanya investment, $6B buyback) collectively make the case for a permanent re-rating of the franchise.
1. New Business Models (NBMs) — The Structural Transformation
The NBM framework is the most important strategic disclosure in SanDisk's standalone history. 5 signed agreements include $42 billion of RPO from 3 Q3 contracts alone (disclosed for first time as RPO metric), $11+ billion in financial guarantees, and quarterly volume commitments through 5-year horizons. Over 1/3 of FY27 bits already under firm customer commitments.
"These partnerships are structured to lock in committed supply for our customers and committed financials for Sandisk Corporation. Our customers' commitments are backed by firm financial guarantees. These partnerships support durable, structurally higher earnings and a significantly more predictable and less cyclical business for Sandisk Corporation. We believe this marks a fundamental evolution of our business centered on deeper customer alignment, enhanced visibility, and long-term value creation."
— David Goeckeler, CEO
Assessment: NBMs are the structural margin protection mechanism. By contracting both demand and pricing multi-year with financial guarantees, SanDisk has converted the NAND cyclicality bear case into a defensive bull case. Even if spot pricing corrects, the NBM-locked portion of revenue (eventually >50% of bits) sustains earnings power.
2. Data Center +233% QoQ to $1.467B
Data center revenue grew 233% sequentially to $1.467B (was $440M in Q2). YoY growth of +645%. Now 25% of total revenue (vs 14% prior quarter). The growth is driven by TLC enterprise SSD across hyperscaler qualifications and broadening customer base. Stargate (BiCS8 QLC) begins shipping for revenue in Q4 FY26.
"Data center is a clear example of this strategy in action, with revenue growing 233% sequentially. This milestone reflects years of preparation and our deliberate shift toward what is now the most strategic and fastest-growing end market."
— David Goeckeler, CEO
Assessment: Data center at 25% of revenue with +645% YoY growth is the most extraordinary segment performance in NAND industry history at this scale. The Stargate ramp in Q4 + sustained TLC qualifications across additional hyperscalers + AI inference demand (KV cache, RAG, agentic systems) all stack as multi-quarter growth contributors. Data center likely exceeds $2B quarterly run-rate by exit FY26.
3. $6 Billion Share Buyback Authorization
SanDisk's board authorized a $6 billion share repurchase program, effective immediately, with no expiration. Combined with the debt-free balance sheet, $1B Nanya DRAM investment, and $3.735B cash, capital allocation is now structured for sustained shareholder return.
"Today we are announcing that our board of directors has authorized a $6 billion share buyback program of outstanding shares of common stock. The repurchase authorization is effective immediately with no expiration date."
— Luis Visoso, CFO
Assessment: The $6B authorization is sized appropriately — represents ~2-3% of current market cap and ~50% of expected FY26 FCF. The "no expiration" structure gives management flexibility to deploy opportunistically. At post-rally $1,200 stock price, even modest buyback deployment (~$1B over next 12 months) reduces share count by ~1% and adds $1-2 to EPS.
4. NAND Becomes "Only Economically Viable Solution" for AI Inference
Goeckeler's prepared remarks made the strongest structural-demand claim in SanDisk's history: NAND flash is "emerging as the only economically viable solution to deliver the capacity, performance, and efficiency required to keep models accessible for real-time inference at scale." This is the long-term demand thesis embedded in current pricing.
"Inference optimizations such as KV cache, along with workloads like RAG, require substantial high-performance, low-latency flash to deliver real-time responsiveness and quality of user experience. These workloads expand the amount of data that now needs to be stored on low-latency flash well beyond the model itself, as systems must retain context, intermediate data, and large external data sets. As a result, NAND flash is emerging as the only economically viable solution to deliver the capacity, performance, and efficiency required to keep models accessible for real-time inference at scale."
