Overview
Micron Technology (Nasdaq: MU) reports fiscal fourth-quarter results after the close on September 30, and the question in front of the market is not whether the company clears its own bar. It is how much further the curve can run. Over four quarters, quarterly revenue has climbed from $13.6 billion to $41.5 billion, non-GAAP gross margin has gone from under 40% to 84.9%, and company guidance for the quarter now being reported is $50.0 billion. When a semiconductor business is running at that slope, investors are not pricing current profit. They are pricing the moment the slope flattens.
Narrow the question one more turn and it lands on HBM. High-bandwidth memory is not simply DRAM sold at a higher price. It consumes several times the wafer area per bit, which makes it simultaneously the amplifier of Micron's margin and the amplifier of the industry's shortage.
NVIDIA posted $89.0 billion of data center revenue last quarter and guided the next one to $108.0 billion, so accelerator shipments set the cadence of HBM offtake. Meanwhile Micron has locked roughly a fifth of its DRAM volume and a third of its NAND volume into five-year strategic customer agreements. Together those variables mean the signal in this print sits in capacity, price bands and the fiscal 2027 supply roadmap, not in whether earnings per share land a few cents above consensus.
Key Takeaways
Last quarter was a record and a punishing comparison base. Fiscal Q3 2026 revenue was $41.46 billion, up 74% sequentially and 346% year over year, with non-GAAP EPS of $25.11, an operating margin of 81.2% and $18.3 billion of free cash flow.
The Q4 guide already contains a deceleration signal. Micron guided revenue to $50.0 billion plus or minus $1.0 billion, gross margin of approximately 86% and non-GAAP EPS of $31.00 plus or minus $1.00, while stating explicitly that the margin outlook reflects a meaningful moderation in the rate of price increases.
The wafer trade ratio is the key to understanding this shortage. Each wafer allocated to HBM displaces several wafers of conventional DRAM output, and that ratio rises with every generation from HBM4 to HBM4E, which makes the tightness self-reinforcing.
Strategic customer agreements have changed the shape of the earnings cycle. Sixteen signed agreements cover roughly 20% of DRAM volume and a third of NAND volume, with 14 of them carrying about $100 billion of minimum-price contract revenue and roughly $22 billion of customer deposits and related financial commitments.
The capacity answer does not arrive until mid-2027. Idaho ID1 is targeted for first wafer output in mid-calendar 2027, the Tongluo site in Taiwan for meaningful shipments around the same time, and Singapore advanced packaging to contribute to HBM packaging capacity from the first half of calendar 2027.
Valuation and positioning are the real risk into the print. MU is up close to 280% this year and closed at $1,053.29 on September 28, not far below its $1,255 record, with options pricing a move approaching double digits in percentage terms.
Starting From a Record Quarter and a Very High Bar
What Actually Happened Last Quarter
Micron's
fiscal third-quarter results showed revenue of $41.46 billion, GAAP net income of $28.24 billion, diluted EPS of $24.67 on a GAAP basis and $25.11 on a non-GAAP basis, and operating cash flow of $25.39 billion. Net capital expenditures were $7.1 billion and adjusted free cash flow reached a company record $18.3 billion.
The product-level detail matters more than the headline. According to the
prepared remarks for the earnings call, DRAM revenue was $31.3 billion, or 76% of the total, with bit shipments up only a low-single-digit percentage while prices rose in the low-60s percentage range. NAND revenue was $9.9 billion, with bit shipments up a mid-single-digit percentage and prices up in the mid-80s. That structure deserves a second reading: almost none of the revenue surge came from volume. The same document shows Cloud Memory revenue of $13.77 billion at an 83% gross margin, Core Data Center at $11.52 billion and 87%, Mobile and Client at $11.52 billion and 87%, and Automotive and Embedded at $4.63 billion and 79%. Four business units posting margins near 80% at the same time has no precedent in memory.
