Industry Analysis
The 46% discount is a residual artifact of the market's cyclical bias toward DRAM makers, not a genuine mispricing. The real inflection: HBM4 qualification windows at NVIDIA and AMD will reshuffle supplier shares within two quarters. If Micron's 1β node yield lags SK Hynix by even one quarter, its AI memory narrative collapses from 'credible third source' to 'fallback option.' CoWoS packaging capacity remains the binding constraint—TSMC's allocation decisions determine whether HBM3E translates into actual shipments or stays a roadmap artifact. Strategically, Samsung will almost certainly compress Micron's HBM3E margins by 5-8% in H2 2025, forcing an accelerated HBM4 differentiation push. The genuine risk is not demand—hyperscaler capex cycles extend at least through 2026—but supply: all three players expanding HBM simultaneously creates a structural oversupply risk by 2026. The 12-24 month litmus test is binary: can Micron push AI SKU revenue mix above 30% (from under 15% today) and establish NAND as a second growth vector? If neither materializes, the 46% 'undervaluation' is simply a cyclical peak wearing a value stock's costume.
This page displays AI-generated summaries and metadata for research purposes. Original content belongs to the respective publishers.