Schedule a talk with one of our advisors to learn more about Summitry and how we can help you get a foothold on your financial life. For career opportunities please visit careers at Summitry.
“Early innings” has become the go-to phrase for describing where we are with artificial intelligence. It’s a comforting idea for investors: if the game just started, there’s plenty of upside left to capture. But at a recent Summitry client webinar, our investment team pushed back on the idea that there’s one single answer.
Chief Strategy Officer, Kurt Hoefer, put it simply: he sees AI the way he remembers the early internet, and if he had to pick an inning, he’d call it the second. Director of Research, Michael Kon, offered a more layered view, and it’s worth breaking down because it changes how we think about positioning.
Three Different Games, Three Different Innings
Michael separates AI into three distinct layers:
Infrastructure, the data centers and compute that make AI possible, is not early stage. The industry will spend roughly $750 billion this year alone building it out, with several more years of heavy investment likely ahead. We’re somewhere in the middle innings here, not the first.
The model layer, the large language models themselves, has already come a long way. Comparing today’s leading models to what existed in 2022 is, in his words, night and day. There’s still room to improve, but this layer has moved well past its opening stages.
Diffusion across the economy is where things genuinely are early. Programmers use AI daily, and there’s a lot of experimentation happening across corporate America, but broad adoption and mature use cases are still developing. This is the layer where “early innings” most accurately applies.
What This Means for Valuations
This nuance matters because of how concentrated the market has become around AI. Using the S&P 500 as a proxy, the Shiller P/E ratio, a cyclically adjusted measure of valuation, is now near 41x. The only time it’s been higher was the dot-com bubble, when it peaked at 44x.
That doesn’t mean a crash is coming. It means that if you’re betting on the index at this level, you need real conviction that the largest companies driving it can keep growing well beyond the next year or two. Year-to-date, nearly all of the S&P 500’s gains have come from a small group of AI-related and energy stocks. Most of the other companies in the index have posted negative returns in aggregate.
We don’t invest in the index itself. We invest in individual businesses that meet strict criteria around competitive advantage, durable free cash flow, and reasonable valuation, regardless of how a broader index is priced.
The Bottom Line
AI is not one uniform story. Infrastructure is maturing, the models are advancing quickly, and real-world adoption is still catching up. Understanding which inning applies to which layer is what allows us to separate genuine opportunity from hype, and to stay disciplined about where and how we invest in this theme.
This post reflects perspectives shared at Summitry’s Q3 2026 Investor Update. The views expressed are those of the Summitry investment team and are intended for informational purposes only.
GET THE NEXT SUMMITRY POST IN YOUR INBOX:
