Micron Just Printed the Best Quarter in Memory History. The Stock Got Cheaper Without Moving.

Record revenue, an 87% gross margin and a guide to $61.5 billion. The share price barely twitched, and the valuation reset anyway. Here is the measure that did the work, and why it matters more than the forward multiple everyone is quoting.

HeyTheo Research · Thursday, October 1, 2026

Quick Read

• The quarter was not a beat, it was a different category of number. Revenue of $54.23 billion, up 31% sequentially and 379% on the year, with a non-GAAP gross margin of 87.0% and non-GAAP EPS of $33.42 against a $31.16 estimate.

• The guide went higher, not lower. Q1 FY2027 revenue of $61.5 billion plus or minus $1.5 billion, and non-GAAP EPS of $38.15 plus or minus $1.00.

• The stock did nothing. It closed at $1,066.10 on the day, slipped 0.78% after hours, and sat at $1,065.11 the next session. It remains 15.1% below a 52-week high of $1,255.00 that was set before this print.

• Price-to-book fell from 12.1x to 8.7x in six days without the share price moving. The quarter added roughly $37.6 billion to shareholders' equity. Book value per share is now $122.46 and the multiple reset itself.

• The forward P/E of 6.6x is the most quoted number and the least useful one. In memory, a low forward multiple at a cycle peak has historically been a countdown rather than a discount.

Micron reported fiscal fourth-quarter results on Wednesday evening and the numbers do not look like a semiconductor company's. They look like a royalty.

Revenue was $54.23 billion for a single quarter, against $41.5 billion in the prior quarter and a company guide of $50 billion. Gross margin was 86.8% on a GAAP basis and 87.0% adjusted. Operating margin was 80.7% GAAP. Full-year revenue came to $133.19 billion with net income of $84.97 billion, which is a net margin of 63.8%.

For scale, Micron's previous cycle peak came in fiscal 2022 on revenue of $30.8 billion, net income of $8.7 billion and a gross margin of 45.2%.

And then the stock closed up 0.1% and gave a little back after hours.

That gap between the result and the reaction is the whole story, and the explanation is not sentiment. It is arithmetic.

What the print actually said

Source: Micron fiscal Q4 2026 results release and earnings call, September 30, 2026.

Two details inside that table deserve more attention than they got.

The first is that the revenue increase came from price, not volume. DRAM bit shipments grew in the mid-single digits sequentially while DRAM prices rose by a high-teens percentage. NAND bits grew about 10% while NAND prices rose roughly 30%. Revenue climbed 31% in the quarter, and the large majority of that came from the price line.

That matters because price and volume behave differently over a cycle. Volume compounds. Price mean-reverts.

The second is that gross margin is guided down. The 87.0% in the quarter becomes roughly 86.25% in the guide, a sequential decline of about 75 basis points, even as revenue is guided up 13.4%. Management characterised that level as the floor for gross margins in fiscal 2027, which is a fair and important framing. It is also a company assertion about the most cyclical line item in technology.

The measure that moved while the price stood still

On September 25, Micron traded at $1,082.28 and 12.13 times book value. That implied book value per share of about $89.22 and total equity near $100.8 billion.

On October 1, Micron trades at $1,065.11 and 8.70 times book value. Book value per share is $122.46 and total shareholders' equity is $138.38 billion.

The share price fell 1.6% over those six days. The book multiple fell 28.3%.

The difference is retained earnings. Fourth-quarter net income of roughly $37.1 billion went almost entirely into equity, because Micron pays a dividend yielding 0.06%. The company earned its way to a cheaper valuation without the market doing anything.

This is the part of a memory cycle that is genuinely hard to hold in your head. On an earnings multiple, the stock looks cheap because the earnings are enormous. On a book multiple, it was expensive and is now merely high. And book value is the only one of the two that does not require you to forecast where in the cycle you are standing.

Run that forward. Consensus has Micron earning $161.49 per share in fiscal 2027 on revenue of $250.52 billion. If the company retains essentially all of it, as it has, book value per share reaches roughly $283 by the end of August 2027. At today's price, that is about 3.8 times the book value the consensus already assumes.

A stock at under four times its own forecast book value is not an expensive stock. It is also not what a bubble looks like.

The two cycles that ended differently

The bear case does not need the guide to be wrong. It needs the year after the guide to be wrong, and there is precedent in both directions.

Source: Micron fiscal results history as reported in cycle reviews published August and September 2026.

One cycle was slow and shallow and never produced a loss. The other took revenue down 49% in a single year and turned an $8.7 billion profit into a $5.8 billion loss, because the industry ended up selling chips for less than they cost to make.

Both share a feature. Most of the peak's earnings were gone within roughly two years of the top.

Apply the gentler of those two patterns to fiscal 2026 and net income falls to the mid-teens of billions by fiscal 2028. On 1.13 billion shares that is an EPS in the low teens, and at $1,065 a share the multiple on it is in the seventies. Apply the harsher pattern and there is no multiple, because there are no earnings.

That is the entire argument against a 6.6 times forward multiple, and it cannot be dismissed by pointing at the guide. The guide covers one quarter.

What the sell-side disagreement is really about

Forty-nine analysts cover Micron. Thirty-six rate it Strong Buy, nine Buy, four Hold, and none Sell. The consensus target is $1,534.

The highest target is $2,200. The lowest is $361.

A $361 target on a $1,065 stock is a 66% decline. It sits in the same coverage list as a target implying a 107% gain. That is a 6.1-fold spread between the extremes, which is not a rounding difference in assumptions. It is two different models of what memory now is.

Translate the targets into book multiples and the disagreement becomes legible:

Source: analyst target range and consensus as compiled October 1, 2026; book value from the Q4 FY2026 balance sheet; forecast book from consensus FY2027 EPS retained in full.

