Business Overview

Micron Technology, Inc. (NASDAQ: MU)

6 September 2026

Evidence base: Micron's SEC filings held in the MU research folder — FY2021–FY2025 Forms 10-K and Forms 10-Q through the quarter ended 28 May 2026 (Q3 FY2026) — supplemented by independent industry sources for market structure, competitor economics and price history. Every company figure is traceable to a filing; every industry figure carries its outside source. The company's fiscal 2026 ended in early September 2026 and had not been reported as at this date.

This document explains how the business works and what drives its economics. It is not a valuation and not a recommendation.

1. Executive Snapshot

2. What the Company Does

Processors compute faster than they can be fed data. DRAM holds working data close to the processor so it can be read in nanoseconds; NAND stores data durably and cheaply when the power is off. Every computer, phone, car and AI accelerator needs both, no volume-production substitute exists for either, and only three companies manufacture both at the leading edge. Micron is one of them.

The unit: one gigabit

Micron buys a 300mm silicon wafer and runs it through a fab operating twenty-four hours a day, seven days a week (FY2025 10-K). The wafer yields several hundred dies, each holding a fixed number of bits. The economics turn on how many good bits come off one wafer: the filings state that the primary determinants of manufacturing cost are “process line-width, 3D non-volatile layers, NAND cell levels, process complexity … and manufacturing yield.” Cost per bit therefore falls with technology transitions rather than with production volume, which is why the industry re-engineers its process every eighteen months or so and why depreciation, not labour or materials, is the dominant cost.

Dies are then packaged — for High-Bandwidth Memory, stacked twelve high and wired vertically through the silicon — tested, shipped, and collected shortly afterwards. Historically, substantially all contracts were short-term at fixed negotiated prices with payment due shortly after delivery (FY2025 10-K), and customers were “generally reluctant to enter into long-term, fixed-price purchase contracts.” That is the arrangement that changed in FY2026, and Section 5 sets out what replaced it.

One disclosure gap organises everything downstream. Micron never publishes the absolute price or cost of a gigabit, nor the absolute number of bits it ships. It reports only percentage changes in average selling price and bit shipments. An outside investor can therefore measure the direction and magnitude of unit economics with unusual precision, but never their level — and so cannot compute where the company's cost per bit sits against Samsung's or SK hynix's, which is the number that decides who survives a trough intact.

Products and business units

DRAM was 76% of revenue in the most recent quarter ($31.3bn), NAND 24$9.9bn), and NOR the remainder. The four reporting units are organised by customer type rather than by product, and were reorganised in the fourth quarter of FY2025 with prior periods restated.

Source: FY2026 Q3 10-Q. Business units were reorganised in Q4 FY2025; all prior-period segment figures were retrospectively restated, so segment data before that date is not comparable to earlier filings as originally reported.

The line being entered is the data centre, and the line being vacated is not a product but a customer set. CMBU grew from 12% of revenue in FY2023 to 36% in FY2025 while MCBU fell from 48% to 32%. The mechanism was deliberate allocation, not new capacity: Micron “shifted a portion of our DRAM supply to the data center and hyperscale cloud markets,” while MCBU “product supply was constrained to meet demand from higher-value segments” (FY2025 10-K). Mobile customers received fewer bits because data-centre customers paid more for them. The margin consequence is visible in the FY2025 unit results, where CMBU earned a 45% operating margin against MCBU's 17%.

Concentration is high and rising on two axes at once. One customer was 17% of FY2025 revenue and the top ten roughly half; the data-centre end market was roughly half of FY2025 revenue and 61% of the most recent quarter.

3. Industry, Competitive Position and Moat

The memory industry sells interchangeable bits built to a common JEDEC standard. A Samsung DDR5 die and a Micron DDR5 die at the same specification do the same job in the same socket. Because the product is fungible, price is set by whichever supplier is willing to sell the marginal bit, not by any individual producer — and because fixed costs dominate cash costs, every producer keeps its fabs running even when price falls below full cost. That combination is the whole reason memory cycles are as violent as they are.

