New Semiconductor Fund Surges 145 Percent in Two Months on Artificial Intelligence Demand

New Semiconductor Fund Surges 145 Percent in Two Months on Artificial Intelligence Demand

2026-06-02 companies

New York, Tuesday, 2 June 2026.
Driven by artificial intelligence, a new memory chip fund skyrocketed 145 percent since April, highlighting a massive market shift toward critical data storage investments despite looming cyclical risks.

The AI Memory Boom and Unprecedented Gains

The Roundhill Memory ETF (NASDAQ: DRAM) has captured the attention of retail and institutional investors alike following a staggering market rally. On June 1, 2026, the fund closed at $68.00, marking an approximate 145 percent increase from its initial public offering closing price of $27.76 on April 2, 2026 [2]. Trading activity has been intense; on June 1 alone, the ETF saw a trading volume of 41.73 million shares, generating a turnover of roughly $2.8 billion [2]. Artificial intelligence models rely heavily on processing massive datasets, making rapid data retrieval essential for computational efficiency [GPT]. This recent market surge reflects a critical pivot in infrastructure demands, as the tech industry’s focus expands from graphics processing units to the memory components—such as high-bandwidth memory, NAND, and solid-state drives—that currently act as data center bottlenecks [2].

Concentration Risks and the Cyclical Trap

Despite the impressive headline figures, the DRAM ETF presents profound structural risks due to its hyper-concentrated portfolio. The fund functions as a pure-play tactical instrument, with 73.04 percent of its exposure tied to just three companies: Samsung Electronics at 24.99 percent, SK hynix at 24.22 percent, and Micron Technology (NASDAQ: MU) at 23.83 percent [4]. These three corporate giants collectively control approximately 95 percent of the global memory market [5]. While this oligopolistic dominance allows for massive upside during supply shortages, it leaves investors highly vulnerable to sharp downturns during synchronized memory cycle reversals, NAND oversupply, or falling average selling prices [4].

Sources


Semiconductors Exchange-traded funds