David Tepper’s Appaloosa Management reshaped its tech exposure late in the year, leaning heavily into memory chips by boosting Micron and adding a Korea-focused ETF packed with Samsung and SK Hynix. Those moves signal a clear bet that the AI-driven hunger for fast memory will keep lifting prices and profits for longer than many expect. This article unpacks what Tepper bought, why the trio of manufacturers matters, and what the risks are for investors considering a follow-along play.
Appaloosa now reads like David Tepper’s personal scoreboard, since he’s returned most outside capital and runs the fund mainly with his own money. That change matters because when Tepper moves, he’s not juggling client mandates; he’s placing high-conviction bets with his own wealth. Following those bets can reveal where a seasoned risk-taker thinks returns will come from next.
In the fourth quarter Tepper expanded his Micron position substantially and used options to amplify exposure. The disclosed activity included adding roughly one million shares of Micron Technology and call options tied to another quarter-million shares. Alongside that, he put nearly two million shares into the iShares MSCI South Korea ETF, a play that brings Samsung and SK Hynix along for the ride.
The Korea ETF is not a broad, diversified basket the way it might look at first glance; it’s dominated by two giants. Samsung Electronics and SK Hynix together represent a huge share of the fund’s value, so buying the ETF is effectively exposure to those firms plus Micron. Put another way, Tepper’s three big holdings now cover the bulk of the world’s advanced memory output.
Market moves this year have rewarded that exposure, with both Micron and the Korea-focused fund rising sharply. Strong quarterly results and a tighter memory market pushed prices higher as customers competed for limited capacity. That run-up reflects real demand pressure, but it also raises the question of how much of the rally is sustainable versus driven by temporary tightness.
Tepper’s bigger bet suggests he sees the current cycle stretching out, not ending in a single year. If demand for high-bandwidth memory remains hot, margins and earnings could surprise on the upside for longer than consensus models expect. Investors should note, though, that conviction from a well-known manager is not a guarantee—cycles turn and capacity responds to profits.
For investors who want exposure to memory without picking a single company, the Korea ETF looks convenient but carries concentration risk. Because Samsung and SK Hynix dominate it, the ETF behaves almost like a direct bet on the memory duopoly plus Micron. That can be efficient in a bull cycle but painful if a supply wave or demand dip hits the sector.
The technological driver behind the frenzy is high-bandwidth memory, which gets packaged with GPUs and AI accelerators to reduce data bottlenecks during training and inference. As models get larger and applications multiply, systems need more fast local memory to keep GPUs fed. That technical reality explains why chipmakers are racing to convert and expand capacity right now.
Capacity additions, however, take time—measured in years, not quarters—so the current shortage dynamic could last into 2027 as new fabs come online. That lag encourages some buyers to over-order or hoard stock ahead of expected price hikes, creating a temporary spike that can unwind later. If supply finally catches up in late 2027 and beyond, the tightness driving today’s pricing could fade.
Valuation matters in this backdrop. Micron’s forward multiple sits noticeably above some peers, while SK Hynix looks cheaper by comparison and Samsung occupies an intermediate spot. Those price-to-earnings differentials reflect varied investor expectations about how long elevated margins will persist. For anyone tempted to mimic Tepper, the sensible move is to watch shipment trends, fab ramp schedules, and pricing indicators rather than chase gains at stretched valuations.
