This post is a preview of my CNBC appearance tomorrow morning at 6:50 AM ET — tune in to catch me live.

I run a long volatility, long stock book. That combination shapes how I read weeks like this one. I don’t need every stock to go up in a straight line — I need to own the right exposure, size it properly, and be positioned to benefit when things move, in either direction. With that lens, here’s what I’m watching across Meta, semis, Apple, and Nvidia.

Meta: Betting on More Than One Frontier:

The Meta news caught my attention, and not because I think Muse Glimmer is going to be crowned the smartest model on earth. That’s not really the game Meta is playing. What I see is a company spreading its chips across smaller models, open-weight releases, agentic tools, and AI built to run on your phone rather than in someone else’s data center.

I like that. I don’t want to live in a world where one company owns the only AI “brain” and everyone else just rents access to it. Competition — across models, platforms, approaches — is how you get real innovation and real choice. Meta pushing into on-device AI and broadening the menu is good for the whole ecosystem, whether or not it ever wins the frontier race outright.

Semiconductors:

This Is What Volatility Trading Looks Like Semis are a wild ride right now, and if July rattled you, I’d gently suggest this sector isn’t for you. I say that as someone who trades volatility for a living. I am long volatility, I am long equities, I love capitalism, I love America — and I still believe the long-term AI and semiconductor story plays out. But months like July are part of the deal. The size and speed of that move shocked a lot of people who hadn’t lived through a real drawdown in this sector.

That doesn’t invalidate the thesis. It’s a reminder that expectations, positioning, and valuation can get way out ahead of fundamentals, and they need to come back to earth sometimes. That’s not a bug in this trade — it’s the trade.

One more thing I’d hammer home: semiconductors is not one stock. It’s not just Nvidia. You’ve got GPUs, memory, networking, foundries, advanced packaging, power management, cooling, analog chips, and equipment makers — different customers, different cycles, different risk profiles. Know what you actually own before you decide how scared to be.

Apple:

Still Waiting for AI to Show Up in the Box The Jefferies downgrade gets at something real. AI hasn’t shown up yet in a form that gives Apple a clean, obvious upgrade-cycle catalyst. Jefferies’ worry is that Apple may have let the window close on a truly differentiated, premium-priced iPhone — right as AI-driven memory and component costs are climbing on them.

For Apple, AI can’t just be a nicer Siri. It has to be the thing that gets someone to upgrade sooner, pay more, and spend more across the ecosystem. I still think that cycle comes. But when it does, Apple won’t be running unopposed — other device makers and model providers aren’t waiting around. Apple has the installed base and the ecosystem moat. It still has to prove it can turn that into incremental dollars and margin, not just incremental headlines.

Nvidia:

August 26 Is the Number Everyone’s Circling Nvidia reports on August 26, and that print is going to move a lot of books, mine included. I don’t have much doubt about underlying data center demand — that part isn’t the question. The questions I actually care about:

  • Is custom silicon starting to chip away at real share, not just headline share?

  • Can Nvidia and its partners finally clear the supply, packaging, power, and networking bottlenecks that have been throttling shipments?

  • How much of the next two years of growth is already sitting in the stock price?

Expectations are stretched, and at this valuation, a “good quarter” isn’t automatically a good stock reaction. Nvidia has to beat what’s already priced in, not just beat the consensus number.

Bottom Line The fundamentals of this AI trade are still strong.

What’s changing is how forgiving the market is willing to be about it. Meta is diversifying its bets instead of chasing one frontier crown. Apple has to prove AI can actually drive an upgrade cycle without wrecking margins. Nvidia has to prove infrastructure demand holds up, competition stays contained, and the bottlenecks actually clear.

From here, I think this trade gets won on earnings and return on invested capital, not on the next big capex headline. I remain constructive on the long-term story. But constructive isn’t the same as smooth. I’m long volatility for a reason — this trade comes with turbulence built in, and you should size your exposure like you know that going in.