Sandisk’s NAND upcycle is being repriced too slowly: Q1 beat is pricing power, not just bit recovery

The print changes the debate on Sandisk Corp from “can NAND recover?” to “how much of the pricing cycle has already been captured in estimates?” The Street was set for $2,152.1 million of revenue and $0.88 of EPS, so the reported $2,308.0 million and $1.22 of EPS were not merely cleanup beats; the surprises were +7.2% on revenue and +38.2% on EPS. What was priced in was sequential improvement from the prior trough, because revenue had already rebounded from $1,695.0 million in Q3 FY2025 to $1,901.0 million in Q4 FY2025, and gross margin had recovered from 22.5% to 26.2%. What was not priced in was that Q1 FY2026 revenue would step to $2,308.0 million with gross margin at 29.8%, while the company’s own next-quarter framework calls for $2,550 million to $2,650 million of revenue and non-GAAP EPS of $3 to $3.40. The variant perception is that the stock should not be analyzed as a low-quality memory beta off the last reported gross margin; the relevant earnings power is already being pulled forward by price, mix, and cost normalization that are visible in management’s guide.

That distinction matters because the quarter’s upside was not dependent on one accounting line or a one-off reconciliation item. CFO Luis Visoso put the company’s own reporting basis plainly: “Revenue for the first quarter was $2,308 million, up 21% quarter-over-quarter and up 23% year-over-year.” The Street comparison basis shows the same direction with $2,308.0 million of actual revenue versus $2,152.1 million estimated, and the historical series shows the inflection was real after a downcycle that took EPS to -$13.33 in Q3 FY2025 and -$0.16 in Q4 FY2025. The company’s reported diluted EPS in the history table is $0.75 for Q1 FY2026, while the print’s Street-comparison EPS is $1.22; those are different reporting bases, so the investment point is not to reconcile them into one number, but to recognize that both show the same turn from losses to positive earnings. The market expected a return to profitability; it did not expect profitability to arrive with enough force to produce a +38.2% EPS surprise.

The revenue trajectory explains why the margin discussion should be framed around supply tightness rather than simple volume absorption. Revenue moved from $1,901.0 million in Q4 FY2025 to $2,308.0 million in Q1 FY2026, a +21.4% QoQ increase, while gross margin moved from 26.2% to 29.8%. That gross margin was not yet at the prior Q1 FY2025 level of 38.6%, which is precisely why the print is still actionable: the recovery has crossed back into earnings, but the margin base has not fully normalized even before Q2 pricing. Visoso’s Q2 guide is the crux because it explicitly attributes the next revenue step to “double-digit price increases and mid-single-digit bit growth.” That wording is more useful than generic demand language: it says the next leg is price-led, not just units, and it arrives immediately in Q2 FY2026 with revenue expected between $2,550 million and $2,650 million.

The operating leverage is also cleaner than the headline gross margin suggests, because Q1 still carried costs that management says are fading. In Q1, Sandisk incurred $61 million in start-up costs and $11 million in underutilization charges, and non-GAAP operating expenses were $446 million versus guidance of $415 million to $430 million. That OpEx miss is the main bear argument in the print, but it is outweighed by the pricing setup because the Q2 non-GAAP operating expense guide is $450 million to $475 million against revenue of $2,550 million to $2,650 million and EPS of $3 to $3.40. The cost bridge matters more than the absolute expense number: Visoso said start-up costs went from “$60 million last quarter to call it, $30 million this quarter to pretty much 0 going forward.” The phrase “pretty much 0” deserves attention because it is not a long-cycle ambition; it is a near-term cost removal claim tied to the same Q2 frame as double-digit price increases.

Cash flow reinforces the view that this is not just a P&L recovery being financed by working capital. Management reported $448 million in adjusted free cash flow, equal to 19.4% free cash flow margin, and closed with $1,442 million in cash and cash equivalents against $1,351 million in gross debt. CEO David V. Goeckeler’s framing is worth quoting because it commits the balance sheet point alongside the cash generation: “We generated $448 million in adjusted free cash flow and closed the quarter in a net cash position of $91 million.” That net cash position matters for a memory name because cycles usually punish companies that must invest into supply constraints with levered balance sheets. Here, the company also paid an additional $500 million of its TLB and reduced inventory days from 135 to 115 as demand exceeded supply. The market may focus on the 29.8% gross margin as still below prior peaks; the better signal is that inventory days fell while revenue beat by +7.2%, suggesting pricing and allocation are doing more work than discounting.

The product and end-market mix make the pricing call more credible, because the strength was not confined to a consumer replenishment channel that could fade after one quarter. Edge revenue was $1,387 million and up 26% sequentially, Consumer revenue was $652 million and up 11% quarter-over-quarter, and data center revenue was $269 million and up 26% sequentially. The second-order read is that Edge and data center are moving at the same sequential rate, while Consumer is positive but slower, which supports the company’s claim that enterprise and infrastructure demand are influencing supply allocation. BiCS8 adds another layer to that mix argument: Goeckeler said “BiCS8, which delivers industry-leading capacity, I/O performance and energy efficiency, accounted for 15% of total bits shipped and is expected to reach majority of bit production exiting fiscal year '26.” The word “majority” is not quantified in the data pack, but the 15% starting point is; the investment significance is that the current quarter’s gross margin of 29.8% is being earned before BiCS8 becomes the dominant bit base exiting fiscal year '26.

