Micron Technology ($MU) — Living Research
Stronger Commitments. A Lower Share Price. The Earnings Floor Still Matters.
Current Hayabusa Read
Our October 3 question still holds: has AI changed Micron’s earnings floor, or extended an unusually profitable memory cycle? The evidence since then gives us more reason to take the structural case seriously. It does not give us permission to treat today’s margins as permanent.
We would revise the thesis incrementally, not reverse it. The annual filing makes the protections behind customer commitments more tangible, and the technology roadmap adds detail to the opportunity beyond high-bandwidth memory, or HBM. We are modestly more constructive on business durability. We remain conditional on the stock because normalized earnings and a defensible entry valuation are still unresolved.
There is also a more attractive starting price for that work. Micron closed October 23 at $877.55, versus $1,074.89 on October 2, the last trading session before our prior update—a decline of about 18.4%, calculated from closing prices and excluding dividends.[9] A lower price improves prospective returns if future cash earnings hold up. It cannot tell us whether they will.
Our stance: keep Micron on the research watchlist and advance the normalized-earnings work. Stronger contracts support the business case; they do not establish that the shares are cheap. This update does not set a buy range, price target or portfolio-specific sizing instruction.
What Changed Since October 3
The most useful new document is the October 9 Form 10-K. Its balances describe the fiscal year ended September 3; they are newly reviewed disclosure, not business booked after October 3. The October 21 technology presentation is newer roadmap evidence. We separate both from the September 30 financial results, which remain our latest reported quarterly baseline.
Contracts Now Have More Visible Protection
The filing reports about $134 billion of remaining performance obligations, roughly one-quarter expected within twelve months and the balance within five years. These use committed volumes and minimum pricing; variable-price and short-duration contracts are excluded. They are not a forecast of total revenue.[6]
Micron received $12.75 billion in customer deposits during fiscal 2026, recorded as financing cash inflows. Nearly all are scheduled for repayment in 2029–2031. A separate $7 billion of issued or committed letters of credit provides conditional protection upon customer default or breach.[6]
Our interpretation: these arrangements make customer commitments more credible and can help fund expansion. But deposits are not earned revenue or free cash flow. Letters of credit are not cash already collected. Adding either to contracted revenue would turn a useful protection into an accounting magic trick.
We need to see the contracts work when customers have less urgency to secure supply. Minimum prices, enforceable volumes and recovery rights matter; so do product costs, renegotiations and customer credit quality. The filing improves visibility into the protection. It does not reveal a reliable company-wide trough EPS.
The AI Memory Story Extends Beyond HBM
Micron’s October 21 presentation outlined higher-capacity server memory, DDR6 and new NAND architectures. Its targets include a 1TB server-memory module and second-half-2027 ramps for 1-delta DRAM and G11 NAND. The DRAM roadmap calls for a mid-20s percentage increase in bit density.[7] These are management roadmap targets, not evidence that every product is qualified or generating revenue today.
The investment implication runs both ways. More memory and storage around each AI workload can broaden Micron’s opportunity. More bits per wafer can improve its costs—but industry-wide improvements also expand supply. A better manufacturing process does not come with an exemption from the memory cycle.
TrendForce’s October 22 research forecasts more than 40% growth in server CPU shipments in 2027 as inference and AI agents expand.[8] We read that as support for a broader server-memory opportunity, not a direct forecast for Micron’s revenue. Memory content, customer platform choices and Micron’s share still determine the dollars it captures.
Capital Allocation Adds a New Checkpoint
On October 8, the board raised total buyback authorization to $35.16 billion, with repurchases under the revised program from December 9, subject to CHIPS funding conditions. Authorization is discretionary, not completed buying. The filing also projects approximately $25 billion of net capex in the first half of fiscal 2027, with more in the second half.[6]
Our test is capital discipline: what remains for shareholders after productive investment, obligations and future deposit repayments? We will measure actual repurchases and the diluted share count rather than assume the authorization automatically lifts per-share value.
Micron’s October 20 investment in Indium Corporation’s New York operations supports domestic critical-materials processing.[10] It is relevant to supply-chain resilience. It does not, by itself, justify changing earnings estimates.
Why Micron, Why Living Research?
The Micron thesis cannot be settled by one quarter. The company sits at the intersection of several forces moving on different clocks: AI demand can accelerate quickly, memory pricing can turn quickly, customer contracts last for years, and new fabrication capacity can take years to become productive.
That makes Micron ideal for Living Research. Readers should be able to return to one page and see whether the evidence is moving toward old Micron—a familiar boom-and-bust memory cycle—or new Micron—a business with structurally higher memory intensity, more contracted demand and a higher normalized earnings floor.
