On Wednesday, 30 September, after the close, Micron reports its fourth quarter of fiscal 2026. The investor call is set for 2:30 p.m. Mountain time, which is 4:30 p.m. in New York.
And to see why this report is interesting, you first have to look at the previous quarter - because it does not resemble anything that has happened in the memory industry before.
Every figure in this article was taken from the company's own releases on its investor relations site.
Errors or inaccuracies are possible. Spotted something that looks wrong? Write to me and I will correct it.
In brief:
- Revenue grew 4.5 times in a year - from $9.3 billion to $41.5 billion.
- Gross margin: 84.6% - in an industry used to single-digit margins in bad years.
- And the company's own guidance for the coming quarter is higher still - $50 billion and a margin near 86%.
- And yet the forward multiple is 7.4 - the market is not pricing continuity.
What Happened in the Third Quarter
The quarter ended on 28 May 2026, and the report was published on 24 June.
| GAAP | Q3 2026 | Q2 2026 | Q3 2025 |
|---|---|---|---|
| Revenue | 41,456 | 23,860 | 9,301 |
| Gross margin | 35,056 | 17,755 | 3,508 |
| Percent of revenue | 84.6% | 74.4% | 37.7% |
| Operating expenses | 1,738 | 1,620 | 1,339 |
| Operating income | 33,318 | 16,135 | 2,169 |
| Percent of revenue | 80.4% | 67.6% | 23.3% |
| Net income | 28,243 | 13,785 | 1,885 |
| Diluted earnings per share | 24.67 | 12.07 | 1.68 |
In millions of dollars, except per-share amounts.
And the operating expense line explains the size of the number at the bottom. Revenue grew from $9.3 billion to $41.5 billion - 4.5 times - while operating expenses grew from $1.34 billion to $1.74 billion, that is, by 30%. This is a cost base that barely moves with sales, so almost every additional dollar of revenue lands intact at the bottom.
And the cash: operating cash flow of $25.39 billion in a single quarter, against $4.61 billion in the same quarter a year earlier. After net capital expenditure of $7.1 billion, adjusted free cash flow was $18.3 billion, and the company ended the quarter with $30.2 billion in cash and marketable investments.
Why 84.6% Is an Unusual Number
Memory is the classic commodity of the chip industry. The product is standardised, the makers are few, and the industry's history is one of brutal cycles: shortage sends prices up, everyone builds fabs, the supply arrives two years late, and prices collapse.
The number that illustrates this best is in the table above, in the rightmost column: in exactly the same quarter, one year earlier, the same company's gross margin was 37.7%.
Same company. Same fabs. Same product. A 47 percentage point gap in a year.
And the business unit breakdown shows this is not only data centres:
| Q3 2026 | Revenue | Gross margin | Operating margin |
|---|---|---|---|
| Cloud Memory | 13,769 | 83% | 78% |
| Core Data Center | 11,524 | 87% | 83% |
| Mobile and Client | 11,521 | 87% | 86% |
| Automotive and Embedded | 4,634 | 79% | 75% |
In millions of dollars.
The surprising line is "Mobile and Client" - memory for phones and PCs, not directly tied to artificial intelligence - at an 87% gross margin. A year earlier that same unit posted 24%.
And that says something important about what is happening here: if only data centres were strong, the story would be demand. When memory for phones also sells at these margins, the story is supply - global manufacturing capacity is being directed to the most profitable products, and everything else gets more expensive because of what is left.
What the Company Itself Says About Wednesday's Quarter
The guidance given on 24 June for the fourth quarter:
| GAAP | Non-GAAP | |
|---|---|---|
| Revenue | $50.0 billion ± $1 billion | $50.0 billion ± $1 billion |
| Gross margin | approximately 86% | approximately 86% |
| Operating expenses | approximately $1.86 billion | approximately $1.65 billion |
| Diluted earnings per share | $30.73 ± $1 | $31.00 ± $1 |
So the company itself expects margins to keep widening - from 84.6% to 86% - and revenue to grow another 20% or so in one quarter.
And for a sense of scale: earnings per share of $30.73 in a single quarter, against $1.68 in the same quarter a year earlier.
