Occam’s Razor & Michael J. Burry MD
The Simplest Short in the Market Might Be Hidden in the Workforce
Over the last 12 months:
Palantir’s headcount grew about 12–14%.
Nvidia’s headcount grew 10–12%.
Now that we have that baseline… know this:
Engineering stayed the dominant slice through the summer, and then both firms hit the brakes into fall. *Peak-to-late drawdowns in Engineering: Palantir -70% (June - October), Nvidia -79% (January - October).
That’s what cooling looks like once you strip the hype: headcount up, net-new external requisitions down hard, engineering-heavy mix even as the whole shebang tightens. If you’re looking for where a depreciation story leaks into labor reality, this just might be it.
We will dive into PALANTIR and NVIDIA workforce data in a moment, because I’ve got some interesting takes for you.
“Prices can catch up whenever they want… labor already has.”
Because the market still sees AI as capital-driven servers, chips, data centers… but the real stuff is happening underneath that flashy story, in the people. The slowdown I’m seeing isn’t layoffs, it’s sequence. Inside the files, the footprints are obvious: engineering still dominant, operations never materializing.
Paid readers: I’ll show you exactly where that break is showing up first. I want to what those cuts imply for capex cadence, asset useful lives, the likely pivot from “build” to “maintain,” how this maps to Burry’s recent X posts and where the first workforce breakpoints show up.