The balance sheet repaired in parallel. Micron ended the quarter with $30.2 billion in cash and investments against $5.7 billion of debt, for a net cash position of $24.4 billion, and inventory of $8.6 billion at 120 days, with DRAM inventory below that level. The
fiscal Q3 Form 10-Q shows cost of goods sold at 15% of revenue for the quarter, against 62% a year earlier.
The Gap Between Guidance and Consensus
Guidance calls for revenue of $50.0 billion plus or minus $1.0 billion, non-GAAP gross margin of approximately 86%, operating expenses of roughly $1.65 billion and non-GAAP EPS of $31.00 plus or minus $1.00 on about 1.15 billion diluted shares. The sell side sits slightly above the midpoint.
Consensus compiled from 33 analysts points to EPS of $31.56 on revenue of $51.20 billion, with estimates ranging from $28.04 to $37.44, a notably wide dispersion.
The qualifier management attached to that guide is the part worth marking. The prepared remarks state that the fiscal Q4 gross margin outlook reflects a meaningful moderation in the rate of price increases. In other words, 86% is not evidence of acceleration. It is what deceleration looks like once it has been priced in. A gross margin materially above 86% would say price pass-through is running ahead of the company's own assumptions. A margin merely in line would make the next quarter's guide far more consequential than this quarter's result.
Why HBM Is the Variable That Decides This Print
HBM Is Not Just Expensive DRAM
Treating HBM as memory with a higher price tag leads to the wrong conclusion. Conventional DDR5 and LPDDR5X are planar, single-die products with mature process flows, high yields and fast throughput. HBM stacks multiple DRAM dies vertically using through-silicon vias, packages them with a logic base die, and places the resulting module on an interposer alongside the accelerator. That path adds TSV etch, extreme wafer thinning, multi-layer bonding and high-precision packaging, and every one of those steps introduces a new point of yield loss.
The roadmap keeps getting steeper. Micron disclosed that HBM4, built on 1-beta DRAM, is in high-volume shipments for its lead customer's platform, that the 12-high ramp is tracking roughly twice as fast as HBM3E 12-high, and that HBM4 revenue has already exceeded $1 billion. HBM4E, built on 1-gamma, is in development with volume production expected in calendar 2027. Each generation raises stack count and data rate, and with them the demands on bonding and thermal design.
Why HBM Consumes So Much More Wafer Capacity
This is where the trade ratio enters. Producing a given quantity of HBM bits consumes far more wafer area than producing the same quantity of conventional DRAM bits. According to
reporting on Micron management's public comments, CFO Mark Murphy described a ratio that historically ran around three to one for HBM3 and that increases as the industry moves to HBM4, HBM4E and HBM5. Micron's own fiscal Q3 remarks list HBM's growth and its rising trade ratio with each generation among the structural factors pressuring non-HBM supply.
Put that mechanism into an industry model and the conclusion is direct. With cleanroom space fixed, one additional HBM wafer means several fewer wafers of DDR5, LPDDR5X or server RDIMM. The stronger AI demand runs, the more capacity HBM absorbs, the fewer bits are left for PCs, smartphones and traditional servers, and the higher those prices are pushed. That is why this pricing cycle is not confined to AI products but has spread across every category. Micron notes a parallel effect in NAND, where suppliers redirecting cleanroom space from NAND to DRAM, combined with limited overall cleanroom availability, constrains NAND bit supply growth.
How Accelerator Demand Reaches the Memory Line
The upstream chain is clear enough. NVIDIA's
fiscal second-quarter results for the period ended July 26, 2026 showed revenue of $96.2 billion and data center revenue of $89.0 billion, up 117% year over year, with next-quarter revenue guided to $108.0 billion plus or minus 2% and Vera Rubin described as being in full production. Every accelerator carries its HBM stacks, so accelerator unit growth converts into HBM demand on close to a linear basis, and each new platform generation typically raises per-system capacity and bandwidth requirements as well.
