MU · Forward model · Contracted case
The Contracted case, 20 quarters out
Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.
Quarters are fiscal, the way Micron labels them: 2026 Q3 is the quarter ended 28 May 2026. Micron reports four business units and this model carries exactly those four — Cloud Memory, Core Data Center, Mobile and Client, Automotive and Embedded — with no sub-splits invented underneath them. Every revenue point is reported, so nothing is marked estimated. They do not quite add up, and that is Micron's arithmetic rather than this model's: the four units sum to $41,448M against a reported $41,456M in the basis quarter, an $8M residual Micron does not assign to any unit, and the same $1M-$8M gap appears in each earlier quarter. What IS assumed, and heavily, is the split of each unit's revenue into bits and price. Micron publishes no bit volumes, so the petabyte figures are revenue divided by an assumed price per gigabyte: the LEVEL is calibrated and the claim is the SHAPE. That shape changed on 18 August 2026. The model used to revert prices from the first projected quarter, which put it ~30% below the street for fiscal 2027 and below Micron's own next-quarter guide. Two disclosures in the 2026 Q3 10-Q say that is wrong. First, the strategic customer agreements are TAKE-OR-PAY with price bands: 'The largest agreements generally have a ceiling price for existing products that approximates the market price in the second calendar quarter of 2026, and a floor price through the term of the agreement,' and Micron expects gross margins at those FLOOR prices to be 'well above our peak quarterly margins in any past cycle.' That is a floor under the downside AND a ceiling on the upside, and the model now carries both. Second, customers are prepaying to secure supply: $22B of cash deposits and related financial commitments on agreements concluded to date, about $18B of it cash. Price now starts strongly positive and glides negative rather than falling from day one, so revenue peaks in fiscal 2028 and declines after, as the fabs being built now land. Fiscal 2027 lands at about $233B against a street range of roughly $225-240B.
Micron's own framing, taken at face value: take-or-pay agreements with floor prices make the trough shallower than any prior cycle, and Micron says margins at floor pricing beat the peak of every past cycle. What it still does not reach on a 3.5x multiple is today's share price — getting there needs the multiple to re-rate to 6x as well, which is the real content of the argument the market is having about Micron.
Latest: $81.50B (2031Q3E)
| Period | Value |
|---|---|
| 2025Q3 | $9.30B |
| 2025Q4 | $11.31B |
| 2026Q1 | $13.64B |
| 2026Q2 | $23.86B |
| 2026Q3 | $41.45B |
| 2026Q4E | $48.76B |
| 2027Q1E | $54.87B |
| 2027Q2E | $59.84B |
| 2027Q3E | $63.88B |
| 2027Q4E | $67.20B |
| 2028Q1E | $70.00B |
| 2028Q2E | $72.10B |
| 2028Q3E | $73.34B |
| 2028Q4E | $74.44B |
| 2029Q1E | $75.47B |
| 2029Q2E | $76.46B |
| 2029Q3E | $77.44B |
| 2029Q4E | $78.44B |
| 2030Q1E | $79.45B |
| 2030Q2E | $80.49B |
| 2030Q3E | $81.56B |
| 2030Q4E | $82.67B |
| 2031Q1E | $82.20B |
| 2031Q2E | $81.81B |
| 2031Q3E | $81.50B |
What drives each segment
Cloud Memory
Capacity × utilisation × priceHBM and high-capacity DRAM sold to the hyperscalers and the AI accelerator makers. The line the whole AI thesis rests on, and the only one here where supply, not price, is the binding constraint — HBM4 is in high-volume shipment and the capacity is contracted ahead.
