Micron's next quarter is 14 weeks, not 13. Twelve filed lines to read before the print.

Every figure below was read from the primary document on sec.gov and is cited with its accession. Guidance figures are Micron's own, published in a Form 8-K exhibit; they are labelled as company guidance and are not our estimate. We show no analyst estimates and no consensus — not because they do not exist, but because they are not filed, and this page carries only what is. Ratios Micron does not print (days of inventory, days sales outstanding, per-week revenue, per-share averages) are arithmetic we did on filed values and are labelled where they appear. 13F figures were read out of each filing's information-table XML and reconciled to its cover-page total; nothing is summed across filers. Research, not investment advice.

Micron $MU reports fiscal fourth-quarter and full-year 2026 results tomorrow, 30 September, after the close — scheduled per the company's announcement. That announcement is not an SEC filing, so we will not quote a time for it.

Almost everything else that will be argued about tomorrow is already written down. The last 10-Q was filed on 25 June; the FY2025 10-K last October. Between them they fix the revenue split, the contract structure, the capex plan, the share count and the debt. The print does not create those facts. It confirms them or breaks them.

If you want a forecast, wrong page. Twelve lines already on file, each with its accession and the question it sets up.

1. The quarter is 14 weeks long, and it ended on 3 September

The line most write-ups will get wrong is in the first paragraph of MD&A (accession 0000723125-26-000015, quarter ended 28 May 2026):

Fiscal 2026 contains 53 weeks and fiscal 2025 contains 52 weeks. The third quarter of 2026 contains 13 weeks and the fourth quarter of 2026 contains 14 weeks.

The cover XBRL puts fiscal year end at 3 September 2026. (Ours: 28 May plus 98 days is 3 September, exactly fourteen weeks.) So the quarter being reported is 7.7% longer than the one before it and a week longer than the year-ago quarter it will be compared against. Micron's guidance table does not mention the week count; only the 10-Q does.

Check: the period-end date on the cover page, and whether any percentage you read has been adjusted for the extra week.

2. The guidance is Micron's own, filed, and specific

Micron publishes its outlook as a table in the earnings release furnished as Exhibit 99.1 to a Form 8-K — here, accession 0000723125-26-000013, 24 June 2026, headed "the following table presents Micron's guidance for the fourth quarter of 2026":

That is the whole table. There is no capex line in it; Micron's filed FY2026 capex figure lives in the 10-Q instead, and it is line 10 below. For scale, against the FY2025 10-K (accession 0000723125-25-000028): guided revenue for this one quarter is 33.8% above Micron's entire fiscal 2025 revenue of $37,378 million. (Ours.)

Check: reported revenue, GAAP gross margin, GAAP operating expenses and GAAP diluted EPS against those four figures. Any "versus expectations" framing tomorrow measures against something that was never filed.

3. Four quarters running, the print landed above the top of Micron's own range

This is the only guidance record that can be built entirely from filings, because both sides are filed: the guide in one quarter's Exhibit 99.1, the result in the next one's.

Ours, from those filed pairs: revenue came in above the top of the range by $315m, $843m, $4,760m and $7,206m. Against the midpoint the gap ran 5.7%, 9.1%, 27.6%, 23.7%. GAAP gross margin came in 3.7, 5.5, 7.4 and 3.6 points above the guided midpoint. We are not extrapolating that. The narrower point is that for a year the gap has been large and one-directional, which makes this guidance a floor-shaped object rather than a centre of gravity. Tomorrow is the fifth reading.

4. The revenue line is two numbers, and the sequential pair is the one to read

The year-over-year decomposition has been quoted everywhere, and we went through it on 26 September: DRAM up 343% on a low-260% range price rise and a low-20% range bit rise. From Revenue by Technology in the same 10-Q: DRAM $31,328 million against $7,071 million, NAND $9,943 million against $2,155 million, total $41,456 million against $9,301 million, income before taxes $33,212 million against $2,113 million.

The pair almost nobody quotes sits two paragraphs down the same page:

Sales of DRAM products increased 67%, primarily due to a low-60% range increase in average selling prices and a low-single-digit percentage range increase in bit shipments.

Sales of NAND products increased 99%, primarily due to a mid-80% range increase in average selling prices and a mid-single-digit percentage range increase in bit shipments.

Quarter on quarter, bits did almost nothing. (Ours, on the guide: $50.0bn over $41,456m is +20.6% sequentially, but per week it is $3,571m against $3,189m, or +12.0% — the extra week is about a third of the headline step.)

