Nasdaq up 3.3% in a single session and I still haven't seen the Strait of Hormuz reopen

Nasdaq up 3.3% in a single session and I still haven't seen the Strait of Hormuz reopen. I've seen a Treasury Secretary say "there's a chance" on cable news. That's the entire catalyst. Scott Bessent telling Andrew Ross Sorkin there might be a deal tomorrow, or the day after, and $600 billion of market cap gets bolted onto US equities before lunch. Somewhere a quant desk is running a regression between Bessent's Q-scores on CNBC and intraday beta. I'd short that trade too.

Let's be honest about what actually happened. The S&P closed at 7,736, a record. The Dow tacked on 907 points to 54,085, also a record. Palantir ripped 29% because Karp's revenue print beat by a margin that would embarrass a SPAC promoter, and everyone decided that was proof the AI trade still has legs. Fine. Palantir prints what Palantir prints. But look one desk over — AMD reported after the close and the stock didn't move the way a "state of the AI boom" data point should move when the state of the boom is allegedly unimpeachable. Guidance came in and nobody was excited. That's the tell nobody wants to write about, because it doesn't fit the "everything's ripping" copy that's already been filed.

Then there's SpaceX. First earnings print as a public company, and the headline isn't the backlog — $47.5 billion, genuinely enormous — it's that AI segment spend came in hotter than modeled and the stock, already down more than 17% from its June IPO price, didn't get a bid on the print. Retail's been buying the dip every single session since the float, per Vanda. Every session. That's not conviction, that's a religion with a Robinhood login. Meanwhile the actual capital markets verdict on SpaceX since IPO is a $500 billion-plus destruction in market value, and I have people in this building telling me the AI capex story is "just getting started." Just getting started at what, exactly — burning cash faster than backlog can absorb it?

And can we talk about Bezos. Amazon crosses $3 trillion market cap for the first time on Monday, a genuine watershed, the kind of number that gets a chyron. Tuesday, Bezos files to sell $4 billion of stock and the thing gives back 2%. I'm not saying insiders always nail the top. I'm saying insiders selling into a fresh all-time high the day after it's set is not information I discard, and neither should you, and neither will the guy who runs my compliance desk when he asks why I trimmed the position an hour after the filing hit.

Here's my actual problem with today, and it isn't any single data point — it's the correlation structure. Oil down 1.2% to $78.43 on Hormuz hope. Equities up on the same hope, plus earnings, plus whatever's left of momentum chasing itself into August. Treasuries — barely moved, ten-year sitting right around where it's been sitting, which tells me the rates market isn't buying the "risk-off resolved" story with anywhere near the conviction equities are pricing. If the Iran deal is real, oil should be down more than a point and yields should be doing something interesting on lower geopolitical risk premium and improved inflation optics. They're not doing much of anything. Bond guys have seen this movie — a Bessent soundbite, a Qatari draft proposal, a "maybe Tuesday or Wednesday" — enough times this cycle to not chase it. Equity guys apparently haven't.

I keep coming back to the FOMC minutes from three weeks ago. Nine-to-three, three dissents in favor of a hike, in favor of tightening, while energy and food prices are still doing exactly the thing central bankers are trained to look past and increasingly can't. Warsh's own dot plot has the terminal rate drifting toward 4.25% by 2027. September hike odds were sitting near 60% right after that meeting, down from near 80% pre-meeting but still very much alive. None of that has gone away just because Palantir beat and a cabinet official floated an interim Iran framework on television. The macro backdrop hasn't shifted. What's shifted is the market's willingness to look at it.

So what do I actually do with this. I trim SpaceX exposure into retail's buying, because a stock down 17% from IPO with worsening burn and no earnings-day bid is not a stock I want to own on hope. I hold the semis I like — TSMC, SK Hynix, the actual picks-and-shovels — because Asian trade tonight, Nikkei up 3.3%, Kospi up 4.4%, Taiex up 3.1%, tells me the supply-chain story is intact even if the US large-cap AI trade is getting frothy at the edges. I don't chase the Iran-deal beta because a Treasury Secretary saying "there's a chance" is not a signed agreement, and I've been burned by "there's a chance" before, in this exact administration, on this exact conflict.

And I write down, for the file, that August 4th was the day the market decided a rumor and a beat from Palantir were enough to erase three weeks of hawkish Fed signaling. History will tell us in about six weeks whether that was conviction or amnesia. My money's on amnesia. It usually is.