Latest: $40.21B (2031Q3E)
| Period | Value |
|---|---|
| 2025Q3 | $3.39B |
| 2025Q4 | $4.54B |
| 2026Q1 | $5.28B |
| 2026Q2 | $7.75B |
| 2026Q3 | $13.77B |
| 2026Q4E | $16.51B |
| 2027Q1E | $18.91B |
| 2027Q2E | $21.00B |
| 2027Q3E | $22.82B |
| 2027Q4E | $24.41B |
| 2028Q1E | $25.83B |
| 2028Q2E | $27.13B |
| 2028Q3E | $28.34B |
| 2028Q4E | $29.48B |
| 2029Q1E | $30.56B |
| 2029Q2E | $31.61B |
| 2029Q3E | $32.63B |
| 2029Q4E | $33.63B |
| 2030Q1E | $34.61B |
| 2030Q2E | $35.57B |
| 2030Q3E | $36.52B |
| 2030Q4E | $37.46B |
| 2031Q1E | $38.38B |
| 2031Q2E | $39.30B |
| 2031Q3E | $40.21B |
Assumptions & reasoning
- Modelled on capacity rather than growth because that is Micron's actual constraint: HBM is sold out ahead of production, so what ships is what the fabs and the packaging lines can make. The 10-Q says it plainly — demand is growing 'at a rate greater than our ability and the industry's ability to increase supply'.
- The petabyte figure is NOT disclosed. It is $13,769M divided by an assumed $6.04 per gigabyte, roughly a 2x premium to commodity DRAM in the same quarter. Micron gives bit-growth commentary in prepared remarks but no volumes in the filings, so treat the level as a calibration and argue with the price instead.
- Price rises 9% a quarter and glides to -4%, rather than falling from the first quarter as this model used to assume. Industry reporting through August 2026 has 2027 DRAM and HBM capacity at Micron, Samsung and SK Hynix already sold out under multi-year contracts, and Meritz put supplier fulfilment at only 75-80% of demand in the second half of 2026, possibly falling into the 60s during 2027. New capacity is scheduled but late: SK Hynix's first Yongin cleanroom opens in February 2027, its second fab not until 2029, and Samsung's first Yongin fab starts up in 2029.
- The upside is capped as well as floored, and by the same disclosure. Micron's largest agreements carry a ceiling price approximating the CALENDAR Q2 2026 market price, which is roughly where price already is. So the contracted book cannot ride a further spike — only the uncontracted minority and mix can. That is why the drift decays fast rather than compounding at 9% for years.
- Bit supply still nearly triples over the horizon while price ends below where it starts, so revenue grows about 80% and Cloud Memory goes from a third of Micron to about 43% of it. That mix shift is the strongest claim in this model and it survives every scenario.
- Utilisation glides DOWN from 95% to 88%, which reads oddly for a sold-out product. It is the cycle: capacity brought on for a peak runs slack when the peak passes, and the last three memory downturns were all made worse by fabs that could not idle cheaply.
- The reversion is timed to that build schedule rather than to a view about prices. Additional wafer starts from the second half of 2027 through 2028 are what turn the cycle, and the same forecasters who call 2027 the tightest year on record put conventional DRAM into excess capacity from about 2029 — which is where this model has revenue rolling over. HBM itself is expected to stay tighter for longer, which is why Cloud Memory keeps the gentlest price decline of the four lines.
Core Data Center
Units × priceConventional server DRAM and enterprise storage — the memory that goes into everything in the rack that is not an accelerator. Revenue went from $1.5B to $11.5B in four quarters on almost no volume growth, which tells you exactly what it is: price.
