Notes from the build.
Coverage is the constraint
A middle market private fund receives roughly 1,000 deals a year and reviews about 400. The other 600 were not rejected. They were never opened.
What we built in Palantir Startup Fellowship - Cohort 002
Antonine got into Palantir Startup Fellowship - Cohort 002. What a solo founder shipped on Foundry in a few months, and why the ontology is the product.
Bain just published Antonine's thesis
The 2026 Global Private Equity Report says 12 is the new 5. The operational bar for every deal just doubled, full potential diligence is the new standard, and most firms can't afford the old way of meeting it.
From plugin to engine: the three levels of AI inside an investment firm
Most firms are using AI. Almost none are compounding from it. Here's why, and what changes when AI becomes infrastructure.
The Harvey thesis for investment firms
Harvey closed the gap between awareness and infrastructure for legal. Now the same playbook is open for investment management. Here is what it looks like and why the firms that ship first compound on the ones that do not.
Origin: why Antonine
The first version of this company tried to sell AI to real estate agents. It did not work. This is the story of the pivot, what I got wrong, and why investment management is the right tribe to build for.
Done reading? Let's get started.
Run a current deal through it yourself, end-to-end, on an instance configured to your mandate — and judge the output for yourself.