Check: whether the sequential DRAM sentence still reads as price with a single-digit bit number attached, or whether bits finally carry the load. That is the difference between a shortage and a boom, in one sentence.

5. Approximately $5 billion, and the sentence printed directly above it

Note 12 of the 10-Q discloses remaining performance obligations of "approximately $5 billion, of which $422 million has been recognized as contract liabilities", against "not material" a year earlier, under take-or-pay agreements whose largest contracts carry "a ceiling price for existing products that approximates the market price in the second calendar quarter of 2026, and a floor price through the term of the agreement". About one-third is expected to become revenue within twelve months.

The preceding sentence decides what the $5 billion means:

Our remaining performance obligations disclosure is based on minimum committed volumes and minimum pricing and is not expected to be indicative of future revenue under these contracts. Agreements without fixed pricing or price bands are not included in the remaining performance obligations disclosure.

Now the FY2025 10-K (0000723125-25-000028), written twelve months ago by the same people:

Substantially all contracts with our customers are short-term in duration at fixed, negotiated prices with payment generally due shortly after delivery. … As of August 28, 2025, our future performance obligations beyond one year were $143 million.

Check: the new RPO balance, the contract liability inside it, the twelve-month share, and whether the qualifying sentence survives unchanged. A number that grows while its own disclaimer softens is a different object from one that grows with the disclaimer intact.

6. $22 billion expected, $422 million booked

In Liquidity and Capital Resources rather than a note, the same 10-Q says Micron expects to receive "cash deposits and related financial commitments of $22 billion for agreements concluded to date", of which "approximately $18 billion … will be in the form of cash deposits."

Expects. What had landed as a contract liability was $422 million. The balance sheet shows where such money appears: other noncurrent liabilities $7,086 million against $1,443 million at the prior year end, and a $5,203 million contribution from that line inside nine-month operating cash flow of $45,702 million.

Check: whether the $22 billion moves, whether the contract liability moves toward it, and what other noncurrent liabilities do. Deposits are the difference between a contract and a press release, and they land in one specific line.

7. Inventory did not move. Receivables tripled.

Balance sheet, 28 May 2026 against 28 August 2025: inventories $8,567 million against $8,355 million; receivables $31,025 million against $9,265 million. Over nine months inventories were a $212 million use of cash and receivables a $19,953 million use.

Ours, since Micron files neither ratio. Days of inventory, on cost of goods sold: 180 in FY2023 (0000723125-23-000054), 166 in FY2024, 135 in FY2025 (0000723125-25-000028), 122 in the quarter ended 28 May. Days sales outstanding: 68 in FQ3-26 against 90 for FY2025. The receivable tripled in dollars while the collection period got shorter — price arriving in the numerator, not a collections problem. A tripling receivable is exactly the line that gets misread as a red flag.

Check: whether inventory dollars stay flat through a 14-week quarter at an 86% guided margin, and whether days of inventory keep grinding down. Inventory is this industry's early-warning instrument in both directions; FY2023 recorded $1,831 million of write-downs to net realisable value on the way down.

8. There is no HBM revenue line anywhere in the filings

"HBM" appears fifteen times in the latest 10-Q and thirty-five times in the FY2025 10-K. Not once attached to a revenue number. HBM sits inside the Cloud Memory Business Unit, defined as "memory solutions for large hyperscale cloud customers, and HBM for all data center customers" — which, per the FY2025 10-K, also carries "DDR, LPDDR, and GDDR".

So the filed HBM facts are narrow. FY2025 10-K: "In the fourth quarter of 2025, HBM3E 12-high represented the majority of our HBM shipments," and HBM4 "remains on schedule for volume production in calendar 2026." FQ3-26 release (0000723125-26-000013): "HBM4 … is in high-volume shipments for our lead customer's platform," with HBM4E "expected in calendar 2027." A roadmap commitment made in October and confirmed in June.

Check: any HBM revenue figure quoted tomorrow, and where it came from. Not from the 10-K, because the 10-K does not contain one.

9. The unit that is not the HBM unit had the better margin

Segment table, Note 17, with margins from the same quarter's Exhibit 99.1:

CDBU — mid-tier cloud, enterprise and OEM data centre, plus data-centre storage — grew fastest of the four and reported four points more gross margin and five more operating margin than the unit carrying HBM. FY2025 in full (0000723125-25-000028): CMBU $13,524m, CDBU $7,229m, so one quarter of CDBU is now 59% above its entire prior year. (Ours.) And from the same note: "Revenue from one customer was 10% and 16% … of total revenue for the first nine months of 2026 and 2025, respectively." Customer concentration fell while revenue tripled.