Structure and where Micron sits

DRAM is a three-firm oligopoly with a fourth, state-funded entrant. In the second calendar quarter of 2026, DRAM revenue share was Samsung 39%, SK hynix 26%, Micron 25% and CXMT 10% (Counterpoint Research, reported August–September 2026). NAND is more fragmented: Samsung $23.1bn, SK hynix and Solidigm $14.3bn, Micron $11.9bn, Kioxia $10.7bn and SanDisk $9.0bn in the same quarter, the top five together up 77% sequentially (TrendForce, 18 August 2026). In HBM, the highest-value product in the industry, SK hynix held 50%, Samsung 33% and Micron 18% — against 62%, 17% and 21% respectively a year earlier (Counterpoint). Micron is third in DRAM, third in NAND and third in HBM.

The barriers that actually bind are capital, lithography and qualification. A leading-edge DRAM fab costs on the order of $10bn: Nanya's new EUV fab is budgeted at roughly $10.7bn and Micron's Hiroshima modernisation at roughly ¥1.5 trillion. EUV lithography is denied to Chinese producers by export control, leaving CXMT on DUV at 193nm, two to three process generations behind and carrying a cost-per-bit disadvantage estimated above 30% (Counterpoint, reported September 2026). HBM must additionally be qualified by the accelerator vendor, which takes multiple quarters. What does not bind is intellectual property in the exclusionary sense: Micron holds more than 60,000 granted patents and roughly 15,000 active US patents, but the industry litigates over them rather than being kept out by them.

The margin test

Micron earned an 80% operating margin in the quarter ended May 2026. This is the single most important thing to establish about it, because in the same window SK hynix reported a 76% operating margin, Samsung's Device Solutions division 69.9%, Kioxia 75% on an adjusted basis and SanDisk 69.1% on GAAP. Kioxia's own decomposition settles the mechanism: blended average selling price rose approximately 70% year over year on low-single-digit bit growth. The margin is coming from price, and the price is coming from the industry.

The consequence is that Micron's current economics are an industry condition, not a company-specific advantage. They are being earned simultaneously by every producer, including the smallest and the least technically advanced, which means they are reproducible by competitors and reversible by the same mechanism that created them. Nothing in the peer data supports treating an 85% gross margin as evidence of a moat.

What is genuinely Micron's

Three things are specific to this company. It is the only leading-edge memory manufacturer headquartered in the United States, which converts industrial policy into cash: up to $6.4bn of CHIPS Act grants and a 35% investment tax credit on qualified US semiconductor investment, plus a non-binding New York State term sheet worth up to $5.5bn over twenty-plus years. It shipped the industry's first 1-gamma DRAM node, its first using EUV lithography, in FY2025. And it carries roughly $24bn of net cash against $5.7bn of gross debt, which is the balance sheet a memory company needs to keep investing when prices collapse.

Against that, the geography of its assets is far more concentrated than the geography of its politics: 40% of property, plant and equipment sits in Taiwan and only 18% in the United States, against 65% of FY2025 revenue coming from US-headquartered customers (FY2025 10-K). The American fabs that would change this do not produce a bit before mid-calendar 2027.

What could weaken the position

CXMT went from 4% to 10% of global DRAM revenue in twelve months. It cannot win on economics — it is generations behind and more than 30% more expensive per bit — and it does not need to. With roughly 36% state-linked ownership before its IPO, about $8.6bn raised, and a stated target above 600,000 wafer starts per month by 2030, it can add supply on a policy signal rather than a price signal. Oligopoly discipline, which is what made the pricing of the last two years possible, cannot be enforced against a participant that is indifferent to returns.

The industry rewards the lowest-cost, highest-share producer with the balance sheet to keep investing through a trough. In the 2018–19 downturn, Samsung and SK hynix gained DRAM share while Micron lost it. Micron now has the balance sheet, and it has closed much of the technology gap, but it remains the third-largest producer in all three of its markets. It is closer to being the kind of company that wins here than at any point in the period examined; it is not yet that company.

