1 No-Brainer Artificial Intelligence (AI) ETF to Buy With $40 During the S&P 500 Sell-Off

The S&P 500 (^GSPC -5.97%) is still in a bull market, which stretches back to October 2022. In fact, it's coming off back-to-back annual gains of more than 25% in 2023 and 2024, a feat it has only achieved one other time since 1957.

However, the S&P is currently in sell-off mode, declining more than 9% from its recent all-time high so far. Many of the stocks that drove the index's incredible gains over the last two years are currently down even more than that, namely those leading the artificial intelligence (AI) race. Since the U.S. stock market always trends higher over the long term, this might be a great time for investors to scoop up some bargains.

Rather than picking winners and losers, it might be a good idea to buy an exchange-traded fund (ETF) that focuses on high-growth AI stocks instead. The Roundhill Generative AI and Technology ETF (CHAT -4.84%) offers diversified exposure to the industry through its portfolio of 37 top AI names, and investors can pick up a single share for under $40.

Image source: Getty Images.

Investing across the AI spectrum

The Roundhill Generative AI and Technology ETF invests in companies developing the platforms, infrastructure, and software that are driving the AI industry forward. It's an actively managed fund, so a team of experts at Roundhill selects which stocks to buy and sell based on their presence in AI and their potential to generate returns for investors.

The ETF holds many of the leaders in the AI race, including Nvidia, which supplies the world's best data center chips, and Microsoft, Alphabet, and Amazon, which are the biggest providers of AI services in the cloud. Those four stocks alone account for 23.1% of the total value of the fund's portfolio, giving investors a very healthy exposure to some of the highest-quality names in the industry.

But the Roundhill ETF also holds a number of other AI stocks that sometimes fly under the radar:

Data source: Roundhill Investments. Portfolio weightings are accurate as of March 30, 2025, and are subject to change.

Nvidia is the most influential chip company in the AI industry, but it owes a lot of its success to its key manufacturing partner, Taiwan Semi. The Taiwan-based fabricator produces Nvidia's industry-leading Blackwell graphics processors (GPUs) for the data center, which are the preferred choice among AI developers. But according to Statista, Taiwan Semi has a whopping 64.9% market share in the foundry business overall, so it's the most important chipmaker in the world.

Oracle, on the other hand, is a major buyer of chips from the likes of Nvidia. It operates some of the best data centers in the world for AI training and AI inference, thanks to its enormous GPU clusters and random direct memory access (RDMA) networking technology, which is faster than traditional Ethernet networks. Top AI developers like OpenAI, Meta Platforms, and Elon Musk's xAI are using Oracle's Gen2 data centers, and the company is struggling to keep up with demand.

Then there is Salesforce, which operates an industry-leading customer relationship management (CRM) platform. Businesses use it to store their clients' details and capture sales opportunities, and the new Agentforce layer of the platform is helping them accelerate those workflows with AI. It allows businesses to create virtual agents that can autonomously reach out to sales prospects, resolve customer issues, and even set appointments.

Like Microsoft, Alphabet, and Amazon, Snowflake offers a suite of cloud services, but it's quickly pivoting toward AI. Developers can access ready-made large language models (LLMs) through its new Cortex AI platform, which they can combine with their internal data to create custom AI software. Cortex also features powerful AI tools like Document AI, which can extract valuable data from unstructured sources like contracts and invoices.

Finally, Micron Technology is a leading supplier of memory and storage chips, which are becoming increasingly important in AI workloads, not just in the data center but also in computers and smartphones. Nvidia is using Micron's HBM3E (high-bandwidth memory) in its Blackwell GPUs, and it's now completely sold out until 2026. Micron stock trades at an attractive valuation relative to its peers, which could pave the way for strong upside in the long run.

The Roundhill ETF is sinking right now, but it has major long-term potential

The Roundhill ETF was established in 2023, so it doesn't have a very long track record for investors to analyze. Nevertheless, it did soar by almost 31% last year, comfortably beating the S&P 500, which rose by 25% (including dividends). That's because of incredible gains in some of its top holdings like Nvidia, which soared by 171% during the year, and Meta, which was up by 65%.

However, those stocks have started 2025 on the back foot, which has triggered a 16% year-to-date decline in the ETF. While a correction of that magnitude can be unsettling, it does create a buying opportunity for investors seeking exposure to the AI industry. But it also serves as a warning that investors should only buy this ETF as part of a diversified portfolio because its high concentration of AI stocks can result in significant volatility.

Citing a study by Goldman Sachs, Roundhill thinks AI can create $7 trillion in global economic activity by 2032, and a lot of that value will come from the companies in this ETF. Plus, Goldman isn't the only bullish firm -- PwC estimates AI will add $15.7 trillion to the global economy by 2030, and Cathie Wood's Ark Investment Management places that number at a staggering $200 trillion.

As a result, the sell-off in the broader market might prove to have been the ideal time to buy the Roundhill ETF when investors look back on this moment in a few years.