The trillion-dollar investment cycle into generative artificial intelligence is only just beginning, and is already starting to pay off, according to John Belton, portfolio manager at Gabelli Funds. Tech companies are set to invest roughly $1 trillion in capital expenditures into AI infrastructure over the coming years, causing some investors to worry the businesses may be spending too much for not enough payoff. In 2025 alone, Google parent Alphabet, Amazon, Meta Platforms and Microsoft are planning to spend as much as $320 billion combined on AI technologies and data center buildouts. However, early signs suggest investors have yet to grasp the full potential of generative AI, given that some initial investments are already starting to bear fruit, according to Belton. “This is a trillion-dollar investment cycle,” Belton told attendees during a panel on AI at the Morningstar Investment Conference in Chicago. “Adoption and usage is really only now starting to hit this very steep part of the so-called S-curve, another two and a half years after the launch of ChatGPT.” “The largest investors in AI infrastructure are generating the most attractive returns today,” he said. The portfolio manager of the Gabelli Growth Fund has significant exposure to the “Magnificent Seven” companies, with Microsoft, Nvidia and Amazon being its top three holdings. This year, the fund has outperformed, sitting in the top 12% of funds in its category, according to Morningstar data . Yet, there are two reasons to remain confident in the potential for AI, Belton said. Specifically, the cost of the technology is coming down, while its capabilities are improving, meaning the number of use cases for companies is growing. Belton said AI capabilities are already superior to human capabilities across a wide range of disciplines, including reading comprehension, science and math. “As AI is becoming cheaper and more capable, the vision is that it’s just going to be used in more and more corporate productivity initiatives,” Belton said. “And it’s going to be used as in many cases, a labor replacement across enterprises, in all different sectors, in all different parts of the enterprise.” Some examples include using AI to cut supply chain costs, or to generate revenue across marketing and operations departments. Meta Platforms , for example, is already using AI for targeted advertisements, a change that has helped the tech company boost sales. Meanwhile, tech companies such as Alphabet and Microsoft have recently divulged that AI is already generating roughly 30% of their internal database, suggesting the role of a software engineer and developer will change. Other use cases are still emerging, including AI used in autonomous vehicles such as at Tesla , as well as drug discovery in health care. Given this, here are some of the companies that can tap into AI potential, Belton said. NOW YTD mountain ServiceNow shares year to date ServiceNow is one holding with a roughly 2% weighting in the Gabelli Growth Fund. The enterprise software firm sells agentic technology to help companies to automate their businesses. Agentic revenue is already tracking close to 10% next year, Belton said, adding it’s a small but promising sign of a “big, exciting” new area for growth. Broadcom shares are another holding that has exposure to the AI theme. The stock is up more than 16% this year, and is a consensus buy on the Street, according to LSEG. GE Vernova and Applied Materials are two other holdings that can benefit from the AI theme, according to the investor. AVGO YTD mountain Broadcom shares year to date.