If the S&P can manage new records up near 7,825 with 10-year Treasury yields at 5.28% and 10-year reals near enough to 3%, one has to ask: Where would the index trade if yields were to fall, say, 100bps?

The answer depends in no small part on the reason for a hypothetical bond rally. If yields drop suddenly in response to another brush with the existential, stocks would very likely swoon. Note that both Ebola and the plague were in the news this week. And we live at a moment in history when you can threaten one day that “a whole civilization will die tonight” and be nominated, again, for the Nobel Peace Prize the next.

In the same vein, a left-tail macro catalyst that puts a US recession in play could pull the rug from beneath the loftiest earnings growth expectations since the dot-com bubble. That’d be bad for stocks but probably good for bonds.

If, on the other hand, yields fell due to some combination of cooler (but not negative) US growth, a long-end-friendly QRA next month and some manner of resolution to the Iran conflict that allows crude to mover durably lower, equities have meaningful room to re-rate despite having hit new highs during Tuesday’s US session.

There’s the chart the Trump critic in me loves to hate, but the apolitical, long-only investor in me just plain loves. Whenever the next record close comes, whether it’s today or tomorrow or next week, it’ll mark the 67th new high for the index since Trump’s second inaugural.

Suffice to say the AI trade’s air cover for Trump’s (melo)dramatics. I’ve said it again and again: It’s very difficult to fade a rally that’s built on some of the strongest earnings growth, both realized and expected, in modern history excluding recession recoveries.

To be sure, red flags abound. Concentration risk is the most acute on record, for example, with the largest 10 companies accounting for ~40% of S&P 500 market cap and nearly that much of overall index earnings.

As the figure above shows, those metrics are off the charts. Figuratively and literally.

But thanks to the profit boom, and to the restrained nature of the rally since June (rising yields effectively capped index upside, resulting in a three-month, sideways meander), the index multiple’s de-rated to “just” 19x.

Consider this: Top-down consensus for 2027 index earnings is $410. Bottom-up is $421. It’s not especially difficult (read: not difficult at all) to imagine a three-turn re-rating in the event of a meaningful rally at the long-end of the curve. Slap a 22 multiple on $410 and whaddya get? If you’re curious, the peak NTM P/E during the dot-com boom was 24x.

In a note published late last month, SocGen spelled it out. “AI sets EPS and bonds set the multiple,” the bank’s Manish Kabra wrote, adding that the tails are fat over the next 12 months. If oil goes to $150 and 10-year yields flirt with 6%, the S&P could fall to 6,000. If, on the other hand, oil moves lower to $80 and 10s fall near 4%, the S&P could hit 10,000, he said.