Deep Dive · Companies, Equities, AI infrastructure
Micron Deep Dive: How Much of the Boom Is Permanent?
Blockhaven's full 43-page initiating research on Micron, published November 2, 2026: 2/5 Cautious at the October 30, 2026 close. A public preview of the rating, thesis, selected sections, risks, methodology and disclosures, quoted verbatim from the report.
- By
- Blockhaven Research
- Publication status
- Length
- 43-page report (preview below)
Contents
Blockhaven rates Micron (MU) 2/5 Cautious at the October 30, 2026 close. The thesis: "Structurally stronger, untested by a downturn." The valuation figures are conditional scenarios, not price targets. The sections below are quoted verbatim from the full Deep Dive.
Published November 2, 2026. Price and valuation date October 30, 2026; research cutoff October 31, 2026; fundamentals through fiscal 2026. This is a dated research record: Micron's first-quarter fiscal 2027 results, expected in mid-to-late December 2026, are not reflected.
What the Deep Dive is
The Deep Dive is Blockhaven's full research on Micron: what it sells, how memory makes money, the contracts, the investment programme, three valuation regimes, an AI-structural alternative and what would change our view. It is written for a reader who wants the evidence, the calculations and the counterarguments, not only the conclusion.
Rating, thesis and key metrics
How much of the boom is permanent?
The next year looks strong on every assumption we make. The shares already pay for a spread between memory prices and costs that the old cycle never sustained.
Coverage
- Where earnings are: fiscal 2026 revenue of $133.2bn; first-quarter fiscal 2027 guided at $61.5bn.
- What the shares cost: $1.001 trillion at $884.74, 11.8x fiscal 2026 net income.
- The deciding variable: the spread, 87.1% in fiscal 2026 on our calculation, against an old-cycle peak of 73.0%.
- Where our reference path matches the price: only at a spread of about 78.6% for good from fiscal 2031, at 12%.
- Values at 12%: downside $201, reference $662 and upside $1,820 a share; conditions, not forecasts.
Key metrics
The thesis, and why the stock is here
Investment thesis
Micron has just reported the most profitable year in its history, and the shares price a good part of it as permanent. Fiscal 2026 revenue was $133.2bn, up from $37.4bn, with a fourth-quarter GAAP gross margin of 86.8%. At $884.74 the company is worth $1.001 trillion, which is 11.8x its fiscal 2026 net income and 71x its net income in fiscal 2018, the peak of the previous cycle. Neither multiple settles anything on its own. The low one flatters a peak, and the high one ignores that the business has genuinely changed.
Our work comes down to a single variable: the spread between what a bit of memory sells for and what it costs to make, which we measure as gross margin before depreciation. That is a Blockhaven calculation, not a figure Micron reports: we add back an assumed 90% of depreciation and amortisation to the reported gross margin, because Micron does not say how much of its depreciation sits in cost of goods sold. On that measure the last full cycle averaged 58.0% and peaked at 73.0%, and fiscal 2026 reached 87.1%; Micron's reported GAAP gross margin for fiscal 2026 was 80.7%. Holding the rest of our reference path fixed, at a 12% discount rate our model matches the price only if that spread settles at about 78.6% for good once the new capacity of 2028 to 2030 arrives, 5.6 points above the best year Micron had before this boom. That reference path, in which the old peak becomes the new normal, is worth $662 a share. It is a judgment rather than a measurement, and it already credits the business outside its contracts with a through-cycle spread well above its old average. Put the other way round, the reference matches the price at a discount rate of about 8.6%, against the 12% we require.
There is real evidence for a structural shift: 26 multi-year take-or-pay agreements that management estimates at over 35% of revenue through 2030, $134bn of remaining performance obligations measured at minimum prices, and a product mix now weighted about 60% to data center. The pricing floors in those agreements are undisclosed, and the agreements end around 2030 and 2031. We think the price asks for more durability than the evidence can yet show, so the rating is 2 of 5, Cautious. The judgment turns on the one variable Micron does not disclose, the contract floors and the spread they protect after 2030; the research itself rests on verified filings, and the next year looks strong on every assumption we make.
What is known
Evidence the report measures
- Gross margin before depreciation was 87.1% in fiscal 2026 on our calculation, against 35.8% to 73.0% in the last full cycle.
- 26 strategic customer agreements; remaining performance obligations of $134bn at minimum prices, about a quarter expected within twelve months.
- $12.7bn of customer deposits received in fiscal 2026, repayable 2029 to 2031, and $9.8bn of noncurrent income taxes payable.
- Fiscal 2027 capital spending guided above $50bn; a majority of the increase over earlier plans is construction.
What is inferred
Blockhaven's reading of that evidence
- That supply stays tight through calendar 2028, as management expects, and eases during 2029 as new cleanrooms produce.
