Global First Storage ETF Surges 88% Five Weeks After Listing. Is It Too Late to Buy Now?

TradingKey - Roundhill Memory ETF ( DRAM) has surged 88% in just 25 trading days since its listing on April 2, with assets under management (AUM) surpassing $5 billion, a growth rate second only to the iShares Bitcoin Trust ( IBIT )'s 24-day record.

This pure-play memory chip ETF, which holds only nine stocks, has recorded net inflows for 23 consecutive trading days as of May 8, with a single-day peak of $1.1 billion. The explosive growth of DRAM has also sparked a heated market debate: Is it still worth investing in memory chip ETFs?

The only pure-play memory chip ETF

Prior to the listing of DRAM, the US stock market only featured broad-based semiconductor ETFs such as SOXX and SMH, which had very low memory stock exposure. SMH contains only Micron ( MU) as a memory stock, while the total memory stock weight in SOXX is only 7-8%.

DRAM marks the first time the three HBM giants—Micron, SK Hynix, and Samsung—have been bundled into a high-purity product, with their combined weight nearing 70% and memory stock exposure approaching 100%.

The scarcity of such a precise tool is the primary reason for the massive capital inflows in a short period. Combined with the explosive growth expectations for HBM demand due to AI computing expansion, capital has naturally concentrated on this scarce target, driving up short-term gains.

The DRAM ETF management fee is 0.65%, higher than that of broad-based semiconductor ETFs (like SMH at ~0.35%), but it remains within an acceptable range given the scarcity of its pure-play memory theme.

Fundamentals: Memory supply-demand gap reaches a 15-year high.

According to a Goldman Sachs research report, the global DRAM supply-demand gap will reach 4.9% in 2026, marking the most severe shortage in nearly 15 years. The three major memory manufacturers (SK Hynix, Micron, and Samsung Electronics) have already seen their production capacity for the year fully booked, and it typically takes four to five years for a new wafer fab to go from groundbreaking to mass production. Pricing power in the memory industry has shifted from buyers to sellers, providing fundamental support for DRAM ETFs.

From a valuation perspective, DRAM's current P/E ratio is approximately 6.31x, which is at historically low levels. However, analysts warn that memory chips often trade at extremely low P/E ratios at profit peaks; in 2018, Micron's P/E ratio dropped as low as 5-6x during its profit peak before share prices retreated sharply due to a reversal in supply and demand. Whether the current low P/E is an "undervalued buy signal" or a "cyclical peak warning" remains to be verified by upcoming earnings reports.

Risk: Position Concentration and Contrarian Signals

The fund's holdings are highly concentrated, with the top three positions representing nearly 70% of the portfolio. Critics argue it is a product of chasing the hottest market trends. Once memory prices reverse, its sheer size could exacerbate selling pressure.

BTIG analyst Jonathan Krinsky noted on the day of the DRAM ETF's debut that while the Goldman Sachs Storage Index rose approximately 350% and Micron surged over 700% over the past year, the DRAM ETF was only launched recently—a move often viewed historically as a contrarian signal. He pointed to the Meme Stock ETF as a cautionary tale; launched in December 2021 toward the end of the meme stock frenzy, it plummeted more than 70% post-listing before being liquidated in November 2023.

Analysts also emphasized that the DRAM ETF and the Meme Stock ETF differ fundamentally in their underlying basics and capital structures; historical precedents serve as references rather than inevitable outcomes. The former is bolstered by structural HBM shortages and five-year long-term agreements, whereas the latter was purely sentiment-driven.

Is the DRAM ETF still worth an entry?

Strategies should vary for investors with different styles. For medium-to-long-term investors optimistic about the long-term trend of AI storage, building positions in stages with a small allocation of no more than 3% of the portfolio remains a rational choice. For short-term traders, the risk-reward ratio at the current level has significantly declined.

Micron will release its fiscal third-quarter results on July 1, marking a critical juncture to test whether the HBM price hike narrative can continue. The market will examine the gap between Micron's revenue guidance of $33.5 billion (corresponding to an EPS of approximately $18.9) and its actual performance, while watching if fourth-quarter guidance remains flat sequentially, which could be an early signal of a cyclical turning point. Prior to this, chasing the rally requires extreme caution.

Furthermore, the SK Hynix ADR listing is imminent. The world's largest HBM supplier plans to complete its U.S. ADR listing between June and July 2026, raising approximately $10 billion.

Currently, SK Hynix's forward P/E ratio is only about 3-4 times, far lower than Micron's 8-9 times. After the ADR listing, U.S. investors will be able to trade the stock directly for the first time, and the potential for valuation rerating could serve as another tailwind for DRAM ETFs.