It's easy to rely on something that’s been there, even if just in the mundane moments. Like the same coffee place every morning with the same people who know what you like, or the same friendly lady outside of school drop off that knows your kid’s name. 

Inevitably though, a shift happens. The coffee place has turnover, or goes out of business. The lady at school is replaced by someone less outgoing. These are little shifts that impact your daily experience but, consciously or unconsciously, you note them. 

AI is shifting. While the ping pong shifts in the news are not small—deal announcements or un-announcements like Blue Owl, earnings disappointing or not, and partnerships on chips and content seem to land regularly, circular and straight. But smaller shifts that can amass to bigger trends have also started to establish roots. 

Back in April, we shared that we see the impact of AI in three distinct phases: infrastructure, software, and productivity. We thought then we were moving out of the infrastructure phase and more fully into the software phase. We were clearly a bit early on that call. 

• The infrastructure phase is now well established. This includes semiconductor companies, the massive data centers needed to store and process all that information and the companies to power it all (Broadcom, Nvidia, Constellation Energy, etc.). 

This phase got hot. It will still all be needed but my view now is that the chips and LLMs will eventually be a commodity—like a lot of tech, eventually things get cheaper and more efficient. And mainstream power providers earnings could be hindered in the near term by their inability to quickly “make” more power. We are starting to do work on fusion as a result which is cheaper and less onerous. 

• The 2nd phase is focused on software, where AI becomes usable. At this point software companies and IT providers begin integrating AI into tools that people and businesses can actually use.

This has started happening, though it’s been slower. Watching companies like Salesforce, for example, you know the benefits are there but the value hasn’t started coming to fruition. But I think we are more readily here now than back in the Spring. 

• Finally, there’s the productivity phase—where AI starts to truly reshape industries, replacing tasks, and in some cases, even jobs. 

We aren’t here yet in a material way. There have been some layoffs as a result of AI, according to Challenger, Gray & Christmas. These ranked 6th in number of layoffs or 4.7% of the total for the year, well behind the number from DOGE actions. But the move to robotics is on its way.

We see the shift even more likely now because AI has been around longer. The companies that haven’t been at the center of the boom have had time to learn and figure out how it can best use it to be more efficient.

This phase is defined by chatbots, copilots, and autonomous AI agents for businesses. Tech giants like Amazon and Microsoft are competing, but a growing number of other Tech names in the software space look interesting. 

Those that have a utilitarian purpose—such as sales management and customer service like Salesforce or tax preparation like Intuit—are starting to look closer to benefiting from AI. For investors, this means broadening your attention beyond the suppliers of chips, servers and energy to include the builders of intelligent, usable software. 

PS: We’ll be taking next week off for the holidays and then be back with our full 2026 outlook. We wish you a wonderful season, and a very happy new year. We truly appreciate you.