Bits were flat. Revenue tripled. The entire print is price.
Thesis: Revenue grew 251% year-over-year while bit shipments were flat, meaning 100% of growth came from ASP. The $42B RPO across five NBMs with over a third of FY2027 bits committed is genuinely different from prior NAND cycles and gives real forward visibility. But only $400M sits in prepayments against an $11B headline guarantee, variable clauses let customers capture downside, and 78-81% gross margin in NAND has no precedent. The contracts are real; the margin durability is unproven. Goeckeler explicitly refused to set a margin floor when asked.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Q4 RPO disclosure grows as two more NBMs fold in, plus Stargate QLC revenue begins. If RPO tracks toward 50% of bits with named segment mix, the structural thesis gains proof. Q4 guidance of $7.75-8.25B at 79-81% GM tests whether pricing holds at these levels.
Key Risk: Bit shipments were flat YoY and down high-teens sequentially while inventory days rose during peak pricing. If hyperscaler CapEx normalizes, variable-pricing clauses trigger downside, margin reverts toward historical 40-60% range, and the $42B RPO becomes a floor at minimum prices rather than a margin guarantee.
The Tell: When Newman asked directly whether the stock prices margin reversion to the 40s, Goeckeler deflected: '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.' He volunteered the reversion framing himself but refused to commit to any floor, which means management cannot or will not guarantee where margins settle.
Friction Level: MODERATE_FRICTION — Both sides agree the quarter is clean and the RPO is real. Disagreement is whether 78-81% gross margin reflects a structural repricing of NAND or a cycle peak that variable-pricing NBM clauses will unwind. Goeckeler refuses to commit to a margin floor.
2026Q3 · Apr 30, 2026
3 of 5 readings agreed3 HOLD · 2 LONGRead the full call
Revenue Up 251% On Zero Volume Growth
Revenue grew 251% year-over-year while bit shipments were flat, meaning 100% of growth came from ASP. The $42B RPO across five NBMs with over a third of FY2027 bits committed is genuinely different from prior NAND cycles and gives real forward visibility. But only $400M sits in prepayments against an $11B headline guarantee, variable clauses let customers capture downside, and 78-81% gross margin in NAND has no precedent. The contracts are real; the margin durability is unproven. Goeckeler explicitly refused to set a margin floor when asked.
When Newman asked directly whether the stock prices margin reversion to the 40s, Goeckeler deflected: '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.' He volunteered the reversion framing himself but refused to commit to any floor, which means management cannot or will not guarantee where margins settle.
Revenue beat high end of guide by 24% ($5.95B vs $4.8B max), GM beat guide by 11.4pts (78.4% vs 67%), EPS beat high end by 77%. Q4 guide above this quarter across all lines: $7.75-8.25B at 79-81% GM, $30-33 EPS.
Revenue up 251% YoY on flat bit shipments. Gross margin expanded 27.3pts sequentially. Mix shift to higher-value customers plus pricing environment drove the entire print. No volume growth required.
$42B minimum contractual revenue from three contracts, $11B in financial guarantees across five agreements, over a third of FY2027 bits committed. Management states normal case is customers asking to increase volume within weeks of signing.
Bit shipments down high-teens sequentially while inventory days rose during a quarter when customers are supposedly desperate for supply. Management attributes buildup to Stargate ramp and NBM prep, but inventory rising at peak pricing is a cycle signature.
$42B RPO is minimum contractual revenue with variable pricing components and undisclosed customer concentration. Only $400M in actual prepayments on balance sheet. Remaining $10.6B sits in unnamed third-party instruments. Enforcement untested.
Management calls 78-81% GM 'sustainable' while acknowledging flat bits and admitting 'next conversions will be a little more expensive on a dollar basis.' Sector has never sustained above 60% through a full cycle. Goeckeler refused to commit to a margin floor.
+75
CONFIDENT"Direct answers, specific numbers ($42B RPO, $11B guarantees, $400M prepayments), forward commitments on buyback and NBM pipeline. Adjectives run hot: extraordinary, spectacular, phenomenal. Single qualifier: 'it pays to be a bit conservative' on Q4 pricing."
Specific financials, RPO disclosure, and guarantee details are precise. Customer names, segment mix behind $42B, counterparty custodians, and margin floor all withheld. Balance of transparency and strategic omission.
Q4 RPO disclosure grows as two more NBMs fold in, plus Stargate QLC revenue begins. If RPO tracks toward 50% of bits with named segment mix, the structural thesis gains proof. Q4 guidance of $7.75-8.25B at 79-81% GM tests whether pricing holds at these levels.