Latest: $21.07B (2031Q3E)
| Period | Value |
|---|---|
| 2025Q3 | $1.53B |
| 2025Q4 | $1.58B |
| 2026Q1 | $2.38B |
| 2026Q2 | $5.69B |
| 2026Q3 | $11.52B |
| 2026Q4E | $13.76B |
| 2027Q1E | $15.60B |
| 2027Q2E | $17.06B |
| 2027Q3E | $18.20B |
| 2027Q4E | $19.11B |
| 2028Q1E | $19.84B |
| 2028Q2E | $20.44B |
| 2028Q3E | $20.97B |
| 2028Q4E | $21.43B |
| 2029Q1E | $21.86B |
| 2029Q2E | $22.27B |
| 2029Q3E | $22.67B |
| 2029Q4E | $23.05B |
| 2030Q1E | $23.44B |
| 2030Q2E | $23.82B |
| 2030Q3E | $24.21B |
| 2030Q4E | $24.60B |
| 2031Q1E | $23.37B |
| 2031Q2E | $22.19B |
| 2031Q3E | $21.07B |
Assumptions & reasoning
- This is the line that reverts hardest, and it is still the line that grew from $1.5B to $11.5B in four quarters without a new product cycle. Gross margin went from 38% to 87% over that stretch, which is the definition of a price spike, and the price of a commodity that has spiked has exactly one long-run direction.
- Bits times price, with price up 10.5% a quarter at first and gliding to -6%. Revenue peaks in fiscal 2028 around $17-18B a quarter and then declines. The old model had this line falling from the first projected quarter, which put the whole company ~30% below the street for fiscal 2027 — that was the error, and it was an artefact of a price drift that could not change sign.
- The volume ceiling is 9,800 PB a quarter, a little over three times today's implied output, and the line reaches it near the end of the horizon. Wafer capacity Micron redirects into HBM cannot also serve this line — the two compete for the same fabs, and this is the one that loses when HBM is short.
- The first projected quarter is $48.0B of company revenue against Micron's own $50.0B guide — still about 4% light. A quarterly step this large is mostly contract resets landing on a specific date, and a smooth glide cannot reproduce a step. The gap is stated rather than tuned away.
- Terminal margin is 50%, raised from 38%, because 38% sat below the floor Micron describes: gross margins on the banded agreements are expected to be 'well above our peak quarterly margins in any past cycle' EVEN AT FLOOR PRICING, and the pre-AI peak was about 61%. Only part of the book is contracted, which is why this does not go higher.
Mobile and Client
Units × priceLPDDR for phones and DRAM and NAND for PCs. Structurally the weakest position Micron holds — the customers are price-sensitive, the volumes are set by handset and PC cycles that are not growing, and the current margin is entirely borrowed from the shortage.
Latest: $8.46B (2031Q3E)
| Period | Value |
|---|---|
| 2025Q3 | $3.25B |
| 2025Q4 | $3.76B |
| 2026Q1 | $4.25B |
| 2026Q2 | $7.71B |
| 2026Q3 | $11.52B |
| 2026Q4E | $13.44B |
| 2027Q1E | $14.90B |
| 2027Q2E | $15.96B |
| 2027Q3E | $16.68B |
| 2027Q4E | $17.16B |
| 2028Q1E | $17.46B |
| 2028Q2E | $17.32B |
| 2028Q3E | $16.51B |
| 2028Q4E | $15.69B |
| 2029Q1E | $14.88B |
| 2029Q2E | $14.09B |
| 2029Q3E | $13.33B |
| 2029Q4E | $12.60B |
| 2030Q1E | $11.91B |
| 2030Q2E | $11.25B |
| 2030Q3E | $10.63B |
| 2030Q4E | $10.04B |
| 2031Q1E | $9.48B |
| 2031Q2E | $8.95B |
| 2031Q3E | $8.46B |
Assumptions & reasoning
- Gross margin here was 24% a year ago and 87% in the basis quarter, on a product line whose end markets — phones and PCs — grew by neither 24% nor 87%. Every point of that came from price, and this model gives most of it back: revenue ends the horizon about 40% below its 2027 peak.
- Volume grows 5.5% a quarter until it hits the 6,000 PB ceiling around the middle of the horizon, then stops. Handset and PC unit volumes are flat, so all the growth is gigabytes per device, and this line gets what is left of the fab after the data centre is served.
- This is where the China risk lands. CXMT now holds roughly 7% of global DRAM revenue and is expanding into exactly the mainstream DRAM the incumbents vacated to chase HBM; Apple has been testing its parts for devices sold in China. Mainstream mobile and client DRAM is the most substitutable thing Micron sells, which is why this line carries the fastest price decline and the lowest terminal margin of the four.
