Micron FQ2 FY2026: The Best Quarter in Memory History, and Still a Hold

Results vs. Consensus

Quality of Beat

• Revenue: Entirely price. DRAM revenue rose 74% sequentially on bit shipments up only mid-single-digits and prices up in the mid-60s percent; NAND rose 82% on low-single-digit bits and prices up in the high-70s. A near-tripling of revenue year-on-year driven almost wholly by price is the signature of a market at the absolute peak of its pricing leverage, not of unit demand.
• Margins: The 74.9% gross margin rose 18 points sequentially and roughly doubled year-on-year, a company record by a wide margin. There is no one-time item; it is pricing, pure and simple. The same purity is what makes it cyclically fragile, because price is the most reversible input in the model.
• EPS: The beat was almost entirely operational, with a clean 15.1% tax rate and a near-zero gap between GAAP ($12.07) and non-GAAP ($12.20). Cash quality was pristine: $11.9B operating cash flow, $6.9B record free cash flow, and a return to a $6.5B net cash position, the strongest in the company's history.

Segment Performance

Revenue by Technology

Revenue by Business Unit (all records)

The whole portfolio at once-in-a-cycle margins

Every business unit set a revenue record and posted a gross margin in the high-60s to high-70s. Mobile & Client, the consumer-facing unit, earned a 79% gross margin, the highest in the company, on pricing alone with bit shipments actually down. Core Data Center's margin jumped 23 points to 74%. The driver across all four was the same: price, not volume.

"Mobile and Client Business Unit revenue was a record $7.7 billion... MCBU gross margins were 79%, up 25 percentage points sequentially, driven primarily by higher pricing and favorable mix." — Mark Murphy, CFO

Assessment: A 79% gross margin in the consumer unit, achieved while volumes fell, is the clearest possible statement that pricing has overshot what end demand can absorb. When the lowest-value segment earns a near-record margin on declining bits, the margin is being set by scarcity, not by value, and scarcity-driven margins are precisely the kind that revert hardest when supply finally arrives.

Key Topics & Management Commentary

Overall Management Tone: Maximally confident, and structurally so. Management framed the quarter as the product of an AI-driven multiyear investment cycle "most of which is ahead of us," leaned into the first signed five-year SCA, raised the dividend, and guided to an 81% gross margin without hedging. The one notably candid admission, buried in a margin answer, was that at these levels incremental price increases have less effect on margin, a quiet acknowledgment that the easy part of the margin expansion is over.

1. An 81% Gross-Margin Guide

Fiscal-Q3 gross margin was guided to ~81%, up roughly 600 bps sequentially, on price, cost, and mix. Management declined fiscal-Q4 guidance but pointed to tight conditions beyond 2026, while acknowledging that at these margin levels, further price increases move the margin less.

"At these gross margin levels, incremental increases in price are going to have less of an effect on gross margin." — Mark Murphy, CFO

Assessment: This is the most important sentence on the call. An 81% gross margin is mathematically near a ceiling, and management said as much. From here the margin cannot expand the way it has; it can only hold or fall. For a stock that has been driven by relentless upward margin revisions, the disappearance of that lever, even at a spectacular level, changes the forward setup materially.

2. The First Five-Year SCA

Micron signed its first strategic customer agreement, a multiyear (five-year) contract with specific commitments from both sides, distinct from the typically one-year LTAs. Management framed SCAs as bringing stability and visibility, is in discussions with multiple customers across markets, but declined all specifics, citing confidentiality.

"We are excited to have signed our first five-year SCA... these SCAs are multiyear and they have specific commitments in them. These are robust agreements, and these are meant to give us the visibility and stability toward our business model." — Sanjay Mehrotra, CEO

Assessment: This is genuinely the most important structural development of the quarter and the strongest argument for a higher rating. A five-year contract with binding commitments is exactly the kind of mechanism that could de-cyclicalize memory. The catch, examined below, is that management would not say whether the agreement protects margins on the downside, which is the only thing that would actually change the cyclical math.

3. Pricing Up 60-80% in a Single Quarter

DRAM average prices rose in the mid-60s percent sequentially and NAND in the high-70s, the dominant driver of the entire revenue and margin beat. Management attributes this to structural supply constraints meeting AI-driven demand.

Assessment: Sequential price increases of this magnitude are without precedent and, by definition, cannot repeat many times. They are wonderful while they last and they are the very thing that, historically, triggers the demand destruction and supply response that ends a cycle. A 65-80% quarterly price jump is not a sign of a durable plateau; it is a sign of a market that has gone vertical.

4. Non-HBM Margins Now Exceed HBM Margins

In a striking reversal, management confirmed that non-HBM (commodity DRAM) margins are now higher than HBM margins, a complete inversion of the cycle's earlier logic where HBM was the premium product carrying the blended margin.

"It is correct that the margins for non-HBM today are higher than HBM margins." — Sanjay Mehrotra, CEO

Assessment: This is a subtle but powerful late-cycle signal. When plain commodity DRAM out-earns the most advanced product in the portfolio, it means the commodity is being priced by acute scarcity rather than by value. That is the textbook condition that precedes a supply response, because every competitor can see that the easiest money is now in the most replicable product.

5. Demand Destruction Begins

For the first time, management guided that memory supply constraints, and the resulting high prices, could push PC and smartphone units into a low-double-digit percentage decline in calendar 2026, even as per-unit memory content rises.

"In calendar 2026, a number of factors including DRAM and NAND supply constraints could cause PC and smartphone units to decline in the low double-digit percentage range." — Sanjay Mehrotra, CEO

Assessment: This is the cycle showing its first visible crack. Prices have risen so far that they are now destroying end-unit demand in the consumer markets. Demand destruction is the mechanism by which every memory boom eventually self-corrects: high prices ration the product until buyers walk, and then the correction follows. Management framed it as a content-growth offset, but the admission itself is significant.

6. Capex to Above $25B, Fiscal 2027 Higher Still

Fiscal-2026 capex was raised again to above $25B (from $20B), driven mostly by cleanroom and facilities, including the newly closed Tongluo acquisition from Powerchip and U.S. fab construction. Fiscal 2027 capex will "step up meaningfully," with construction alone rising more than $10B year-on-year plus higher equipment spend.

"We expect fiscal 2026 CapEx to be above $25 billion... We project our fiscal 2027 CapEx to step up meaningfully... construction-related CapEx to increase by over $10 billion year over year in fiscal 2027." — Sanjay Mehrotra, CEO

Assessment: Here is the supply response, accelerating. A capex bill above $25B rising materially again in fiscal 2027 is the industry building the capacity that today's 81% margins demand, and that capacity lands in 2027-2028. Micron is funding it from a position of strength, but the aggregate industry build is the single clearest leading indicator that the back half of this cycle is being constructed right now.

7. The Balance Sheet and the Cash Question

Record free cash flow drove a return to $6.5B net cash, the strongest ever, alongside two credit-rating upgrades to a solid triple-B, a 30% dividend increase, and a $350M buyback. Management flagged return on capital above 30% heading toward 50% and meaningful future capacity for buybacks.

"We received two credit upgrades in the quarter, so we are now a solid triple-B... we believe we will have significant capacity for returning cash to shareholders through repurchase." — Mark Murphy, CFO

Assessment: The cash story is genuinely excellent and is the reason a downgrade here would be unwarranted; this is a Hold, not a sell. A net-cash, triple-B Micron generating tens of billions of free cash flow can fund the capex, raise the dividend, and stockpile buyback capacity for the other side of the cycle. The capital-return optionality is real and is a meaningful support under the stock.

8. HBM4 Ramps Into Vera Rubin

Micron began volume shipment of HBM4 36GB 12-Hi for NVIDIA's Vera Rubin, sampled a 16-Hi (48GB) part, and expects HBM4 to reach mature yields faster than HBM3E, with HBM4E (on 1-gamma) ramping in calendar 2027. The secular HBM roadmap continues to execute.

Assessment: The HBM franchise remains the durable core that justifies owning Micron through a cycle, and the faster-than-HBM3E yield ramp is a real execution win. It is the foundation of why we would re-engage at the right price, even as the near-term cyclical math keeps us on the sidelines today.

Guidance & Outlook

The fiscal-Q3 guide is, on its face, the most aggressive in the company's history: revenue up 40% sequentially to $33.5B, gross margin to ~81%, and EPS to $19.15. Management noted that single-quarter guided revenue now exceeds Micron's full-year revenue for every year through fiscal 2024, a useful measure of how vertical the trajectory has become.

Implied trajectory: Annualizing the $19.15 fiscal-Q3 EPS yields a run rate near $77. The market is being handed an explicitly guided peak, not asked to forecast a recovery.

The margin ceiling: Management's own admission that incremental price now moves margin less, combined with the refusal to guide fiscal Q4, signals that the rate of upward revision is about to slow even if the absolute level holds. For a stock propelled by ever-rising estimates, the second derivative matters.

Capex offset: The above-$25B capex bill and the fiscal-2027 step-up are the counterweight to the cash flow. The company can afford it, but it is the supply response that the current margins are financing.

Analyst Q&A Highlights

The sustainability of an 81% gross margin

The opening question pressed directly on whether the 81% margin can hold, particularly as more HBM4 enters the mix. Management pointed to tight conditions beyond 2026 and an AI investment cycle still ahead, but made the key concession that incremental price increases now move the margin less and declined to guide fiscal Q4.

Q: "The 81% gross margin guide is very impressive. Just kind of curious how to think about the sustainability of gross margins, especially as you bring more HBM4 into the mix... how to think about gross margins in the August quarter and beyond?"
— Krish Sankar, TD Cowen

A: "We are not going to provide the fourth quarter gross margin guidance. However, we have indicated that we expect the market conditions to remain tight beyond 2026... keep in mind that at these gross margin levels, incremental increases in price are going to have less of an effect on gross margin."
— Mark Murphy, CFO

Assessment: Management defended the level but conceded the slope. "Tight beyond 2026" supports the absolute margin; "incremental price has less effect" concedes the expansion is nearly done. For a stock driven by upward revisions, that concession is the tell that the easy part is behind us.

Whether the SCAs protect margins on the downside

The most important question of the call asked whether the SCAs contain a mechanism that would limit gross margin on the downside when the cycle eventually turns. Management declined to confirm any such mechanism, citing confidentiality, and reiterated only that the agreements are multiyear with specific commitments.

Q: "We are all trying to think to the other side of the cycle and hope that these SCAs provide some mechanism that will kind of limit your gross margin on the downside to a certain number... is it fair to say that there is a mechanism in these SCAs that would limit your gross margin on the downside when things do finally roll back over?"
— Timothy Arcuri, UBS

A: "We are certainly not getting into the specifics of these SCAs for the obvious reasons of confidentiality... these SCAs are multiyear and they have specific commitments in them. These are robust agreements... Beyond that, I cannot get into any specifics at this point."
— Sanjay Mehrotra, CEO

Assessment: This is the exchange that keeps us at Hold. The single thing that would justify de-cyclicalizing the multiple, a contractual floor under margins, is exactly what management would not confirm. Until an SCA is shown to protect the downside, it improves visibility on volume without changing the cyclical risk to price, which is what actually drives the earnings.

Demand destruction and end-market allocation

A question asked whether management worries about demand destruction in PCs and smartphones and how it balances allocation across customers. Management acknowledged price-sensitive consumer demand is being impacted by higher prices, while reaffirming a diversified-supplier strategy and noting it still fills only 50-66% of several key customers' demand.

Q: "Obviously, AI is the area that has the most urgency. But do you worry about demand destruction for things like PCs and smartphones? ... how are you thinking about that allocation process?"
— Joseph Moore, Morgan Stanley

A: "Price-sensitive markets such as the consumer example that you gave may have some demand that is getting impacted due to the higher prices. But overall demand in those markets as well stays pretty strong... we want to maintain that well-diversified mix for our end markets."
— Sanjay Mehrotra, CEO

Assessment: Management confirmed demand destruction is already occurring in consumer markets, the first such admission of the cycle. It is being masked by content growth and data-center reallocation for now, but a market acknowledging it is rationing demand by price is a market closer to its top than its middle.

What to do with a $50B-plus cash pile

A question probed capital allocation given the prospect of $35-40B of fiscal-year free cash flow and a $50B-plus cash balance, including the CHIPS-related buyback restrictions. Management prioritized the balance sheet and organic investment, cited two credit upgrades to triple-B and a 30% dividend raise, and flagged significant future buyback capacity.

Q: "You are going to generate $35 to $40 billion in free cash flow this fiscal year... what do you do with this? Are you planning to set aside a bunch of it to buy back a bunch of stock on the other side?"
— Timothy Arcuri, UBS

A: "Balance sheet is always going to be a priority along with organic investment... we are generating return on capital at this point over 30%, headed towards 50%... we believe we will have significant capacity for returning cash to shareholders through repurchase."
— Mark Murphy, CFO

Assessment: The capital-return optionality is real and is the strongest support under the stock. A net-cash company stockpiling buyback capacity "for the other side" is well-armed to defend the equity in a downturn, which is a genuine reason this is a Hold rather than a sell.

HBM versus non-HBM margins and bit allocation

A question asked whether the HBM TAM outlook had changed and whether any players are shifting bits to DDR5 over HBM given higher commodity margins. Management held the prior ~$100B-by-2028 HBM TAM, and confirmed that non-HBM margins are currently higher than HBM margins while demand for both remains strong.

