Micron FQ3 FY2026: The Take-or-Pay Floor Changes the Math, and We Upgrade to Outperform

Results vs. Consensus

FQ3 FY2026 Scorecard

Year-Over-Year Comparison

Quarter-Over-Quarter Comparison

Quality of Beat

• Revenue: Still overwhelmingly price. DRAM revenue rose 67% sequentially on low-single-digit bit growth and prices up in the low-60s percent; NAND rose 99% on mid-single-digit bits and prices up in the mid-80s percent. A 74% sequential revenue jump on flattish-to-modest bit growth is, again, the signature of pricing leverage at a cyclical extreme, not unit demand. What is new is that a growing share of that revenue is now contracted under multiyear floors rather than left to spot.
• Margins: The 84.9% gross margin rose 10 points sequentially and roughly 46 points year-on-year. There is no one-time item; it is price, cost, and mix. Operating margin of 81.2% on $1.5B of operating expense (up only $97M sequentially, on higher variable comp) shows the operating leverage that flows when revenue near-doubles against an almost fixed cost base.
• EPS: Clean and operational. The 14.9% tax rate is in line with the ~15% run rate, the GAAP-to-non-GAAP gap is narrow ($24.67 vs. $25.11), and cash quality is pristine: $25.39B of operating cash flow, $18.3B of record free cash flow after $7.1B of capex, and a record $24.4B net cash position after $4.4B of debt reduction.

Segment Performance

Revenue by Technology

Revenue by Business Unit (all records)

Data center is now the engine, at a $100B annualized run-rate

The single most important segment fact on the call: data center revenue exceeded $25B in the quarter, an annualized run-rate above $100B, and data center SSD revenue alone exceeded $5B after more than doubling sequentially. Cloud Memory and Core Data Center together (the two purest data-center units) were $25.3B, 61% of company revenue, with gross margins of 83% and 87%. Two years ago Micron was a memory commodity supplier whose fortunes tracked PCs and phones. Today a majority of its revenue, at its highest margins, is sold into AI infrastructure.

Assessment: The data-center concentration cuts both ways, and the read has flipped from last quarter. The bear framing is single-end-market dependence. The bull framing, now better supported, is that this is precisely the revenue that the SCAs are contracting: take-or-pay agreements with the hyperscalers and large data-center buyers convert the most cyclically-feared concentration into the most visible, floor-priced revenue in the model. Concentration plus a contract is a different risk than concentration alone.

Mobile, client and auto: records on price, units still soft

Mobile & Client posted an 87% gross margin, a once-unthinkable figure for the consumer-facing unit, on a 49% sequential revenue gain driven by pricing even as bit shipments fell. Auto & Embedded grew 71% sequentially to a record $4.6B at a 79% margin. The driver across every unit is identical and it is price.

Assessment: An 87% gross margin in the mobile and client unit, achieved while bits declined, remains the clearest evidence that consumer pricing has overshot end demand; that has not changed since last quarter. What has changed is the framing management now puts around it: PC and smartphone industry revenue is expected to grow despite unit declines, and the bits not going to price-sensitive consumers are being redirected to floor-priced data-center contracts. The scarcity is being monetized deliberately, not absorbed passively.

Key Topics & Management Commentary

Overall Management Tone: Confident and, for the first time in this cycle, structural rather than merely cyclical. Where prior calls defended the durability of a spot-price boom, this one spent its airtime on a new business model: take-or-pay contracts, customer deposits, remaining performance obligations, and a margin floor above prior peaks. The tone was that of a company arguing it has changed what it is, not just how much it earns. The one candid concession, consistent with last quarter, was that the rate of price increase is now moderating even as the absolute level steps higher.

1. The Take-or-Pay SCAs: A Contractual Margin Floor

This is the disclosure that changes the investment case. Micron has now signed 16 strategic customer agreements spanning data center and automotive, typically five-year terms (calendar 2026 through 2030; auto generally three years). They are structured as take-or-pay agreements with binding volume commitments. The largest carry a ceiling price at current CQ2 market levels and a floor price through the term; some carry fixed prices. Critically, management quantified the floor in margin terms.

