The AI Trade Is Getting Harder to Pick, But These 3 ETFs Take a Different Route

Following a 7% pullback earlier in June, the tech-heavy Nasdaq-100 has bounced back more than 5% as bullish investors demonstrate that they believe the artificial intelligence (AI) trade still has plenty of room to run. But with volatility remaining elevated, investors are finding it increasingly difficult to identify potential individual winners.

The Magnificent Seven hyperscalers have seen wildly disproportionate performances. Alphabet (NASDAQ: GOOGL), for example, is up more than 100% in the past year, but Microsoft (NASDAQ: MSFT) has lost more than 20% over the same period.

Meanwhile, pure-play AI stock Palantir (NASDAQ: PLTR) has fallen nearly 36% from its all-time high in November 2025, but flash storage solutions provider Sandisk (NASDAQ: SNDK) has gained over 700% since the start of 2026 and is beginning to look like one of the most overbought stocks in the space.

Still, it would be imprudent for growth-focused investors to not have some degree of exposure to AI. But rather than trying to find a needle in a haystack, the following three thematic exchange-traded funds (ETFs) can offer a targeted, multi-faceted, pick-and-shovel approach to the ongoing megatrend.

1. The Memory Chip Shortage

AI has an insatiable appetite for memory, which has forced the market to shift its focus from raw processing power to the memory and storage systems that make AI workloads usable at scale. That paved the way for companies specializing in high-bandwidth memory (HBM), DRAM, and NAND flash memory chips to handily outperform hyperscalers and pure-play AI leaders over the past year.

Now, a severe global shortage has created supply constraints, leading to price surges as chipmakers can charge a premium for their products. That shortage is expected to last at least through 2028. For now, companies cannot keep up with demand.

The three major memory chip makers—SK Hynix, Micron Technology (NASDAQ: MU), and Samsung (OTCMKTS: SSNLF)—have already sold the entirety of their 2026 HBM production capacity with orders on the books reaching well into 2027. In turn, their respective stocks have ballooned, with each joining the trillion dollar market cap club in May.

According to industry consultancy firm Grand View Research, those conditions are unlikely to abate any time soon. The global semiconductor memory market, which had an estimated value of more than $111 billion in 2023, is forecast to grow to more than $240 billion by 2030—a compound annual growth rate (CAGR) of 11.6%—due to the increasing adoption of components across various industries, including automotive, consumer electronics, IT and telecom.

The memory chip market in the United States—which accounts for nearly 20% of the entire global industry—is expected to grow at an even faster pace, with a forecast CAGR of 12.2% through 2030.

The Roundhill Memory ETF (BATS: DRAM) is a thematic, sector-specific investment vehicle that offers investors a way to access concentrated exposure to memory chips, cyclicality, and innovation in a single ETF..

The fund focuses on firms operating across the memory chip supply chain, including companies involved in design and development of DRAM and NAND memory, wafer fabrication, packaging and testing, and the manufacture of semiconductor capital equipment and materials.

With a focus on the memory market segment rather than the broader semiconductor industry, DRAM provides targeted exposure to companies whose primary business activities are connected to memory chips and related technologies. Since its inception on April 2, the ETF is up more than 145%.

2. AI Power Infrastructure

While the real estate investment trusts specializing in AI data centers offer a pick-and-shovel AI option, investors can get more nuanced exposure within that segment by turning to the providers of power infrastructure services. Data centers require hardware, maintenance, and electricity—including transformers, turbines, power distribution lines, and HVAC systems—in order to achieve operational efficiency.

Individual hyperscale data centers are now requesting grid connections that range from 300 megawatts up to 1,000 gigawatts. To contextualize those power needs, a 1-gigawatt facility would require the same output as a large nuclear reactor, with the capacity to power a small city. That electric power consumption is forecast to skyrocket, too, growing from roughly 4% of all U.S. power usage in 2023 to as much as 12% by 2028–2030.

The Defiance AI & Power Infrastructure ETF (NASDAQ: AIPO) seeks to provide investors with exposure to companies operating at the intersection of AI and power infrastructure. The ETF is up more than 40% year-to-date (YTD).

The fund targets firms involved in the development, production, and deployment of AI technologies alongside businesses that design, build, or operate electric power systems and related infrastructure.

Holdings typically span the tech and industrials sectors, including companies that supply AI hardware and software, semiconductors and data center infrastructure, as well as utilities, grid modernization firms, renewable energy developers and companies focused on electric transmission, distribution, and energy storage.

3. High-Demand AI Commodities

Over the past year, broadly, prices for raw materials have outperformed the S&P 500.

Copper prices, for instance, are up more than 35% versus the benchmark index’s gain of nearly 25%. Similar to memory chips, copper is undergoing a global shortage caused by supply disruptions at major mines around the globe.

Despite the tightening market, demand for the metal isn’t going away. Copper’s properties make it a critical conductor and the most commonly used metal for electrical wiring and electronics. With the highest electrical conductivity of all industrial metals—and second only to silver, a precious metal—copper is essential to electrification, renewable energy, AI and data center expansion.

Beyond its high conductivity, copper is a cost-effective metal known for its superior pliability, durability, and corrosion resistance.

Together, those properties are driving a global market that was valued at nearly $242 billion in 2024 and is projected to undergo a compound annual growth rate of 6.5% through 2030, when it reaches nearly $340 billion, according to industry analysis firm Grand View Research.

The Global X Copper Miners ETF (NYSEARCA: COPX) tracks the Solactive Global Copper Miners Index, which is designed to reflect the performance of the copper mining industry.

It is composed of common stocks, American Depository Receipts (ADRs) and Global Depository Receipts (GDRs) of global companies engaged in some aspect of the copper mining industry (i.e., exploration, mining, or refining). The ETF has a YTD gain of about 20%, and over the past year it has gained more than 100%.

The article "The AI Trade Is Getting Harder to Pick, But These 3 ETFs Take a Different Route" first appeared on MarketBeat.