SK Hynix Stock Debuts for U.S. Investors Tomorrow. The DRAM ETF Could Be the Biggest Loser.
By Maksym Misichenko · Yahoo Finance ·
By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel is divided on the DRAM ETF's outlook, with concerns about concentration risk, high expense ratio, and memory cycle volatility, but also potential opportunities from HBM scarcity and widened access through SK Hynix's US ADS listing.
Risk: Concentration risk and memory cycle volatility
Opportunity: Potential HBM scarcity premium through 2025
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Since its blockbuster debut on April 2, the Roundhill Memory ETF (DRAM) has shattered every historical funding record. It crossed the $1 billion mark in just its first 19 days and rocketed past $6.5 billion in assets over its first few weeks. That's because it's devoted specifically to the hottest segment of the hottest trade on Wall Street – memory solutions for AI.
Look at where it sits today, about 14 weeks after its launch. More than $22 billion in assets. Not bad for an ETF which is about 75% allocated to just three stocks. And when we add in the next four by weight within DRAM, we get more-than-90% allocation.
There's also a lot not to like here. For starters, I think the stocks it holds are likely to crash sooner rather than later.
Then, you can look at investors paying a 0.65% expense ratio for this ETF when you can essentially replicate its performance, fee free, by outright holding its three biggest holdings.
Until this week, its one silver lining was that DRAM provided rare access to SK Hynix for U.S. investors. The leading memory company has a thinly traded over-the-counter listing, and you can access it through a single-country South Korea ETF. But nothing quite compared to DRAM.
Before DRAM launched, if you wanted direct exposure to the high-bandwidth memory chips powering the AI boom, your choices were highly limited. You either had to put all your eggs into one basket with a U.S. stock like Micron (MU), or wade through complex international accounts to buy Asian giants like Samsung or SK Hynix. DRAM wrapped the entire specialized hardware pipeline into a single, easy-to-buy U.S. ticker.
Its top holdings are a powerful triumvirate of Samsung, Micron, and SK Hynix. DRAM is far from the traditional "diversified" ETFs we are used to. That allows investors to truly target this trio of HBM leaders.
But the memory landscape is about to dramatically change. Later this week, SK Hynix will list American Depositary Shares (ADSs) under the ticker SKHY, opening the door wide to U.S. investors. Does that make the DRAM ETF irrelevant?
Plus, the memory chip market is famous for its extreme price swings, and the fund's honeymoon phase is facing its first major reality check.
The primary catalyst for this recent bump came straight from South Korea, where tech giant Samsung Electronics reported preliminary quarterly financial results. On paper, the numbers looked incredible. Samsung's operating profits jumped significantly, driven by skyrocketing global memory demand and higher microchip selling prices. Yet, instead of zooming higher on the stellar news, major memory stocks immediately faced a wave of aggressive profit-taking.
This dramatic "sell the news" reaction highlights a common trap in cyclical tech investing. Because investors had spent months aggressively buying up chip stocks ahead of these earnings releases, the blowout numbers were already fully priced into the market. The sudden drop proves that Wall Street is no longer questioning whether these chip companies are highly profitable today. Instead, big money is actively worrying about how long this massive AI demand cycle can last before supply catches up.
Adding fuel to the fire, the explosive popularity of the DRAM fund has triggered the creation of even riskier trading tools built right on top of it. Wall Street issuers recently launched (RAM) and (DRAL), a pair of 2x leveraged versions of DRAM's portfolio. Of note, while there is an application filed for an inverse DRAM ETF, it has yet to reach the market.
While these hyper-aggressive tools are pulling in massive short-term trading volumes, they add an immense layer of risk for normal investors. When an elite, go-go market theme starts to experience heavy profit-taking and daily price swings get choppy, the daily math inside a leveraged fund can erode your cash surprisingly fast.
The SK Hynix U.S. listing creates a, dare I say, dramatic situation for holders of DRAM. It is not exactly "why buy the cow if you can get the milk for free," but more like "why own DRAM when you can just own SKHY?" Not to mention, Micron is easy enough to buy, other than the fact that it takes about $950 to own one full share. That's half of DRAM right there.
That said, investors have a psychological attraction to what has worked for them. So my suspicion is that DRAM's asset base is not likely to flow out so quickly, simply because SKHY now exists. (QQQ) and (SPY) holders can similarly own 5-10 stocks and get a similar vibe to those market-trackers, and they still pile in.
At the end of the day, then, DRAM's recent post-earnings stumble is a clear reminder that the memory space remains a highly volatile arena where yesterday's winners can cool off in a hurry. SK Hynix's U.S. listing only complicates that.
Rob Isbitts is a semi-retired CIO, former fiduciary investment advisor, and Barchart columnist. Check out his other work at ETFYourself.com (featuring the Fresh Charts weekly trading post), and ROAR.PiTrade.com, helping investors to better-manage their own portfolios.
On the date of publication, Rob Isbitts did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com
Four leading AI models discuss this article
"The DRAM ETF's primary risk is not the arrival of direct SK Hynix shares, but the inevitable cyclical mean reversion of HBM pricing as supply catches up to AI-driven demand."