Check: whether CDBU crosses CMBU, and whether the margin ordering holds. A story about one product is not a story about four units, and the filing gives you four.

10. $27 billion of capex, against an $18 billion baseline filed in October

The 10-Q: "We estimate capital expenditures for property, plant, and equipment, net of proceeds from government incentives, to be approximately $27 billion in 2026." The FY2025 10-K, filed 3 October 2025, said: "approximately $4.5 billion in first quarter of 2026 … this level serves as a reasonable quarterly baseline for the planned capital expenditures for 2026." (Ours: four such quarters is roughly $18 billion.)

Actual nine-month spend: $19,602 million against $10,199 million a year earlier, offset by $2,989 million of government-incentive proceeds; property, plant and equipment, net, $56,426 million against $46,590 million. The programme is dated in both filings: Singapore HBM advanced packaging from the first half of calendar 2027, a second Singapore cleanroom in the second half of 2028, and the Tongluo fab in Taiwan acquired from Powerchip for $1.8 billion in March 2026.

Check: the FY2026 capex figure, any FY2027 figure, and those ramp dates. Supply arriving in 2027 and 2028 is the mechanism by which a price cycle ends, and it is on a published schedule.

11. Zero interest expense, $8.5 billion of debt retired, no new debt

The income statement shows interest expense as a dash for the quarter — zero — against $(123) million a year earlier, with $215 million of interest income. Debt carrying value fell to $5,722 million from $14,577 million at the prior year end. (Ours: down $8,855 million, 61%.)

Note 9 lists twelve prepayments between 24 October 2025 and 3 April 2026: principal down $(8,511) million, cash out $(8,985) million, and "we recognized losses in other non-operating income (expense) of $323 million and $500 million for the third quarter and first nine months of 2026, respectively." Proceeds from issuance of debt: zero, against $4,430 million a year earlier. Cash, equivalents and restricted cash closed at $25,022 million; total equity $100,724 million against $54,165 million.

Check: whether interest expense is still zero, whether new debt appears, and the closing cash number once deposits are inside it. Retiring debt at a premium during a price spike says something about a company's own view of duration.

12. The share count went up, and the buyback stopped

Shares outstanding 1,129 million, up from 1,122 million at the prior year end, on 1,275 million issued. Weighted-average diluted shares 1,145 million against 1,125 million.

No shares were repurchased in the third quarter of 2026. We repurchased 2.5 million shares of our common stock for $650 million in first nine months of 2026. Through May 28, 2026, we had repurchased an aggregate of $7.84 billion under the authorization.

$2.16 billion of the $10 billion authorisation remains. (Ours: $650 million for 2.5 million shares is about $260 a share on average, and the programme then went quiet.) Withholdings on vesting employee awards took $762 million over the same nine months — more than the repurchase programme itself.

One claim worth carrying into the print, from MD&A: Micron expects gross margins from the price-band agreements, "even at floor pricing levels, to yield gross margins well above our peak quarterly margins in any past cycle." The filing does not quantify that peak. (Ours, from Micron's own XBRL as filed in the FY2019 10-K, accession 0000723125-19-000094: the highest quarterly gross margin Micron has ever reported is 61.0%, for the quarter ended 30 August 2018 — $5,146 million on $8,440 million.) The FQ4-26 guide is approximately 86%.

What changed in the risk factors between the last two 10-Qs

We diffed Part II Item 1A of 0000723125-26-000006 against 0000723125-26-000015 in full. Four changes are substantive.

The ASP-volatility risk factor now carries the up-cycle inside it: new sentences state that "DRAM average selling prices increased approximately 140%" and NAND "approximately 130%" for the first nine months of 2026. The five-year range around them did not change.

A new paragraph makes data-centre construction an explicit dependency: build-out "requires significant energy capacity, water, and capital," and "any stakeholder opposition to data center development" or constrained "access to capital for our customers" could "delay customer and partner deployments."

A new passage handles the other side of take-or-pay: "if customers fail to meet their purchase commitments, we may need to enforce our contractual rights, which could result in litigation or disputes."

And one hedge was dropped: "we have incurred and may continue to incur costs in anticipation of demand that ultimately may not materialize" became "expect to".

Check: the 10-K's risk factors against this list. Risk-factor edits are the cheapest forward-looking signal in the document set, because counsel only adds them when the exposure has become real.

