David Tepper's Appaloosa Bets Big on AI Memory, Pivoting to Micron and South Korean Chips

Investor David Tepper's fund, Appaloosa, has strategically shifted its portfolio to capitalize on the AI boom, significantly increasing investments in memory chip maker Micron and a South Korea-focused ETF.

This move is driven by the belief that the current bottleneck in AI development is not computing power itself, but the supply of high-bandwidth memory (HBM), which is dominated by Korean firms like Samsung and SK hynix.

By trimming positions in companies like AMD, which are more exposed to US-China policy risks, Appaloosa is concentrating its capital on the memory sector, further supported by South Korea's favorable 'Value-Up' corporate governance program.

Famed investor David Tepper's fund, Appaloosa, is making a significant strategic shift in its portfolio, betting heavily on the future of AI.

The core of this bet is a focus on what's called the AI 'memory bottleneck.' Think of it this way: while powerful GPUs from companies like Nvidia and AMD are the 'brains' of AI, they need incredibly fast, high-capacity memory to function effectively. This special memory, called HBM (High-Bandwidth Memory), is in extremely high demand and short supply. Google DeepMind's CEO even called it the industry's 'choke point,' and Appaloosa is investing in the companies that control this critical resource.

So, how is Tepper making this play? First, the fund tripled its investment in Micron (MU), a leading U.S. manufacturer of memory chips. Micron has reported blowout earnings, indicating that its HBM supply for 2026 is already effectively sold out. Second, Appaloosa made a major new investment in the iShares MSCI South Korea ETF (EWY). This is a clever way to invest in the two undisputed leaders of the HBM market, Samsung Electronics and SK hynix, which are the top holdings in this ETF.

This move also involves reducing risk in other areas. Appaloosa trimmed its stake in AMD, a company more directly exposed to the risks of U.S.-China chip export restrictions. By focusing on memory instead of compute (GPUs), the fund sidesteps some of that geopolitical tension. It also exited positions in regional banks, likely due to lingering concerns about credit quality in that sector.

In essence, Tepper's strategy is a clear rotation. He is moving capital away from areas with higher policy and funding risks and concentrating it on the tightest, most profitable point in the current AI supply chain: high-performance memory. This is further supported by South Korea's 'Value-Up' program, a government initiative to improve corporate governance and shareholder returns, which has been drawing significant investment into the Korean market.

HBM (High-Bandwidth Memory): A type of high-performance memory crucial for AI processors (GPUs) that enables faster data processing by stacking memory chips vertically.
ETF (Exchange-Traded Fund): An investment fund traded on stock exchanges, much like stocks. It holds assets such as stocks or bonds and typically tracks a specific index.
Bottleneck: A point of congestion in a system that limits overall throughput. In the AI supply chain, the limited supply of HBM is slowing down the production of AI systems.

David Tepper's Appaloosa Bets Big on AI Memory, Pivoting to Micron and South Korean Chips

Famed investor David Tepper's fund, Appaloosa, is making a significant strategic shift in its portfolio, betting heavily on the future of AI.

HBM (High-Bandwidth Memory): A type of high-performance memory crucial for AI processors (GPUs) that enables faster data processing by stacking memory chips vertically.
ETF (Exchange-Traded Fund): An investment fund traded on stock exchanges, much like stocks. It holds assets such as stocks or bonds and typically tracks a specific index.
Bottleneck: A point of congestion in a system that limits overall throughput. In the AI supply chain, the limited supply of HBM is slowing down the production of AI systems.