The market's AI winners are now so dominant that they're literally holding up an index while the median stock bleeds out beneath the surface.
The Summary
- AI-focused tech stocks are propping up major indices even as most individual companies face steep declines, creating what Schwab's Kevin Gordon calls "increasingly narrow market leadership"
- The real watch: whether companies can sustain their massive AI capital spending without clear ROI timelines
- Fed tightening plus a stable labor market could keep the AI trade alive, but the rest of the market is already pricing in a different reality
The Signal
We're watching a market bifurcate in real time. A handful of AI infrastructure companies are carrying the weight of entire indices while thousands of other stocks trade like we're already in a recession. This isn't new, but the gap is widening at an accelerating pace.
Schwab's macro team is flagging what every portfolio manager already feels: if you're not in the AI capex story, you're fighting a different market. The S&P might be flat or up, but the median stock is down double digits. Market breadth is collapsing. Only the compute providers, the infrastructure builders, and the companies promising to monetize AI at scale are getting funded.
"Market leadership is narrowing to companies that can credibly claim they're building or profiting from AI infrastructure."
The sustainability question is the right one. Companies have spent hundreds of billions on AI compute, data centers, and talent. Revenue growth tied directly to AI products is still mostly speculative for anyone not selling chips or cloud services. At some point, boards will want to see margin expansion, not just bigger training runs.
Here's what makes this moment different from past tech bubbles:
- The capital spending is real, not vaporware. Data centers exist. Chips are shipping.
- The labor market hasn't cracked. No mass layoffs means consumer spending holds, which keeps the rest of the economy from tipping.
- The Fed is tightening gradually, not slamming the brakes. That gives AI companies time to prove out revenue models before funding costs spike.
But the narrowness is a warning sign. When only a handful of stocks drive index performance, you get fragility. One earnings miss from a mega-cap AI name and the whole market reassesses. One breakthrough in open-source models that commodifies what was supposed to be proprietary, and valuations reset hard.
The companies that survive the next phase won't be the ones with the best demos. They'll be the ones that figured out how to turn AI compute into margin. That means verticalized models, proprietary data moats, or platforms that other businesses have to plug into. Generic AI infrastructure plays are already overvalued. The winners will be the companies that make AI boring and profitable.
The Implication
If you're building in AI, the message is clear: ship revenue, not roadmaps. Investors are still funding infrastructure, but the window for "trust us, monetization is coming" is closing fast. The market is pricing in massive future cash flows. Deliver them or get repriced.
If you're watching this from outside the AI trade, the play is patience. A market this narrow eventually broadens or breaks. When it does, the stocks that got left behind become interesting again. Until then, track AI capex commitments in every earnings call. The moment those numbers plateau, the narrow leadership trade is over.