Neoclouds After the Jensen Huang Exit and Why I Still Believe in the AI Build-Out
Jensen Huang is out. I still can't quite type that without checking the news again. Thirty-three years running Nvidia, and it ends over a personal scandal and a national security probe, with Colette Kress as interim CEO and a board that says (as of Nov 16) it isn't thinking about bringing him back. Trump says he wouldn't mind. Musk too. Doesn't matter. The board decides.
My first reaction last week was the obvious one. Sell everything with "AI" in the name. My book is semis and memory, and the neoclouds were already down roughly 50% from early October before any of this landed. CoreWeave, Nebius, IREN, all of them. Huang was their biggest fan. Who talks them up now?

Then I actually looked at the numbers, and wrote them up properly on marketanalysis.com: Neocloud Stocks After the Jensen Huang Exit Price In a Bust the Numbers Do Not Show.
Short version. CoreWeave grew revenue 168% last quarter and is sitting on a $132.5 billion backlog. Nebius grew 571%. These are not companies whose customers just vanished. OpenAI, Microsoft and Meta don't stop renting GPUs because Nvidia's CEO had an affair.
Huang was the cheerleader. He wasn't the customer.
What actually scares me is older than the scandal. Money (these guys spend years of revenue up front). Power, permits, delayed sites. A handful of model labs filling most of the backlog. And yes, a new Nvidia boss maybe giving the hyperscalers first dibs on chips. That last one is the only risk Huang's exit really adds, and I don't think a sane CEO hands the pricing power back to Microsoft and Amazon on purpose.
The bigger thing for me: we're still early. Like, really early. AI software revenue is around $200 billion this year. Nvidia talks about $3 to $4 trillion of infrastructure by 2030. Most companies are still running pilots. Agents barely work in production yet. That's not the end of a cycle. That's year two or three of something long.
So I'm not panic-selling. Not adding with both hands either, leverage has taught me humility the hard way. Some of these neoclouds will fail, Musk is probably right about that much. The ones with power contracts, funded pipelines and real backlogs, I want to own through this, not sell into it.
Nvidia reports tonight. I'll be listening for one thing: does the post-Huang company still talk about customers outside the big four the same way. If yes, last week was noise.
My opinion on my own money, not advice for yours.