I just finished looking at the charts, and my take is this: the Magnificent 7’s pullback isn’t ordinary valuation digestion—it’s the market repricing the “AI premium.” Five stocks have shed a combined $3.54 trillion from their highs, and Google alone is down 30.8%. This isn’t a normal correction; capital is rotating into a new narrative. You can feel it in the market action: what’s falling isn’t “poor earnings,” but “a story that isn’t new enough.” Advertising and cloud revenue remain resilient, and capex is still elevated, but the market now cares only about “where the next growth driver will come from,” not “how much money you’ve made in the past.”

I’ve been keeping a close eye on changes at the margin in the giants’ capex for the past two years, because that’s the hardest anchor for the AI narrative. The case is simple: when a giant’s capex shifts from “revised higher beyond expectations” to “in line with expectations or even slightly below,” the stock’s reaction to the AI premium can be more dramatic than its reaction to the earnings numbers. Conversely, as long as advertising and cloud continue to provide resilient cash flow, a pullback doesn’t necessarily mean the fundamentals have collapsed; it may mean pricing power is shifting from “old growth” to “new growth.” The problem now is that the options chains on the Seven Sisters are packed to the brim. Record open interest often signals that a reversal is imminent—and once the market picks a direction, the move will be more forceful than most people expect. That’s especially true when near-term contracts, weekly contracts, and the put-call ratio are all crowded at once. The market isn’t slowly digesting things; it’s waiting for a trigger.

The same logic behind auto-compounding in yield aggregators applies here: money finds the most efficient place to go on its own; it won’t stick with an old story. Over the past two years, the AI premium led investors to pay upfront for “future cash flows.” Now the market is asking where the next leg of growth will come from: models, compute, applications, or greater efficiency in cloud and advertising. Whoever can offer a clearer path to returns as capex changes at the margin will be more likely to earn a new valuation. If you think I’m talking nonsense, take a screenshot and save it.