Tepper Puts $3.1B—40% of Appaloosa—Into Three AI Infrastructure Names

David Tepper's Appaloosa Management disclosed a $7.7 billion equity portfolio in its latest 13F, with $3.1 billion—roughly 40%—concentrated in three positions: Amazon, Micron Technology, and Taiwan Semiconductor Manufacturing. The filing shows Tepper sold 41% of his Micron stake during the quarter even as the stock rose 242%, yet Micron remains his second-largest holding. The move is less about skepticism and more about rebalancing after violent appreciation in a name that delivered exactly what he bet on.

Appaloosa entered Micron heavily in prior quarters, riding the memory semiconductor thesis as high-bandwidth memory became the chokepoint for AI training clusters. The 41% trim came after Micron reported record datacenter revenue and raised full-year guidance twice. Tepper sold into strength but kept enough on to maintain a top-three position, signaling he still expects datacenter memory to remain supply-constrained through 2025. Amazon and Taiwan Semi round out the trio: Amazon for its AWS infrastructure dominance and internal chip development, Taiwan Semi for its monopoly on leading-edge logic production. All three sit at different choke points in the AI stack—hyperscale cloud, memory, and foundry—giving Tepper exposure to infrastructure spend without direct model-layer risk.

The concentration matters because Appaloosa has historically run a diversified book across sectors. A 40% allocation to three names in a single thematic vertical is unusual for a fund that managed $20 billion at its peak and has spent two decades navigating macro volatility. Tepper is not making a six-month trade. He is positioning for a multi-year capital cycle where AI infrastructure spend becomes the largest incremental driver of semiconductor revenue and cloud margin expansion. The trim in Micron reflects profit-taking discipline, not thesis abandonment. When a position doubles in a quarter, selling half still leaves you with the original dollar amount at a much higher cost basis. Tepper kept $600 million to $800 million in Micron after the sale, more than he held in most other disclosed names.

Allocators should watch Micron's forward guidance in its March earnings call, specifically any commentary on high-bandwidth memory supply agreements with Nvidia and hyperscalers. Taiwan Semi's Arizona fab ramp and any incremental Arizona-based orders from U.S. hyperscalers will signal whether domestic supply chain redundancy is moving from political theater to actual capital deployment. Amazon's Q1 AWS revenue growth and operating margin will show whether infrastructure spend is accelerating or plateauing after two years of heavy buildout. If AWS margin expands while revenue growth stays above 15%, it confirms that utilization is catching up to capacity additions, which would validate Tepper's timing on the trim.

The filing also revealed Appaloosa added new positions in several Chinese ADRs, though none reached top-ten size. Tepper has historically used China exposure as a macro hedge, buying during dislocation and exiting on policy relief. The AI infrastructure trio remains the core bet. The next 13F, due in mid-May, will show whether he used the recent Taiwan Semi pullback to add or held position size flat. The dollar amount he kept in Micron after a 242% run is the opinion.