Memory & HBM leader. AGI Score 10/10. Revenue up 7x from trough. HBM revenue $13.5B and accelerating. | Analysis: 2026-03-13
Micron is THE memory play for AGI. HBM (High Bandwidth Memory) is physically required for every GPU cluster — no software substitute exists. HBM revenue went from ~$1.9B to $13.5B in two years, a 7x increase driven entirely by AI demand. At $472B market cap, Micron is no longer cheap — but it's deeply cyclical. The question is: what price makes this a 10x candidate? Memory cycles create opportunities; we need to understand the floor.
Micron manufactures DRAM (including HBM), NAND flash, and NOR memory. It serves data centers (AI/ML training), mobile, client computing, and automotive markets. Advanced nodes include 1-gamma DRAM with EUV lithography, G8/G9 NAND, and HBM3E 12-high stacks.
Physical manufacturing moat. Building a memory fab costs $15-20B and takes 3-4 years. Only 3 companies worldwide can produce DRAM: Micron, Samsung, SK Hynix. Only 4 can produce advanced NAND. This oligopoly has existed for decades and barriers keep rising with EUV lithography requirements.
Technology lock-in: HBM requires CoWoS packaging and integration with GPU architectures. Once qualified with NVIDIA/AMD, switching costs are enormous. Design wins are sticky for 2-3 year product cycles.
Micron is a pure AGI infrastructure play. AI training and inference consume enormous memory—HBM for GPU clusters, DDR5 for servers, NAND for storage. HBM revenue exploded from $1.87B (2023) to $13.52B (2025), a 7x increase driven entirely by AI. AGI scaling requires exponentially more memory bandwidth and capacity. Micron's HBM3E/HBM4 and high-capacity DDR5 are irreplaceable for next-gen AI. Manufacturing scale (400GB+ of advanced process capacity) takes years to replicate, creating supply bottleneck. Memory is a physical necessity for compute—no software substitute exists. Innovation risk low because new memory types deploy over decades. Textbook AGI winner.
AGI is pure demand for memory. Every AI training run and inference cluster requires massive memory bandwidth and capacity. HBM demand scales directly with GPU deployments. AGI development means exponentially more compute, which means exponentially more memory.
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Revenue | — | $30.8B | $15.5B | $25.1B | $37.4B |
| Net Income | — | $8.7B | ($5.8B) | $778M | $8.5B |
| Operating Income | — | $9.7B | ($5.4B) | $1.3B | $9.8B |
| Operating Cash Flow | — | $15.2B | $1.6B | $8.5B | $17.5B |
| CapEx | — | ($12.1B) | ($7.7B) | ($8.4B) | ($15.9B) |
| Free Cash Flow | — | $3.1B | ($6.1B) | $121M | $1.7B |
| Buybacks | — | ($2.4B) | ($425M) | ($300M) | $0 |
| Item | Value | Notes |
|---|---|---|
| ASSETS | ||
| Total Assets | $82.8B | |
| Cash & Equivalents | $9.6B | |
| Net PP&E | $47.3B | Physical assets |
| Goodwill | $1.1B | Acquisition premium |
| Inventory | $8.4B | |
| Current Assets | $28.8B | |
| LIABILITIES | ||
| Total Liabilities | $28.6B | |
| Current Liabilities | $11.5B | |
| Long-Term Debt | $11.5B | |
| Total Debt | $12.5B | |
| EQUITY | ||
| Stockholders' Equity | $54.2B | Book value |
| Tangible Book Value | $53.0B | Equity - Goodwill - Intangibles |
| Book Value / Share | $52.23 | |
| Debt / Equity | 21.2% | |
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Diluted Shares | — | 1122M | 1093M | 1118M | 1125M |
Share count trend: N/A
Current market cap: ~$480B. For 10x, we need Micron at ~$48B market cap, or roughly $43/share.
Micron traded at $49 in September 2024 and $61 at 52-week low. The stock was at $43 in late 2022 during the memory downturn (revenue dropped to $15.5B, net loss of $5.8B).
Is $43 possible again? Only in a severe memory downcycle where:
Key insight: Memory is the most cyclical semiconductor segment. Micron has gone from $48B market cap to $480B and back. The 10x entry zone is $40-50/share, which requires a major downcycle. Given memory cycle history, this will happen — the question is when. Keep on watchlist for the next downturn.
Data sources: SEC EDGAR XBRL, yfinance, 10-K filings, AGI scoring model (Claude Sonnet 4.5). Analysis date: 2026-03-13.