Autonomous Intellects (AI)
AI operates a portfolio of companies — building at the frontier of capital markets, compute, and physical autonomy.
Six fields where autonomy is no longer optional.
We invest where general intelligence meets a real-world physical or financial substrate — places where the cost of a wrong decision is high and the value of a right one is compounding.
Finance
Autonomous capital allocation, market microstructure, and risk systems that operate continuously and self-improve.
Cloud
Inference fabric, model orchestration, and developer surfaces purpose-built for agentic and continuous workloads.
Robotics
General-purpose physical agents — perception, manipulation, and locomotion stacks that ship.
Drones
Aerial autonomy at scale — long-duration flight, swarm coordination, and contested-environment navigation.
Silicon
Next-generation compute — domain-specific accelerators, packaging, and the manufacturing flow behind them.
BioTech
Computational biology and lab automation — closing the loop between in-silico design and physical experiment.
Companies we operate.
A small number of focused bets, each with its own product, team, and market. Built to stand alone — designed to compound together.
CapitalsAI
Finance · AIAutonomous research and execution platform for capital markets. Continuous signal extraction across structured and unstructured data, with risk-aware deployment.
Atomius
Compute · InfrastructureAtomic-scale compute and orchestration. Building the substrate that lets autonomous systems think faster than the world they operate in.
Operating, not just allocating.
We are operators first. We embed alongside founding teams, share an engineering bench, and treat each portfolio company as a focused expression of a shared autonomy thesis.
Identify the substrate
We start with industries where autonomy unlocks order-of-magnitude improvements — and where the underlying physics or economics already favors machines.
Compose the company
Founding team, product wedge, infrastructure stack. We bring a shared platform — models, data, tooling — so each company starts a year ahead.
Run the loop
Ship, measure, retrain. Every portfolio company runs its own autonomy flywheel; the holding company connects the lessons across them.