Read that way, nobody is arguing about this year. They are arguing about what multiple of equity a memory maker deserves once the cycle turns, and the honest answer is that it depends entirely on whether the earnings that built the equity were a one-off or a new baseline.

The non-obvious reasons this cycle might not rhyme

Three things are structurally different from 2018 and 2022, and they all cut toward a longer cycle rather than a shorter one.

Supply is committed rather than sold spot. More than 75% of fiscal 2027 output is already committed through customer agreements, and Micron has agreements covering the vast majority of calendar 2027 high-bandwidth memory bits at significantly higher year-on-year prices. In prior cycles, the collapse arrived through spot pricing.

The balance sheet cannot force a capitulation. Total debt is $5.18 billion against $138.38 billion of equity, a debt-to-equity ratio of 0.04, with $73.48 billion in cash, marketable investments and restricted cash. In 2023 the industry cut output partly because it had to. This balance sheet does not have to do anything.

New capacity arrives late. Idaho's first fab begins wafer output in mid calendar 2027, the second in late 2028, Singapore's high-bandwidth memory facility in early 2027, and the New York fab not until calendar 2030. Supply responses in memory are historically what ends the party, and most of this one shows up after 2027.

Against that, the single most reliable fact in the sector's history is that an 87% gross margin is not a resting state. The question is not whether it normalises. It is whether normalisation means 60% or 20%, and the answer to that question is worth roughly the entire gap between the $2,200 target and the $361 one.

The rules HeyTheo tracks

• Basket: hold Micron against the memory complex and against the hyperscaler capital-expenditure names at the same time. It is a supplier to the AI build and a commodity producer, and those two identities get priced very differently when the cycle turns.

• Triggers: flag each DRAM and NAND contract price print, Micron's quarterly gross margin against the stated fiscal 2027 floor, and any change to the committed-output percentage.

• Money flow: watch whether rotation into the name comes with the book multiple rising or falling. A rising price on a falling book multiple is the cycle paying for itself. A rising price on a rising book multiple is re-rating, and re-rating is what reverses.

• Ask Theo: pull the gross margin history for any covered semiconductor name next to its current price-to-book before treating a low forward multiple as cheapness.

• Check the rule behind any trigger before acting on it. You trade through your own broker; HeyTheo helps you decide.

The Denominator Problem

A forward price-to-earnings ratio is a fraction, and nearly all of the argument about Micron lives in the bottom half of it.

At 6.6 times fiscal 2027 consensus, the stock is one of the cheapest large caps in the US market. At 8.7 times book, it is priced like a high-quality compounder. At 3.8 times the book value the same consensus says it will hold next August, it is priced like something the market does not quite believe.

All three are true at once, and they are true because the earnings number is the one input nobody can anchor. Equity is a fact. Peak EPS is a forecast about where you are standing in a cycle, made by people who have been wrong about exactly that twice in a decade.

Our September 28 value screen flagged Micron as a name that passed one valuation test and failed the other. That was the right read, and the quarter has narrowed the gap rather than closing it. The test it failed has moved from 12.1 times book to 8.7, not because the price corrected but because the company retained $37.6 billion in ninety days.

If you want one thing to watch, it is not the forward multiple. It is whether the next two quarters hold the margin floor management has now put a number on. That floor is the only thing standing between a six-times earnings multiple and a seventy-times one.

FAQs

Why did Micron shares not rise on a record quarter and a higher guide?

Because the result was close to what the market had already priced. The stock had run to a 52-week high of $1,255.00 before the print and sat 15% below that level going in, which tells you expectations were elevated and had already begun to deflate. A 7% EPS beat against that backdrop is not new information. Price reactions measure surprise, not quality.

Is a forward P/E of 6.6x cheap for a semiconductor company?

It is low in absolute terms and normal for memory at a cycle peak. The sector has repeatedly traded at single-digit forward multiples on peak earnings shortly before those earnings fell by 80% or more. A low forward multiple in memory is therefore closer to a statement about where the cycle is than about how much value is on offer.

Why does price-to-book matter more than usual here?

Because book value does not require a forecast. Micron's equity of $138.38 billion is a reported figure, while fiscal 2027 earnings of $161.49 per share are an estimate that assumes the cycle holds. When the two measures disagree as sharply as they do now, the one that depends on fewer assumptions is the more useful anchor, even though it is the less flattering one.

What would signal that the cycle is turning?

Contract pricing is the leading indicator rather than the share price. Specifically, a quarter in which bit shipments grow but revenue does not, which means price has started falling faster than volume is rising, and gross margin coming in below the fiscal 2027 floor management has guided to. Both would appear in Micron's own reporting before they appear in consensus estimates.

Does Micron's contract book protect it in a downturn?

Partly. More than 75% of fiscal 2027 output is committed and a majority of the longer agreements carry pricing bands with floors. That is genuine protection and is new relative to previous cycles. The limits are that commitments are not cash until delivered, that a floor only holds if the customer honours it when spot prices sit far below, and that fiscal 2028 and beyond are far less covered than fiscal 2027.

Sources

Sources: Micron Technology investor relations, GlobeNewswire, Investing.com, Benzinga, CNBC, StockAnalysis, Yahoo Finance, The Motley Fool, FXStreet.

Disclaimer

Disclaimer: HeyTheo is a research and education platform, not an investment adviser or broker-dealer. Nothing here is advice to buy, sell, or hold any security. You trade through your own broker; HeyTheo helps you decide. Backtested results are hypothetical and do not guarantee future returns. References to governments, officials, or policies are for market context only and are not political endorsements. All investing involves risk, including loss of principal. Data is as of the dates noted.