— David Goeckeler, CEO
Assessment: The "only economically viable solution" framing is the strongest structural-demand statement in NAND industry history. If accepted, it implies AI inference demand can absorb essentially unlimited NAND supply additions — supporting the multi-year tightness thesis. The contrarian risk is that newer storage architectures (CXL, in-memory compute) could displace some NAND demand, but the current architecture lock-in is strong.
5. Q4 FY26 Guide $30-33 EPS — Annualizes to $120+
The Q4 guide of $30-33 EPS (midpoint $31.50) on $7.75-8.25B revenue (midpoint $8.0B) implies an annualized EPS run-rate of ~$126. Q4 GM guide 79-81%. The guide assumes continued bit growth + sustained pricing + Stargate revenue contribution beginning.
Assessment: The Q4 guide is the third consecutive quarterly guide blowout (Q2 +94%, Q3 +135%, Q4 +60-70% above Street). The decelerating "above-Street" magnitude is itself a positive signal — Street is finally calibrating to SanDisk's earnings power. FY27 EPS at sustained run rate likely $80-120 range; even with moderating pricing assumptions, FY27 EPS lands $50-70.
6. Yokaiichi JV Extension Through 2034 (Reaffirmed)
Yokaiichi extension through 12/31/2034 + Kitakami JV aligned to same date secures multi-year supply continuity. The $1.165B payment to Kioxia 2026-2029 flows through COGS over 9 years (~$130M/year impact).
Assessment: The Yokaiichi extension is the most important supply-side strategic move. Eliminates the 2027 expiration overhang that was the only material long-term structural risk to the SanDisk franchise.
7. Nanya DRAM Investment ($1B)
SanDisk invested approximately $1B in Nanya to secure long-term DRAM supply. This extends SanDisk's strategic supply chain diversification beyond pure NAND into DRAM — relevant for combined NAND + DRAM enterprise SSD architectures and HBF.
Assessment: The Nanya investment is forward-positioning for HBF and combined NAND/DRAM enterprise products. Modest in scale ($1B vs $3.7B cash position) but strategically important for long-term product roadmap.
8. CapEx Discipline Maintained (Third Consecutive Quarter)
Despite the most aggressive pricing environment in NAND history, management maintained the mid-to-high teens bit growth framework with no CapEx expansion. CapEx ratio at 4% of revenue (vs 16.8% Q1) reflects revenue scaling rather than CapEx cutting.
"Any material increase in capital deployment would require high confidence that demand at attractive pricing levels is durable over a several-year horizon with financial commitments."
— Luis Visoso, CFO
Assessment: Continued CapEx discipline — now with NBM-backed multi-year demand visibility — sets up the framework for potential incremental CapEx in FY28 only if NBMs continue to expand and financial commitments justify expansion. Until then, supply tightness sustains.
9. Inventory Build (Intentional Under-Ship Q3)
Bit shipments down high-teens sequentially in Q3 (intentional under-ship to build inventory for BiCS8 QLC Stargate ramp in Q4 + recently signed NBM ramp). Fiscal YTD bit growth 18% — within the mid-to-high teens long-term framework.
Assessment: Intentional Q3 under-ship is strategically correct. Building inventory ahead of high-margin Stargate launch + NBM ramp captures more value than maximizing Q3 unit volumes at lower margins. The Q4 guide bit growth (mid-single QoQ implied) reflects the inventory release into the higher-priced market.
10. Data Center Exabyte Demand Revised UP Again
Fourth consecutive quarterly upward revision to CY26 data center exabyte demand: mid-20% (4 quarters ago) → mid-40% (3 quarters ago) → high-60% (2 quarters ago) → mid-70% (current). The relentless upward revisions reflect the AI infrastructure CapEx cycle accelerating faster than any modeler had calibrated.
Assessment: The pattern of consecutive upward revisions is itself the structural-tightness validation. Each quarter the data center demand goalposts move higher, supply remains capped at mid-to-high teens bit growth, and pricing has to clear the gap. Until demand growth visibly moderates, the bear thesis cannot crystallize.
