The S&P 500 is within 1% of its record high, but the typical member of the index has fallen 14% from its peak. That gap, according to Schwab's head of macro research and strategy, is the clearest sign yet that a handful of AI-linked names are holding up the entire market.
Kevin Gordon said on Bloomberg This Weekend that the 14% figure is the average peak-to-trough drawdown among S&P 500 members since early August — not their loss on any given day. The index itself looks calm because it's weighted by market value, so the largest companies dominate its moves.
The index and the average stock have parted ways
Gordon's point is structural. A market-cap-weighted index can keep climbing even when most of its components are falling, as long as the biggest names keep rising. That's what happened this summer: tech sat out much of the rally, but AI-related stocks picked up the slack.
The result is a benchmark that looks healthy on a chart and a median stock that doesn't. Gordon said the divergence is wide enough that investors watching only the headline index are getting a distorted picture of what's happening underneath.
Capex season, not earnings season
Gordon relayed a line that captures how the market's attention has shifted: "it's no longer earning season, it's CapEx season." The comment reflects the degree to which AI capital spending budgets — not traditional quarterly profit reports — now drive sentiment in the largest stocks.
That focus creates a specific vulnerability. Gordon warned that one capex miss, paired with pared-back budgets, is where earnings disruption could begin. If a major AI spender cuts or delays its plans, the effect wouldn't stay contained to that company; it would ripple through the suppliers and peers that the index's largest weights depend on.
Big Short investor Steve Eisman made a related point in July, saying the whole market has become one AI bet. Gordon's version is narrower but similar in spirit: the index's health is now tied to a spending cycle that a small number of firms control.
Valuation worries are fading — for now
Gordon added that two earlier drags on AI names — valuation concerns and public pushback against the technology — are beginning to fade. That shift helps explain why the largest stocks have been able to keep climbing despite stretched multiples.
The fading pushback doesn't erase the underlying risk. Apollo Global Management chief economist Torsten Slok issued an AI debt warning last month, saying the cost of insuring cloud giants' debt against default signals risk in debt-funded AI spending. That's a different channel of exposure than an equity capex miss, but it points at the same dependency.
FactSet projects 32.4% S&P 500 earnings growth for 2026, up from about 15% at the start of the year. Those estimates assume the spending holds.
What Gordon wants from the Fed
On monetary policy, Gordon said Federal Reserve hikes at roughly every other meeting would be the best case for stocks. That pace, in his view, wouldn't aim to slow growth sharply or hit the labor market — the outcome equity investors would prefer if the central bank has to tighten again.
The scenario matters because AI capital spending is sensitive to financing conditions. A Fed that moves gradually leaves room for budgets to stay intact; a Fed that moves faster could force the kind of pullback Gordon is watching for.
The test starts this month
Third-quarter earnings season opens this month, and it will show whether AI capital spending budgets hold. That's the near-term checkpoint. If the budgets stay firm, the index's narrow leadership can persist a while longer. If they don't, the 14% average drawdown already visible beneath the surface stops being a curiosity and becomes the main story.