"The largest agreements generally have a ceiling price for existing products at the current CQ2 market price and a floor price through the term of this agreement... For our SCAs' midsize band, the floor price enables a very robust gross margin for Micron, well above our peak quarterly margins in any past cycle." — Sanjay Mehrotra, Chairman, President & CEO

The scale: the 16 agreements cover roughly 20% of DRAM volume and a third of NAND volume over the period, and management expects approximately half or more of total company revenue to fall under SCAs once the targeted set is complete. When all planned SCAs are executed, agreements with fixed prices or ceilings at or near current levels are expected to be ~40% of revenue. Fourteen of the sixteen carry cumulative revenue at minimum (floor) price of approximately $100B over the remaining term.

Assessment: In March we wrote that an SCA would change our rating "the day an SCA is shown to protect downside margins, and not before." That day is today. A take-or-pay contract with a floor price that holds gross margin above any prior-cycle peak is, mechanically, the de-cyclicalization device the bull case always needed and could never point to. The honest caveat is that "above prior peak" is a floor near the low-60s percent, far below today's 84.9%, so the floor protects against a collapse rather than guaranteeing current earnings. But protecting the downside is exactly what de-risks the multiple, and it is what we required.

2. The Beat-and-Raise: $41.5B at an 84.9% Margin

Revenue of $41.46B grew 74% sequentially, the largest dollar step-up in company history at $17.6B, eclipsing the prior record of $10.2B set just last quarter. Gross margin of 84.9% was a new company record and more than doubled year-on-year. Operating margin reached 81.2%.

"With significant records in revenue, gross margin, and EPS, all exceeding the high end of our guidance... our data center revenue exceeded $25 billion in fiscal Q3, on an annualized run rate of over $100 billion." — Sanjay Mehrotra, CEO

Assessment: The print itself is almost beside the point for the thesis now, because it is so far beyond any sustainable mid-cycle level that arguing about its exact magnitude misses the question. The relevant question shifted this quarter from "how high does the peak go" to "what is contractually protected on the way down," and that is the right question to be debating. The beat earns the company the benefit of the doubt; the SCAs earn it the rating change.

3. HBM4 Shipping Ahead of Schedule

Micron has already shipped over $1B of HBM4, and its HBM4 12-high volume ramp is tracking twice as fast as the HBM3E 12-high ramp, with mature yields expected significantly faster than HBM3E. The 1-gamma DRAM and G9 NAND nodes are both ramping toward becoming the highest-volume nodes in company history, and the next-generation DRAM and NAND nodes are on track for volume production in the second half of calendar 2027.

"HBM4 12-high volume ramp is tracking twice as fast as HBM3E 12-high, and we have already shipped over $1 billion in HBM4 revenue. We expect to reach mature yields on HBM4 12-high significantly faster than HBM3E 12-high." — Sanjay Mehrotra, CEO

Assessment: The HBM franchise remains the durable technology core that justifies owning Micron across a cycle, and a yield ramp twice as fast as the prior generation is a concrete execution win, not a slogan. HBM is also where the SCA logic is strongest: it is the product customers must engage on years early, and the agreements deepen that lock-in. This pillar is squarely on track.

4. The FQ4 Guide, and the First Sign of a Moderating Price Slope

Management guided fiscal Q4 to a record $50B of revenue (plus or minus $1B), ~86% gross margin, ~$1.65B of operating expense, and a record $31 of EPS (plus or minus $1) on ~1.15B diluted shares. The gross-margin guide is notable for the qualifier attached to it.

"Our fiscal Q4 gross margin outlook reflects a meaningful moderation in the rate of price increases." — Mark J. Murphy, CFO

Assessment: Even at a guided 86%, the second derivative is turning. Margin is still rising, but the rate of price increase that drives it is decelerating, exactly as the company foreshadowed last quarter when it said incremental price now moves margin less. For a stock that has run on relentless upward revisions, a moderating slope matters. The reason it is not a downgrade trigger this time is that the SCA floor now backstops the level even as the slope flattens.

5. $22B of Customer Deposits: A New Line on the Balance Sheet

Under the SCAs signed so far, Micron projects $22B of cash deposits and related financial commitments, of which roughly $18B is cash deposits (the remainder largely letters of credit). About $0.5B was received in FQ3, with roughly $10B more expected in FQ4. The deposits flow through financing cash flows, not free cash flow, and are returned to customers over time toward the latter half of each agreement term.