Analyst Q&A Highlights
NBM Coverage Trajectory
The dominant topic on the call. A recurring line of questioning probed the trajectory of NBM coverage from current 1/3 of FY27 bits to potential 50%+. Management's framing: aggressive desire to expand coverage; multiple conversations ongoing.
Q: "You said one-third of bit growth next year is contracted. Where do you see this going based on your conversations? It is one-third next year, but can this be above 50% where you know how much is kind of done going into the year?"
— Analyst, Melius Research
A: "We are in active conversations for our supply going forward—that includes next year all the way through the next five years. We said at least a third; we are over a third, and I expect that number to go up over the next several quarters. Where can it get to? I definitely think it can get above 50%, and we have a desire to drive it quite high."
— David Goeckeler, CEO
Assessment: NBM coverage scaling to >50% of FY27 bits represents the structural transformation completion. At >50% NBM coverage, the cyclicality bear case dies — half or more of the business runs on multi-year contracts with financial guarantees, and only the remaining portion is exposed to spot pricing dynamics.
Q4 Pricing Moderation
An exchange on the Q4 guide implications for pricing trajectory revealed the most quantitative forward commentary. Mark Newman noted that the Q4 EPS guide implies slowing rate of price increases.
Q: "The EPS guidance you have given, $30 to $33—these are fantastic numbers. It does imply that the rate of price increase is slowing a bit into the current quarter. I wondered if that is either being conservative on your side because we are still quite early in the quarter, or is that related to some of these very long-term agreements that you signed?"
— Mark Newman, Bernstein Research
A: "On next quarter and pricing, we do not really guide pricing, but I think you saw in FQ3 rather extraordinary pricing acceleration across the business, so we are very happy about that. And you are right, it is early in the quarter and it is an extremely dynamic market, so it pays to be a bit conservative when you are going down that path. But we are very confident in the numbers."
— David Goeckeler, CEO
Assessment: The Q4 guide deceleration in pricing rate of change reflects (a) conservatism on partial-quarter visibility, (b) NBM-locked components within the revenue mix that don't reprice quarterly, and (c) "we are still in the early quarter" framing. The pricing trajectory hasn't necessarily peaked — it's just less visible to model.
Enterprise SSD Mix Evolution
A question on the enterprise SSD opportunity scaling produced the clearest forward framing on data center mix evolution. Goeckeler confirmed data center could be "a significant amount" of the long-term portfolio.
Q: "Where do you see that in a few years? It seems like hyperscalers want everything you can make and then some. Just how much of the business could be enterprise in the long term?"
— Joe Moore, Morgan Stanley
A: "It could be a significant amount. You have known for a long time we value a balanced portfolio. We want to mix in the way that gets the best financial return for us, and that changes quarter over quarter. The key point is we are in a position where we can mix into data center in a way that we have never been able to do before. I expect that number to keep rising over the next several quarters and the next several years."
— David Goeckeler, CEO
Assessment: Data center mix evolution from 25% currently toward 40-50% by exit FY28 supports a structurally higher gross margin profile. Enterprise SSDs carry premium pricing vs. consumer/edge — the mix shift alone supports 50-60% steady-state gross margins.
NAND Industry Supply-Demand Balance
A question on the broader NAND industry supply-demand outlook surfaced the structural framework. SanDisk sees CY26 data center growth revised UP again to mid-70% (from high-60% last quarter) — the fourth consecutive upward revision.
Q: "How are you thinking about when the industry might get into balance? We are getting only limited greenfield, mostly layer count driving bit growth, while new greenfield is being prioritized for DRAM. With that construct and the agentic AI incremental growth, how are you thinking about when the industry might get into balance?"
— C.J. Muse, Cantor Fitzgerald
A: "My point of view is the industry is always in balance—markets always balance supply and demand… We would raise our calendar year '26 data center growth number to the mid-70s from the 60s just three months ago, which is up from the 40s three months before that and the 20s three months before that. Very strong growth in data center."
