Micron’s beat was not the surprise; the supply-duration reset was

The thesis from this earnings event is that Micron is no longer just a cyclical memory recovery story with HBM upside attached; it is becoming a capacity-allocation story where customers are forcing investment forward faster than the Street had modeled. What was priced in was a material beat: the Street estimate sat at $12,906.7 million of revenue and $3.96 of EPS, and Micron delivered $13,643.0 million and $4.78, a +5.7% revenue surprise and +20.7% EPS surprise. What was not priced in, in our view, was the combination of a record gross margin at 56.1%, a guided next-quarter revenue level of $18.7 billion, plus or minus $400 million, and a CapEx lift to approximately $20 billion versus the prior estimate of $18 billion. Those three numbers change the debate from “how high is peak HBM pricing?” to “how long does undersupply persist if Micron is still spending into the second half of fiscal 2026?” The print argues that memory supply discipline has not broken yet, even as profitability is already at a level that historically would have invited an aggressive supply response.

That distinction matters because the quarter’s upside was not a narrow accounting beat. On the Street-comparison basis, revenue of $13,643.0 million versus $12,906.7 million was a +5.7% surprise, while EPS of $4.78 versus $3.96 was a +20.7% surprise, which tells us operating leverage, mix, and pricing mattered more than units alone. The quarterly history makes the inflection visible: Micron’s gross margin moved from 44.7% in Q4 FY2025 to 56.1% in Q1 FY2026, while revenue moved from $11,315.0 million to $13,643.0 million. That is the market’s first clue that the cycle is not merely recovering from the 2023 trough, when gross margin was -17.8% in Q3 FY2023 and revenue was $3,752.0 million; it is already generating margins above the prior recovery levels of 38.4% in Q1 FY2025 and 37.7% in Q3 FY2025. The second clue is that Q1 FY2026 revenue was up +20.6% QoQ and +56.7% YoY, which is not the shape of a one-quarter HBM pull-forward if inventory were already building.

The financial trajectory also explains why the EPS surprise was larger than the revenue surprise. Micron’s reported gross margin step from 44.7% in Q4 FY2025 to 56.1% in Q1 FY2026 absorbed operating expenses that CFO Mark Murphy said were $1.3 billion, up $120 million quarter over quarter and in line with guidance. The company’s own call basis had Murphy state, “Total fiscal Q1 revenue was $13.6 billion, up 21% sequentially and up 57% year over year, setting a quarterly record for the third consecutive quarter.” That wording matters less for the record claim than for the cadence: management is identifying the current run as consecutive, not episodic. Operating income of $6.4 billion and operating margin of 47% reinforce that interpretation, because the business is already converting the mix shift into cash rather than merely carrying higher WIP and receivables. Fiscal Q1 operating cash flows were $8.4 billion, capital expenditures were $4.5 billion, and free cash flow was $3.9 billion, so the near-term spending step is being funded from the cycle rather than from a balance-sheet stretch.

The reason this print should alter portfolio positioning is that Micron tied the margin and revenue acceleration to a larger and earlier HBM market, not just to a quarter of DRAM pricing. CEO Sanjay Mehrotra said, “We forecast an HBM TAM CAGR of approximately 40% through calendar 2028, from approximately $35 billion in 2025 to around $100 billion in 2028.” The phrase that changes the model is not the CAGR alone, but the timing: he also said the “$100 billion HBM TAM milestone is now projected to arrive two years earlier than in our prior outlook.” That is the variant perception. Investors already know HBM is scarce; the more important point is that the industry’s revenue pool is being pulled forward by two years while Micron is committing incremental capacity dollars now. If the $100 billion opportunity arrives in 2028 rather than later, the risk shifts from “Micron over-earns in 2026” to “customers under-secure supply into 2027,” which is a different multiple discussion for a memory name.

