AI Memory Chip Surge Pushes DRAM ETF Past BlackRock IBIT Record

The artificial intelligence infrastructure boom is rapidly reshaping the exchange-traded fund market, and one semiconductor-focused fund is now rewriting ETF history. The Roundhill Memory ETF (DRAM), built around global memory chip manufacturers, has officially become the fastest-growing ETF ever, overtaking the explosive early growth of BlackRock’s iShares Bitcoin Trust (IBIT).

Fuelled by soaring demand for AI memory chips, high-bandwidth memory (HBM), and data center expansion, investors are pouring billions into semiconductor-linked ETFs at a pace Wall Street has rarely seen. The rise of DRAM highlights how AI infrastructure investing is becoming one of the strongest themes in global markets during 2026.

DRAM ETF Breaks Growth Records Amid AI Memory Demand

The Roundhill Memory ETF reached nearly $6.5 billion in assets under management in just over a month after launch, surpassing previous ETF growth records held by BlackRock’s spot Bitcoin ETF, IBIT. Analysts say the rally reflects massive investor appetite for exposure to the AI semiconductor supply chain.

Launched in April 2026, DRAM focuses specifically on memory chip companies benefiting from the AI computing race. The ETF includes major industry leaders such as Micron Technology, Samsung Electronics, SK Hynix, and SanDisk, all of which have experienced significant stock price gains this year.

Unlike broader semiconductor ETFs, DRAM gives investors concentrated access to the memory segment, which many analysts now consider the biggest bottleneck in AI development.

Why AI Memory Chips Are Suddenly Dominating Markets

The latest wave of AI expansion has shifted attention away from graphics processing units alone and toward memory infrastructure. As AI models become larger and inference workloads increase, data centers require substantially more high-bandwidth memory to process information efficiently.

Industry experts say memory shortages and limited HBM production capacity are driving both chip prices and investor enthusiasm higher. According to MarketWatch, DRAM has surged nearly 100% since launch, while several underlying holdings posted triple-digit gains in 2026.

SK Hynix, one of the ETF’s largest holdings, has emerged as a major AI beneficiary because of its leadership in HBM production. Samsung Electronics and Micron have also benefited from increased AI server spending and cloud infrastructure demand.

The growing importance of memory technology in AI systems is also being supported by ongoing academic research into next-generation memory architectures and scalable computing systems.

Retail Investors Fuel Massive ETF Inflows

Retail traders have played a major role in DRAM’s explosive rise. Research firms tracking investor behaviour noted that self-directed traders poured money into the ETF faster than they did into many pandemic-era thematic funds or even Bitcoin-related products.

Reuters reported that DRAM attracted nearly $1 billion in inflows during a single trading session earlier this month, an unusually large figure for such a newly launched fund.

Market analysts believe investors are using the ETF as a simplified way to gain diversified exposure to the booming AI infrastructure market without selecting individual semiconductor stocks.

The ETF’s international exposure has also attracted attention because many competing U.S.-listed semiconductor funds do not offer substantial access to South Korean memory giants like SK Hynix and Samsung Electronics.

Risks Remain Despite Strong Momentum

While enthusiasm surrounding AI memory stocks continues growing, some market experts warn that valuations may be overheating. Short sellers have already questioned whether certain smaller memory-related companies are truly benefiting from AI demand or merely riding speculative momentum.

Volatility has also increased across semiconductor markets. DRAM recently experienced sharp pullbacks following rapid gains, showing that investor sentiment can shift quickly in high-growth sectors.

Still, many analysts believe the broader trend remains intact as hyperscalers, cloud providers, and AI developers continue investing heavily in memory-intensive computing infrastructure.