HBM is not the only transmission path, though. Micron observes that agentic AI is reshaping data center infrastructure beyond accelerator-only racks to include CPU racks for the agent control plane and program execution, plus storage racks for a rapidly expanding context store. On that basis the company raised its calendar 2026 industry server unit growth expectation to a high-teens percentage from low double digits, while noting the upgrade is enabled by a modest reduction in average server DRAM content growth as customers prioritize unit shipments under tight memory allocation. Micron's fiscal Q3 data center revenue exceeded $25 billion, an annualized run rate above $100 billion, with data center SSD revenue above $5 billion after more than doubling sequentially.
Where Pricing Stands and How Long Margins Can Hold
Contract Price Increases Are Converging
Price has been the engine of this earnings cycle, which makes the second derivative more informative than the level.
TrendForce reported in its
third-quarter 2026 memory pricing survey that conventional DRAM contract prices are forecast to rise 13% to 18% sequentially and NAND contract prices 10% to 15%, a noticeably slower pace than prior quarters, as consumer customers in PCs and smartphones reach the limit of what they can absorb and the comparison base rises. Its
fourth-quarter outlook holds that cloud service provider demand will keep DRAM and NAND prices rising while capacity reallocation toward servers and HBM continues to squeeze consumer segments, producing divergent trends across applications.
Align that cadence with Micron's fiscal calendar and the framework is clean. Last quarter captured something close to the peak rate of price increase. The quarter being reported enters the deceleration zone. The first half of fiscal 2027 will show whether increases stay positive at all. A converging rate of increase is not the same as falling prices, but for a company already guiding to 86% gross margin, the two have very different implications for earnings per share.
The Agreements Put a Floor Under Margin, and a Ceiling Over It
The 16 strategic customer agreements disclosed last quarter are the key to the future margin shape. The prepared remarks describe agreements that typically run five years, from calendar 2026 through the end of calendar 2030, with automotive deals generally running three years. Together they cover roughly 20% of Micron's DRAM volume and about a third of its NAND volume over that period, and include four very large customers and three medium-sized ones. They are structured as take-or-pay arrangements with binding commitments to purchase specified volumes.
The pricing terms deserve close reading. The largest agreements generally carry a ceiling price for existing products set at the calendar Q2 market price, together with a floor price that holds through the term, while some agreements use fixed prices or no price bands at all. Once all planned agreements are executed, Micron expects arrangements with fixed prices or ceilings at or close to current calendar Q2 levels to account for roughly 40% of revenue. Management stated that for agreements carrying price bands, the floor supports a gross margin well above the company's peak quarterly margins in any past cycle.
That cuts both ways. On the downside, even if the industry rolls over, a substantial share of revenue cannot price below its contractual floor, which lifts the trough of the earnings cycle. On the upside, if spot and contract markets keep climbing, the capped portion cannot fully participate, which trims the peak. For a financial model, this means Micron's profit curve is migrating away from the classic memory cycle toward something closer to a band-constrained utility. Disclosure is shifting with it: the company began reporting remaining performance obligations in the May quarter, at over $5 billion at quarter end, with roughly $100 billion of RPO across agreements signed to date, alongside an expected $22 billion of customer deposits and related financial commitments, of which approximately $18 billion is cash deposits.
One caveat travels with those figures. RPO is calculated on minimum committed volumes and minimum prices, and Micron has said the estimate is inherently conservative and should not be read as a forecast of future revenue.
Capacity Is the Only Cure, and It Arrives in 2027
The Composition of Capex Matters More Than the Number
Micron guided fiscal Q4 capital expenditures to around $10 billion, bringing fiscal 2026 spending to approximately $27 billion, net of anticipated government incentives. For fiscal 2027, quarterly capex is expected to run above the fiscal Q4 level, with more than half of the year-over-year increase coming from construction capex as the company pulls in cleanroom capacity. Operating expenses are projected to rise by roughly $1 billion in fiscal 2027, weighted to the second half, as research and development expands.
Construction accounting for more than half the increase is routinely misread as evidence that expansion marks a cycle top. The implication is close to the opposite. Money spent on shells and cleanrooms does not become bits in the coming quarters, because building cycles typically run two to three years. In the near term it depresses free cash flow without relieving the shortage. What changes the supply-demand balance on a medium horizon is equipment capex and node transitions.
