Rates Are Spiking. The S&P Is One Big AI Bet. What If Your Biggest Risk Is What You Don't Own?

The wrap centered on how investors can misunderstand risk by focusing on labels instead of exposures. Andy Constan argued that Treasury buybacks may matter at the margin, but long-term rates are still driven mainly by growth and inflation expectations, so the effect is likely only a few basis points rather than a decisive turn in yields. He also made the case that bonds are no longer the terrible proposition they were in 2020, but they still need to be judged against cash and volatility, not treated as a simple yield call.

Jason Hsu’s point was that index investors may be far less diversified than they think. He argued that the S&P 500 today is highly exposed to a single AI theme, so passive ownership can hide concentration risk even when the number of holdings looks large. That led to a broader discussion of factor diversification and why investors should think in terms of what actually drives returns inside the index.

Jack Raines emphasized that investors need real, small losses early enough to learn what risk feels like before the stakes matter. Gary Mishuris added that great investors should be studied, not cloned: the real task is understanding their ideas, applying them, and adapting them to your own temperament and circumstances rather than cargo-culting Buffett-style habits.

The episode also touched on China’s energy buildout and how AI is turning power infrastructure into a strategic advantage, with the hosts arguing that the U.S. may be forced to rethink energy investment and grid capacity.