"Under the SCAs we have signed so far, we project to receive cash deposits and related financial commitments of $22 billion... These customer deposits will show up on our balance sheet more in fiscal Q4. The cash flows associated with customer deposits appear in financing related cash flows, and will not affect our free cash flow." — Mark J. Murphy, CFO

Assessment: A customer putting cash on deposit against a take-or-pay commitment is the strongest possible signal that the commitment is real. It is skin in the game from the buyer side, and it is the kind of structural feature that simply does not exist in a pure spot-priced commodity. The mechanics matter for modeling (this is financing cash, not operating, and it reverses out later), but the signal matters more than the mechanics.

6. Supply Stays Tight "Beyond 2027," and the Capex That Builds the Other Side

Management now expects DRAM and NAND to remain tight beyond calendar 2027, with industry DRAM bit growth in the low-to-mid-20s percent and NAND around 20% in calendar 2026. Funding the supply to meet that demand is an accelerating capital bill: FY26 capex of approximately $27B (FQ4 around $10B), with FY27 quarterly capex stepping above FQ4 levels and more than half of the year-on-year FY27 increase coming from construction as Micron pulls in cleanroom capacity.

"We expect quarterly CapEx in fiscal 27 to be above fiscal Q4 levels, with more than half the increase year over year in fiscal 27 from construction CapEx as we pull in clean room capacity required to address long term demand." — Mark J. Murphy, CFO

Assessment: This is the bear case that survives the SCA disclosure intact. A capex bill rising toward and beyond $40B annualized is the industry building the capacity that today's margins demand, and that capacity lands as supply in 2027-2028. The SCAs change how exposed Micron is to that supply when it arrives, by contracting volume and flooring price, but they do not repeal the cycle for the unconstrained portion of the market. This is the single line item we will track most closely from here.

7. The Balance Sheet and the Capital-Return Pivot

Micron ended the quarter with a record $30.2B of cash and investments and a record $24.4B net cash position after reducing debt by $4.4B (including a $4.3B tender offer). It received credit-rating upgrades from all three major agencies this fiscal year, including to BBB+, and declared a $0.15 dividend. Most significantly, management laid out a capital-return pivot tied to the expiry of CHIPS-related restrictions.

"From 12/09/2026, the second anniversary of the signature of our definitive CHIPS agreements, we intend to increase our capital return. Over time, we expect to return 100% of our excess cash to shareholders." — Mark J. Murphy, CFO

Assessment: A net-cash, BBB+ Micron generating tens of billions of free cash flow, committing to return 100% of excess cash from December, is a materially different capital-allocation story than the company that spent the last cycle deleveraging. The buyback capacity being built "for the other side" is real downside support under the equity, and it is part of why the risk/reward has tilted.

8. Demand Destruction, Reframed

Last quarter's first visible crack, the warning that memory scarcity could push PC and smartphone units down by low-double-digits in calendar 2026, was reframed rather than retracted. Management now expects PC and smartphone industry revenue to grow despite unit declines, and actually raised its calendar-2026 server unit growth expectation to high-teens (from low-double-digits), citing AI-accelerator and traditional-server strength.

Assessment: The demand-destruction signal is being offset, not resolved. Units are still expected to fall in consumer end-markets; the scarcity is simply being routed to higher-value data-center and contracted demand, where it earns more. That is a rational allocation response and it pushes the demand-destruction risk further out, but it does not eliminate the underlying fact that prices are high enough to suppress end-unit demand. We keep this on the watch list, downgraded in urgency but not dismissed.

Guidance & Outlook

The FQ4 guide extends the trajectory: revenue up another 21% sequentially to $50B, gross margin to ~86%, and EPS to a record $31. Annualizing the FQ4 EPS guide yields a run-rate near $124. The market is, once again, being handed an explicitly guided peak rather than asked to forecast a recovery.

Implied trajectory: A $50B FQ4 would put FY26 revenue near $129B and FY26 EPS near $73 (FQ1 $4.78 + FQ2 $12.20 + FQ3 $25.11 + FQ4 $31). For context, that single fiscal year of EPS is more than eight times the company's entire FY2025 EPS of $8.29.