- There is a feedback loop worth naming: NVIDIA's own commentary blames slower consumer PC demand on elevated memory prices. Micron's peak price is suppressing the volume of the very market this line serves, which is the mechanism by which memory cycles turn.
Automotive and Embedded
Growth pathMemory designed into cars, industrial equipment and robots. Small, slow and by far the most stable line Micron has — design wins last for years and prices move on contracts rather than on spot.
Latest: $11.76B (2031Q3E)
| Period | Value |
|---|---|
| 2025Q3 | $1.13B |
| 2025Q4 | $1.43B |
| 2026Q1 | $1.72B |
| 2026Q2 | $2.71B |
| 2026Q3 | $4.63B |
| 2026Q4E | $5.05B |
| 2027Q1E | $5.45B |
| 2027Q2E | $5.83B |
| 2027Q3E | $6.19B |
| 2027Q4E | $6.53B |
| 2028Q1E | $6.87B |
| 2028Q2E | $7.20B |
| 2028Q3E | $7.52B |
| 2028Q4E | $7.84B |
| 2029Q1E | $8.17B |
| 2029Q2E | $8.49B |
| 2029Q3E | $8.82B |
| 2029Q4E | $9.16B |
| 2030Q1E | $9.50B |
| 2030Q2E | $9.85B |
| 2030Q3E | $10.21B |
| 2030Q4E | $10.58B |
| 2031Q1E | $10.96B |
| 2031Q2E | $11.36B |
| 2031Q3E | $11.76B |
Assumptions & reasoning
- The one line carried on a plain growth rate, and the only one where that is the honest choice: automotive memory is sold on multi-year design wins at contracted prices, so there is no meaningful bit-versus-price tension to model.
- It has still quadrupled in four quarters, which means it is not as insulated from the spot market as the design-win story suggests. Terminal growth of 1% a quarter says the model does not believe that rate survives.
- At 11% of revenue this line cannot decide the valuation. It is carried in full because it is a disclosed business unit, and because it is the only part of Micron that would still be growing in a downcycle.
- Margins here glide down least — 79% to 45% — because contract pricing falls slower than spot in both directions. That is also why this unit's margin was the lowest of the four on the way up.
Where each case comes from
Bear case — primary sources
The primary sources this case is built from — filings, calls and posts. Where they are claims by an interested party rather than disclosures, every number in the Bear column is what happens if they are taken at face value.
What turns the cycle: new fabs, and China
- Aug 16, 2026 Memory shortage could turn into a glut by 2028, says Kwon Seok-joon... conventional DRAM could face excess capacity beginning in 2029. China, he said, is the largest uncertainty.
- Aug 17, 2026 CXMT sits on the other side of this trade. Running around 300,000 wafers a month, close to MU's total DRAM capacity, CXMT still channels almost all its silicon into conventional bits, swallowing the market space vacated by the top three.
Bull case — primary sources
The primary sources this case is built from — filings, calls and posts. Where they are claims by an interested party rather than disclosures, every number in the Bull column is what happens if they are taken at face value.
2027 capacity is reported sold out across the industry
- Aug 18, 2026 All 2027 DRAM & HBM capacity at Samsung, $SKHY & $MU sold out... Samsung says shortage persists through to 2028 and gets tighter in 2027.
- Aug 11, 2026 Meaningful new industry capacity is not expected until late 2027 or (more likely) 2028. Supply is meeting only ~60-70% of requested volumes in many cases for 2027.
Contracted case — primary sources
The primary sources this case is built from — filings, calls and posts. Where they are claims by an interested party rather than disclosures, every number in the Contracted column is what happens if they are taken at face value.
The agreements, from the 2026 Q3 10-Q
- Jun 25, 2026 Strategic customer agreements are structured as take-or-pay agreements, with binding commitments for specific volumes over the multi-year contract terms.