Bit shipments were flat YoY and down high-teens sequentially while inventory days rose during peak pricing. If hyperscaler CapEx normalizes, variable-pricing clauses trigger downside, margin reverts toward historical 40-60% range, and the $42B RPO becomes a floor at minimum prices rather than a margin guarantee.
“For the third quarter, we expect revenue between $4.4 billion and $4.8 billion.”
“Revenue for the third quarter was $5.95 billion, up 97% sequentially and up 251% year-over-year.”
Companies tend to guide low: the median company we cover reports +1.7% against its guidance midpoint, so a small beat is the norm. A transcript can drop part of a guided range, so read both quotes.
“For the fourth quarter, we forecast revenue between $7.75 billion and $8.25 billion from both bits growth and higher pricing.”
Changed less than 52% of calls
Wording compared with the Jan 29, 2026 call, ranked among 97 calls reporting on Q1 2026. A measure of how much the story moved, not a signal about the stock.
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Language change is computed from word frequencies, with no model involved. Descriptive only: none of this feeds the verdict.
Beat-and-raise at extreme scale. Revenue $5.95B vs $4.8B high end, GM 78.4% vs 67% guide, EPS $23.41 vs $14 max. Q4 guide above this print: $7.75-8.25B, 79-81% GM, $30-33 EPS. $42B RPO from three contracts, $11B guarantees, over a third of FY2027 bits locked. Data center +233% sequential to $1.467B, calendar 2026 growth forecast raised to mid-70s from 60s three months ago. FCF margin 49.7% on $240M CapEx (4% of revenue). BiCS 8 drives mid-to-high-teens bit growth without proportional spend. $6B buyback with no expiration, net cash position. Five NBMs with customers asking to increase volume within weeks of signing. Goeckeler directly addresses the margin-reversion narrative: 'We are not interested in trading away that value for certainty.' Street models this as cyclical memory at peak multiples. The contracts say otherwise.
Revenue up 251% YoY on zero volume growth. Bits flat YoY, down high-teens sequentially. 100% of growth is price. Inventory days rose during peak pricing against a narrative of desperate scarcity. 78.4% GM, guided to 79-81%, is the highest in NAND history with no precedent for durability. CapEx starved at 4% of revenue to manufacture heroic FCF; deferred BiCS 8 conversions hit when demand normalizes. Only $400M in prepayments against $11B headline guarantee. Remaining $10.6B in unnamed third-party instruments, custodians undisclosed. Variable pricing clauses explicitly let customers capture downside if prices fall. Data center $1.47B vs edge $3.16B: edge units declining, 2027 bounce-back is hope not contract. RPO has undisclosed customer concentration. Goeckeler refuses to commit to any margin floor. Stock priced for perfection on peak pricing with commodity underneath.
Unsustainable Trend Confidence
78-81% gross margin in NAND has no historical precedent for full-cycle durability. Management calls it sustainable while admitting flat bits, rising inventory, and more expensive upcoming conversions. Goeckeler explicitly declined to state a margin floor.
Backlog Quality
$42B RPO is minimum contractual revenue with variable pricing that lets customers capture downside. Only $400M in actual prepayments. Remaining $10.6B in unnamed third-party instruments with undisclosed custodians. Segment mix and customer concentration withheld when asked directly.
Inventory Divergence
Inventory days rose during a quarter when bits were down high-teens sequentially and pricing was parabolic. Management blames Stargate ramp and NBM prep. Building inventory at peak pricing is historically a cycle-top signature regardless of stated reason.
High Utilization as Ceiling
Running full capacity with inventory building leaves zero buffer against demand shock. CapEx at 4% of revenue means no greenfield supply buffer. If hyperscaler CapEx normalizes, there is no lever but price to manage utilization.
Hold. The print is clean: no channel stuffing, clean non-GAAP bridge, operating cash flow confirms earnings. $42B RPO is contractual minimums, not vapor. But 100% of growth from price with flat volume and rising inventory at record margins is a cycle signature, and Goeckeler will not commit to a margin floor. The asymmetry favors waiting for bits to grow alongside price. If Q4 shows sequential bit growth with GM holding 79%+, the structural thesis is proven and you buy. If bits stay flat and inventory builds again, this is a peak print. Entry price for the long is higher RPO disclosure with segment mix. Entry price for the short is bits growing while price holds being tested, likely FY2027.