— David Goeckeler, CEO
Assessment: The four consecutive upward revisions to CY26 data center demand (mid-20% → mid-40% → high-60% → mid-70%) is the strongest demand-side signal. As long as demand continues to revise upward and supply caps at mid-to-high teens, the tightness extends.
Margins Target Range
A question on long-term gross margin targets surfaced management's reluctance to commit to a specific number while the structural transformation is in early innings.
Q: "With regard to margins, when you do these kinds of agreements, can you lock in margins? Is there a target margin range you are comfortable talking about?"
— Analyst, Melius Research
A: "I do not think we are there yet to talk about a target. When we get a little further along, we will wrap this all up in a new model for everybody… We are very proud of our technology. I think for the first time in decades in this business we are getting to the point where the value of our technology is getting recognized by producers."
— David Goeckeler, CEO
Assessment: Management's reluctance to lock in a margin target preserves optionality. Q4 guide at 79-81% GM is unprecedented; even moderating to 50-60% steady-state through cycle would represent a permanent structural step up from the historical 25-30% mid-cycle level.
Market Reaction
- Pre-print setup: Stock ~$1,000-1,100 (rallied from $245 post Q2). YTD CY26 +500%+; trailing 12M from spinoff +3,000%+. Analyst PT range $1,300-2,350.
- After-hours: +10-12% on print + Q4 guide + $42B RPO + $6B buyback.
- May 1 session: Closed approximately $1,200 — +8%. Volume ~5M shares, ~3x average.
- Sector read-across: Memory peers MU +4%, STX +3%, WDC +5%. AI infrastructure broadly higher on $42B RPO disclosure.
The +8% reaction is materially smaller than the +25% (Q1 FY26) and +33% (Q2 FY26) reactions despite an arguably more positive print. Three reasons: (1) expectations bar fully calibrated after two consecutive triple-digit guide beats, (2) stock had rallied 300%+ entering print, (3) Mark Newman pricing-deceleration framing interpreted as approaching cyclical peak. At post-rally $1,200, stock trades ~10x FY27 annualized EPS — still inexpensive in absolute terms but the easy multiple expansion is done.
The investment setup has now evolved from "structural re-rate trade" to "execution compounder." Forward returns depend on (a) sustainability of pricing through FY27, (b) execution on $42B RPO conversion, (c) Stargate ramp delivery, (d) buyback execution, (e) NBM coverage expansion beyond 50%. All five are achievable but require time. Initial positioning: hold existing positions; selectively add on consolidations below $1,000.
Street Perspective
Debate: Can Q4 Guide Sustain Into FY27?
Bull view: The $42B RPO + $11B financial guarantees + multi-year NBM coverage locks in roughly $30+ EPS quarterly run-rate through FY27. Even if spot pricing moderates 20-30%, NBM-locked bits at contracted pricing sustain GM in the 60-70% range. FY27 EPS lands $100-130.
Bear view: Q4 guide represents cyclical peak. Spot pricing corrects 40-50% in CY27 as Korean producers add capacity (Samsung announced 2H26 expansion; SK Hynix considering). NBMs may have variable pricing components that re-set lower. FY27 EPS lands $35-50.
Our take: The structural tightness extends 4-6 quarters beyond Q4 (through end of CY26 / H1 CY27). Korean CapEx response takes 18-24 months minimum to manifest. NBM coverage scaling to >50% protects against the worst of any spot pricing correction. FY27 EPS in the $60-90 range is the realistic outcome.
Debate: Is the $42B RPO Risk-Free?
Bull view: $11B in financial guarantees (prepayments + third-party financial instruments) protect SanDisk if customers fail to perform. The structure is rigorously designed: $400M already in prepayments on Q3 balance sheet. The $42B represents minimum contractual revenue — actual revenue likely higher as volumes increase over contract life.