The downcycle is already written down, by the same filer

No projection is needed for the symmetric case; it exists in the same form from the same company. The FY2023 10-K (accession 0000723125-23-000054) records DRAM sales down 51% on a high-40s percent range price decline, cost of goods sold at 109% of revenue, $1,831 million of inventory write-downs, $382 million of fabrication-facility underutilisation costs, a headcount reduction expected "to approach 15% by the end of calendar 2023" — and the sentence that has aged most interestingly: "Due to volatile industry conditions, our customers are generally reluctant to enter into long-term, fixed-price purchase contracts." The full comparison is in the 26 September note; the downstream version is in the 18 June note on server OEMs.

The ownership view stops short of this quarter

There is no filed ownership picture of this quarter. Form 13F arrives 45 days after quarter end, so the most recent snapshot is 30 June 2026 and the September one is not due until mid-November. 3,927 filers reported a Micron position at 30 June.

What that snapshot shows is the same price-versus-bits pattern inside a portfolio. Appaloosa LP reported 975,000 Micron shares worth $1,125,432,750 at 30 June (0001656456-26-000003; 27 rows summing to the cover-page total of $7,725,383,349), or 14.57% of its reported book. At 31 March it held 1,665,000 shares worth $562,503,600, or 9.48% (0001656456-26-000002). (Ours: the share count fell 41% while the reported value doubled and the weight rose five points; implied per-share value $337.96 to $1,154.29.)

Situational Awareness LP (0000935836-26-000418) reports 26 positions totalling $20,242,292,228, equal to its cover page to the dollar, of which Micron is row 14 at 4,828,786 shares / $5,573,819,392 (27.54%) and SanDisk row 18 at 28.03% — ours: 55.57% of the book in two memory names. And a mechanical warning: Coatue Management (0000919574-26-005478; 211 rows, $48,629,053,706 matching its cover page) files seven separate Micron rows, totalling 3,142,269 shares and $3,627,089,685, or 7.46%. Read one line and you understate the position sixfold.

(What a 13F is not: US-listed long equity and options only — no shorts, no cash, no bonds, no non-US listings — and nothing after 14 August.)

Run it yourself

Start with the statements, not the headline. Micron's financials give you the standardised statements with the quarterly history attached, and /filings/MU takes you to the documents. Open the 10-Q, search the page for "average selling prices", and read the sentence next to it — that is line 4 above, and it takes ninety seconds.

/earnings/MU holds the filed quarterly series, so you can put tomorrow's four headline numbers beside the guided four from Exhibit 99.1. /holders/MU shows who reports owning the stock and /holders/MU/owns turns it around; every concentration figure above reconciles to its own 13F cover page to the dollar.

Then write down what you believe before the print. Our Thesis Monitor reads each new filing against what you already believed — approximately $5 billion of RPO with its own disclaimer, $22 billion of expected deposits against $422 million booked, approximately $27 billion of capex, 1,129 million shares — and tells you which of those moved. A quarterly headline cannot, because it does not know what you thought last quarter. If you would rather have the filings inside an agent, the same search and statement tools are on the EvidInvest MCP server. It is free to start, credit packs begin at $10, and there is no subscription.

The first thing we will read tomorrow is not the revenue number. It is the period-end date on the cover page, and then the sequential DRAM sentence in MD&A. One tells you whether the comparison is honest. The other tells you whether this is still a price cycle.

Sources. Micron FQ3 2026 10-Q, quarter ended 28 May 2026: accession 0000723125-26-000015 (sec.gov). FQ3 2026 Exhibit 99.1, furnished: 0000723125-26-000013 (sec.gov). FQ2 2026 10-Q: 0000723125-26-000006. Earlier guidance and results exhibits: 0000723125-26-000004, 0000723125-25-000044, 0000723125-25-000024, 0000723125-25-000019. FY2025 10-K: 0000723125-25-000028. FY2023 10-K: 0000723125-23-000054. FY2019 10-K, for the 2018 quarterly gross-margin comparatives: 0000723125-19-000094. Form 13F-HR, 30 June 2026: Appaloosa 0001656456-26-000003 and 0001656456-26-000002, Situational Awareness 0000935836-26-000418, Coatue 0000919574-26-005478. No analyst estimate or consensus figure appears anywhere on this page. Days of inventory, days sales outstanding, per-week revenue, per-share averages, percentage gaps to guidance and the 13F weights and share-count changes are arithmetic on those filed values, labelled as ours where they appear. Research, not investment advice.