4. Growth Engine

There is no acquired growth to separate out. Micron has purchased no revenue over the period examined; its only material transaction was the March 2026 acquisition of a wafer fabrication facility in Tongluo, Taiwan from Powerchip for $1.8bn in cash, which adds capacity from mid-calendar 2027 and contributed nothing to reported revenue. Every point of the 203% nine-month revenue increase is organic. The decomposition that matters is therefore price against volume, and Micron discloses it precisely.

Source: FY2025 10-K and FY2026 Q3 10-Q, MD&A. Micron expresses price and volume changes as approximations and ranges; the figures above are as filed. Bit shipments made up the remainder of each revenue change.

Roughly four-fifths of the nine-month revenue increase came from average selling prices. Bit volume — the variable Micron actually controls — grew about 30% in DRAM and in the low twenties in NAND, rates a memory company would consider unremarkable. FY2024 is the instructive contrast: revenue rose 62% that year on a reversed composition, with DRAM bits up in the mid-40% range against a low-teens price increase. Same company, same fabs, opposite engine.

Drivers, ranked

1. DRAM price per bit — cyclical. Roughly 140 of the 211 points of nine-month DRAM revenue growth. Set by industry supply and demand; Micron is a price-taker.
2. NAND price per bit — cyclical. Roughly 130 of 183 points. Kioxia's parallel disclosure of a ~70% blended ASP increase on low-single-digit bit growth confirms the same industry mechanism.
3. Data-centre mix shift — management-driven. CMBU rose from 12% of FY2023 revenue to 36% of FY2025 by reallocating existing DRAM supply to the highest-priced end markets while deliberately constraining mobile bit shipments. This raises revenue per bit without adding a wafer.
4. Bit supply growth from node transitions — structural. About 30% DRAM bit growth over nine months, generated almost entirely by putting more bits on existing wafers through 1-gamma DRAM and G9 NAND. No new fab produces before mid-calendar 2027.
5. HBM — structural, and the reason the four above are happening. HBM consumes roughly three times the wafer area per bit of standard DDR5 (Micron, Hot Chips 2026, reported August 2026), so each bit of HBM demand mechanically removes about three bits of conventional DRAM supply from the market. This is why AI demand tightens PC and phone memory pricing without any change in PC or phone demand. Micron does not disclose HBM revenue separately, so its direct contribution cannot be measured from the filings (NOT FOUND).
6. Take-or-pay strategic agreements — temporary in their growth effect, potentially structural in their stability effect. Signed from the third quarter of FY2026 onwards; terms and cash effects are in Section 5.

The honest summary: this is a price event occurring on a capacity base that has barely grown. Management's contribution has been to aim existing supply at the highest-priced demand, which is a real skill with a real margin consequence, but it is not the same thing as having created the growth.

5. Margin, Cash and Capital Allocation

Memory is a near-fixed-cost business, and two numbers carry the point. In the quarter ended May 2026, cost of goods sold rose 5% sequentially while revenue rose 74% — $6.4bn of cost against $41.5bn of revenue, producing an 85% gross margin against 74% the prior quarter and 38% a year earlier. Depreciation, $8.28bn in FY2025, is the largest single cost and is fixed the moment the fab is built. Because almost nothing in the cost base moves with price, essentially every incremental dollar of price falls to gross profit.

The same structure runs in reverse with equal force. In FY2023 it produced a negative 9% gross margin, a $1.83bn charge to write inventories down to net realisable value, and $382m of period costs recognised simply for running fabs below capacity. Operating leverage is not a feature of the good years; it is the permanent shape of the business.

Operating expenses have become almost invisible against revenue — R&D was $1.32bn in the latest quarter, 3% of revenue, against 10% in FY2025 and 20% in FY2023, and SG&A 1%. This is entirely a denominator effect: absolute R&D rose 36% year over year as Micron ramps investment, and the ratio fell only because revenue moved further.

