Micron Reports $54.2 Billion in Quarterly Revenue, Beating Estimates, and Forecasts $61.5 Billion for Next Quarter

The Record Quarter That Left the Stock Barely Moving

Micron Technology closed the books on its best quarter ever on September 30, 2026. Revenue hit $54.23 billion for the three months ending September 3 — up 379% from a year earlier and 31% from the quarter before that^[1]. Adjusted profit came in at $33.42 a share, beating the roughly $31.61 that analysts had expected^[2][10]. The company then forecast $61.5 billion in revenue for the current quarter, far above the $56.8 billion Wall Street had penciled in^[5].

By any normal measure, that's a blowout. Yet Micron's stock barely twitched after hours, moving only about 1% to 2% in either direction depending on the moment you checked^[5][6]. Different outlets looked at that same flat trading and wrote completely different headlines. Stocktwits said the stock "drops," pointing to Micron's plan to spend heavily on new factories^[5]. TradingKey said shares "reverse losses to rise 2%," pointing to the same guidance^[6]. Both are describing the identical few minutes of trading.

That gap between the numbers and the reaction is the real story here. A company doesn't post a 379% jump in sales and get a shrug for no reason. The explanation sits in a single word that came up constantly on the earnings call: memory.

What AI Chips Actually Need, and Why Micron Suddenly Has So Much of It

Micron makes memory chips — the parts that store and feed data to a computer's processor, as opposed to the processor itself. For years this was treated as a commodity business, cheap and cyclical. That's changed because of a specific product called HBM, or high-bandwidth memory. HBM stacks memory chips directly on top of each other and places the stack right next to an AI processor, so it can feed data fast enough to keep that processor busy^[4]. Without enough HBM, an expensive AI chip sits idle waiting for data. That makes memory a bottleneck AI companies will pay a premium to avoid, not a cheap commodity they shop around for.

Only three companies make most of the world's DRAM, the chip family HBM belongs to: Samsung, SK hynix, and Micron^[12][15]. As all three shift factory space toward HBM for AI data centers, less ordinary memory is left over for phones, laptops, and other everyday devices^[15][16]. DRAM revenue at Micron hit a record $39.8 billion this quarter, 73% of total sales and up 343% from a year ago^[1]. NAND, the other major memory type Micron makes, actually grew even faster — up 526% to a record $14.1 billion^[1]. That matters because some coverage credited the quarter to "DRAM and HBM," when NAND was the faster-growing piece.

The payoff shows up in Micron's margin — the share of each sales dollar left over after the cost of making the chips. It hit 87.0% on an adjusted basis this quarter, compared with 41.1% for the entirety of fiscal 2025^[5]. That's an extraordinary jump for a business that used to run on thin, boom-and-bust margins.

The Trough That Isn't Supposed to Be a Peak

Here's where the story gets genuinely contested. Micron guided next quarter's margin down slightly, to about 86.25%, and tied the dip to roughly $1 billion in extra costs^[4][5]. CEO Sanjay Mehrotra called that the low point for the entire 2027 fiscal year, not the start of a decline^[4]. Micron also said it has already locked in contracts for the vast majority of its 2027 HBM supply, at sharply higher prices than a year ago^[4].

Skeptics hear something different in that same guidance. Memory has a well-worn boom-and-bust pattern: high prices draw heavy factory spending, that new supply floods in all at once, and prices crash. The Motley Fool put it bluntly — no memory maker has ever held an 86% margin for long^[13]. Micron is now planning to raise capital spending sharply, with roughly $25 billion going into new factories in just the first half of fiscal 2027 alone^[5][21]. Bloomberg data cited in the coverage show expected revenue growth slowing from about 247% this past year to 96% next year and just 12% the year after that^[22]. To skeptics, a shrinking margin and a muted stock reaction look like the market already pricing in a peak^[7].

Both sides are reading the same facts through a different lens on timing. Micron's contracts argument is real: locked-in prices for most of 2027's HBM supply do make next year's revenue more predictable than in past cycles^[4][20]. The skeptics' history argument is also real: the industry has never escaped this pattern before, and Micron and its rivals are now ramping spending together, which is exactly the mechanism that has crushed prices every previous time^[5][22]. Neither side has proof the other is wrong yet, because the test is 2027 and 2028, not today.

Why a Shortage That Enriches Micron Also Raises Your Phone Bill

Lost in most of the earnings coverage is a simpler, more immediate effect: this same memory squeeze is making ordinary consumer electronics more expensive. Axios and The Register, tech-focused outlets rather than markets outlets, reported that AI data centers are absorbing so much memory supply that phones, PCs, and game consoles are seeing prices climb^[15][16]. That's the flip side of Micron's record margin — the shortage that lifts the company's profit is the same shortage pushing up the price of a new laptop.

Mehrotra has pushed back on the idea that memory makers are to blame. He's argued that customers themselves spent years pressuring suppliers for the lowest possible prices, which discouraged the factory investment that might have prevented today's shortage^[14]. In other words, cheap memory in the past meant less capacity built, which now means a tighter market today.

Either way, relief isn't coming soon. New factory capacity takes years to build. Micron's first new fab in Idaho is targeting its first chips in mid-2027, while a planned New York facility isn't expected to produce anything until around 2030^[17][18]. Some projections suggest the DRAM shortage could last until 2028^[23]. Until new supply arrives, AI data centers and consumer electronics makers are competing for the same limited pool of chips.

A Preview for Korea, and a Split Screen in the Coverage

Because Micron reports before its two main rivals, South Korean outlets read its results as a preview of what's coming for Samsung and SK hynix. Korea JoongAng Daily and Herald Business both framed the quarter as evidence that fears of a memory-cycle peak were overblown, and reported that analysts had pushed their combined third-quarter operating profit forecast for the two Korean companies up to roughly 188 to 190 trillion won — something like $140 billion — from an earlier estimate near 180 trillion won^[10][11]. SK hynix still leads specifically in HBM, holding a 56.4% share of that market as of the first quarter of 2026^[10].

The broader coverage split along familiar lines. CNBC led with the straightforward numbers and the data-center revenue jump, in a mostly neutral, investor-focused frame^[2]. ZeroHedge called the 86.25% margin guide a "miss," even though other outlets described the same figure as beating a different consensus estimate^[7]. A Polish brokerage, XTB, described the quarter in the most promotional terms of any outlet reviewed, reporting that Micron's site "crashed" from investor traffic and that shares "surged" — language that doesn't match the roughly 1% to 2% moves others recorded^[8]. None of these framings is factually wrong on its own; they're each picking which true thing to emphasize.

What nobody disputes is the scale of what Micron actually sold: $54.23 billion in memory chips in three months, with 87 cents of every dollar left over after production costs^[1][5]. What's unresolved is what happens once the new factories Micron is building now actually start shipping, sometime in 2027 and beyond, and whether locked-in contracts are enough to break a cycle the industry has never broken before.