MU Investment Research Report

The most recent reported period is fiscal Q3 FY2026, in which Micron Technology, Inc. reported revenue of $41.46B against $23.86B in the prior quarter and $9.30B a year earlier, and guided fiscal Q4 FY2026 revenue to $50.0B +/- $1.0B. [1] [2]

The key analytical issue is whether fiscal Q3 FY2026 earnings power represents a cyclical peak or a more durable reset in profitability, because gross margin and cash generation both moved sharply against their year-ago comparisons. [3] [4] [5]

The report does not issue a rating or price objective; it frames the evidence investors would normally debate before the next earnings event. [6]

Fiscal Q3 FY2026 revenue was $41.46B, versus $23.86B in the prior quarter and $9.30B a year earlier, while operating cash flow was $25.39B. [1] [4]

Reported profitability was 84.6% GAAP gross margin and 84.9% non-GAAP gross margin, with GAAP diluted EPS of $24.67 and non-GAAP diluted EPS of $25.11. [3] [7]

Capital intensity remains central to the debate because the company paired $18.3B of adjusted free cash flow with $7.1B of net capital expenditures and $30.2B of cash, marketable investments, and restricted cash. [8] [9]

The report uses 7 captured source documents and 42 sourced facts for source verification before publication. [6]

Micron Technology, Inc. filed 20 reports with the SEC in the trailing twelve months, with the latest verified filing dated 2026-06-25. [10] [11] [12]

Micron Technology, Inc. filed 10-Q with the SEC on 2026-06-25. [10]

Micron Technology, Inc. filed 8-K with the SEC on 2026-06-24. [11]

Micron Technology, Inc. filed 8-K with the SEC on 2026-06-09. [12]

Micron Technology, Inc. filed SD with the SEC on 2026-05-20. [13]

Micron Technology, Inc. filed S-8 with the SEC on 2026-04-17. [14]

Micron Technology, Inc. filed 8-K with the SEC on 2026-04-01. [15]

Outside institutional holders filed 26 beneficial-ownership schedules (Schedule 13D/13G) naming Micron Technology, Inc. as the issuer in the retrieved EDGAR window, the most recent dated 2026-05-14. These are filed by the holder, not by the company. [16]

Business-unit economics show where the result came from: Cloud Memory, Core Data Center, Mobile and Client, and Automotive and Embedded each reported fiscal Q3 FY2026 gross margin on the disclosed segment basis. [17] [18] [19] [20]

Cloud Memory was the largest segment revenue line at $13.769B with 78% operating margin, followed by Core Data Center at $11.524B with 83% operating margin. [17] [18]

The financial-quality question is whether gross margin of 84.6% in fiscal Q3 FY2026 and guided gross margin of Approx. 86% in fiscal Q4 FY2026 can persist as the company invests in its manufacturing footprint. [3] [21]

Strip away the product names and a memory business is an identity: revenue equals bits shipped multiplied by price per bit, and gross profit equals bits shipped multiplied by the spread between price per bit and cost per bit. [6]

That identity has a hard consequence. There are only three ways to earn more: ship more bits, raise the price of each bit, or lower the cost of producing one. Every strategic claim a memory company makes has to land in one of those three buckets or it does not affect earnings. [6]

The three levers behave very differently through time. Cost per bit falls slowly and fairly predictably with process technology. Bits shipped are capped in the short run by installed capacity, which takes years and heavy capital to expand. Price per bit is the only variable that can move violently within a single quarter, because it clears the gap between fixed short-run supply and moving demand. [6]

So the central question for any quarter is not whether results were strong. It is which of the three levers produced the result, because price is mean-reverting on a supply response while cost and volume are not. [6]

Applying the identity to fiscal Q3 FY2026 settles the question immediately. DRAM revenue moved +67% sequentially on a low-60% range increase in average selling prices against a low-single-digit percentage range increase in bit shipments. NAND revenue moved +99% sequentially on a mid-80% range increase in average selling prices against a mid-single-digit percentage range increase in bit shipments. [22] [23]

Volume contributed almost nothing. The company sold approximately the same quantity of bits it sold in the prior quarter, at a dramatically higher price. This is a price event, not a capacity event, and it should be underwritten as one. [22] [23]

The year-over-year decomposition tells the same story on a longer base: DRAM revenue +343% on a low-260% range increase in price against a low-20% range increase in bits. Over a full year the company expanded volume meaningfully, but price still did the overwhelming majority of the work. [24]

The practical implication is that the durability of these earnings is a question about the durability of memory pricing, and memory pricing is set by the balance between industry supply and demand more than by any single producer's execution. [25]

The cost side confirms the diagnosis. Cost of goods sold and revenue grew at +4.8% vs. +73.7% respectively (COGS $6,105M to $6,400M; revenue $23,860M to $41,456M). A producer that had genuinely expanded output would show cost rising alongside revenue; this cost base barely moved. [26]

Incremental gross margin on the sequential change was 98%, computed as ($35,056M - $17,755M) / ($41,456M - $23,860M). Effectively every additional dollar of revenue arrived as gross profit, which is what happens when the additional revenue comes mostly from price rather than volume. [26]

Cost of goods sold fell to 15% vs. 62% a year ago of revenue ($6,400M / $41,456M vs. $5,793M / $9,301M). No manufacturer reduces unit cost at that rate in twelve months. The ratio collapsed because the denominator expanded, which means the same operating leverage runs in reverse if price retraces. [26]

Guidance extends the same mechanism: gross margin is guided from 84.6% to 86%, and the revenue guide implies a +21% sequential step-up. Both require the price environment to persist, because capacity cannot expand materially inside a single quarter. [2] [21]