Street at: Consensus had FQ3 near $35.7B and $20.49; the company printed $41.46B and $25.11 and guided FQ4 well above where the Street sat. Estimates will again be chasing.

Guidance style: Historically conservative, and the FQ4 GM guide of ~86% with an explicit "moderation in the rate of price increases" qualifier is the most measured framing of the cycle so far, even as the absolute numbers set records.

Analyst Q&A Highlights

How much revenue is actually locked at the floor price

The opening exchange went straight to the heart of the SCAs: of the headline ~$100B figure, how much is genuinely guaranteed at a floor price versus exposed to the market. Management clarified that the ~$100B is the floor-price minimum and that it expects actual revenue to run well above it, while sizing the contracted share of the business.

Q: "I think we are all trying to figure out how much is locked in kind of a floor price scenario over the next 5 years... 14 of the 16 SCAs have $100 billion in cumulative revenue... but then you also said that 40% of revenue will be moving inside of these SCAs. Can you help us think about how much of revenue per year would be guaranteed?"
— Timothy Arcuri, UBS

A: "Under these SCAs that have been completed so far, at the floor price, the revenue is projected to be $100 billion. But, again, we expect revenue to be much higher than that. Note that at the floor price our profitability levels and the gross margins are higher than peak margins at any time in the past... about 20% of DRAM and about 30% of our NAND volume is covered in these SCAs so far. So that jumps to about 25% of our revenue that you can project over the term of these agreements."
— Sanjay Mehrotra, CEO

Assessment: This is the exchange that earns the upgrade. The $100B is a contractual minimum at a floor that holds margins above prior peaks, with realized revenue expected to exceed it. The honest read is that the floor de-risks the tail rather than guaranteeing the peak, but a quantified, contracted floor above past-cycle peak margins is exactly the mechanism we said we were waiting for.

The purpose and mechanics of the customer cash deposits

A recurring line of questioning probed why customers would post billions in cash deposits if the money is neither a prepayment nor revenue. Management framed the deposits as a binding reflection of shared commitment under the take-or-pay structure, held by Micron and returned in the back half of each term.

Q: "What is the role... it seems like they are putting a deposit, and then they get the deposit back. What is the reason for them to commit that cash? Is that something where there is a take or pay that cash is related to? It is not a prepayment. Can you help us understand that?"
— Joseph Moore, Morgan Stanley

A: "It is not a prepayment. It is a separate commitment by the customers, and a reflection of the fact that we have a binding agreement and these are take or pay agreements. And we hold the cash, and it is a reflection of our shared commitment to perform under these agreements... it is good for the customers because they have supply assurance. They have leading technology. So in our view, it is a win."
— Mark J. Murphy, CFO

Assessment: The deposit is collateral for commitment, not a financing trick. A buyer willing to tie up cash to secure supply is making the most credible statement available that the volume commitment is real. It is the feature that most clearly separates this from a handshake LTA, and it is why the SCAs deserve to be treated as structural rather than promotional.

Whether the deposit cash is fungible, and capital-return implications

A question pressed on whether the deposit cash is usable for capex and whether it changes the gross-cash level management is comfortable holding, particularly heading into the post-CHIPS capital-return window. Management said the cash is unrestricted but would not change the near-term liquidity posture.

Q: "When you think about these cash deposits, do you view that as fungible cash and used for CapEx? And when you contemplate capital returns, particularly after December 9... will you include that cash in your gross cash thoughts, or would that cause you to hold more gross cash all else equal?"
— CJ Muse, Cantor Fitzgerald

A: "It is unrestricted... Not in the near term. We are going to have what we deem as adequate liquidity to support the operation of the business that would include over time returning the deposits as customers and Micron perform on the contracts... and then we would hold liquidity to satisfy what investments we believe are important for the business."
— Mark J. Murphy, CFO

Assessment: Management is being disciplined about not treating returnable deposits as permanent capital, which is the right posture. The capital-return story rests on the underlying free cash flow and net-cash balance sheet, not on the deposits, and that is a more durable foundation for the 100%-of-excess-cash commitment.

