Thus far in 2025, the stock market performance went from being very narrow, led by a handful of AI-related infrastructure stocks, too much broader participation throughout the quarter. Ironically, our relative performance during the quarter was initially hampered by investor concerns centered on those industries likely to be disrupted by the implementation of AI, notably enterprise software companies. The thought being that AI would eliminate the need for entry level and mid-level employees across the work force limiting the sales opportunities for these companies but gave no credence to the notion that these same software companies might enhance their own offerings using AI thereby making their products more valuable in turn creating more opportunity. These product enhancements center on embedding intelligent agent technology aimed at making the application much more productive and useful.
At the same time, market participants fretted about the yield curve, particularly about the potential rise in long rates and the possible negative impact on the economy. Recession or no recession, risk on, risk off. The Fed’s actions (the timing and pace of rate cuts) and the impact of tariffs (who pays, one time increases or not) were also hot topics. All the while the greatest transformation of the world in which we live continued apace- the AI revolution. Still very much in formative stages, it is helpful to assemble a roadmap and explore the timing of prospective developments.
2025 is the year of chatbots and natural language models. These products can take unstructured data and use natural language to manipulate and generate content while responding to various queries. These models are largely powered by “Hopper” and “Blackwell” semiconductors provided by Nvidia today.
Later this year and into next, a new phase of AI capability will be available called inference. A dramatic step up in intelligence, here the software can make predictions on new, unseen data, not just relying on pre-written algorithms. These models will also require “Blackwell” chips from Nvidia.
The holy grail of AI, reasoning, essentially simulates human-like logical processes to solve problems. Extremely powerful hardware exemplified by Nvidia’s upcoming “Vera Rubin” family of chips available in 2026 will power this technology.
These software models require enormous infrastructure build outs today in the form of datacenters and power requirements, not to mention enhancements in communication technologies and many others. Other technologies are not standing still. A new form of computing, Quantum Computing, is rapidly developing, promising to tackle these needs as well in a more effective manner.
You can imagine the seismic changes that may be coming in the next 24 months when applying these technologies to every aspect of our lives. This is why there is so much enthusiasm around AI, probably the most disruptive technology in our lifetimes. And as you might expect dramatic investment winners and losers will be created. We invite you to participate along with us on what promises to be a very exciting journey