What We’re Watching
These are proposed research review triggers, not automatic trading rules. Quarterly baselines remain fiscal Q4 2026 and fiscal Q1 2027 guidance. New filing evidence is dated October 9; roadmap evidence is dated October 21.
| Research Question | Starting Evidence | Next Test And Action |
|---|---|---|
| Are pricing and margins durable? | Fiscal Q4 non-GAAP gross margin reached 87%. Micron guided fiscal Q1 2027 non-GAAP gross margin to approximately 86.25%, even as revenue guidance rose to $61.5bn.[2] | At each earnings report, compare DRAM and NAND pricing commentary, gross margin and mix with the prior quarter. Moderating price increases are not automatically bearish; a sustained pricing reversal accompanied by faster margin compression would require a normalized-earnings reset. |
| Are customer contracts changing the cycle? | The October 9 annual filing adds quantified obligations, deposit terms and contingent credit protection. See the contract discussion above.[6] The prior agreement count is historical, not a newly verified October total. | At each filing, reconcile contract balances with delivered revenue, deposits, repayments and credit support. Look for weakened terms, nonperformance or renegotiation. Do not infer a fixed company-wide earnings floor from contracted revenue alone. |
| Is AI broadening memory demand beyond HBM? | The October 21 roadmap spans server DRAM, HBM and NAND; the October 22 server outlook supports broader demand but remains a third-party forecast.[7][8] | Seek customer qualifications, volume shipments, mix and realized margins. Raise confidence when broader adoption becomes profitable sales; reduce it if launches slip or customers achieve the same workloads with less memory content. |
| When does new supply actually arrive? | The original capex baseline now sits alongside a higher forward investment plan and process-density improvements. Construction spending and technology targets remain distinct from saleable output.[6][7] | Track Micron and competitors: qualified wafer starts, yields, bits per wafer and customer shipments. Capacity arriving faster than demand is the warning, even if an individual factory or process is a technical success. |
| What can Micron earn after scarcity fades? | Current quarterly earnings and guidance still describe unusually favorable economics. The October share-price decline changes the entry calculation, not the durability of the earnings denominator.[2][9] | Before a buy call, build downside, base and upside cases for normalized pricing, costs, mix, capex, taxes and diluted shares. Compare current price with that range. No current consensus or investor-expectations comparison has been established in this review. |
| Does capital allocation improve per-share returns? | The new buyback authorization adds a dated monitoring item; it does not establish purchases or value creation.[6] | From December 9, review disclosed purchases, average prices, diluted shares and funding conditions. Compare buybacks with capital needs and future obligations. Repurchasing at poor prices can destroy value despite reducing share count. |
What Would Change Our View?
More constructive: contracts convert into paid deliveries without weaker terms; AI demand broadens into profitable server-memory and storage sales; manufacturing execution reduces costs; and free cash flow stays resilient as supply expands. The strongest confirmation would be a materially higher earnings floor during an actual industry slowdown.
More cautious: pricing falls while productive supply accelerates, customers reduce memory content or seek concessions, product ramps disappoint, or capital spending outruns durable cash generation. Contract protection deserves less weight if enforcement, credit quality or repayment obligations prove more costly than expected.
There is already a reason to resist a one-directional demand story. TrendForce’s September 30 forecast—context available before our last update, not new October evidence—described strong AI demand alongside weak consumer demand and buyers reducing storage capacity to manage costs.[11] High prices can change customer behavior before new factories reach volume production.
The strongest opposing view remains that AI has lengthened a memory upcycle without eliminating its eventual reversal. The new evidence improves the structural case, but it has not yet defeated that explanation.
What’s Actionable Now?
For new money: wait for valuation proof. A lower share price and better contract visibility justify renewed underwriting. They do not replace it. Do not annualize one exceptional quarter and label the resulting multiple a margin of safety.
Model impact: keep the reported quarterly baseline and existing guidance separate from a revised long-term case. The new disclosures justify testing more resilient downside revenue and cash-flow scenarios; they do not yet support a numerical increase to normalized EPS or a higher valuation multiple. No earnings forecast or price target is being raised in this update.
For existing shareholders: compare the new contract, supply and capital-allocation evidence with the assumptions behind the position. This review does not establish a portfolio-specific add, trim or hedge instruction. A weakening contract cushion alongside falling prices would require a fresh downside assessment before additional exposure.
Next Checkpoint
Fiscal Q1 2027 results and filing: compare realized revenue, margins, pricing, contract execution and cash conversion with the September 30 guidance. We did not verify an official reporting date in the accessible investor-relations material as of October 23; third-party estimates should not be treated as a confirmed company date.[2][4]
December 9, 2026: review the buyback program’s permitted commencement and subsequent purchase disclosures. This is a capital-allocation checkpoint, not a promised market purchase or stock-price catalyst.[6]
Second half of calendar 2027: test the announced manufacturing roadmap against customer qualification, production yields and commercial shipments.[7] Earlier material changes in contract terms, prices or supply warrant an interim review.
Research Archive
Read the earlier research to see how our view developed. The archived edition preserves the analysis as of its stated date; this page carries our current assessment.