- Jun 25, 2026 The largest agreements generally have a ceiling price for existing products that approximates the market price in the second calendar quarter of 2026, and a floor price through the term of the agreement.
- Jun 25, 2026 We expect gross margins from our strategic customer agreements with price bands, even at floor pricing levels, to yield gross margins well above our peak quarterly margins in any past cycle.
Customers are prepaying for supply
- Jun 25, 2026 we expect to receive cash deposits and related financial commitments of $22 billion for agreements concluded to date. Approximately $18 billion of these commitments will be in the form of cash deposits.
- Jun 25, 2026 Robust overall DRAM and NAND demand and constrained supply has led to increased pricing and improved the profitability across our portfolio.
From cash flow to fair value
| Present value of free cash flow, 20 quarters | $494.98B |
| Terminal-year revenue | $328.18B |
| Terminal-year EBITDA | $235.71B |
| Exit multiple, on revenue | 6.0x |
| Terminal value | $1.97T |
| Discounted at 10.0% a year, terminal value becomes | $1.22T |
| Enterprise value | $1.72T |
| Net cash | $24.00B |
| Equity value | $1.74T |
| Diluted shares | 1.15B |
| Fair value per share | $1,521.07 |
| Against the current price of $974.33 | +56% |
3.5x terminal revenue, on a terminal year with roughly a 55% EBITDA margin — about 6x EBITDA, which is where memory trades on mid-cycle earnings. The discount rate is 11% rather than NVIDIA's 10% because this is a commodity producer with a cycle, not a platform, and each case discounts at its own rate: 14% in Bear, 9.5% in Bull, 10% in Contracted. The shape the model now carries is a peak, not a plateau — revenue rises through fiscal 2027, peaks in fiscal 2028 around $257B, and declines from there as the fabs being built now come online. That is what the sources support: 2027 capacity sold out, and the new supply landing 2028-2029. Move the slider: at 2x the fair value is about $522 and at 6x about $1,014 — so even a re-rating to 6x on this operating case only just reaches today's price. The honest warning is unchanged — the exit multiple and the price drift together decide almost the whole answer, and both are opinions about when the cycle turns, not facts about it.
Read the other way round: at $974.33 the market is paying 2.9x terminal-year revenue, holding every other assumption on this page fixed. That is the number to argue about.
The projected path
| Quarter | Cloud Memory | Core Data Center | Mobile and Client | Automotive and Embedded | Revenue | YoY | EBITDA | Capex | FCF | R40 | PV of FCF |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 Q4E | $16.51B | $13.76B | $13.44B | $5.05B | $48.76B | +331% | $46.18B | $10.28B | $30.51B | +394 | $29.80B |
| 2027 Q1E | $18.91B | $15.60B | $14.90B | $5.45B | $54.87B | +302% | $50.04B | $12.10B | $32.25B | +361 | $30.75B |
| 2027 Q2E | $21.00B | $17.06B | $15.96B | $5.83B | $59.84B | +151% | $52.75B | $13.71B | $33.18B | +206 | $30.89B |
| 2027 Q3E | $22.82B | $18.20B | $16.68B | $6.19B | $63.88B | +54% | $54.61B | $15.13B | $33.55B | +107 | $30.50B |
| 2027 Q4E | $24.41B | $19.11B | $17.16B | $6.53B | $67.20B | +38% | $55.88B | $16.38B | $33.58B | +88 | $29.80B |
