Micron’s $41 Billion Quarter and the Question of Cycle Timing

Micron’s fiscal third-quarter revenue was $41.46 billion, up 346% from a year earlier and up 74% from the quarter before. Non-GAAP gross margin hit 84.9%, a company record. Micron reports its next quarter, the one management guided to roughly $50 billion, on September 30. That’s days away as I write this, not months, which is part of why the timing question in this post’s title matters right now rather than in some abstract future.

Micron trades around $1,015.80 today, 23.0 times trailing earnings. On forward earnings it’s 6.8 times. A gap that wide between trailing and forward multiples means the market is pricing in a huge step up in profit, and my thesis is that the step up is real but the multiple already assumes it lands close to on schedule, which leaves less room for a slow quarter than the trailing P/E alone would suggest.

Memory chips have burned people so many times that the pattern is almost a running joke among anyone who has followed the sector for more than one cycle.

What actually happened last quarter

Prices spike, every manufacturer adds capacity, supply catches up to demand, prices fall, and companies that looked brilliant a year earlier report losses. Micron’s last full down year was fiscal 2023, when it lost $5.83 billion on revenue of $15.54 billion, a swing from the $8.69 billion it had earned the year before that. I checked that figure against Micron’s fiscal 2023 results filed with the SEC rather than trust it from memory, because a number that dramatic is exactly the kind that gets rounded and repeated until it drifts. It’s accurate. Now the company just posted a single quarter, $41.46 billion, that’s close to triple that entire down year’s full twelve months of revenue.

EPS for the quarter was $25.11, and Micron is sitting on $8.5 billion of net income for the trailing year with a net margin near 23%. Management guided the next quarter to about $50 billion, plus or minus $1 billion, with gross margin near 86%. An 86% gross margin is a number I associate with enterprise software, not a manufacturer that needs billion-dollar fabs to make anything at all. That’s exactly the kind of figure that makes me want to slow down rather than get more bullish, because margins that rich in a hardware business are historically a signal that supply is tight, not a permanent state of affairs.

Annual revenue for fiscal 2025, Micron’s last full reported year before this stretch of quarters, was $37.4 billion, itself up sharply from $25.1 billion in fiscal 2024. Annualize the most recent single quarter instead and you get a run rate near $165.8 billion, more than four times that full fiscal 2025 year. Revenue didn’t creep up over several years the way it usually does in a normal business cycle. It jumped in two or three quarters, which is exactly the kind of move that tends to overshoot in both directions.

Why this capacity cycle looks different so far

The engine behind all of this is high-bandwidth memory, the stacked DRAM that sits next to AI accelerators and feeds them data fast enough to keep expensive GPUs from sitting idle. Nvidia’s chips are the ones creating most of that demand, and HBM is a different product than the commodity DRAM that crashed the industry in past cycles. It requires more advanced packaging, longer qualification cycles with customers, and it locks in design wins a year or more ahead of shipment, rather than getting bid down in a spot market the way standard memory chips historically have.

Micron disclosed on its fiscal third-quarter call that it has signed roughly 16 long-term supply agreements worth close to $100 billion in minimum contracted revenue, based on the investor materials the company posted alongside that report. Contracted revenue isn’t the same thing as booked profit, and terms can include price adjustments, but a multi-year minimum commitment is a meaningfully different starting point than the last cycle, when Micron had almost no visibility past the current quarter.

Why does that matter for timing? Because the last two busts were driven almost entirely by spot pricing on commodity chips nobody had locked in. This one has a contractual floor under a growing share of the business. That floor doesn’t make Micron recession-proof. It does make the drop-off, if one comes, more gradual than the cliff investors remember from 2022 and 2023.

Micron isn’t the only supplier of HBM. Samsung and SK Hynix are both qualifying their own stacks with the same customer base, and I don’t have a verified, current market-share breakdown between the three that I’d trust enough to print here, so I won’t guess at one. What I can say directly from Micron’s own numbers is that its gross margin and revenue growth over the last several quarters suggest it isn’t losing the pricing fight, whatever the exact share split looks like today.

