I don’t love Mag7 reporting days. Or maybe I should say Mag6 reporting days.

Tesla’s its own thing by now. I don’t think it makes much sense to subsume it under a banner. It’s “liquid Musk,” as I’m fond of putting it. Yes, Tesla’s a car company, and sure, there’s obviously an AI connection. But the stock’s not a “pure play” on anything other than Elon himself. Tesla prioritizes whatever he’s obsessed with currently, and as an eccentric obsessive myself, I’d caution against trying to predict what his next obsession will be.

As for the other six names in the septet which together comprise more than a third of S&P 500 market cap, you pretty much know what you’re betting on: AI, Cloud and, in Apple’s case, new colorways in a 15-year-old phone. (I’m just joking, Tim. Everyone loves you, but you do need to start innovating again.)

The problem for someone such as myself — i.e., someone committed to penning all-day market color — is that although you know what to look for in Mag6 results, and while you know those results will be impressive on an objective definition of the term, what you don’t know is how high the subjective bar is. Nor is it possible to know the location of the “too much of a good thing” threshold for AI capex and related spending.

Microsoft and Meta both beat on the top-line this week, and their current quarter sales guides were generally fine, but the market frowned at both companies’ spending outlooks. Traders were particularly suspicious of Meta’s remarks, which I think it’s fair to say counted as “TMI.” And so, both stocks sold off, weighing on the broader market Thursday.

It’d be easy enough to roll that color up with some perfunctory remarks about Jerome Powell’s “pushback” on December rate-cut pricing and add a reference to the “sell the news” trade around Donald Trump’s trade “deal” with Xi Jinping, on the way to crafting a plausible pullback narrative. But I was loath to go that route Thursday afternoon with Apple and Amazon results still looming as potential game-changers.

That’s why I don’t love Mag6 reporting days: Because anything you say can and will be used against you in the court of reader opinion in the event you brave anything that even loosely approximates a prediction and one or more of the mega-caps delivers results which move the market against your narrative during the ensuing cash session.

With that in mind, I’ll let someone else be brave. Below, find a good summary from Nomura’s Charlie McElligott, excerpted from a (much) longer Thursday note. To wit:

With this i) sneaky bit of Fed “not so fast, my friend” reintroduction of rate-path optionality plus ii) the pre-traded Trump / Xi China tariff “kick the can” going out with a whimper and now iii) the clearance of the meat of the AI EPS last night (really good prints, but mixed-bag market reaction which may be showing us that the blind “throw my money at AI capex spend”–theme [will] require greater nuance moving forward), the recent “cutting-into-loose FCI and positive wealth effect impulse” trade is probably going to have to take a step back, especially [if] USD goes bid any further.

Nevertheless, just like the conditioning and muscle memory says, traders keep telling me they want / need a pullback to more comfortably re-engage on better entry points on the long side into year-end. Many [are] still highlighting the AI capex story [with] memory chips ‘sold out’ on too much pipeline.

To me, the 2026 structural downside risk for AI stock leadership and accordingly, all of Equities, is about demand for AI capex potentially overwhelming and burning through currently manageable / acceptable organic cashflow generation.