October 3, 2026 — Initial Living Research
Our starting assessment of whether AI memory demand, customer commitments and constrained supply could raise Micron’s earnings floor after scarcity pricing fades.
September 30, 2026 — The 7× P/E Wasn’t a Typo
The original earnings re-underwrite, preserved as a separate dated article.
Research Timeline
September 30, 2026 — Event Re-underwrite
Starting article: Micron (MU): The 7× P/E Wasn’t a Typo.
Initial view: Micron’s earnings power is extraordinary, but the apparent low multiple is not enough to establish cheapness. The core question shifted toward whether AI memory intensity, strategic customer agreements and slow supply growth can raise normalized earnings after the shortage ends.
Evidence added: $54.23bn of quarterly revenue, 87% non-GAAP gross margin, $33.42 of adjusted EPS and a materially higher fiscal Q1 outlook. The article also moved Micron into Hayabusa Living Research.[1]
October 3, 2026 — Living Research Coverage Established
Change recorded: Micron now has a permanent company coverage page separate from the dated September 30 research article.
Current view: structural uplift is plausible and the evidence is strengthening, but permanence remains unproven. The research priority is normalized earnings power, not the current peak-cycle P/E.
Next test: fiscal Q1 2027 results, followed by evidence from pricing, contracts, supply additions and AI infrastructure demand.
October 23, 2026 — Contract Protection Gets More Concrete
Prior view: a higher earnings floor is plausible, but peak-cycle earnings do not establish cheapness.
Evidence reviewed: October 9 annual filing, October 20 supply-chain investment, October 21 technology roadmap and October 22 industry outlook. Financial balances retain their original reporting dates.
Judgment: modestly strengthen confidence in business durability; keep the security thesis conditional. Deposits and contingent credit support are not revenue, and roadmap targets are not realized returns.
Next test: contract execution, pricing, cash generation, capital allocation and a valuation grounded in normalized earnings.
How This Research Will Evolve
Review Micron after quarterly results and whenever a material pricing, strategic-customer, capacity or AI-demand development changes an assumption. Routine headlines do not need a new article.
The current view at the top may change. Earlier conclusions in the timeline should not. When Hayabusa changes its mind, the page should show the new evidence, the prior view, the resulting judgment and the next test. That preserved history is part of the product.
Sources
- Hayabusa Research — Micron (MU): The 7× P/E Wasn’t a Typo — September 30, 2026 event re-underwrite and the first dated entry in this Living Research record. Includes the original thesis framing, strategic-customer-agreement review and monitoring list.
- Micron Technology — Fiscal Q4 and Full-Year 2026 Results — revenue, margin, EPS, cash flow, capital expenditures, business-unit results, fiscal Q1 2027 guidance and product highlights.
- Micron Technology Form 10-Q, fiscal Q3 2026 — structure of Strategic Customer Agreements, including take-or-pay commitments and pricing mechanisms. Company statements about future margin durability remain management expectations and require subsequent testing.
- Micron — Events and Presentations — checked October 23, 2026; exact fiscal Q1 release date not verified in accessible material.
- Micron Technology — Fiscal Q3 2026 Results — HBM4 high-volume shipment update and HBM4E development timeline used as a starting product baseline.
- Micron FY2026 Form 10-K, filed October 9, 2026 — Fiscal year ended September 3. Note 16: remaining performance obligations and customer protection; liquidity and cash flows: deposit classification and capex; Item 5: October 8 buyback authorization. Filing balances are not post-October 3 orders.
- Micron October 21, 2026 technology presentation — public copy hosted by 6ix — Company-authored roadmap, slides 16–24. Targets and development plans require execution; this review used the publicly hosted presentation copy, not a complete webcast transcript.
- TrendForce — Server CPU outlook, October 22, 2026 — Independent industry research and forecast; memory implications are Hayabusa interpretation, not Micron guidance.
- Micron share-price history — Stock Analysis / S&P Global Market Intelligence — October 2 and October 23 closing prices. Percentage change is Hayabusa arithmetic using unadjusted closes; excludes dividends and after-hours trading.
- Micron — Indium Corporation investment, October 20, 2026 — Company supply-chain announcement; no incremental earnings contribution assumed.
- TrendForce — Q4 memory pricing outlook, September 30, 2026 — Pre-existing context on AI demand, consumer affordability and capacity reductions per device. Forecasts are not realized October contract prices.
Source review: October 23, 2026. This update separates newly available disclosures from their historical reporting periods, company targets from demonstrated results, and industry forecasts from Micron guidance. No independently verified October contract-price series, complete webinar transcript or current consensus dataset was available in this review. Financial baselines remain fiscal Q4 2026 and fiscal Q1 2027 guidance; market prices are dated snapshots.
Disclosure: Hayabusa Research publishes independent investment research for informational and educational purposes. Nothing on this page constitutes personalized investment advice, a recommendation, or an offer to buy or sell securities. Semiconductor and memory markets are cyclical, and views can change materially as pricing, demand, capacity and customer commitments evolve.