Memory Has Become the Bottleneck of the AI Economy
Micron's Q1 FY2026 results suggest that AI is changing memory from a commodity component into what management calls a strategic asset. Revenue reached a record $13.6 billion, up 57% year over year, while gross margin increased by 11 percentage points sequentially to 36.8%.
The company's framing is that memory performance now plays a central role across data centers and edge devices, making AI demand increasingly important to both pricing and the structure of the memory market.
The shortage behind those numbers is the real story. CEO Sanjay Mehrotra said that in the medium term Micron can meet only about half to two-thirds of demand from several key customers, a gap he described as the widest between memory demand and supply the industry has ever seen. Aggregate industry supply, the company believes, will remain substantially short of demand for the foreseeable future, with tight conditions persisting through and beyond calendar 2026.
Where AI Demand Is Showing Up First
- Servers acceleratingMicron now expects calendar 2025 server unit growth in the high teens, up from 10% guided on the prior earnings call, with strength continuing into 2026.
- PCs upgraded againPC unit forecasts were raised to high single-digit growth for 2025 on Windows 10 end-of-life and AI PCs, though memory supply constraints may limit some 2026 shipments.
- Phones getting denserFlagship smartphones with 12 GB of DRAM reached 59% of the mix in calendar Q3, more than double a year earlier, as AI drives mobile memory content.
- Data center SSDs surgingData center NAND revenue exceeded $1 billion in fiscal Q1, and Micron's PCIe Gen 6 SSD is seeing rapidly increasing qualification commitments including at hyperscalers.
- Auto design winsMicron's ASIL-rated LPDDR5X and UFS 4.1 automotive products have secured billions of dollars in design wins, with L2+ and L3 driving demand today.
HBM: Sold Out, and the Market Is Growing Faster Than Anyone Predicted
High bandwidth memory is at the center of the squeeze. Micron said its 2026 HBM supply is sold out, with negotiations for both volume and pricing already completed for the calendar year.
The dramatic growth in HBM demand is also compounding the wider shortage: each HBM wafer carries a 3:1 trade ratio against DDR5, meaning every unit of HBM capacity crowds out three units of conventional server DRAM, and that ratio only worsens with future generations.
The demand picture keeps being revised upward. Micron raised its HBM revenue forecast and now expects the HBM market to reach a $100 billion TAM by 2028, two years ahead of its prior outlook. Mehrotra noted that customers across the entire AI ecosystem, from custom ASIC programs to established accelerators, are requiring more and more HBM as their architectures evolve, because memory increasingly determines whether AI functionality can be delivered at all.
Micron's competitive bet is HBM4, which it says offers industry-leading performance of over 11 gigabits per second and is ramping in the second calendar quarter of 2026 with a faster yield ramp than HBM3E achieved. Management declined to specify HBM revenue or share targets, emphasizing only that both HBM and non-HBM products carry strong profitability and that mix will be managed across the portfolio.
The Financials: Records Across Every Line
The quarter set records at both the company and business-unit level. DRAM revenue was a record $10.8 billion, up 69% year over year, with prices up roughly 20% on tight supply and favorable mix. NAND revenue hit a record $2.7 billion, up 22% year over year. Every business unit (cloud memory, core data center, mobile and client, and automotive and embedded) posted record revenue, and non-GAAP EPS of $4.78 grew 167% versus the year-ago quarter.
Free cash flow of $3.9 billion broke the company's prior quarterly record (from fiscal Q4 2018) by more than 20%, and Micron returned to a net cash position after cutting debt by $2.7 billion.
The momentum is expected to steepen: fiscal Q2 guidance calls for record revenue of $18.7 billion, gross margin of 68%, and EPS of $8.42, with management indicating margins should continue expanding beyond Q2, though more gradually than the last two quarters' surge.
The Capacity Battle: $20 Billion and a Race Against Cleanroom Time
Micron is spending aggressively but disciplinedly against the shortage. Fiscal 2026 capital expenditures were raised to approximately $20 billion from a prior $18 billion estimate, weighted to the second half of the year and aimed primarily at HBM capability and 2026 NAND supply. Even so, bit shipment growth in calendar 2026 is expected to be constrained by supply (both Micron's and the industry's) at around 20% growth in DRAM and NAND.
The constraint is physical: cleanroom buildout takes time, and lead times are lengthening across geographies. Micron is pulling in its first Idaho fab to first wafer output in mid-2027, beginning construction on a second Idaho fab in 2026, breaking ground on its first New York fab in early 2026 for supply in 2030 and beyond, and investing in Japan, Singapore's HBM packaging facility, and an Indian assembly plant ramping in 2026.
Customers, aware of the shortage, are locking in access. Micron is negotiating multi-year contracts with specific commitments (a far stronger structure than the one-year LTAs of prior cycles) spanning DRAM, NAND, and data center SSDs across multiple market segments.
Meanwhile, AI demand keeps rippling outward: flagship smartphones with 12 GB of DRAM more than doubled their shipment mix year over year, and memory content growth in PCs and edge devices continues even as some unit demand may soften under higher prices.
Charts & Visual Insights
Micron Fiscal Q1 2026 Revenue by Business Unit ($ billions)
All four business units posted record revenue in fiscal Q1 2026, led by the cloud memory unit at $5.3 billion.
| Business unit | Revenue ($B) | Source |
|---|---|---|
| Cloud memory | $5.3B | |
| Mobile and client | $4.3B | |
| Core data center | $2.4B | |
| Automotive and embedded | $1.7B |