The Timeline for New Supply
The prepared remarks put specific dates on the map. Idaho ID1 is on track for first wafer output in mid-calendar 2027 and ID2 in late calendar 2028, while the first fab in the New York cluster broke ground in January 2026. At the newly acquired Tongluo site in Taiwan, the existing 300,000-square-foot fab is expected to support meaningful product shipments in mid-calendar 2027, about a quarter earlier than previously expected, and construction has begun on a similarly sized second cleanroom there that will support EUV equipment. The Manassas, Virginia fab has started first production of 1-alpha DDR4 for legacy needs in automotive, industrial, medical, aerospace and defense. Micron also concluded a multi-year EUV supply agreement with
ASML to support adoption at the 1-delta node and beyond. On packaging, the Singapore site becomes a second center of excellence after Taiwan and is expected to contribute meaningfully to HBM packaging capacity from the first half of calendar 2027.
The industry view is worth writing down. Micron expects calendar 2026 industry DRAM bit shipments to grow in the low-to-mid 20s percentage range and NAND around 20%, with its own DRAM supply growing roughly in line with the industry and NAND somewhat below it. On balance, management expects tight conditions to persist beyond calendar 2027, and while industry supply should improve gradually in 2028, the company says it has no line of sight on when supply catches up with demand. The constraints it cites include long fab construction lead times worldwide, shortages of workers with critical trade skills, complex permitting and regulation, and the need for enhanced energy infrastructure.
Valuation, the Chart and How Traders Are Positioned
Cheap or Expensive Depends on Which Year of Earnings You Use
Summing Micron's reported non-GAAP EPS across fiscal 2026, namely $4.78 in the first quarter, $12.20 in the second, $25.11 in the third, and the $31.00 guidance midpoint for the fourth, gives roughly $73 for the year. Against the September 28 close of $1,053.29, that is about 14 times current fiscal-year earnings. The multiple is exceptionally low relative to the rest of the semiconductor complex, but the historical pattern in memory is precisely that cycle tops carry the lowest multiples, because the denominator is at its peak.
That is exactly where the disagreement lives.
Fifty-seven analysts cited by The Motley Fool carry a 12-month median target of $1,600, while
49 analysts tracked by stockanalysis.com average $1,515 with a consensus rating of Strong Buy. The bull case holds that contractual price floors mean fiscal 2027 and 2028 earnings no longer obey the old cycle assumptions, so 14 times should not be read as a peak signal. The bear response is equally coherent: ceilings cap the upside, and the new capacity eventually arrives.
The Levels That Matter
MU is up close to 280% year to date.
TradingKey's market tracking notes the stock has rebounded roughly 17% from a September 16 low near $926, closed at $1,082.28 on September 25 and cleared the September 9 high of $1,042, with near-term resistance around $1,120 and room toward $1,200 and the $1,255 record if that level holds.
Robinhood's quote page shows a September 28 close of $1,053.29 within a 52-week range of $155.18 to $1,255.00.
For traders, that distribution makes the risk-reward into the print asymmetric. Consensus already sits above guidance, the stock is within reach of its record, and options are pricing an event move approaching double digits in percentage terms. In that setup, position size and entry timing carry more weight than the directional call.
Beyond the EPS Print: What Actually Deserves Attention
Three Numbers That Outrank the Beat
The first is fiscal Q1 2027 guidance. Revenue and gross margin guidance will reveal the real slope of price deceleration. A margin guide flat to slightly lower in the 84% to 86% range is consistent with what management has already signaled. A guide meaningfully below 84% would suggest the drag from contractual ceilings on revenue mix is arriving sooner than expected.
The second is the specific fiscal 2027 capex figure and its composition. So far the company has offered only the qualitative statement that quarterly spending will run above fiscal Q4 levels. A full-year guide well above $40 billion, with a rising equipment share, would indicate Micron is betting on supply release beyond 2028 and could pull forward the point at which the market rebalances.
The third is progress on the strategic customer agreements. Micron expects that approximately half or more of company revenue will eventually sit under these contracts once all planned agreements are executed. Updates to the count, the share of revenue covered and the scale of customer deposits are the most direct evidence on whether earnings stability warrants a re-rating.