Bear view: Hyperscaler customers are credit-worthy but could renegotiate if pricing moderates dramatically. The financial guarantees protect SanDisk for principal but not for opportunity cost if SanDisk could have sold to higher-paying spot customers.
Our take: The $42B RPO is highly de-risked. Financial guarantees + hyperscaler counterparty quality + multi-year visibility all support the RPO conversion to actual revenue. The variable-pricing component within NBMs is the design feature that prevents SanDisk from being locked in at uncompetitive pricing if spot rises further.
Debate: Capital Allocation at Peak Valuation
Bull view: $6B buyback at post-rally $1,200 stock price is reasonable given $40-50 FY27 EPS (10x P/E). The deployment over 12-18 months allows opportunistic buying on consolidations. Combined with potential dividend initiation, capital return supports 15-20% IRR.
Bear view: Buybacks at peak-cycle pricing risk being mistimed. Capital should be retained for M&A (consolidate NAND industry) or capacity expansion. The Nanya investment is a better use of capital than buybacks at this stock level.
Our take: $6B buyback is appropriately sized. Deployment over 12-18 months at $1,000-1,400 range is value-accretive. Dividend initiation likely Q4 FY26 or FY27.
Model Update Needed
| Item | Prior Model (Q2 FY26) | Suggested Change | Reason |
|---|---|---|---|
| FY26 Revenue | $15.5B | $21B | Q3 + Q4 guide ahead |
| FY26 EPS | $28 | $66 | Q3 actual + Q4 guide |
| FY27 Revenue | $22B | $30B | NBM-locked revenue + sustained tightness |
| FY27 EPS | $45 | $95 | Margin step-up + NBM coverage scaling |
| FY27 FCF | $9B | $15B | FCF margin 50%+ |
| FY28 EPS | n/a | $60-80 | Moderation but NBM floor supports |
Valuation impact: Fair value range raised to $1,400-1,800 (~15-19x FY27 EPS of $95). Post-rally close $1,200 within reach. Catalysts: Q4 FY26 print Aug 2026; FY26 EPS confirms $60+; NBM coverage expansion announcements.
Thesis Scorecard
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: Structural NAND tightness multi-year | Fully Confirmed | $42B RPO from 3 contracts locks in tightness |
| Bull #2: CapEx discipline preserves pricing | Maintained 3rd quarter | Mid-to-high teens bit growth cap reaffirmed |
| Bull #3: BiCS8 mix accelerates margin | Confirmed | GM 78.4% Q3; 79-81% Q4 guide |
| Bull #4: Data center largest NAND market 2026 | Confirmed | 25% of SNDK revenue Q3; trajectory continues |
| Bull #5: Multi-year LTAs / NBMs locking in pricing | FULLY Confirmed | 5 NBMs signed; $42B RPO; $11B guarantees |
| Bull #6: Yokaiichi extension through 2034 | Confirmed | JV supply continuity secured |
| Bull #7: KV Cache / AI inference demand | Strengthened | "Only economically viable solution" framing |
| Bull #8: Capital return acceleration | NEW — Confirmed | $6B buyback authorization announced |
| Bear #1: NAND cyclical correction | Mitigated | NBM coverage protects against worst case |
| Bear #2: Korean CapEx response | Open | Watch Samsung 2H26 expansion announcements |
| Bear #3: Stock has rallied 3,000%+ — valuation | Open | Still ~10x FY27 EPS at $1,200 |
Overall: Thesis fully materialized. Eight bull points confirmed/strengthened (one new this quarter — capital return); two bears mitigated or pushed out. The structural transformation from cyclical NAND vendor to multi-year-contracted strategic input supplier is complete.
Action: Maintaining Outperform with High conviction. Fair value $1,400-1,800. Post-rally close $1,200. Hold existing positions; selectively add on consolidations.