There's an old adage we kept coming back to on our latest podcast: “you never stand in the same river twice”. In plain terms, the river looks the same every time you step in but the water is always new. Markets have spent the past three months proving the point, swinging hard in both directions only to end up close to where they started. Underneath those round trips, the moves in individual names were enormous. Two of the largest companies in the world added roughly $1.2 trillion of combined market value in just five trading days.
In our view, a lot of this comes down to market structure. The ten largest names now make up nearly 40% of the S&P 500, and four of them are semiconductor companies. Roughly a fifth of "the market" now trades like a semiconductor index, about 4x its share in 2020. Add record leverage on top and momentum begets momentum: by some measures, the first half of the year hit extremes not seen in more than two decades (cough, 2000). Then in July, a lot of that unwound because leverage cuts both ways. The same borrowed conviction that amplified the ride up accelerated the way back down. Moves this large and this mechanical can't be timed, not by us, not by anyone. Which is why we believe diversification, across factors and themes - not just sectors - matters more right now than at any point we can remember. That’s why we keep talking about cake and veggies.
The cake is the AI build-out, and it has five layers:
No single layer matters more than the rest; each depends on all the others. Within the cake, we gravitate toward the toll roads: businesses that collect from everyone building, whichever builder ultimately wins. The veggies are everything else worth owning: world-class franchises in themes AI doesn't dominate, quality that compounds on a different clock.
In the second quarter the cake sprinted while the veggies sat still. In July it flipped, and the veggies provided exactly the ballast they're there for. That balance is the point. When one side gets marked down, it's what lets you be a buyer instead of a forced seller.
From here, we're watching interest rates, inflation, credit spreads, and crowding out. With the 10-year near 4.7%, we've likely crossed the point where rising rates become a real headwind for stocks. $80+ oil is complicating the inflation picture. And next year's data-center spending is estimated at $1.4 trillion, a build-out surpassed only by the interstate highway system and the railroads, one big enough that it's already crowding out labor, materials, and components across the rest of the economy. The river has moved; we don't expect a rerun of Q2. Different variables equal different outcomes - and there is a lot of sorting out to do.
Our approach remains the same: deep research, structural understanding, and long-term thinking in an increasingly short-term world. As always, reach out to anyone at Doxa to discuss markets or plan out your financial goals for 2027 and beyond. Enjoy the last stretch of summer!
Click the link below or search "Doxa Takes" on Apple Podcasts, Spotify, or Overcast. Enjoy!