Q: "Are you seeing any sort of preference for perhaps moving bits to DDR5 over HBM by any industry players given, today, the higher margins that we see there?"
— C.J. Muse, Cantor Fitzgerald

A: "It is correct that the margins for non-HBM today are higher than HBM margins. Demand for HBM, of course, continues to be strong... we continue to manage the mix of the business as the data center AI demand continues to grow."
— Sanjay Mehrotra, CEO

Assessment: The confirmation that commodity DRAM out-earns HBM is the most economically revealing data point of the call. It both demonstrates how extreme standard-DRAM pricing has become and creates an industry incentive to pour capacity into the most replicable product, the classic setup for the next oversupply.

SCAs and earlier customer engagement

A question explored whether the multiyear SCAs are partly driven by the need for earlier, deeper engagement on custom HBM base-die designs. Management would not tie SCAs to specific customers but confirmed the agreements bring closer R&D collaboration and roadmap planning.

Q: "How much of these multiyear SCA agreements is due to the inherent requirements for earlier and longer-term engagement with your GPU/XPU chip customers, just due to the customization of their next-generation HBM architectures, especially around the base die?"
— Harlan Sur, JPMorgan

A: "These SCAs really bring us closer to the customer in terms of partnership. That partnership extends into bringing us closer in terms of R&D collaboration and roadmap planning, both ours as well as for customers." — Sanjay Mehrotra, CEO

Assessment: The SCA-as-deeper-partnership framing is a genuine positive for customer stickiness and the HBM franchise's durability. It strengthens the long-term moat even as it leaves the near-term cyclical question, what happens to commodity DRAM pricing, untouched.

What They're NOT Saying

1. Whether the SCAs protect margins on the downside: Asked directly and repeatedly, management would only call the agreements "robust" with "specific commitments," never confirming a margin floor. This is the single most important omission, because it is the only thing that would actually de-cyclicalize the earnings.
2. Any SCA specifics: Counterparty, volume, pricing structure, duration mechanics, and cancellation terms are all confidential. The market is asked to take the visibility benefit on faith.
3. Fiscal-Q4 gross margin: Declined, with only "tight beyond 2026" offered. At an 81% guide, the refusal to extend the trajectory is conspicuous.
4. An updated HBM TAM: Management held the prior ~$100B-by-2028 figure rather than raising it, notable in a quarter where everything else was revised up.
5. Any framing of the down-cycle: At a 75% printed and 81% guided margin, management offered no view on what normalization looks like or when industry supply rebalances. The absence of any peak discussion is itself a data point.

Market Reaction

• Pre-print setup: MU closed at $461.73 on March 18, up 61.8% year-to-date and 353.9% over the trailing 12 months, sitting at a 52-week closing high. Expectations were as elevated as a chart can make them.
• After-hours / open: The stock gapped down about 8% to open at $424.97, an unusual reaction to a record beat-and-raise.
• Reaction session: Shares recovered through the day to close -3.8% at $444.27, on 74.6M shares (2.1x the 30-day average). The S&P 500 was -0.3%.

A 3.8% decline on the best quarter and guide in the company's history is the most important signal of the day, more important than any single financial metric. It is the market doing what it always does near a cyclical peak: looking across the guided earnings to the normalization beyond, and declining to pay up even for a historic print. The intraday recovery from down 8% shows the bulls are not gone, but the inability to rally a stock that just guided to $19 of quarterly EPS is the clearest evidence yet that the easy expectations gap has not merely closed, it has inverted. When good news stops working, the tape is telling you something the income statement is not.

Street Perspective

Debate: Do the SCAs de-cyclicalize memory, or just smooth a still-cyclical business?

Bull view: Five-year contracts with binding commitments fundamentally change the model, giving Micron visibility and stability that warrant a structurally higher, less cyclical multiple. This is the end of memory as a pure commodity.

Bear view: Management would not confirm any downside-margin protection, so the SCAs likely lock volume and visibility but leave price, and therefore margin, exposed. A volume contract at a peak price does not save you when the price normalizes.

Our take: Decisive for the rating, and it lands with the bear until proven otherwise. Visibility on volume is valuable, but the earnings are driven by price, and nothing disclosed suggests the SCAs floor the price. We would re-rate the stock the day an SCA is shown to protect downside margins, and not before.

Debate: Was the 3.8% decline the top, or a pause?

Bull view: The stock is up over 350% in a year; a single-digit dip on profit-taking after a vertical run is healthy consolidation, and the next leg comes as the $19 fiscal-Q3 EPS prints and estimates chase it.