The breadth of demand reduces the risk that the whole thesis rests on one accelerator customer or one package generation. DRAM revenue was a record $10.8 billion, up 69% year over year, and represented 79% of total revenue, but NAND was not absent from the acceleration: NAND revenue was a record $2.7 billion, up 22% year over year, and represented 20% of total revenue. Management also put data center NAND portfolio revenue above $1 billion in fiscal Q1, which matters because it gives Micron an enterprise-storage leg alongside HBM and DDR5 rather than a pure AI-memory exposure. Business-unit mix supports the same conclusion: Cloud Memory Business Unit revenue was $5.3 billion and 39% of total company revenue, Core Data Center Business Unit revenue was $2.4 billion and 17% of total company revenue, Mobile and Client Business Unit revenue was $4.3 billion and 31% of total company revenue, and Automotive and Embedded Business Unit revenue was $1.7 billion and 13% of total company revenue. The market may be over-ascribing the quarter to a single data center product when 31% of revenue still came from Mobile and Client and 13% came from Automotive and Embedded, both of which benefit if DRAM supply is diverted toward AI.

That capacity diversion is exactly why the CapEx guide is not a negative signal, even though memory investors are trained to sell rising supply. Micron plans to increase fiscal 2026 CapEx to approximately $20 billion versus the prior estimate of $18 billion, and Murphy framed it with a conditional but important duration marker: “To address tight supply-demand conditions, if extending beyond 2026, we now project our capital spending in fiscal 2026 to be approximately $20 billion, weighted to the second half of the fiscal year.” The conditional language is the right way to read it. Management is not saying the industry is already flooded with economics that justify indiscriminate wafer starts; it is saying tightness may extend beyond 2026, and the spending is weighted to the second half of the fiscal year. That timing should make portfolio managers less worried about a near-term margin break in Q2 FY2026, where the guide is revenue of $18.7 billion, plus or minus $400 million, gross margin of 68%, plus or minus 100 basis points, operating expenses of $1.38 billion, plus or minus $20 million, and EPS of $8.42 per share, plus or minus $0.20.

The balance sheet and inventory numbers also do not support a cycle-top call yet. Ending inventory was $8.2 billion, down $150 million sequentially, with days of inventory at 126. In an overheated memory cycle, one would expect the bearish case to point to inventory rising faster than revenue or to customers visibly building buffers ahead of price increases. Instead, inventory was down sequentially while revenue was up +20.6% QoQ on the reported quarterly-history basis. Cash and investments were $12 billion at quarter-end, and liquidity was $15.5 billion including the untapped credit facility, so the $20 billion CapEx plan is not a forced bet that depends on immediate debt-market access. Fiscal Q1 taxes were $977 million on an effective tax rate of 15.1%, which also means the EPS bridge is being taxed at a normal-looking rate rather than being flattered by an obvious one-off tax benefit in the data pack.

The read-through for the supply chain is therefore more specific than “AI is good for semicap.” For customers, NVIDIA and AMD are the named beneficiaries and competitors for scarce HBM/DRAM supply, with Micron explicitly discussing HBM3E and HBM4 and saying it will begin to ramp HBM4 in the CQ2 time frame in line with customer demand. Apple and Intel sit on the other side of the same allocation problem as memory customers, while Nanya Technology has a Micron DRAM technology license and is exposed to the same DRAM pricing environment. For suppliers, the cleanest read-through is the fiscal 2026 CapEx increase to approximately $20 billion from $18 billion: ASML is tied to EUV and DUV lithography scanners, Applied Materials to etch, CVD, and PVD for DRAM/NAND fabs, Lam Research to etch and deposition for 3D NAND/DRAM, KLA to inspection and metrology, Advantest to T5500 memory test, Axcelis to Purion implanters, Tokyo Electron to coater/developer, etch, and deposition, TSMC to advanced DRAM/HBM controller fabrication, HPSP to high-pressure annealing tools, and Hanmi Semiconductor to TC bonders with approximately 50 units ordered. The magnitude to use is not a guessed allocation by vendor; it is the $2 billion CapEx increase and the $20 billion fiscal 2026 budget, weighted to the second half of the fiscal year.

The competitive context supports the idea that Micron is participating in a broad memory profit pool, not uniquely over-earning in isolation. In the peers table, MU’s latest reported quarter shows $41,456.0 million of revenue, 84.6% gross margin, and +345.7% revenue YoY. Nanya Technology shows 49,086.9 million of revenue, 67.9% gross margin, and +582.9% revenue YoY, while SNDK shows $5,950.0 million of revenue, 78.4% gross margin, and +251.0% revenue YoY. The right inference is that memory pricing and mix are lifting the group, but Micron’s scale and HBM exposure put it in the cohort where gross margin can detach meaningfully from traditional commodity-memory ranges. That also means the risk is group-level supply response rather than Micron-specific demand failure. If peers keep posting gross margins such as 67.9% or 78.4%, supply discipline will be tested, but Micron’s second-half-weighted CapEx language argues that the physical response is not immediate enough to break the next-quarter guide.