The financial spine

Sources: FY2023 and FY2025 Forms 10-K; FY2026 Q3 10-Q. Comparability cautions: business units were reorganised in Q4 FY2025 with prior periods restated; FY2023 predates HBM at scale, so the product mix behind these lines is not the same business; the 9M and Q3 FY2026 columns are unaudited interim periods and are not annual figures; capital expenditure is gross of government incentives, which were $2.01bn in FY2025 and $2.99bn in the nine months to May 2026. FY2023 and FY2025 are chosen to show the trough and the last complete year; FY2021, FY2022 and FY2024 appear in the discussion where the shape of the change requires them.

Cash conversion

Nine-month operating cash flow of $45.7bn against $47.3bn of net income is close to one-for-one, which is unusual given that receivables ballooned from $9.3bn to $31.0bn over the same period. The offset came from a $3.3bn increase in payables and accrued expenses, higher income-tax payables, and higher consideration payable to customers for pricing adjustments. Inventory tells the sharpest story: $8.57bn at May 2026 against $8.36bn nine months earlier — effectively flat while revenue rose four and a half times. Micron is selling everything it makes, and that fact, more than any margin figure, is what defines the current moment.

Capital expenditure

Capex was $19.6bn gross in nine months against $2.99bn of government incentives received, and is guided to approximately $27bn for FY2026 net of incentives — against $15.86bn gross in FY2025 and $7.68bn at the FY2023 trough. Micron does not disclose a maintenance-versus-growth split (NOT FOUND), but it does disclose the destination, and the timing matters more than the amount.

• Boise, Idaho fab one: first DRAM wafer output projected mid-calendar 2027.
• Boise fab two: construction beginning 2026, initial output expected late calendar 2028.
• Clay, New York: first fab broke ground January 2026, supplying 2030 and beyond; up to four fabs over twenty-plus years.
• Singapore: HBM advanced packaging from the first half of calendar 2027; an additional wafer fab cleanroom in the second half of calendar 2028.
• Tongluo, Taiwan: acquired March 2026 for $1.8bn cash; meaningful shipments from mid-calendar 2027.

Not one dollar of the current capital programme produces a saleable bit before mid-2027. Management has separately indicated FY2027 capital expenditure above $45bn (EVP Sumit Sadana, Six Five Summit, August 2026, reported in trade press) — a figure that post-dates every filing in the evidence base and is not contained in any document within it.

Where the cash went

The ranking is the finding. In the largest cash-generating period in the company's history, management repaid $9.4bn of debt, took gross borrowings from $14.6bn down to $5.7bn, and let cash build to $30.1bn — while buying back $650m of stock against $47.3bn of earnings. Total equity reached $100.7bn. Buybacks are restricted under the CHIPS Act funding agreements, which is part of the explanation, but the pattern is consistent with a management team that is fortifying the balance sheet for a downturn rather than distributing a windfall. That is the behaviour of people who have run this business through FY2023.

The strategic customer agreements — announced, and post-dating most of the reported figures

From the third and fourth quarters of FY2026, Micron began entering multi-year take-or-pay agreements with binding volume commitments — a structural departure from the short-term negotiated pricing that had defined the business. Most are fixed-price or carry floor and ceiling bands; the largest have a ceiling for existing products approximating the market price in the second calendar quarter of 2026, and a floor running through the term. A minority have no price bands at all. Micron expects $22bn of cash deposits and related financial commitments from agreements concluded to date, roughly $18bn of it in cash, and states that gross margins at floor pricing would be “well above our peak quarterly margins in any past cycle” (FY2026 Q3 10-Q).

Two things belong beside that claim. The disclosed remaining performance obligation is only about $5bn, because it is measured at minimum volumes at minimum prices and excludes the agreements without price bands — it is a floor on a floor, not a backlog, and roughly one third of it is expected to convert to revenue within twelve months. And the floor prices, counterparties, contract lengths and cancellation provisions are not disclosed (NOT FOUND). The most consequential claim in Micron's current story is, from the filings, untestable.