Bear view: A decline on the best news imaginable, after demand destruction was confirmed and margins were called near their ceiling, is the classic anatomy of a cyclical top. The marginal buyer is exhausted.

Our take: We do not claim to call the exact top, which is why we are at Hold and not Underperform. But the burden of proof has shifted: when a stock cannot rally on a historic print, forward returns tend to be made on the short side or not at all. We would rather watch than chase.

Debate: Is the $25B-plus capex a rational response or the seed of the next glut?

Bull view: The build is demand-driven and disciplined, funded by record free cash flow and a net-cash balance sheet, securing Micron's supply leadership into a durable AI build-out and largely tied to long-dated facilities.

Bear view: Every supplier is responding to 75-80% margins at once; the aggregate industry capex now being committed lands as supply in 2027-2028 and is the most reliable leading indicator of the next down-cycle.

Our take: Micron's own spend is prudent and well-financed; the risk is the industry aggregate. Capex peaks have preceded every memory glut, and the combination of an 81% margin and a doubling construction bill is the market building the conditions for its own normalization. The most important line item to track from here.

Model Update & Valuation Framework

Valuation framework: At the $444.27 reaction close, the fiscal-Q3 guide of $19.15 EPS annualizes near $77, leaving the stock around 6x that run rate with a net-cash balance sheet. On the bull framing, single-digit forward multiples with a five-year contract signed are absurdly cheap. On the bear framing, a 6x multiple on an explicitly guided peak is the market's own forecast that the earnings normalize sharply, and paying it is a bet on multiple expansion at the very top of the most extreme margin cycle in the industry's history. The stock fell on the print, which is the market casting its vote on which framing it believes.

Where we land: We maintain Hold, and we are candid that this Hold was early: we downgraded in December at $248, and the stock has since nearly doubled. That is the cost of the call, and we own it. But the question is forward, not backward, and the forward case at $444 is harder, not easier, than it was at $248: margins are more extreme, the stock is more extended, demand destruction has begun, and the supply response is accelerating. We would return to Outperform on (1) a meaningful pullback that resets the risk/reward, or (2) confirmation that an SCA contractually floors downside margins, which would genuinely change the cyclical math. We would move to Underperform on the first concrete evidence of pricing rolling over or order cancellations. Absent those, Hold is the honest rating.

Thesis Scorecard Post-Earnings

Overall: The thesis is fully realized on the bull side and the SCA marks real structural progress, but the new and escalating bear pillars (demand destruction, a near-ceiling margin, an accelerating capex build) define the forward risk. The decisive open question is whether the SCAs floor downside margins; until answered, the cyclical risk is intact. Note continuity: our December re-upgrade trigger of "signed multiyear LTAs" technically fired, but the agreements as disclosed do not yet protect margins, and our other trigger (a pullback) moved the wrong way, so the rating holds.

Action: Maintain Hold. Do not chase the apex. Re-engage at Outperform on a pullback or on confirmed downside-margin protection in an SCA; move to Underperform on the first signs of pricing rolling over.

Bottom Line

Micron just printed the best quarter the memory industry has ever produced: revenue nearly tripled to $23.9B, gross margin reached 74.9%, EPS hit $12.20, free cash flow set a record, the balance sheet turned to $6.5B net cash, and management guided fiscal Q3 to an 81% gross margin and $19.15 of EPS. It signed its first five-year strategic customer agreement and raised the dividend 30%. On the fundamentals, it is difficult to imagine a better quarter, and the SCA is a genuine structural step forward.

We are maintaining Hold, and we will be honest about the scoreboard: we downgraded in December at $248, the stock has since nearly doubled to the mid-$400s, and our caution has cost return. Being early is a real error and we own it. But a rating is a forward judgment, and the forward case is harder now than it was then, not easier. The signed SCA delivered on a trigger we named, yet management would not confirm it protects margins on the downside, which is the only thing that would actually de-cyclicalize the earnings. Meanwhile every cyclical-peak marker has brightened: commodity DRAM now out-earns HBM, prices rose 60-80% in a single quarter, consumer units are about to decline on memory scarcity, capex is past $25B and climbing, and most tellingly, the stock fell on the best news it will ever get. When good news stops working, the market is discounting the far side of the peak. We respect how extraordinary this business has become, and we hold the gain we did capture from $94 to $248 earlier in our coverage. But we will not pay up for an explicitly guided peak at 6x on $77 of annualized peak earnings. We would rather re-engage on a pullback, or on proof the SCAs floor the downside, than chase the most vertical move in the company's history at its apex. Hold.