What a data-center SCA actually looks like

A question sought to characterize the typical large SCA, and where the ~$100B figure sits relative to the large, medium, and small customers. Management confirmed the large agreements span data center, consumer, and automotive, and reiterated that the large deals carry a floor-to-ceiling price band with the ceiling set at CQ2 levels.

Q: "You have mentioned 4 large and 3 medium-sized customer agreements... how many of them are related to the data center? The $100 billion, does that align with large and medium size, or the smaller size customers? Have you given enough breadcrumbs for us to figure out what a data center SCA looks like over the next few years?"
— Vivek Arya, Bank of America Securities

A: "Our large customers include data center, and the large and medium customers... go across data center, consumer and automotive markets. The large agreements generally have a ceiling price and a price band which has a floor as well as a ceiling, and the ceiling is established at the CQ2 price levels... those price bands also provide for floor prices which where the gross margins are well above the peaks at any past cycle."
— Sanjay Mehrotra, CEO

Assessment: The structure is consistent across the large agreements: a ceiling at today's elevated prices and a floor above prior-cycle peak margins. That band is the whole story. The ceiling caps how much further the contracted book benefits if spot keeps rising; the floor is what protects the book when spot eventually falls. For a cyclical, trading away some upside for a hard floor is a trade that lowers the through-cycle risk, which is what the multiple keys off.

Where gross margins settle on the other side

The dominant modeling question was how to think about normalized gross margin once the SCAs layer in and price increases moderate, with one analyst floating a mid-70s range between today's level and prior peaks. Management declined to guide beyond FQ4 but pointed to tight conditions beyond 2027 and the operating leverage from new volume in 2027-2028.

Q: "As your long term investors build their models for 2027, 2028, should they be assuming a normalized gross margin range somewhere in the mid-70s, kind of the range between where you are today versus the prior peaks?"
— Vivek Arya, Bank of America Securities

A: "We are not providing guidance beyond the fourth quarter. But we are at margin levels that, as we have talked about before, incremental price yields less in gross margin expansion... we expect the market to remain tight beyond 2027... we will get additional volume materially beginning midyear 2027 that will grow into 2028... over time, we will get that operating leverage."
— Mark J. Murphy, CFO

Assessment: Management would not put a number on normalized margin, which is the most important omission of the call. The refusal is understandable but it leaves the central modeling question open: the SCA floor sits well below 84.9%, so a normalization toward the floor is a large earnings decline even though it never breaches the contract. The floor caps catastrophe, not normalization. That distinction is the reason this is an Outperform and not a table-pounding call.

Quantifying the 2027 supply-demand imbalance

A closing question asked whether the undersupply in 2027 could be quantified relative to 2026. Management would not size it but reiterated that tightness persists beyond 2027 because new capacity takes years to build and technology transitions yield less incremental supply per node.

Q: "You mentioned how DRAM bit should grow low to mid twenties, NAND in the 20% range this year, and clearly we are undersupplied on both. Is there a way to quantify what happens in 2027? Is the undersupply going to be double what it is this year in 2027?"
— Krish Sankar, TD Cowen

A: "We see 2027 overall tight. We have said we see tightness continuing beyond 2027... it takes a long time to bring up the additional capacity that is needed... and even in 2028, when supply begins to improve gradually, we see that the demand will continue to be on a robust trajectory because these AI trends are very long term trends."
— Sanjay Mehrotra, CEO

Assessment: Tightness beyond 2027 is the demand-side support for the whole thesis, and management is consistent on it. The unhedged honesty is the 2028 line: supply "begins to improve gradually" in 2028, which is the same window the accelerating capex points to. The bet the SCAs let an investor make is that the contracted, floor-priced book carries the earnings through that 2028 supply improvement, which is precisely the bet the upgrade expresses.