And What the Market Says About It
As of Friday's close, 25 September:
| Share price | $1,082.28 |
| Market value | $1.22 trillion |
| 52-week range | $154.65 - $1,255.00 |
| Revenue, trailing 12 months | $90.27 billion |
| Net income, trailing 12 months | $50.47 billion |
| Trailing price-earnings multiple | 24.4 |
| Forward price-earnings multiple | 7.4 |
| Annual dividend | $0.60, a 0.06% yield |
And the gap between the two multiples is the whole story.
A forward multiple of 7.4 against a trailing multiple of 24.4 says one thing: the market takes the coming year's earnings forecast and prices it at a third of the multiple it is willing to pay for earnings already reported. That is not the pricing of a growth company, and it is not the pricing of a company in distress - it is the pricing of a profit regarded as temporary.
And the dividend tells the same story from the company's side: $0.15 per share per quarter, on about 1.13 billion shares, is roughly $170 million - less than one percent of a single quarter's adjusted free cash flow. A company that believed this level of profit was permanent would raise the dividend. Micron is putting almost all of it back into capacity.
Four Things That Get Measured on Wednesday
1. Whether margins actually reached 86%. That is the company's own guidance, and it already assumes expansion. A miss here is not merely a missed quarter - it is the first sign that the cycle's turning point is close.
2. What exactly is in the Strategic Customer Agreements. The chief executive described them in June as multi-year, and said they would "significantly enhance the durability and predictability" of performance. That is a claim about reduced cyclicality - precisely what a multiple of 7.4 does not believe. The more that is disclosed about duration, volume and pricing mechanism, the more the claim can be tested.
3. Capital expenditure. $7.1 billion in a quarter, and rising. In this industry today's profit funds the supply that breaks the price two years out - so the investment pace of the industry as a whole, not Micron alone, is the variable that determines how long this lasts.
4. Guidance for the first quarter of 2027. This is where the real difficulty begins: the comparable quarter a year ago was already strong, so this is the first time the year-on-year comparison is not against a low base.
And What Is Already Known About the Products
Per the company's June release: HBM4, the fourth generation of high-bandwidth memory, is in high-volume shipments to the lead customer, and qualification samples have shipped to several end-customers. HBM4E is in development, with volume production expected in calendar 2027.
Also reported: qualification samples of 256GB memory modules, high-volume production of PCIe Gen6 solid-state drives, and the start of shipments of a 245TB drive.
And beyond products, three moves from recent months point the same way: strategic memory supply agreements with General Motors and with Ford, a strategic agreement with Anthropic, and an investment of up to $3 billion in strengthening the American supply chain. All of them try to convert cyclical demand into long-term commitment.
הזווית שלי
דעה אישית של אילן אברמוב - לא ייעוץ ולא המלצה
What I find interesting about Micron is that it has become the cleanest test of whether artificial intelligence changed an industry's structure or merely stretched a cycle.
In other companies the question is blurred. At an advanced logic foundry there is a technological moat that makes it hard to separate what comes from the product and what from the cycle. In memory there is no such blurring - the product is standardised, and the price is the price. So an 84.6% gross margin in memory is either a genuine change in market structure, or the peak of a cycle.
And what strikes me most is that the market refuses to rule in favour of the first. A forward multiple of 7.4 on a $1.2 trillion company whose profit grew fifteenfold in a year is an explicit statement: investors are paying for the profit that exists, and refusing to pay for its continuation. After decades of memory cycles, that is a position hard to call unreasonable.
And the thing I will be looking for on Wednesday is not the revenue line. Revenue will almost certainly be large - guidance said $50 billion, and the company's recent record is one of meeting its guidance. What I will be looking for is how much substance is given on the multi-year agreements - because that is the only thing that can settle the argument between 7.4 and 24.4.
And where I am careful: a big number is not a claim about direction. In a cyclical industry, a record profit is also the point from which the next comparison is hardest - and that is true regardless of what the price does.
(It is important to stress: this is my personal opinion only, nothing herein constitutes a recommendation to take any action, and there is no forecast here of the report's results.)
Ilan Abramov, the author, or anyone on his behalf may hold some of the securities or funds mentioned in this article.