- That the spread settles at the old peak of 73.0%: our judgment, not something the contracts show.
- That floor-and-ceiling bands cover about 13% to 26% of revenue, read from management's words.
What remains unproven
What the evidence does not yet show
- The contract floors, which Micron does not disclose, and management's expectation of margins at the floors.
- Whether customers have committed to more memory than they need.
- Whether demand keeps pace with the industry's capacity from late 2028.
- Whether a rising share of HBM lifts margins; in the fourth quarter it did not.
Why the stock is here
Micron's shares closed 2024 at $84.16 and 2025 at $285.41, and reached their highest month-end close, $1,154.29, at the end of June 2026. They fell to $823.03 a month later, recovered to $1,065.11 at the end of September, and closed at $884.74 on 10-30-2026, 23% below that month-end high and about 11x their level at the end of 2024.
| Month-end | Close |
|---|---|
| December 2024 | $84.16 |
| June 2025 | $123.25 |
| December 2025 | $285.41 |
| March 2026 | $337.84 |
| June 2026 | $1,154.29 |
| July 2026 | $823.03 |
| September 2026 | $1,065.11 |
Month-end closes from Yahoo Finance's public monthly history, retrieved 10-22-2026; the price we use is the 10-30-2026 close, checked against Nasdaq.
The rise followed the earnings. Revenue went from $25.1bn in fiscal 2024 to $133.2bn in fiscal 2026, and at the dated price the company is valued at 11.8x its fiscal 2026 net income, a low multiple that shows the market already treats much of today's profit as cyclical. The swings since June show how quickly the market reprices the expected length of the boom: the debate is no longer whether fiscal 2027 will be strong, but how long the spread lasts after it.
A sample of the analysis
The assessment in four parts
A single rating can blur four different questions together, so we answer them separately.
| Question | Our assessment | What it means |
|---|---|---|
| How reliable is the evidence? | High for the reported record; limited for the decisive future variable | Historical figures come from Micron's filings and releases and were reconciled across sources; the model's arithmetic reproduces. What the evidence cannot show is the contract floors and the post-2030 spread |
| How uncertain is the value? | Wide | The upside regime is worth 9.1x the downside, and one assumption, the spread path including its long-run level, explains 78% of that gap |
| What does the rating rest on? | One contested judgment | Where the spread settles after the industry's next capacity wave, which the evidence brackets without picking a centre. The rating section sets out the strongest competing reading and the evidence or price that would change our view |
| What is the near-term business outlook? | Constructive | Guidance, HBM agreements for calendar 2027 and management's supply outlook all point to a very strong fiscal 2027 |
A highly reliable record can still leave the value very uncertain, and a strong near-term outlook can coexist with weak evidence about profitability after 2030. None of the four is a forecast of the share price.
Principal risks and what would change the view
Risks
The obvious risk is the cycle itself, and the less obvious one is that the contracts could make it sharper rather than softer: if customers have over-committed to secure supply, the inventory overhang arrives all at once. Concentration is high, with approximately one-half of revenue from the top ten customers in each of the last three years. Micron's ramp involves new fabs on three continents, and a delay or yield problem would raise costs exactly when prices are falling. Government support is conditional: CHIPS Act grants of up to $6.4bn and a 35% investment tax credit come with spending and operating conditions, and the grant agreements restrict buybacks in their later years by a free-cash-flow test. The board raised the buyback authorisation to $35.16bn, usable from 12-09-2026; an authorisation permits purchases and does not mean any have been made. Finally, the Singapore tax incentives, worth a net $2.21bn to fiscal 2026 after the minimum-tax top-up, expire in stages through 2034.
What would make us upgrade, downgrade or reconsider
A trigger starts a review; it does not decide its outcome. We distinguish evidence that would revise the economics from changes that move the rating without changing the business, and we say plainly when evidence leaves the question open.
| Kind | What we would watch | Where it would show | What it would do |
|---|---|---|---|
| Fundamental revision | Disclosed floor prices or floor margins, or agreements covering more of fiscal 2028 to 2031 revenue | Micron's filings and earnings remarks | Strengthen the structural case; could lift the reference spread and the rating toward 3 / 5 |
| Fundamental revision | Prices holding above the reference path as new supply lands in 2029 | Micron's gross margin by quarter; industry contract prices | Strengthen; the reference timing or level would be revisited |
| Fundamental revision | Competitors deferring or accelerating cleanroom and tool schedules | SK hynix and Samsung results and capital plans | Deferral strengthens the reference timing; acceleration weakens it |
| Fundamental revision | Customer inventories building beyond production needs, or buyers cutting memory content as costs rise | Customers' reports and Micron's commentary | Weaken; would bring the turn forward |
| Fundamental revision | Capital spending or depreciation materially above the model | Micron's guidance and cash-flow statements | Weaken the free cash flow path at a given spread |
| Rating trigger | A material change in the share price | The market | A share price around $662, where our reference matches the price at the hurdle rate, would move the rating toward 3 / 5 with no change in the business; a materially higher price without new evidence, toward 1 / 5 |
| Rating trigger | A change in the cost of capital we judge appropriate | Interest rates and Micron's risk | Moves every regime's value and the rating with it |
| Unresolved | Results in line with guidance, with no new contract disclosure | Quarterly results | Confirms the near-term view and leaves the long-term question where it is |
We do not promise to monitor these continuously. A later edition will keep this one intact and attribute every change to its cause: reported results, management guidance, our own judgment, a model correction or the price.