What They're NOT Saying

1. The actual floor price or floor-margin number: Asked directly to confirm a low-teens-to-mid-$20s per-gigabyte range, management declined all specifics, offering only that floor margins are "well above" prior peaks. We know the floor exists and beats past peaks; we do not know how far below 84.9% it sits, which is the number that determines through-cycle earnings power.
2. Normalized gross margin: Management explicitly would not put a figure on where margins settle in 2027-2028, deflecting the mid-70s framing offered to them. At a printed 84.9%, the refusal to anchor a normalized level is the most consequential omission for the model.
3. SCA counterparties: The four very large and three medium customers are unnamed, as is the data-center-versus-auto split of the $100B. The market is asked to take the contracted book on faith, supported by the deposits.
4. The spot-versus-contracted mix within the FQ4 guide: How much of the $50B FQ4 guide is already under SCA versus left to spot was not disclosed, which matters for judging how protected the next print actually is.
5. An updated HBM TAM: Management did not refresh the prior ~$100B-by-2028 HBM TAM, again notable in a quarter where nearly everything else was revised higher.
6. FY27 revenue or margin framework: Beyond "tight beyond 2027" and the capex shape, no quantified FY27 outlook was offered, leaving the bridge from a guided FQ4 peak to a contracted-but-normalizing FY27 entirely to the analyst.

Market Reaction

• Pre-print setup: MU closed at $1,048.51 on June 24, up 267.4% year-to-date and 719.7% over the trailing 12 months, having rallied 39.6% in the prior 30 days. Unlike last quarter, the stock entered this print roughly 13% below its 52-week closing high of $1,211.38, recovering toward that high rather than sitting on it. The S&P 500 was up 7.5% year-to-date.
• After-hours move: Shares jumped roughly 13-14% in extended trading following the print and FQ4 guide, toward a post-market level near $1,196. This is the first clearly positive print reaction in several quarters, a direct reversal of last quarter's 3.8% decline on a record beat-and-raise.

The reaction is the most important non-financial signal of the day. For three quarters, Micron's defining tell was that the stock could not rally on extraordinary news, the classic anatomy of a market discounting the far side of a cyclical peak. That pattern broke tonight. The market is rewarding not the size of the beat, which it has shrugged off before, but the structural disclosure that accompanied it: take-or-pay contracts, customer deposits, and a margin floor above past peaks. When the tape stops fading good news and starts paying up for a change in business model, it is casting a vote on the durability of the earnings, not just their level. We note this is the extended-hours move on the print date; the regular-session reaction had not closed as of writing.

Street Perspective

Debate: Do the SCAs actually de-cyclicalize memory, or just smooth the top?

Bull view: Take-or-pay contracts covering ~40% of revenue at fixed or ceiling prices, with floors above prior-cycle peak margins and billions in customer deposits behind them, structurally change the business and warrant a higher, less cyclical multiple. This is the end of memory as a pure commodity.

Bear view: The floor sits well below today's 84.9% margin and only ~40% of revenue is fixed-or-ceiling priced, leaving ~60% exposed to spot. A floor near prior-cycle peaks still permits an enormous earnings decline from current levels without ever being breached. The SCAs cap the disaster, not the normalization.

Our take: The bear is right about the magnitude and the bull is right about the direction, and the direction is what changes the rating. A contracted floor above past peaks is a genuine, quantified reduction in tail risk, which is exactly what we said we required. It does not protect current peak earnings, which is why we upgrade to Outperform rather than to maximum conviction. The single most valuable disclosure of the next several quarters will be the floor-margin number management would not give.

Debate: Is the post-print jump the start of a re-rate or a late-cycle blow-off?

Bull view: The first positive print reaction in quarters, after the best news repeatedly failed to lift the stock, marks the moment the market accepts the structural story; estimates are still chasing and the stock trades below 9x annualized peak EPS.

Bear view: A 14% pop on a stock up 720% in a year is the kind of exhaustion move that prints near tops, and the moderating price slope plus accelerating capex are the textbook late-cycle tells.

Our take: We lean bull but size for volatility. A change in how the tape treats good news is a real signal and it aligns with the structural change in the business, but a stock this extended will trade violently in both directions. We would not chase the after-hours print; we would own the position through the next pullback, which the SCA floor makes a buyable dip rather than a falling knife.

Debate: Is the accelerating capex the seed of the 2028 glut?

Bull view: The build is demand-driven, SCA-backed, and funded from a record net-cash balance sheet; securing supply leadership into a durable AI cycle is the right use of capital, and the contracted book absorbs much of the new volume.

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

Our take: This is the dominant residual risk, and it is unresolved by anything on this call. The difference from last quarter is that Micron now enters that supply response with a contracted, floor-priced book rather than fully exposed to spot. The capex is still building the other side of the cycle; the SCAs are what let us own the stock into it. We track FY27 capex as the key line item.