Three Categories of Risk
Pricing risk is the most immediate. TrendForce already sees consumer price tolerance near its limit, and if PC and handset makers respond with downgraded configurations, delayed procurement or reduced builds, non-HBM demand could soften before AI demand does, and that segment still contributes a substantial share of Micron's high-margin revenue.
Execution risk concentrates on HBM4 and HBM4E. Higher stack counts and data rates amplify yield variability, and the packaging ramp carries its own uncertainty. Any slippage in customer qualification affects both share and pricing power in a market still led by SK hynix.
Cycle risk is about timing. Micron itself expects industry supply to improve gradually in 2028, and several of its new sites release capacity from mid-2027 onward. In memory history, the overlap of concentrated supply release with decelerating demand growth is where price inflections occur. Contractual floors raise the trough. They do not repeal the cycle.
Exclusive View from James Mitchell
For James Mitchell, the easiest thing to misread in this print is the 86% gross margin guide, taken as confirmation of strength while the qualifier in the same document goes unread: that the outlook already reflects a meaningful moderation in the rate of price increases. Last quarter DRAM bit shipments grew only a low-single-digit percentage while prices rose in the low 60s, so revenue elasticity came almost entirely from price. Any earnings curve driven by price alone gives its signal in the second derivative long before the first. The information density in this report therefore sits in what the fiscal Q1 2027 margin guide implies about the slope of pricing, not in the quarter just closed.
The second thing the market may be misreading is the nature of the strategic customer agreements. They are widely treated as an unambiguous positive; in fact they are a symmetric exchange of risk. The floor genuinely lifts the trough of earnings to what management describes as well above peak quarterly margins in any past cycle, which is a structural change for this industry. But once roughly 40% of revenue carries fixed prices or ceilings near current levels, that revenue cannot fully follow the spot market higher. Micron is trading away part of its upside convexity for a lower volatility of earnings. For long-horizon investors seeking a re-rating, that is a good trade. For traders betting on peak-cycle extremes, it means peak earnings may land below what a pure price-extrapolation model produces.
The variable most worth tracking from here, in James Mitchell's view, is the relative contribution of bit shipment growth versus price to revenue growth. When bits begin to contribute more than price, capacity release has started and the cycle has entered its second half. While price remains the dominant driver, the shortage has not eased. Two companions belong alongside that measure. Days of inventory stood at 120 at quarter end with DRAM below that level, and two consecutive quarters of increase would be early evidence of a turn. So would a shift in the mix of equipment versus construction capex, since construction spending becomes capacity in two to three years while equipment spending converts into bits within a few quarters.
The cross-asset lesson runs beyond Micron. When memory moves from commodity to contractually governed strategic input, the cost structure of the entire AI infrastructure chain is rewritten: bill-of-materials costs rise for accelerator vendors, unit compute costs rise for cloud providers, and memory's share of device BOM rises for hardware makers. That explains why this pricing cycle has spilled into PC and smartphone pricing. It is equally worth noting that AI hardware profits currently driven by price still require downstream AI application revenue to validate them. Until that chain is verified end to end, unusually high profitability at any single link should be read as a position within a cycle rather than a new steady state.
FAQ
When does Micron report fiscal Q4 results?
Micron reports fiscal fourth-quarter 2026 results after the US market close on September 30, 2026, with an analyst call the same day. Company guidance calls for revenue of $50.0 billion plus or minus $1.0 billion, non-GAAP gross margin of approximately 86% and non-GAAP EPS of $31.00 plus or minus $1.00. Consensus sits modestly above the midpoint at roughly $51.2 billion of revenue and $31.56 of EPS.
What is the difference between HBM and conventional DRAM?
Conventional DRAM is a planar, single-die product with mature manufacturing and high yields. HBM stacks multiple DRAM dies vertically using through-silicon vias, packages them with a logic base die and places the module on an interposer beside the accelerator. The extra thinning, bonding and precision packaging steps mean HBM consumes far more wafer area per bit, which lifts price and margin while consuming disproportionate capacity.