Catalysts: what must happen
Mid-to-late December 2026, date not announced — Micron's first-quarter fiscal 2027 results and second-quarter guidance
- What is known:
- Micron guides $61.5bn of revenue for the 13-week quarter at a GAAP gross margin of 85.95%.
- Why it matters:
- it is the first quarter of the year every regime treats as exceptional, and the first chance to see new agreements or contract terms disclosed.
- What would support the upside:
- results above guidance, and disclosed floors or volumes for new agreements.
- What would leave the question open:
- results in line with guidance and no new contract terms, which confirms the near-term view and says nothing new about the long term.
- What would support the downside:
- a miss against guidance, or guidance that points to price increases ending sooner than management expected.
Late January 2027, dates not announced — SK hynix's and Samsung's fourth-quarter 2026 results
- What is known:
- both reported record third quarters in October 2026; SK hynix expects its 2027 investment to increase significantly, and neither gave a price guide for the fourth quarter.
- Why it matters:
- they are the other two suppliers adding cleanroom space against the same demand, so their capacity plans and pricing bear directly on when the spread turns.
- What would support the upside:
- capacity plans for 2027 and 2028 described as disciplined, and memory prices guided flat or up.
- What would leave the question open:
- no capacity or pricing guidance.
- What would support the downside:
- cleanroom or capacity expansion for 2027 and 2028 beyond the plans described in October, HBM price concessions, or memory price declines reported or guided.
Through 2027 — Micron's first new capacity ramps
- What is known:
- HBM packaging in Singapore is targeted for early 2027, the first Idaho fab and the Taiwan Tongluo site for mid-2027, and the 1-delta DRAM and G11 NAND nodes for the second half.
- Why it matters:
- gross margins through the ramps show whether depreciation arrives ahead of profitable output.
- What would support the upside:
- margins that hold as the new capacity arrives.
- What would support the downside:
- delays, yield problems or margins falling as depreciation rises.
Late 2028 to 2029 — The industry's new cleanrooms produce
- What is known:
- Micron targets its second Idaho fab and the Japan expansion for late 2028 and Singapore NAND for the second half of 2028, and expects meaningful output a few quarters after first output; competitors are building too.
- Why it matters:
- this is when the reference path lets the spread ease, so it is the period that tests the assumption the value rests on.
- What would support the upside:
- prices that hold above the reference path through fiscal 2029.
- What would support the downside:
- a spread falling faster than the reference path.
Blockhaven view
What supports it, and what weighs against it
- For: fiscal 2026 gross margin before depreciation of 87.1% on our calculation, 26 take-or-pay agreements and $134bn of obligations at minimum prices.
- For: every regime models fiscal 2027 free cash flow of $126.2bn or more, and the balance sheet adds $40 a share after the obligations it carries.
- Against: our reference matches the 10-30-2026 close only at a spread of 78.6% for good, against an old-cycle peak of 73.0% and an average of 58.0%.
- Against: the contract floors are undisclosed, the agreements end around 2030 and 2031, and three suppliers are adding capacity from late 2028.
The Blockhaven Rating is our long-term assessment of the risk and reward at the stated price, on Blockhaven's own five-level scale. It is not a share-price forecast or individualized advice, and it can change when the price changes.
How the research was done
Methodology and limitations
- Reported figures are from Micron's fiscal 2026 results release and prepared remarks of 09-30-2026, its fiscal 2026 Form 10-K filed 10-09-2026, its technology webinar presentation of 10-21-2026, earlier 10-K filings through SEC XBRL company facts, Samsung Electronics' guidance of 10-08-2026 and the U.S. Treasury daily par yield curve; prices are Nasdaq and Yahoo Finance closes. The third-quarter reports of SK hynix, Samsung, Sandisk and Amazon of 10-28-2026 and 10-29-2026 inform our judgments and are not model inputs. The sources appendix gives each document's address and the pages we cite.