The call delivery was consistent with a management team raising the revenue and CapEx ceiling while becoming less promotional about AI language, which is useful because it lowers the risk that the message was only narrative inflation. In the tone history, Q1 FY2026 sentiment was 0.58 versus 0.55 in Q4 FY2025, guidance_tone was 0.49 versus 0.50, tone_confidence was 0.61 versus 0.59, prepared_sentiment was 0.02 versus 0.02, qa_sentiment was 0.56 versus 0.54, ai_optimism was 0.45 versus 0.56, uncertainty was 33.2 versus 34.4, and qa_evasiveness was 33.3 versus 34.5. The combination is telling: sentiment and confidence improved slightly, uncertainty and evasiveness fell slightly, but ai_optimism declined. That is exactly the tone profile investors should prefer in a memory upcycle, because the numbers are moving up without the transcript becoming more AI-promotional by the model score. The conflict in the broader tone table is that later Q3 FY2026 shows guidance_tone at 0.72 and tone_confidence at 0.74 but also uncertainty at 42.6 and qa_evasiveness at 42.8; for this Q1 FY2026 event, however, the call statistics align with the thesis rather than complicate it.

That measured tone also helps separate company-reported basis from the Street-comparison basis. The print’s EPS actual is $4.78 versus the $3.96 estimate, and the print’s revenue actual is $13,643.0 million versus $12,906.7 million. On the call, Murphy used the company’s own phrasing that non-GAAP diluted EPS in fiscal Q1 was $4.78, with 58% sequential growth and 167% versus the year-ago quarter. Those are not numbers to blend with the quarterly-history diluted EPS of $4.60; they serve different purposes in the data pack. For investment work, the Street-comparison figures answer whether estimates were too low, while the call figures answer how management wants investors to understand the company’s own operating trajectory. Both point the same way on this event: the quarter beat, but the more durable estimate revision should come from the Q2 FY2026 guide and the pulled-forward HBM TAM.

The main bear case is no longer that demand was overstated in Q1 FY2026; it is that the new $20 billion fiscal 2026 CapEx plan becomes the seed of a late-cycle oversupply problem. That risk is real because memory has always converted high margins into new capacity, and Micron’s own gross margin moved to 56.1% in Q1 FY2026 with a Q2 FY2026 guide of 68%, plus or minus 100 basis points. But the available numbers argue the oversupply break is not yet in the next quarter. Inventory was $8.2 billion, down $150 million sequentially, days of inventory were 126, Q2 FY2026 revenue guidance is $18.7 billion, plus or minus $400 million, and the CapEx increase is weighted to the second half of the fiscal year. The market may sell the CapEx increase reflexively; we would treat that as the wrong reaction unless order visibility weakens or inventory reverses higher before the capacity spend begins translating into output.

What to watch next is precise. The thesis is confirmed if Q2 FY2026 revenue lands at or above the $18.7 billion guide, plus or minus $400 million, gross margin stays inside or above the 68%, plus or minus 100 basis points range, and EPS tracks the $8.42 per share, plus or minus $0.20 guide on the approximately 1.15 billion share count. It is also confirmed if inventory remains near or below $8.2 billion, days of inventory does not rise materially from 126, and management repeats that fiscal 2026 CapEx is approximately $20 billion and weighted to the second half of the fiscal year rather than pulling more spending into the near term. The thesis breaks if Q2 FY2026 gross margin falls below the guided 68%, plus or minus 100 basis points range while inventory rises from $8.2 billion and days of inventory rises from 126, because that would indicate pricing power is fading before the new capacity even arrives. The next date to anchor is fiscal Q2 FY2026, where the company has already put $18.7 billion of revenue, 68% gross margin, $1.38 billion of operating expenses, and $8.42 of EPS in front of investors; those numbers, not the headline beat, will decide whether Micron is still a duration-of-tightness story or just another memory cycle peaking faster than the bulls expect.