One Huge Fund Just Bought Taiwan Semiconductor Manufacturing and Dumped Micron Stock. Which Stock Is the Superior Choice?

The article highlights Appaloosa Management’s Q2 move, selling over 40% of its Micron Technology stake while adding nearly 25% to Taiwan Semiconductor (TSMC), lifting TSMC to ~11% of the fund (third-largest). It argues Micron could benefit most over the next 1-year window from a memory supply shortage driving soaring prices, but that the tailwind may ease as soon as 2028. For a 3–5 year horizon tied to the AI build-out, it favors TSMC as the more resilient option, while noting the trade-off in return magnitude versus potential downturn risk in memory capacity.

Analysis

Near term, the cleaner earnings-revision setup is still MU: when a commodity supply squeeze is this tight, equity upside comes from incremental pricing power, not just volume. The catch is that this is a classic mean-reversion business; if capacity additions from Samsung or SK Hynix arrive faster than expected, the same operating leverage that drives upside will reverse hard. That makes MU a stronger 1-2 quarter trade than a 3-5 year compounding story.

TSM is the higher-quality asset, but the market already pays for that quality, so the marginal upside depends on AI capex staying hot and on continued share gains in advanced nodes and packaging. The second-order risk is that TSM can be fundamentally right and still lag if investors decide the AI supply chain is over-owned or if Taiwan geopolitical premium widens. In other words, TSM is the safer business, not necessarily the better stock over the next 6-12 months.

The consensus is missing that these are different factor exposures: MU is a cyclical pricing beta trade, while TSM is a durability/sovereign-risk compounder. If memory ASPs stay elevated through the next earnings cycle, OEMs and server builders absorb the pain first, which can pressure downstream hardware margins before it shows up in the chip names. Falsifiers: DRAM/HBM price declines, inventory days rising, or any AI capex guide cut that implies TSM utilization is peaking.

Key Decisions for Investors

• Prefer a 3-6 month relative-value long MU / short TSM only if next earnings commentary confirms memory pricing remains tight; target 15-20% spread capture, stop if MU gross margin guidance rolls over.
• Do not chase TSM at current levels; use 8-10% pullbacks to build a 12-18 month core long, with thesis invalidation on AI capex deceleration or a material widening of Taiwan risk premium.
• If memory shortage persists, consider a basket long in semiconductor equipment (AMAT, LRCX, KLAC) as a second-order beneficiary; equipment orders can lag pricing by 1-2 quarters but often rerate faster than chip names.
• Set an alert on MU inventory days and DRAM/HBM pricing into the next print; any inflection there is the earliest signal to take profits or flip the trade.