AI agents take centre stage at the third quarter earnings results
Investor focus in the third quarter earnings season will naturally be on different indicators, including margin resilience, cost pressures from higher oil and commodity prices, and funding due to the recent sharp rises in bond yields. However, a key focus will be how much companies are spending on AI infrastructure, whether those investments are starting to generate good returns, and what impact the increasing use of agentic AI will have.
Agentic AI is where the AI tool used can take actions rather than just giving answers to queries. The agent can plan steps, use appropriate tools and systems, conduct tasks and give results, depending on what it has been authorised to do. For example, agentic AI can actually book a holiday or a restaurant for the consumer instead of just providing helpful information. This is increasingly the future of AI, and how fast it takes off has implications for the earnings, and indeed business models, of many companies.
In the corporate world AI agents can run multiple tasks and checks across the various systems employed by that company, helping to generate efficiencies and increase productivity. In the consumer world personal AI agents can help with bookings, diary and email management and subscription renewals. Just a few weeks ago Meta launched its personal AI agent called Muse, which surpassed ChatGPT in downloads over the first 12 days after launch. Other tech giants such as Alphabet (Google) and Apple will likely launch their own personal AI agents in the near future.
As companies report third quarter earnings, investors will ask who the winners and losers from agentic AI are. Generative AI has already shown it can impact business models, and agentic AI could be even more transformative. For example, will there be the same need for online consumer marketplaces if an agent can do the searches and comparisons instead? Agents care about price and delivery times but not about advertising, so will business models that are reliant on advertising see an impact? Are companies getting a return yet by using more agents, or is it just costing them more as agents tend to use AI tokens more rapidly than consumers do? Are consumers the real winners here, as a proliferation of agents will likely drive down prices of different goods and services? These questions won’t be answered overnight but will remain a key focus for the quarters and years to come as companies adopt and adapt to this new technology.
Analysts currently expect third quarter earnings growth from US companies of about 25%, with the full year now likely to be in excess of 30%. Expectations are high, but deservedly so given the steep rises in earnings revisions so far in 2026. Technology companies have led the charge, but other sectors such as healthcare, industrials and energy have also seen positive revisions in recent months. With the US market trading at a 20% premium to its 20-year median valuation, based on price-earnings ratios, and with bond yields having risen sharply, continued strength in earnings is important if equity markets are to make further progress.
Recent market strength has been driven largely by the Magnificent Seven, despite fears about potential over-spending on AI infrastructure such as data centres by those same companies. Under the surface of the market, the picture is less rosy, as breadth has deteriorated quite sharply. In fact, almost 60% of the S&P 500 constituents are more than 20% below their highs, showing how concentrated the market has become once again, with large technology stocks doing the heavy lifting. Investors typically view such a narrow market as being unhealthy, with broader participation from the majority of companies usually a better sign for a continuation of the bull market. This makes the forthcoming earnings reports of the largest technology companies in the US even more important.