Micron stock looks like a contradiction. The shares have multiplied many times over in the AI memory boom, yet the forward price-to-earnings ratio sits in the single digits. A stock that has run that hard and still trades that cheap on next year's earnings either is a rare bargain or is telling you something about where it sits in its cycle.
This article works through how to value Micron without being fooled by either signal. Micron is one of the most cyclical large companies in the market, so the method matters more than the headline multiple. If you have not read it yet, how to value memory and cyclical stocks lays out the framework this article applies to Micron specifically.
What Micron Actually Sells
Micron is one of only three companies that make the bulk of the world's DRAM, and one of a handful that make NAND flash. DRAM is the fast working memory in servers, PCs, and phones. NAND is the storage in solid-state drives and devices. That makes Micron a full-line memory maker rather than a specialist in one product.
The part of the business driving this cycle is data-center memory. The high-bandwidth memory used alongside AI accelerators is the tightest corner of the market, and Micron reports it inside its Cloud Memory and Core Data Center business units rather than as a separate line. So Micron is a direct beneficiary of AI infrastructure spending, but it is a commodity supplier into that boom, not a designer of the accelerators themselves. It sells what everyone needs and no one can easily differentiate.
That distinction is the whole reason valuation is hard here. A company with pricing power earns steady margins. A commodity maker earns whatever the supply and demand balance allows, which right now is a great deal and at other times is very little.
The Latest Quarter
Micron's most recent reported quarter, FQ2 of fiscal 2026 ended February 26, 2026, was the strongest in its history. These figures come from the company's SEC earnings filing.
| Metric | FQ2 FY2026 | Year earlier |
|---|---|---|
| Revenue | $23.86B | $8.05B |
| Revenue growth | +196% year over year | |
| GAAP gross margin | 74.4% | 36.8% |
| Non-GAAP diluted EPS | $12.20 | |
| Operating cash flow | $11.9B | |
| Adjusted free cash flow | $6.9B |
Revenue nearly tripled in a year and gross margin roughly doubled. The business units tell you where it came from:
| Business unit | Revenue | Gross margin |
|---|---|---|
| Cloud Memory | $7.75B | 74% |
| Core Data Center | $5.69B | 74% |
| Mobile and Client | $7.71B | 79% |
| Automotive and Embedded | $2.71B | 68% |
Management guided the next quarter higher still, to about $33.5B in revenue at roughly 81% gross margin. These are extraordinary numbers for a commodity business, and that is exactly the problem a valuation has to confront. Margins this far above the long-run average tend to come back down, because high margins invite the capacity additions that eventually erase them.
Why MU's Forward PE Looks So Low
As of late May 2026, Micron traded near $900 a share for a market capitalization around $1 trillion, and the forward price-to-earnings ratio sat in the single digits. On its face a single-digit forward multiple on a growing company is absurdly cheap.
It is also the classic signature of a cyclical at the top. The forward PE divides today's price by next year's expected earnings, and next year's earnings are being forecast at or near a cycle peak. When the denominator is a peak number, the ratio looks small even if the price already reflects how unsustainable that peak is. This is the trap covered in detail in : a multiple is only as good as the earnings figure underneath it, and for a cyclical that figure is the least stable number you can pick.
So the low forward PE is not evidence that Micron is cheap. It is evidence that the market expects current earnings to fall. The real question is how far they fall, and what the business earns on average once the cycle turns.
A Normalized Valuation Framework
To value Micron sensibly you have to replace peak earnings with normalized earnings, meaning what the business earns on average across a full cycle. The single most important input is normalized free cash flow per share, and reasonable people disagree about it by a wide margin. The table below is illustrative. It shows how the intrinsic value moves with the assumption, using a standard two-stage discounted cash flow.
| Scenario | Normalized FCF per share | Growth, years 1-5 | Discount rate | Illustrative DCF value |
|---|---|---|---|---|
| Bear | $5 | 2% | 12% | roughly $60 |
| Base | $12 | 4% | 11% | roughly $165 |
| Bull | $22 | 6% | 10% | roughly $340 |
Two things stand out. First, the value swings enormously with the normalized cash flow assumption, which is why this number deserves most of your attention. Second, every scenario here sits well below the late-May share price near $900. On a historically grounded normalization, even the bull case does not reach the current price.
That does not automatically make the stock overvalued. The bull argument is precisely that AI has permanently raised the floor under memory demand, so the true normalized free cash flow is far higher than any past cycle would suggest, perhaps close to the current run rate. If you believe that, you would plug a much larger normalized figure into the model and reach a much higher value. The point of the framework is to force that belief into the open. At the current price the market is implicitly assuming the elevated earnings last longer than a normal cycle would allow, whether because demand stays high for years, because the supply response is slower this time, or because the right discount rate is lower than history suggests. The framework does not tell you which of those is right. It just makes the assumption visible so you can weigh it for yourself. The discount rate that anchors each scenario can be reasoned through with , and the two-stage mechanics are the same ones in .
The Risks Are Real
The largest risk is the one the whole article circles. Micron is being valued on earnings that have never been sustained at this level for long. If AI memory demand cools, or if Micron and its two large competitors add capacity faster than demand grows, gross margins can fall from the mid-70s toward the long-run average in the 20s to 30s, and free cash flow can fall much faster than revenue because the cost base is heavy and fixed.
A second risk is capital intensity. Memory requires constant multibillion-dollar investment just to stay competitive, so reported free cash flow at the peak overstates what shareholders keep across a full cycle. A third is concentration. A few large customers and a few end markets drive most of the upside, and they can pull back together.
None of this says the stock falls. It says the margin of safety matters more here than almost anywhere, because the range of outcomes is so wide. That discipline is the subject of . For a comparison with a pure-play that carries even more cycle risk, see .
How to Run Your Own Micron Valuation
The honest output of this exercise is a range, not a price target, and the range depends mostly on the normalized cash flow you are willing to assume. The best way to use it is to test your own assumption rather than accept anyone else's.
Open the and replace the peak free cash flow with a mid-cycle figure you can defend, then watch how the intrinsic value responds. Compare it against the to see the same business through the multiple the market is quoting. If you want to start from a blank model and build the scenario yourself, the lets you enter any normalized input directly.
Frequently Asked Questions
Why does Micron have such a low forward PE if it is growing? Because the forward PE divides the price by next year's expected earnings, and those earnings are forecast at or near a cycle peak. A peak denominator makes the ratio look small even when the price already reflects an expected decline in earnings. For a cyclical, a low forward PE is often a warning rather than a bargain.
Is Micron a good AI stock? Micron is a direct beneficiary of AI memory demand, especially high-bandwidth memory for accelerators. But it is a commodity supplier rather than a differentiated designer, so its earnings rise and fall with the memory supply and demand balance rather than holding steady the way a company with pricing power would.
What is the right way to value Micron? Replace peak earnings with normalized earnings, meaning the average across a full cycle, and run a discounted cash flow on that figure rather than on the latest quarter annualized. The intrinsic value is highly sensitive to the normalized cash flow assumption, so that input deserves the most scrutiny.
Does Micron report HBM revenue separately? No. High-bandwidth memory is reported inside the Cloud Memory and Core Data Center business units rather than as its own line, so any specific HBM revenue figure is an estimate rather than a disclosed number.
Ready to test your own assumption? Open the and enter a normalized free cash flow instead of the peak figure.