High prices are a signal, and in a commodity industry the signal is answered with capital. The mechanism that has historically ended memory upcycles is visible inside the same cash flow statement that reports the profits. [27]

Capital expenditure is up +92% year to date ($10,199M to $19,602M). That spending does not add supply this quarter or next; memory capacity arrives with a lag measured in years, which is precisely why the industry systematically overshoots. [27]

Capex is running at 2.9x depreciation ($19,602M capex / $6,862M depreciation and amortization). An asset base expanding this much faster than it is written off does two things at once: it builds the supply that will compete with today's prices, and it builds a larger fixed cost base that must be absorbed at whatever price prevails when the capacity arrives. [27] [28]

Roughly 43% of operating cash flow is being redeployed into capacity ($19,602M capex / $45,702M operating cash flow). Peak-cycle cash is funding the assets whose output will meet a market that may no longer be short. [27]

This is the feedback loop that defines the industry: shortage lifts price, price lifts profit, profit funds capacity, capacity arrives together across producers, and the resulting supply resets price. The loop is slow, which is why it is repeatedly underestimated at the peak. [27]

The strongest argument against treating this as an ordinary cyclical peak is contractual, and it deserves to be tested rather than accepted. [29]

Micron Technology, Inc. disclosed multi-year take-or-pay agreements with binding volume commitments, where pricing is either fixed or bounded by floor and ceiling prices, and stated that margins at floor pricing would still exceed prior peak quarterly margins. [29]

From first principles this changes the shape of the distribution rather than its centre. A floor price transfers downside risk from the producer to the customer, which genuinely raises trough earnings. A ceiling price does the opposite at the top, capping realised price on existing products near the market level of the second calendar quarter of 2026. [29]

Both tails are therefore truncated. The correct reading is not that the cycle has been abolished, but that its amplitude has been contracted for the covered volume, and that the the rest is still exposed to market pricing. The share of volume under contract, and the level of the floors, are the facts that determine how much this matters. [29]

The disclosure does not quantify what proportion of output is committed, so the magnitude of this protection cannot be verified from public filings. That is a genuine limit on this analysis and is recorded as an open evidence gap. [29]

Customers are being invoiced faster than they are paying. Across the first nine months of the fiscal year, unpaid customer bills grew by $19,953M — equal to 44% of the $45,702M of cash the business generated over the same period. [30] [27]

Both figures are year-to-date, so they should not be read against a single quarter's revenue. A build of this size is what a steep revenue ramp normally produces and is not by itself evidence of a problem. It is worth watching because sales are recorded when goods ship, while the cash arrives later, so a downturn in demand can show up here before it shows up in revenue. [30]

A view that cannot be proven wrong is not analysis. The following are the specific, observable facts that would falsify the reading above, listed so that the next update can be scored against them rather than rewritten around them. [6]

If bit shipments begin growing faster than price, the cycle has turned from a price event into a volume event, and the incremental margin recorded in this quarter will not repeat. [22]

If cost of goods sold starts rising materially faster than revenue, new capacity is being absorbed and the fixed cost base built by current capital spending has begun to bite. [26]

If the fiscal Q4 FY2026 result lands below the revenue or gross-margin guide, the pricing assumption embedded in the guide was too strong, and the peak-versus-plateau question resolves toward peak. [2] [21]

If receivables continue growing faster than revenue, the gap between recognised profit and collected cash is widening rather than normalising. [30]

If industry capital expenditure across producers continues compounding at current rates, the supply response is being built now and will arrive regardless of how demand evolves. [27]

The next reporting event is the fiscal Q4 FY2026 result, which tests whether the $50.0B +/- $1.0B revenue guide and Approx. 86% gross-margin guide remain credible as demand and supply conditions evolve. [2] [21] [31]

The most important watch item is the bridge from fiscal Q3 FY2026 revenue of $41.46B to the fiscal Q4 FY2026 guide of $50.0B +/- $1.0B, which implies another material sequential move. [1] [2]

A second watch item is whether operating expenses remain contained relative to gross profit, because fiscal Q4 FY2026 operating expense guidance was $1.86B on a GAAP basis and $1.65B on a non-GAAP basis. [32] [3]

A third watch item is capital allocation: the declared quarterly dividend of $0.15 per share sits alongside continuing manufacturing investment, so shareholder returns and capex needs have to be compared directly. [33] [8]

Investor-relations prepared remarks and slides could not be retrieved from this host during the run, so the report relies on the official SEC exhibit release for its financial spine. [34]

The valuation debate should be framed around normalized earnings power rather than annualizing fiscal Q3 FY2026, because the reported quarter and the fiscal Q4 FY2026 guide both reflect exceptional conditions. [1] [2] [3]

A constructive scenario depends on guided revenue and gross margin converting into durable free cash flow, while a more cautious scenario assumes margins normalize as supply, pricing, or capital requirements change. [2] [21] [8]

The report avoids unsupported recommendation language; instead, it lays out what the next update must prove about demand durability, segment mix, and cash conversion. [6]

The largest fundamental risk is cyclicality: if fiscal Q4 FY2026 revenue or gross margin comes in below the guide, the market may reassess whether fiscal Q3 FY2026 represented peak conditions rather than a durable earnings base. [2] [21] [1]

Execution risk is tied to capital intensity, because the company is investing in a global manufacturing footprint while reporting large free cash flow in the current upcycle. [8] [5]

Micron investor-relations pages were unavailable, so prepared remarks and presentation retrieval are listed as a gap. [34]

Source notes are preserved as reader citations while operational retrieval details remain outside the public report narrative. [6]