- The question-and-answer part of the 09-30-2026 call was read from a published third-party transcript, not checked against the recording, and we did not hear the 10-21-2026 webinar. The SK hynix and Samsung calls and the Sandisk call were read from third-party transcripts and a third-party summary, labelled as such where we cite them.
- Industry figures from TrendForce and the press are third-party estimates and reports, labelled as such; the capacity articles were published before our cutoff, retrieved on 10-23-2026 and are context, not model inputs.
- Morningstar's estimates are cited from its publicly indexed note summaries; Blockhaven has not seen its model.
- Management's statements about contracts, pricing and capacity are reported as management's; the regimes, the spread paths and the valuation are Blockhaven's analysis.
- The model is a driver-based discounted cash flow over fiscal 2027 to fiscal 2033 with a terminal value; revenue is written as bits times the industry's cash cost per bit over one minus the spread, so the three move together by construction.
- The industry's cost per bit changing by -7% a year, the same in every regime, is a Blockhaven assumption, not a company forecast. The AI-structural alternative is a sensitivity of the reference on assumptions we chose, with product weights and spreads that are not measured; it is not a scenario, a forecast or a probability. The cost and HBM reports it cites were published before our cutoff and retrieved on 10-23-2026.
- Customer deposits and the noncurrent tax payable are deducted once, at face value, in the equity bridge, and kept out of free cash flow.
- The three regimes are conditions, not forecasts, and carry no probabilities by our choice of method; model figures are conditional on the stated assumptions.
- Gross margin before depreciation adds an assumed 90% of depreciation and amortisation back to gross margin; Micron does not report the split.
- The return-on-capital cross-check shares the near-term cash flows, the capital base, the discount rate and the bridge with the main model; only its long-run construction differs.
As of
What the full report contains
34 sections, 26 data tables and 7 figures in the full report.
- Investment thesis
- The assessment in four parts
- The question
- Why the stock is here
- What Micron sells
- How memory makes money
- The memory industry and how prices form
- What AI changes, and what it does not
- Customers, contracts and what they protect
- Financial history: what changed in fiscal 2026
- Fiscal 2026 in cash, and the part that is owed
- The investment programme and the supply response
- Capital allocation and the balance sheet
- The next year and the long run
- Valuation
- The model, year by year
- Terminal economics: what the steady state assumes
- From the business to the share
- The range, the reference and the price
- An AI-structural alternative
- What moves the value
- The bull case and the bear case
- What the latest results from peers and customers show
- The rating and the alternatives we considered
- What would make us upgrade, downgrade or reconsider
- Catalysts: what must happen
- Decision rules
- Risks
- Blockhaven view
- Appendix: the model in figures
- As of
- Methodology and limitations
- Appendix: sources
- Disclosure
Disclosures
Disclosure
Blockhaven ™ research is for information and is not individual investment advice. The Blockhaven Rating is a research judgment at the stated price; it is not a recommendation to buy, hold or sell.
Important disclosures
Blockhaven research is provided for general informational and educational purposes. It is not individualized investment advice and does not take into account any reader's financial circumstances, investment objectives, risk tolerance, tax situation or portfolio.
Nothing in this report is an offer or a solicitation to buy or sell a security or other financial instrument. Reading or accessing Blockhaven research does not by itself create an investment-advisory, fiduciary, brokerage or client relationship.
Investing involves risk, including the possible loss of principal. Scenario values, terminal values, valuation ranges, sensitivities and forward-looking assumptions are analytical tools, not guarantees or promises of future performance, and actual results may differ materially.
Blockhaven uses information from company filings, company publications, market-data providers and other sources it believes to be reliable, but that information may be incomplete, delayed, revised or incorrect.
Opinions, ratings and interpretations reflect Blockhaven's research judgment as of the report date and may change as new evidence becomes available.
Positions and conflicts — as of publication, November 2, 2026
At the time of publication, Blockhaven's owner held no MU shares, options, short positions, swaps, or other MU derivatives; reported no other material MU interests through controlled entities, family accounts, or related parties; and had no advisory, banking, consulting, or sponsorship relationship with Micron. The owner had received no payment or compensation from Micron.
These disclosures reflect circumstances at publication and do not represent future holdings.
© 2026 Blockhaven LLC. All rights reserved.
What a Blockhaven Rating means, and what it is not: Research disclosures.
Sources
- 1.
Micron Technology, Inc.
Micron Technology, Inc. Form 10-K for the fiscal year ended September 3, 2026 (filed October 9, 2026)(opens in a new tab)Accessed November 1, 2026
- 2.
Micron Technology, Inc.
Micron fiscal fourth-quarter 2026 prepared remarks (September 30, 2026)(opens in a new tab)Accessed November 1, 2026
Corrections and updates
No corrections or updates have been made to this piece.
Blockhaven research is published for information and education. It is not personalized investment advice.