• AI investment potential remains high, akin to 1996 in the internet's rise, says Bank of America.
• Adoption is expected to accelerate, with early-stage use from 44% of S&P 500 firms that touted AI.
• Investors can explore AI through semiconductors, software, ETFs, or top-rated stocks.

If you think it's too late to invest in AI because the industry is too rich, think again.

The technology is expected to change the global economy. One way to think about it is that we are in about 1996 relative to the internet, according to Bank of America. Considering how much the web has grown since then, this analogy means the potential gains from AI stocks aren't yet fully priced in.

The trajectory of gains is steep: in the past, disruptive technologies took 15 to 30 years to reach widespread adoption. For AI, this process is expected to be accelerated over the next 3-5 years, read an August 5 note. As for the widespread use of AI, the bank forecasts early-stage adoption will begin in 2025 for 44% of S&P 500 companies that discussed AI during their second-quarter earnings.

There's a wide range of ways investors can gain exposure to artificial intelligence, including by buying up semiconductor or software stocks or utility companies that supply data centers, among other things.

Investors can also choose to take the stock-picking route or buy ETFs packaged with AI stocks.

The bank favors three funds in the latter option. They include the Global X Robotics and Artificial Intelligence ETF (BOTZ) for pure AI stocks, the KraneShares Artificial Intelligence and Technology ETF (AIQ) for diversified exposure, and the iShares US Tech Breakthrough Multi-Sector ETF (TECB) for broader exposure.

As for single stocks, Bank of America compiled feedback from 130 fundamental equity analysts to get a list of their top "buy-rated" American companies with AI exposure. These names are expected to see market-value expansion as they integrate the technology into their businesses. They are listed below in alphabetical order.