The arrival of SK Hynix (SKHY) ADSs on U.S. exchanges is a structural challenge for the DRAM ETF, but the article overestimates the impact on retail flows. While 0.65% is a steep expense ratio for a concentrated basket, the ETF offers a 'set-it-and-forget-it' vehicle for thematic investors who lack the appetite for currency risk or individual stock volatility management. The real risk isn't the SKHY listing; it's the HBM (High Bandwidth Memory) supply-demand equilibrium. If HBM capacity expansion outpaces AI server deployment by Q4, the 15-20x forward P/E multiples currently assigned to these memory giants will compress rapidly, regardless of whether you hold them via DRAM or direct equity.
The DRAM ETF may actually benefit from the SKHY listing, as increased U.S. liquidity for the underlying asset reduces the fund's tracking error and execution costs, potentially justifying the fee for passive allocators.
"N/A"
[Unavailable]
"DRAM's real risk is cyclical (memory oversupply in 2026) not structural (SKHY competition), but the article provides no data on HBM supply timelines or demand elasticity to adjudicate which dominates."
The article conflates two separate issues: DRAM's structural vulnerability (concentration, fee drag, redundancy post-SKHY listing) with cyclical memory-chip weakness. The 'sell the news' reaction after Samsung's blowout earnings is real and concerning—it signals the market has front-run the AI demand narrative. However, the article undersells a critical counterpoint: if HBM supply truly lags demand through 2025-26, current prices may not reflect the scarcity premium yet. DRAM's $22B inflow despite 90% concentration in three names suggests retail is buying conviction, not chasing momentum. The leveraged ETFs (RAM, DRAL) are a red herring—they're symptoms, not causes. The real question: is memory pricing peaking now, or are we in year-two of a multi-year upcycle?
SKHY's listing doesn't kill DRAM if the fund's appeal is convenience and thematic exposure, not arbitrage—and the article's own comparison to QQQ/SPY holders suggests sticky capital. More damaging: if memory demand accelerates further, DRAM's concentrated bet on Samsung/Micron/SK Hynix outperforms a diversified chip ETF, making the fee-drag complaint moot.
"The ETF’s fate is a pure play on a narrow memory cycle; if AI demand cools or supply accelerates, its heavily concentrated exposure could lead to sharper drewdowns than a more diversified tech sleeve."
DRAM is a highly concentrated, momentum-driven bet on AI memory demand. The Roundhill DRAM ETF is ~75% Samsung, Micron, and SK Hynix, with >90% in the top nine names—exposed to a single cycle rather than a broad tech theme. SK Hynix’s US ADS listing could sustain flows by widening access, potentially strengthening the ETF’s case even as it duplicates exposure. Yet the article glosses over critical risks: (1) 0.65% expense versus near-zero replication costs from owning a few stocks, (2) concentration risk if any top name underperforms, (3) memory-cycle volatility and abrupt profit-taking, and (4) leveraged products magnifying drawdowns. If AI demand fades or supply outruns demand, DRAM’s key bets could unwind quickly.
Counterpoint: SKHY’s US listing could actually boost DRAM ETF’s relevance and liquidity, and a concentrated trio can outperform if the memory cycle stays hot, so the article’s ‘biggest loser’ take might be overstated.
"Thematic ETFs holding cyclical commodities face catastrophic retail-driven outflows the moment the AI narrative shifts, regardless of long-term HBM supply fundamentals."
Claude, you’re missing the capital allocation reality: retail investors in thematic ETFs are rarely macro-cyclical thinkers; they are momentum chasers. If the memory cycle peaks, these investors won't wait for a 2026 supply-demand rebalance—they will rotate into the next narrative. The real risk isn't just the expense ratio or SKHY listing; it's the 'ETF-ification' of a cyclical commodity. When the HBM scarcity premium evaporates, these retail-heavy vehicles will face massive, indiscriminate outflows, regardless of the underlying fundamentals.
[Unavailable]
"DRAM's fee is only a problem if memory demand disappoints; concentration risk is real but secondary to the cyclical timing question."
Gemini conflates retail behavior with fund mechanics. Yes, momentum chasers will flee when the cycle peaks—but that's true whether they own DRAM or Samsung directly. The real arbitrage Gemini misses: DRAM's 0.65% fee is painful *only if* the underlying thesis is wrong. If HBM scarcity holds through 2025, concentration becomes a feature, not a bug, and the fee drag disappears into rounding error. The article's 'biggest loser' framing assumes the cycle is already pricing in the upside. That's unproven.
"Concentration risk plus a 0.65% fee remains material, and SKHY's listing does not eliminate idiosyncratic risk or cushion a demand-cycle reversal."
One gap I’d push on is the concentration risk stressed by Claude. Even if HBM scarcity lasts, the ETF’s fate is a swing of three stocks; SKHY's US listing may aid liquidity, but it won’t diversify idiosyncratic risk or cushion a sudden demand collapse. The 0.65% fee remains a drag vs owning a few names, and a snap reversal in the memory cycle could slam the ETF despite scarcity pricing.
The panel is divided on the DRAM ETF's outlook, with concerns about concentration risk, high expense ratio, and memory cycle volatility, but also potential opportunities from HBM scarcity and widened access through SK Hynix's US ADS listing.
Potential HBM scarcity premium through 2025
Concentration risk and memory cycle volatility