The piece was inspired by confusion around good macro but struggling market. Problem is a chunk of market depended on two arguments for higher valuation: 1) Mag7 as giant free cash flow generators (worth more per dollar earnings), but they're now spending 80-90% of cash flow on AI, losing that free cash flow premium. 2) Asset-light software businesses (like Block) face margin pressure as AI becomes competitor instead of software engineers.

There will still be a premium. Cut long-term multiple estimate from 22 to 20 vs. rest of world ~18. Exaggerated to make a point; sooner or later Big Mag7 will generate cash flow—either AI works and they have more profits, or it doesn't and they pull back. Their base businesses are still very profitable; they can return to that. Premium probably lower than before, but still a reason to have one.

Nvidia isn't a completely asset-light business; it's a high free cash flow generator. It's the primary beneficiary of AI buildout. Debate around AI centers on Nvidia: if Mag7 companies are in a one-off spending cycle, where does Nvidia sell to after replacing last year's sales? If AI goes on forever (10 years), it's a screaming buy. Real debate around Nvidia—we don't hate it, still involved. It won't get cannibalized.

We've been signaling a violent rotation, don't want investors to panic. Not in panic; mostly tilted where we want to be. From history of economic development, every major invention disrupts status quo, economy grows from efficiencies/productivity. Somehow excess personnel get redeployed to higher value-add activities. I believe entrepreneurs will figure out how to use high-talent white-collar workers to do more. Within our business (Federated) and others, we're seeing that.