That mix shift also gives a sharper read-through to named partners than a generic “NAND is tight” conclusion. Kioxia is both a customer of Sandisk through JV NAND fabs at Yokkaichi and Kitakami and a supplier to Sandisk through the same JV NAND fabs. Sandisk’s inventory days falling from 135 to 115 while demand exceeded supply implies the JV ecosystem is operating into a tighter allocation environment, and that matters for Kioxia because the peer table shows Kioxia with revenue of ¥1,002,852.0 million, gross margin of 64.1%, and revenue YoY of +188.9%. The magnitude is not subtle: Sandisk’s own latest peer-quarter revenue is $5,950.0 million with gross margin of 78.4% and revenue YoY of +251.0%, while Kioxia’s reported gross margin is 64.1% and revenue YoY is +188.9%. For PMs long or short NAND supply-chain exposure, Sandisk’s double-digit price increase guide is a positive read-through to Kioxia’s JV economics, but the comparative margin table also says Sandisk is capturing the cycle at a higher reported gross margin than Kioxia in the latest peer set.

The peer context is important because it prevents over-crediting Sandisk for what is plainly an industry upcycle, while still showing why Sandisk’s setup is distinct. MU reported revenue of $41,456.0 million, gross margin of 84.6%, and revenue YoY of +345.7%, which is above Sandisk’s latest peer-quarter $5,950.0 million revenue, 78.4% gross margin, and +251.0% revenue YoY. WDC, by contrast, reported $3,337.0 million of revenue, 50.2% gross margin, and +45.5% revenue YoY, while STX reported $3,112.0 million of revenue, 46.5% gross margin, and +44.1% revenue YoY. The comparative point is that Sandisk is not the fastest or highest-margin memory print in the table, but it is much closer to the high-end memory recovery cohort than to the storage hardware cohort. That matters for valuation debate: if investors anchor Sandisk to WDC and STX because of storage history, they miss that the reported peer-quarter gross margin and revenue YoY sit far nearer the memory winners.

The call delivery supports the same thesis, with one caveat: management sounded more confident on guidance, but uncertainty also rose in the transcript scoring. The tone history shows sentiment improved to 0.44 in Q3 FY2026 from 0.34 in Q2 FY2026, and guidance_tone rose to 0.51 from 0.23, with tone_confidence up to 0.61 from 0.42. That is the constructive side. The caution is that uncertainty increased to 50.0 from 42.2, and qa_evasiveness moved to 2.5 from -11.4. The conflict is interpretable rather than disqualifying: prepared_sentiment actually eased to 0.70 from 0.76, while qa_sentiment stayed at 0.19, so the improvement in the tone series is not from promotional prepared remarks alone. It is more likely the model is capturing a call where management had better near-term numbers but still faced more complex questions about cycle durability and supply discipline.

That tone nuance is why the stock reaction should not be judged only by whether Q1 gross margin exceeded guidance. Visoso said Q1 gross margin “compares favorably to our guidance of 28.5% to 29.5%,” but the real debate is whether Q2’s EPS guide of $3 to $3.40 can be sustained beyond a pricing reset. The factors that argue yes are concrete: Q2 revenue of $2,550 million to $2,650 million is tied to double-digit price increases and mid-single-digit bit growth; Q2 non-GAAP operating expenses are expected at $450 million to $475 million; start-up costs are described as going to “pretty much 0 going forward”; and inventory days already fell from 135 to 115. The factors that argue no are also concrete: Q1 non-GAAP operating expenses of $446 million were above guidance of $415 million to $430 million, Q1 included $61 million of start-up costs and $11 million of underutilization charges, and uncertainty in the tone table rose to 50.0. The investment call is that the positive factors are nearer-term and numerically larger in EPS sensitivity than the negatives shown in the data pack.

The balance sheet and capital spending details also complicate the usual memory-cycle skepticism. Gross capital expenditures totaled $387 million and represented 16.8% of revenue, while the Flash Ventures-related cash flow included $337 million in gross CapEx, $107 million funded through depreciation as part of cost of goods sold, and $240 million funded from external sources, mainly subsidies and equipment leasing. Those numbers matter because they show Sandisk is participating in capacity and technology investment without the full cash burden flowing through its own free cash flow in the quarter. The company still generated $488 million cash from operations and received $10 million cash from activities related to Flash Ventures, offset by $50 million invested in back-end operation and offices. This is not a claim that supply can never catch up; it is a claim that in this quarter, the company delivered $448 million of adjusted free cash flow while demand exceeded supply and inventory days fell to 115.

The next confirmation point is therefore narrow and investable: Q2 FY2026 needs to show that price-led guidance converts into the EPS range management laid out. Watch the next quarter against revenue of $2,550 million to $2,650 million, non-GAAP operating expenses of $450 million to $475 million, non-GAAP interest and other expense of $40 million to $45 million, non-GAAP tax expense of $80 million to $90 million, and non-GAAP EPS of $3 to $3.40 on 155 million fully diluted shares. The thesis is confirmed if double-digit price increases and mid-single-digit bit growth deliver that revenue range while start-up costs move from the Q1 $61 million burden toward the stated “pretty much 0 going forward.” It breaks if OpEx again exceeds its guide after the Q1 $446 million versus $415 million to $430 million miss, or if inventory days reverse from 115 despite management’s statement that demand exceeded supply. For now, the print says the Street priced a NAND recovery, while Sandisk reported a pricing cycle with cash flow, a net cash position of $91 million, and a Q2 EPS guide that resets the earnings base faster than estimates had allowed.