| 2028 Q1E | $25.83B | $19.84B | $17.46B | $6.87B | $70.00B | +28% | $56.79B | $17.50B | $33.40B | +75 | $28.95B |
| 2028 Q2E | $27.13B | $20.44B | $17.32B | $7.20B | $72.10B | +20% | $57.26B | $18.43B | $33.01B | +66 | $27.93B |
| 2028 Q3E | $28.34B | $20.97B | $16.51B | $7.52B | $73.34B | +15% | $57.22B | $19.14B | $32.37B | +59 | $26.75B |
| 2028 Q4E | $29.48B | $21.43B | $15.69B | $7.84B | $74.44B | +11% | $57.20B | $19.78B | $31.81B | +53 | $25.67B |
| 2029 Q1E | $30.56B | $21.86B | $14.88B | $8.17B | $75.47B | +8% | $57.24B | $20.38B | $31.33B | +49 | $24.69B |
| 2029 Q2E | $31.61B | $22.27B | $14.09B | $8.49B | $76.46B | +6% | $57.34B | $20.94B | $30.94B | +47 | $23.81B |
| 2029 Q3E | $32.63B | $22.67B | $13.33B | $8.82B | $77.44B | +6% | $57.52B | $21.47B | $30.64B | +45 | $23.02B |
| 2029 Q4E | $33.63B | $23.05B | $12.60B | $9.16B | $78.44B | +5% | $57.78B | $21.98B | $30.43B | +44 | $22.32B |
| 2030 Q1E | $34.61B | $23.44B | $11.91B | $9.50B | $79.45B | +5% | $58.12B | $22.48B | $30.29B | +43 | $21.70B |
| 2030 Q2E | $35.57B | $23.82B | $11.25B | $9.85B | $80.49B | +5% | $58.53B | $22.97B | $30.23B | +43 | $21.14B |
| 2030 Q3E | $36.52B | $24.21B | $10.63B | $10.21B | $81.56B | +5% | $59.02B | $23.45B | $30.23B | +42 | $20.65B |
| 2030 Q4E | $37.46B | $24.60B | $10.04B | $10.58B | $82.67B | +5% | $59.57B | $23.92B | $30.30B | +42 | $20.21B |
| 2031 Q1E | $38.38B | $23.37B | $9.48B | $10.96B | $82.20B | +3% | $59.08B | $23.94B | $29.87B | +40 | $19.45B |
| 2031 Q2E | $39.30B | $22.19B | $8.95B | $11.36B | $81.81B | +2% | $58.68B | $23.95B | $29.52B | +38 | $18.77B |
| 2031 Q3E | $40.21B | $21.07B | $8.46B | $11.76B | $81.50B | +0% | $58.38B | $23.98B | $29.25B | +36 | $18.16B |
Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.
Model revisions
Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.
| Date | Changed | Fair value then | Note |
|---|---|---|---|
| 2026-08-17 | all | $542.18 | First cut, built on the FQ3-26 press release. Four verticals matching Micron's four business units; bits and price split out under each. |
| 2026-08-18 | coreDataCenter.unitsCeiling, coreDataCenter.unitsGrowthQoQ, mobileClient.unitsGrowthQoQ, scenarios.discountRate, scenarios.terminalGrowthDelta | $542.18 | Core Data Center ceiling raised to 9,800 PB and bit growth to 7% a quarter; Mobile and Client volume growth to 5.5%. Each scenario now discounts at its own rate and carries its own terminal growth delta. |
| 2026-08-18 | priceDriftQoQ, unitPriceDrift, terminalPriceDrift, priceDriftDecay, terminalEbitdaMargin, capexIntensity, autoEmbedded.growthQoQ, scenarios | $706.33 | Price drift now glides instead of holding constant, so the model can carry a spike that is still running: prices rise into 2027, revenue peaks in fiscal 2028, and the reversion follows. Driven by the 2026 Q3 10-Q take-or-pay disclosure (ceiling near calendar Q2 2026 market price, floor through the term, floor-price gross margins above any past cycle peak) and by industry reporting that 2027 capacity is sold out. Terminal margins raised toward that floor: Cloud Memory 62%, Core Data Center 50%, Mobile and Client 40%. Capex intensity raised to match the ~$27B of 2026 capex the 10-Q guides to. Fiscal 2027 now lands at ~$233B against a street range of $225-240B, where the old model sat ~30% low. |