6. Cyclicality, Constraints and What to Monitor

Cyclicality here is high, and the evidence sits inside the document set itself. Between FY2022 and FY2023 Micron's revenue fell 49%, gross margin went from 45% to negative 9%, and the company lost $5.8bn. It suspended the company-wide bonus, cut executive salaries, reduced wafer starts significantly below peak capacity, and cut capital expenditure from $12.1bn to $7.7bn. Industry-wide the collapse was of the same order: DRAM revenue fell from $25.6bn in the second calendar quarter of 2022 to roughly $11.4bn a year later, about 55% in four quarters, with Micron's own DRAM revenue falling 41% in the fourth quarter of 2022 alone (TrendForce). The 2018–19 downturn was milder — industry DRAM revenue fell 18.3% in a quarter and SK hynix's operating margin went from 66% to 58% — which is a useful reminder that not every downturn is a catastrophe, and that the two most recent ones differed by a factor of three.

Where the business sits right now

At an unprecedented peak on price and margin, and at an ordinary point on volume. The 85% gross margin of the latest quarter is 40 points above the previous cycle peak of 45% in FY2022 and 94 points above the FY2023 trough. Bit shipments, by contrast, grew about 30% in DRAM over nine months. A record margin reached on ordinary volume is a price peak, and price peaks in this industry have never been the durable part.

End-market exposure has also concentrated. Data centre, through CMBU and CDBU, was 61% of the most recent quarter's revenue ($25.3bn of $41.5bn); mobile and client 28%; automotive, industrial and consumer 11%. In FY2023 the two data-centre units together were 26%. The business is now roughly twice as levered to a single capital-spending cycle as it was three years ago. By customer headquarters, 65% of FY2025 revenue came from the United States and 7% from mainland China — the latter still depressed by the May 2023 decision of China's Cyberspace Administration barring critical information infrastructure operators from purchasing Micron products, which remains in force.

What is temporary

Supply is genuinely tight today, and the constraint is mechanical rather than rhetorical: SEMI projects 300mm memory wafer capacity rising only from about 4.1 million to 4.2 million wafers per month between 2026 and 2027, even as memory equipment spending rises from $52bn to $57bn, because the capital is going into more expensive nodes and HBM packaging rather than raw wafer starts. But tightness of this kind is capital, not physics. Samsung has committed roughly $73bn across chips in 2026, SK hynix's first-half capital expenditure rose 73% year over year, and Micron's own FY2026 capex is roughly double FY2025's. That is the same synchronised capacity response that ended each of the last two cycles, arriving on the same eighteen-to-twenty-four-month lag — which places it in 2027 and 2028.

A second temporary item is specific to Micron and easy to miss. Singapore tax incentives reduced the FY2025 tax provision by $1.05bn, worth $0.93 of diluted EPS. Singapore's implementation of the Pillar Two 15% minimum tax “largely offsets the benefit” from FY2026 onward, and the effective rate has already moved from 11.6% to 15.0% (FY2025 10-K; FY2026 Q3 10-Q). Government capital incentives are likewise conditional and subject to reduction, termination or clawback.

Durable versus borrowed

Leading indicators, and where they are published

• DRAM and NAND contract and spot prices — TrendForce and DRAMeXchange, monthly. The primary driver of everything above.
• Micron's inventory days and the ASP-versus-bit split in its MD&A — the quarterly 10-Q. Inventory building while revenue holds, or ASP contribution turning negative while bits still grow, is the first mechanical sign that FY2022–23 is repeating.
• Industry wafer capacity and equipment spending — SEMI fab forecasts and quarterly billings reports; this is where 2027–28 supply becomes visible before it becomes price.
• Competitor capital expenditure and capacity announcements — Samsung and SK hynix quarterly releases.
• HBM share by vendor — Counterpoint and TrendForce quarterly trackers. Micron's 18% against SK hynix's 50% is the gap that determines whether it participates in the high end of the next cycle.
• CXMT revenue share and node progress — Counterpoint DRAM tracker. The variable with no precedent in prior cycles.

7. Risks, Unknowns and Questions for Deeper Work

Cyclicality is covered in Section 6 and is not repeated here. What follows is what cyclicality alone does not capture, ordered so that the risks that compound come first.

