I realize we’re all desensitized by now to visuals illustrating extreme market concentration and its corollary, poor breadth.
Just as we no longer register surprise at gratuitous violence — having been exposed on a daily basis to the realities of war in Gaza and Ukraine — we barely bat an eye when someone shows us a chart that would’ve been deemed shockingly perilous in another era.
But on the off chance the tech rally, which is by now synonymous with the AI trade, ever tips over to the detriment of the “broader” market, we should keep one apathetic eye on visuals like those shown below.
The charts are from Goldman, and they give you some context for US equity market narrowness headed into the new year.
Long story short, there’s very little in the way of recent precedent for the breadth metric shown on the left — the spread between the cap-weighted S&P’s distance from its local high and that distance for the median stock — and no precedent at all for the dominance of the 10 largest names in the big-cap index.
An optimistic take says this just means there’s ample scope for a so-called “catch-up” trade as the rally broadens. A pessimistic (or skeptical) assessment wonders if we’re approaching a kind of singularity (no tech joke intended): When 10 names comprise almost half the market cap of an index with 500 stocks, it raises a lot of uncomfortable questions.
Note also from the figure on the right above that the trajectory of the light blue line — the share of S&P 500 earnings attributable to the top 10 names — is running inexorably higher with no sign of letting up. What happens if that converges on the market cap line? What if 10 companies comprise 40% or even 50% of overall index earnings?
In any event, it’s worth noting that fund managers polled for the final 2025 installment of BofA’s monthly survey still viewed the Mag7 as the most crowded trade and said an AI bubble was the biggest tail risk.
More notably, panelists said the second-most likely source of a credit event in 2026 is hyper-scaler capex.
That, at least, suggests the pros aren’t completely oblivious.




Reminds me of the 2000 bubble when Nortel was over 30% of the TSE300 and then mostly gone by 2003 and bankrupt in 2009.
MH – do you remember those great Nortel TV ads back then? “We have one product. One. Quota people, Quota.” There were two I can remember but I never have found them online. (Unlike the equally fun ads where Stewart encourages his boss to “click it in there.” Click it in there has been a staple in our office to this day.)