There are current market concerns about a slowdown in AI innovation, but Bank of America believes that the technological transformation of AI is actually accelerating, and the market focus has entered the second phase of transactions—software stocks and AI agents, Agentic AI.

In a report on November 21, Bank of America pointed out:

AI technology is developing rapidly, with breakthroughs occurring almost daily. The capabilities of AI are extensive, including completing software engineering tasks, providing customer service, booking travel, and even being used in precision surgeries or vehicle assembly. The Agentic AI wave will catalyze the accelerated development and deployment of AI-driven applications, as well as industrial and commercial robots.

Bank of America further points out that as AI applications become more widespread, the market focus also shifts to the second phase of AI beneficiaries:

Do not underestimate the short-term disruptive potential of AI at this stage. The entire AI investment field, especially software stocks, is unlikely to be fully priced by the market. As the software "AI moment" arrives, AI monetization is expected to begin in 2025 and become meaningful in 2026 with the acceleration of enterprise adoption.

Advertisement

AI applications continue to expand from knowledge workers to manual laborers

Specifically, Bank of America points out that AI applications continue to flourish:

New, disruptive AI capabilities seem to be emerging at a rate of weekly, or even daily, and the development of AI capabilities is so fast that AI agents may be able to complete end-to-end software engineering tasks. We expect the latest wave of Agentic AI to catalyze the accelerated development and deployment of AI-driven applications, as well as industrial and commercial robots, which could change the global economy sooner than investors expect.

Bank of America states that the progress of AI agent technology not only improves existing tools but also gives birth to new applications.

The current AI software only shows the tip of the iceberg of future possibilities. Models from OpenAI and Anthropic have recently demonstrated capabilities at the doctoral level, and Mistral's new model has performed well in visual tasks. Microsoft and Google have also launched new products that enhance AI agent capabilities, with Google Gemini AI introducing memory features for personalization, Perplexity partnering with Stripe to launch the Buy With Pro shopping platform, and even benchmarks for evaluating model performance are improving.

Bank of America emphasizes that robots have expanded the scope of AI applications:

Emerging robotic capabilities may have a disruptive impact in various fields such as warehousing, catering, and construction. Although the growth in global industrial robot installations has not yet accelerated, it is expected that general-purpose robots will enter the market at a lower cost by 2026. Among them, China accounted for 51% of new installations in 2023 and 41% of operational robot shares.

Regarding the impact on the labor market, Bank of America states:

The advancement of AI technology may affect a wide range of professions, from knowledge workers to manual laborers. Although large-scale AI-driven labor replacement is unlikely in the short term, competition will become increasingly fierce as AI agents and robots become more widespread. By 2030, agent AI may fill the shortage of 10 million global health workers.

Investor focus shifts to the second phase of AI

Bank of America further points out that as AI applications become more widespread, the market focus also shifts to the second phase of AI beneficiaries:

It's not just semiconductor companies and large cloud service providers, but S&P-indexed companies are increasingly mentioning AI in their earnings reports and conferences. As enterprise AI adoption increases in 2025 and accelerates in 2026, the focus for the coming year will shift to the second phase of AI beneficiaries, especially software companies. Bank of America favors large companies such as Microsoft, Salesforce, and Adobe for their布局 in Agentic AI, while smaller and mid-cap companies like NICE, Informatica, and Zeta are also seen as potential targets.

Bank of America lists the three major development stages of AI beneficiaries:

Phase one: It is expected that AI beneficiaries will expand from Nvidia to a broader semiconductor field and cloud service providers (CSPs), including large cloud service providers, as well as data center real estate investment trusts (REITs) within the communications field.

Phase two: Beneficiaries may expand to software companies that produce AI-driven applications, as well as technology hardware and capital goods companies that produce data center infrastructure, such as servers, networking equipment, and electrical and air conditioning equipment.

Phase three: Beneficiaries may expand to companies that integrate AI products and services, such as consumer service companies like catering and retailers, banks and financial service companies, consumer non-essential companies, media and entertainment companies, as well as biopharmaceutical and medical device companies.