Why does HBM push up prices for ordinary memory?
Because cleanroom space is finite. Micron management has described a trade ratio of roughly three to one in the HBM3 generation, rising with HBM4 and beyond. With total capacity fixed, each additional HBM wafer displaces several wafers of conventional DRAM. The stronger AI demand runs, the more server, PC and smartphone memory is crowded out, and the higher those prices are pushed, producing a shortage that spreads across every category.
Can Micron's gross margin keep rising?
Non-GAAP gross margin was 84.9% last quarter and guidance for the quarter being reported is approximately 86%, but management stated the outlook already reflects a meaningful moderation in the rate of price increases. TrendForce expects conventional DRAM contract price gains to converge to 13% to 18% sequentially in the third quarter of 2026. A high plateau is therefore a more reasonable expectation than continued expansion, and the real signal comes from the next quarter's guide.
What do the strategic customer agreements mean for Micron?
The 16 signed agreements typically run five years, cover roughly 20% of DRAM volume and a third of NAND volume, and are structured as take-or-pay contracts, with 14 of them carrying about $100 billion of minimum-price contract revenue. They set price floors and, for many products, price ceilings. The effect is a higher earnings trough and lower volatility, alongside reduced revenue elasticity when prices rise sharply.
When will new capacity relieve the shortage?
The milestones cluster beyond 2027. Idaho ID1 targets first wafer output in mid-calendar 2027 and ID2 in late calendar 2028, the Tongluo site in Taiwan targets meaningful shipments in mid-calendar 2027, and Singapore advanced packaging is expected to contribute to HBM packaging capacity from the first half of calendar 2027. Management expects tightness to persist beyond calendar 2027, with gradual improvement in 2028 but no line of sight on when supply catches demand.
What is the biggest risk into this report?
Positioning more than fundamentals. MU is up close to 280% this year and closed at $1,053.29 on September 28, not far from its $1,255 record, with consensus already above guidance. In that configuration, an in-line quarter may not move the stock higher, while any negative detail on price deceleration or HBM qualification progress gets amplified. Options are pricing an event move approaching double digits in percentage terms.
Should investors just watch whether EPS beats?
No. Beating consensus is not especially demanding in the current pricing environment. Three things determine what happens next: the pricing slope implied by fiscal Q1 2027 revenue and margin guidance, the size and equipment share of fiscal 2027 capex, and progress on the coverage and deposits attached to the strategic customer agreements. Those govern durability of earnings, which a single quarter's accounting result does not.
Disclaimer
The information above is provided for general market information and analysis only and does not constitute investment advice, financial advice, legal advice, tax advice or a recommendation to trade. Prices of equities, crypto assets and other related financial assets can fluctuate sharply, and moves around earnings announcements are frequently more violent than usual. Past performance, technical indicators, valuation multiples and on-chain data do not guarantee future results. The financial figures, company guidance, industry forecasts and market prices cited here reflect the point at which they were disclosed and may change at any time, so the latest official information from the companies and institutions concerned should be treated as authoritative. Readers should conduct their own research and make decisions based on their own financial circumstances, investment objectives and risk tolerance, consulting a qualified professional where appropriate. The MEXC Crypto Pulse team accepts no liability for any direct or indirect loss arising from the use of this information.
About the Author
James Mitchell specializes in technical analysis, market trends, and trading strategies for both Bitcoin and altcoins. Based in London, he has over 10 years of experience in financial markets. Before joining MEXC Learn, James worked as a senior analyst at a leading European investment firm, where he developed expertise in risk management and quantitative trading. His transition to cryptocurrency markets began in 2017, and he has since become recognized for his data-driven approach. He holds a Master's degree in Financial Economics from the London School of Economics. His analytical approach combines traditional technical analysis with on-chain metrics to provide readers with actionable insights.
Areas of Expertise: Technical Analysis, Market Trends & Cycles, Trading Strategies, Bitcoin & Altcoin Analysis, Risk Management.
Research References