Billionaire David Tepper of Appaloosa Slashed His Micron Stake and Piled Into a "Fabulous" AI Stock for a 6th Consecutive Quarter

Appaloosa’s David Tepper cut his Micron stake by 41% after exiting 690,000 shares, with Micron shares having more than tripled in Q2 to over $1,200 (up ~1,300% since mid-2023). Offsetting this, Tepper added 322,500 shares of Taiwan Semiconductor Manufacturing (TSMC), increasing the position by 24% for a sixth straight quarter, underscoring the AI-infrastructure buildup. The article frames both moves as largely positioning around AI demand and potential cycle/bubble risk rather than a single fundamental break, suggesting modest near-term stock-level impact.

Analysis

This is better read as a relative positioning signal than a sector-wide call: the market is still crowding into AI exposure, but the preferred exposure is shifting from memory beta toward the higher-quality bottleneck in the stack. TSM has the cleaner moat because advanced packaging/foundry capacity is harder to replicate and less exposed to spot pricing, while MU’s economics remain more cyclical and more vulnerable to a pause in incremental AI orders. In other words, the trade is not “AI up or down,” it is “which part of AI gets paid first.”

The immediate implication is for relative performance, not absolute direction. If hyperscaler capex stays strong over the next 1-3 months, TSM should outperform MU as investors reward visibility, utilization, and pricing power; if spending slows, MU is likely to be hit harder because memory demand can correct faster than foundry demand and margin assumptions can compress abruptly. The second-order effect is on suppliers and peers: equipment vendors and packaging-adjacent names tied to advanced-node supply could stay bid, while memory peers with less differentiated HBM exposure may face multiple compression.

The contrarian point is that 13F-driven enthusiasm can overstate signal quality: this is stale data, and some of the MU trimming may simply reflect portfolio rebalancing after a sharp run. The real falsifier for the bullish TSM / cautious MU relative view is a sharp step-down in AI capex or a negative utilization/margin guide from either company over the next quarter. Conversely, if TSM continues to show capacity tightness while MU sees even modest inventory or pricing normalization, the pair can widen materially over 6-18 months.

Key Decisions for Investors

• Long TSM / short MU as a relative-value pair for the next 1-3 months; target outperformance if AI capex remains firm and advanced-node utilization stays tight. Risk: pair reverses if memory pricing re-accelerates or TSM guidance disappoints on capex intensity.
• If already long broad semis, rotate toward TSM and reduce MU weight rather than adding net exposure; TSM offers better earnings visibility and lower cyclicality for similar AI exposure.
• Use SMH as a hedge against an outright AI sentiment drawdown while keeping a TSM overweight; this keeps the portfolio exposed to the bottleneck winner without taking full beta.
• Set an alert on next quarterly guide from TSM and MU: any utilization or margin commentary that implies pricing power is fading would be the clearest signal to exit the relative trade.
• Avoid chasing MU on momentum alone until there is evidence of sustained HBM pricing strength beyond current order book visibility; the setup is more fragile than the market is likely discounting.