The balance sheet looks nothing like the last downturn

Micron held $8.5 billion of net income and closed the quarter with roughly $24.4 billion of net cash, a figure the company reported alongside its fiscal third-quarter results. Going into the last downturn, Micron was still working through debt taken on for prior fab expansions, and it had to keep spending through a period when revenue was cut in half. This time the company is funding its next round of capacity largely from a cash pile it built during the boom rather than debt raised at the top of a cycle, which changes how much pressure a slowdown would put on the balance sheet.

That distinction is worth thirty seconds of your time before you assume this is 2022 again.

A market that has quietly repriced the stock twice

C to A. Those are Micron’s own rating trend letters in the data I track, moving from the middle of the pack toward the top over the period this rally happened, which lines up with how the market has treated Micron’s last two reports. The last earnings move, on 2026-06-24, was +15.7%, against a typical swing of 8.1%, meaning the market rewarded that print by nearly double the usual reaction. The stock sits 19.0% below its 52-week high of $1,255 and is 558% above its 52-week low of $154, a low that dates back to before this cycle got going.

Short interest is only 2.6% of the float. That’s not a stock the bears have given up fighting reluctantly; it’s one where very few professional short sellers currently think the setup favors them, which I read as a data point, not a verdict.

Where the multiple actually sits

Trailing P/E of 23.0 looks unremarkable for a company growing revenue at 49% annually. Forward P/E of 6.8 is where the story gets aggressive: $149.85 of forward EPS implies earnings growth near 239% from here. Price-to-sales tells a similar story from a different angle, at 12.2 now against a five-year average of 5.2, well above Micron’s own history. Price-to-book is even more stretched, 10.9 against 3.2 historically.

Metric	Micron now	Five-year average	Read
P/E (trailing)	23.0	13.5	above average, cycle peak-ish
P/E (forward)	6.8	n/a	prices in a large earnings jump
P/S (TTM)	12.2	5.2	well above average
P/B (TTM)	10.9	3.2	well above average
Net cash	roughly $24.4 billion	n/a	funds capacity without new debt

The 29 analysts covering the stock average a target of $1,566, about 54% above today’s price, and 97% rate it a buy. I’ve made a version of this same argument about the broader semiconductor cycle hiding inside AI capital spending, and Micron is the cleanest single-stock expression of it: the multiple only works if the AI buildout keeps absorbing supply at something close to the current pace.

If the buildout pauses before the contracts do

Here’s my honest uncertainty. I don’t know, and nobody outside the biggest cloud providers’ capital budgets knows for certain, how much of current HBM demand is locked in for years versus how much is orders that could get pushed if one or two hyperscalers decide to slow their AI data center spending for a couple of quarters. Micron’s contracted revenue floor helps, but it doesn’t eliminate that risk, it just changes how the risk shows up.

The clearest way this thesis breaks is if data center capital spending growth slows sharply at Micron’s largest customers before those contracted minimums roll off, because a stock trading at 12.2 times sales against a 5.2 historical average has very little room for a growth scare on top of a valuation that already assumes the good scenario. I’d compare Micron’s setup to AMD and Nvidia sitting side by side on that same question: all three names are pricing in a buildout that keeps running near its current pace for multiple more years, and all three would reprice hard together if that assumption cracked.

The number due in days, not quarters

If Micron delivers close to the guided $50 billion with gross margin near 86% on September 30, the trailing multiple compresses fast because the E in P/E jumps, and 23.0 today could look cheap in hindsight within a single earnings report. If revenue comes in meaningfully under $49 billion or margin guidance slips below the mid-80s, I’d treat that as the first real evidence the supply-demand balance is loosening, and I’d expect the stock’s premium to its own five-year multiples to compress rather than the earnings catching up to the price.

That’s the one number I’m watching next Wednesday, not some vague sense of where the cycle is.

Analysis and opinion only, not investment advice. Figures come from Micron’s fiscal third-quarter 2026 results and SEC filings on SEC EDGAR and its investor site; valuation multiples are approximate and were checked on 2026-09-18 20:02:33.