Model Update & Valuation Framework

Valuation framework: At the $1,048.51 pre-print close, the FQ4 guide of $31 annualizes near $124, leaving the stock around 8.5x that run-rate, or about 14x the implied ~$73 of FY26 EPS, with a record net-cash balance sheet. On the bull framing, a high-single-digit multiple on earnings that are now partly contracted at floors above prior-cycle peaks is too cheap for a business that has demonstrably changed its risk profile. On the bear framing, applying any multiple to an explicitly guided peak is a bet on where normalized earnings land, and the SCA floor (well below current margins) implies that normalized level is far below the run-rate the multiple is being applied to.

Where we land: The valuation question turns entirely on the earnings level the multiple should capitalize. Last quarter we argued the right level was a mid-cycle number far below the peak, and that a single-digit multiple on the peak was the market's own forecast of mean reversion. The SCAs change that argument: they put a contractual floor under a meaningful, growing share of the revenue, which raises the through-cycle earnings the multiple should be applied to and lowers the variance around it. That is what justifies paying up modestly here even after the run. We would revisit if the moderating price slope turns into an outright spot-price decline before the contracted book is large enough to cushion it.

Thesis Scorecard Post-Earnings

Overall: The thesis is strengthened, decisively, on the bull side. The pillar that has gated our rating for two quarters, whether the SCAs protect downside margins, moved from AT RISK to ON TRACK on a confirmed, quantified margin floor above prior-cycle peaks. The bear pillars are intact but reordered: demand destruction is being offset, the margin slope is moderating as expected, and the dominant residual risk is now squarely the 2027-2028 supply response that the accelerating capex is building. The decisive open question shifts from "do the SCAs floor margins" (now answered yes) to "how far below the peak does that floor sit" (unanswered).

Action: Upgrade to Outperform from Hold. The pre-committed re-upgrade trigger fired. Re-engage at higher conviction if management quantifies the floor margin or the contracted book reaches half of revenue; move back to Hold if spot pricing rolls over faster than the contracted book can cushion; move to Underperform only on outright order cancellations or a breach of the take-or-pay commitments.

Bottom Line

Micron just printed $41.5B of revenue at an 84.9% gross margin and $25.11 of EPS, beat its own already-record guide by every measure, and guided fiscal Q4 to $50B and $31. Those numbers are extraordinary, and they are also, by now, almost expected from this company at this point in the cycle. What was not expected, and what actually matters, is the structure that came with them: 16 take-or-pay strategic customer agreements, roughly $100B of remaining performance obligation, $22B of customer deposits, and management's first explicit confirmation that the contracts floor gross margins above any prior-cycle peak.

We are upgrading to Outperform, and we will be candid about the scoreboard first. We downgraded Micron to Hold in December near $248 and maintained that Hold through last quarter at $444, and the stock has since reached the $1,000s. Being early at a peak that kept extending is a real cost, and we have owned it in each of the last two notes. But a rating is a forward judgment, and the forward case genuinely changed tonight. We told readers in March, in writing, that we would re-rate "the day an SCA is shown to protect downside margins, and not before." Management has now shown exactly that: a take-or-pay floor, with deposits behind it, that holds margins above prior peaks. To keep the Hold after the trigger we named has fired would be moving the goalposts, and that is not how we run this book.

The upgrade is to Outperform, not to maximum conviction, and the reasons for the restraint are honest ones. The floor sits well below today's 84.9% margin, so a normalization toward it is still a large earnings decline; roughly 60% of revenue remains spot-priced; the price slope is moderating; and the accelerating capex is building the 2027-2028 supply that ends every memory cycle. What tips the balance is that the business now enters that supply response contracted and floored rather than fully exposed, the balance sheet is a record net-cash fortress moving to return 100% of excess cash, the stock trades below 9x annualized peak earnings, and for the first time in this cycle the tape paid up for the news. We respect how cyclical this business remains, and we are not pretending the floor repeals the cycle. But the one thing we said would change our mind has happened, the valuation does not demand heroics, and the risk/reward over the next twelve months now favors the upside. Outperform.