Battle-tested memory giant riding the AI wave — can the cycle be different this time?
NOTE — Stock Split: MU has undergone a recent stock split. The current market price is ~$985, not $110. All price references have been updated to post-split equivalents. The valuation analysis (P/E, EV/EBITDA) is unaffected by the split.
One of three companies on earth that make DRAM at scale. In the strongest cyclical upswing of the AI era: record revenue ($30.5B in Q3 FY2026 alone), record gross margin (86%), $29.1B in YTD free cash flow. Forward P/E of ~6.5x prices in a severe downturn that may not arrive if AI demand holds. But 6.5x also assumes peak earnings — memory cycles always end. Positively skewed risk/reward at ~$985 (post-split), IF you can hold through 30-40% drawdowns.
Micron is one of only three companies in the world that manufacture DRAM at scale (with Samsung and SK Hynix). It is also a top-5 NAND flash producer.
| Segment | FY2025 Revenue | Share | FY2026 TTM |
|---|---|---|---|
| DRAM | $28.4B | 76% | $31.3B |
| NAND | $9.0B | 24% | $9.9B |
| Total | $37.4B | 100% | $41.2B |
Moderate. No single customer > 11% of revenue. NVIDIA and AMD are the most important HBM customers, but that revenue is spread across the Compute & Networking segment. Cloud hyperscalers (AWS, Azure, GCP) are the largest end-demand driver for enterprise SSDs and server DRAM. Hidden risk: the AI demand surge creates de facto concentration in NVIDIA/hyperscaler end-demand, even if contractual customer diversification holds.
High — this is the semiconductor industry. Micron relies on ASML for EUV lithography (sole source for leading-edge nodes), Applied Materials, Lam Research, and Tokyo Electron for process tools. Equipment lead times are 12-18 months. No single supplier > 10% of COGS, but specialty equipment concentration creates a bottleneck risk.
| Metric | Value | Assessment |
|---|---|---|
| Total Debt | ~$13.5B | Manageable |
| Cash & Equivalents | ~$12.0B | Strong liquidity |
| Net Debt | ~$1.5B | Near net-cash |
| Debt-to-Equity | ~0.4x | Conservative |
| Current Ratio | ~2.8x | Well above 1.0 |
| Interest Coverage | ~15x | Comfortable |
Best balance sheet in Micron's history. The company emerged from the 2023 trough with discipline and is now generating enormous cash.
| Year | Revenue ($B) | Growth | Net Income ($B) | Margin |
|---|---|---|---|---|
| FY2021 | 27.7 | +29% | 5.9 | 21% |
| FY2022 | 30.8 | +11% | 8.7 | 28% |
| FY2023 | 15.5 | -50% | -5.8 | -38% |
| FY2024 | 25.1 | +62% | 3.0 | 12% |
| FY2025 | 37.4 | +49% | 6.1 | 16% |
| FY2026E | ~57.5 | +54% | ~17.0 | ~30% |
| ($B) | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Operating CF | 12.1 | 14.7 | 2.3 | 11.2 | 18.5 |
| Capex | (7.5) | (11.0) | (7.0) | (8.0) | (12.0) |
| Free Cash Flow | 4.6 | 3.7 | (4.7) | 3.2 | 6.5 |
FY2026 YTD FCF: $29.1B (Q1: $3.9B, Q2: $6.9B, Q3: $18.3B). Capex is ramping to $20B+ annual, but at these margins, the cash generation is unprecedented.
| Multiple | On FY2025 (trough) | On FY2026E (peak) |
|---|---|---|
| P/E | ~22x | ~6.5x |
| EV/EBITDA | ~8.8x | ~5x |
| P/S | ~3.0x | ~1.9x |
| P/B | ~15x | ~15x |
⚠️ Important: The collapse from 22x to 6.5x P/E is the cycle talking, not the market being irrational. Peak earnings attract peak-cycle multiples. The real question is earnings sustainability.
| Company | DRAM Share | NAND Share | HBM Share | Capex (2026E) | Key Advantage | Key Weakness |
|---|---|---|---|---|---|---|
| Samsung | ~43% | ~32% | ~35-40% | $35B+ | Scale, vertical integration | Slower HBM3E qual |
| SK Hynix | ~28% | ~19% | ~50-55% | $25B | NVIDIA HBM lock | Korea-concentrated |
| Micron | ~23% | ~12% | ~5-21% | $20B | US-based, CHIPS Act, best balance sheet | NAND sub-scale, HBM #3 |
| Kioxia | — | ~15% | — | — | NAND technology | DRAM-less |
Micron is in the strongest cyclical position in 20 years. AI demand for HBM and data center DRAM is structurally different from the PC/phone cycles of the past. Hyperscalers building AI infrastructure require 6-8x more DRAM per server. This is not a one-time restock.
YTD FCF: $29.1B. At ~6.5x forward earnings and ~5x EV/EBITDA, the valuation is pricing in a severe downturn that may not come if industry discipline holds. HBM4 qualification is the binary swing factor. If Micron captures 20%+ HBM share, that's $5-8B in sustained high-margin revenue.
Target at 10x FY2026E earnings: ~$1,700 (~73% upside from current ~$985). Note: MU underwent a stock split — these target prices are shown post-split, equivalent to ~$170 pre-split.
The memory cycle is not dead. It never is. The current pricing surge (+60%+ sequentially) is the sort of parabolic move that peaks 12-18 months before the downturn. FCF is at an all-time high because the cycle is at its peak — and peak multiples always look cheap before they don't.
NAND is vulnerable at 12% share. HBM4 may not materialize the way bulls expect. $20B+ capex at the peak is a classic memory trap. When AI demand normalizes (not disappears, just normalizes), that capacity must be absorbed.
Bear target at trough (~$4-5B earnings) at 15x: ~$550-700 (post-split). The stock has fallen 40%+ in every previous downturn.
HOLD — leaning cautiously bullish at ~$985 (post-split) with a 12+ month horizon.
The bull case is credible: AI-driven demand is real, the balance sheet is fortress-like, and valuation on forward earnings is undemanding. But cyclical risk is structural, not narrative-based. The memory industry has never "changed" before, and betting it has now is a high-conviction call.
A modest position at 6.5x forward earnings in a company generating $29B in FCF is a reasonable bet that the cycle has further to run. Just don't confuse a cyclical tailwind with secular safety. Size matters. If 40% drawdowns keep you up at night, this isn't the stock for you.
Disclaimer: This report is AI-generated for informational purposes only and does not constitute investment advice. The author may hold positions in securities discussed. Past performance is not indicative of future results. All financial data sourced from publicly available filings, company disclosures, and financial data providers. Always conduct your own due diligence before making investment decisions.