As Agentic AI Use Grows, These ETFs Stand to Profit [Weekend Money]

"Demand for CPUs and cybersecurity set to grow
The more tasks processed, the more important CPUs become
Cybersecurity gains prominence as access permissions expand"

Which companies stand to make more money as agentic AI is used more widely? Among the many related industries, including semiconductors, data centers, and components, industry analysts say demand is likely to grow for CPUs (central processing units) and cybersecurity.

Ko Gyeong-beom, a researcher at Yuanta Securities, said, "As the scope of development and analysis work delegated to AI expands, demand for AI is expected to depend more on the types of tasks and the time spent executing them than on the number of users." He added, "Beyond GPU (graphics processing unit) inference, the importance of CPUs, which read files and run programs, and of input/output systems is also likely to increase." While GPUs handle large-scale, repetitive calculations simultaneously, CPUs rapidly process complex tasks in sequence.

The structure of AI services also supports demand for computing power. AI services can generate more revenue as charges accrue for subscriptions, tokens, API calls, and agent runs. The more tasks each customer runs simultaneously and repeatedly validates, the more revenue and demand for computing power can grow, even if the number of subscribers remains unchanged.

Ko said, "Assigning tasks to agents, searching files, and running programs require CPU and input/output performance in addition to GPU inference." He added, "As more tasks are processed simultaneously, CPU parallel processing and power efficiency become more important, broadening the competitive field for server CPUs, including those based on AMD and ARM designs."

For overseas-listed ETFs related to the theme, he cited the AOT Growth and Innovation ETF (AOTG). The actively managed fund selects stocks and determines their weightings based on assessments of business models with low marginal costs and growth potential. As of October 29, it held AMD at 15.28%, Micron at 11.30%, and NVIDIA at 10.94%, in that order. Ko described it as "a fund that offers exposure to both growing CPU demand and the expansion of AI semiconductors and services."

For a Korea-listed ETF, he named the KODEX U.S. CPU Semiconductor TOP 10. Ko said, "It invests in related supply chains, focusing on companies such as AMD, ARM, and Intel." He added, "Unlike the AOTG ETF, which selects growth stocks across the board, this fund is suited to investors seeking concentrated exposure to key CPU companies."

Alongside CPUs, the spread of agentic AI is also driving greater demand for technology protection and cybersecurity. This is because expanding the scope of agents' work also increases the access permissions granted to files, accounts, and external systems. With U.S.-China technology rivalry and demand for protecting critical infrastructure adding to this trend, cybersecurity is expected to gain prominence as AI services become more widespread.

Ko cited the First Trust Nasdaq Cybersecurity ETF (CIBR) and the TIGER Global AI Cybersecurity ETF as investment options. He said, "These offer ways to invest in corporate access controls and system protection." He added, "Investors can compare the companies and weightings held by each fund to invest in the growth of security software and related infrastructure." CrowdStrike is the largest holding in the CIBR ETF, while Okta is the largest holding in the TIGER Global AI Cybersecurity ETF.

This content was produced with the assistance of AI translation services.