• The take-or-pay agreements may not do what they are described as doing. Floors are struck against calendar Q2 2026 market prices, the highest in the industry's history. If spot prices fall well below those floors, the customer's economic incentive is to renegotiate, defer or dispute, and Micron's remedy is litigation with its largest customers during a downturn. The disclosed remaining performance obligation of about $5bn is a small fraction of the ~$22bn of expected deposits and commitments. A contract that binds only while it is out of the money converts margin protection into receivable and deposit-refund risk.
• Customer and end-market concentration compound rather than diversify. One customer was 17% of FY2025 revenue, the top ten roughly half, and data centre is 61% of the latest quarter. These are not independent exposures: a single hyperscaler pausing capital spending reduces Micron's revenue twice — once through the direct relationship, and again through the industry price that the same order book supports.
• Capacity is being committed at peak prices into a supply response that has already been announced. $19.6bn spent in nine months, approximately $27bn guided for FY2026, and above $45bn indicated for FY2027, against fabs whose first output is mid-2027 at the earliest. The mechanism is the oldest one in the industry: capital committed at peak margin arrives as depreciation and fixed cost precisely when price normalises, so the cost base steps up permanently while the revenue that justified it does not.
• CXMT does not need to earn a return. A producer roughly 36% state-linked before its IPO, two to three nodes behind, more than 30% more expensive per bit, that raised about $8.6bn and targets over 600,000 wafer starts per month by 2030, adds supply on a policy signal. The supply discipline that made the last two years possible has no mechanism for disciplining a participant indifferent to price.
• Asset geography is a live concentration. Forty per cent of property, plant and equipment sits in Taiwan (FY2025 10-K), while the Idaho and New York fabs that would rebalance it do not produce between mid-2027 and 2030. A memory fab cannot be substituted or relocated, so a disruption to Taiwanese operations would remove a large share of output for years rather than quarters.
• Government incentives are conditional and constraining. CHIPS grants of up to $6.4bn and non-US incentives are subject to reduction, termination or clawback if outcomes and compliance requirements are not met, and the CHIPS agreements restrict share repurchases. A shortfall raises net capital cost on projects already committed and irreversible.

What the sources could not answer

• Absolute average selling price or cost per gigabit, and absolute bit volumes. Micron discloses only percentage changes, so unit economics can be tracked in direction and magnitude but never in level — and Micron's cost position against Samsung's and SK hynix's cannot be computed.
• HBM revenue, HBM gross margin, and HBM as a share of DRAM bits. Never disclosed separately, despite HBM being the stated reason for the entire mix shift.
• The maintenance-versus-growth split of capital expenditure, which would separate the cost of standing still from the cost of the expansion.
• Strategic customer agreement counterparties, floor price levels, contract lengths, volume commitments and cancellation provisions.
• FY2026 fourth-quarter and full-year results. Fiscal 2026 ended in early September 2026 and had not been reported as at the date of this document; the most recent audited figures are FY2025 and the most recent reported figures are the nine months to 28 May 2026.

Before an investor could form a thesis, three things would need resolving: the floor-price mechanics of the strategic agreements at least in aggregate; a bottom-up view of industry bit supply growth for calendar 2027 and 2028 built from announced capacity rather than from Micron's characterisation of it; and Micron's cost per bit relative to its two larger competitors, which is the only number that determines who emerges from the next trough with share intact.

8. Investor Takeaways

• Micron is a commodity manufacturer of memory bits — third in scale in each of DRAM, NAND and HBM — whose earnings are set by an industry price it does not control.
• The economic engine is operating leverage on a near-fixed cost base: $6.4bn of cost against $41.5bn of revenue in the latest quarter produced an 85% gross margin; the identical structure produced negative 9% in FY2023.
• The main growth lever in the current period is price, not volume — roughly four-fifths of the nine-month revenue increase came from average selling prices, on bit growth of about 30%.
• What could break the story is the supply response already announced and funded across the industry, arriving on a 2027–28 lag, compounded by a state-funded entrant that adds capacity irrespective of returns.
• What to monitor is DRAM contract prices, Micron's inventory and the price-versus-volume split in its own MD&A, industry capacity additions for 2027–28, and any disclosure that makes the take-or-pay floors testable.