What Is the Palantir Ontology?

Not affiliated with Palantir Technologies Inc. This page is an independent explainer based on publicly available sources.

The Ontology is the structural core of Palantir's platforms — the layer that turns raw, disconnected data into a coherent, navigable map of an organization, sometimes described by Palantir as a "digital twin."

The basic idea

Rather than leaving data scattered across databases, spreadsheets, and software systems, the Ontology organizes it into object types representing real-world entities — a customer, a piece of equipment, a financial transaction. Each object type carries properties (a customer's name, a machine's location) and link types that define how objects relate to one another (a customer places an order; a machine belongs to a facility). Palantir's own documentation describes the Ontology as sitting on top of an organization's integrated datasets and connecting them to their real-world counterparts.

The three layers

Independent technical writeups commonly break the Ontology into three conceptual layers. The semantic layer defines what exists — the entities, their relationships, and their properties, answering "what are the things that matter in our world?" The kinetic layer captures actions: how objects can be modified and how changes ripple through connected systems. A third, dynamic layer handles real-time state and security as data changes. Together, these are what let the Ontology function as more than a static map — Palantir describes it as an operational layer that connects data to real-world action, not just a data catalog or schema.

What it's used for in practice

Once built, the Ontology becomes the foundation other Palantir tools are built on top of: users can create reusable object views, search for specific objects, run analyses, and build applications — all referencing the same underlying map rather than each tool maintaining its own disconnected version of the data. Palantir has stated the goal is to enable better decision-making at scale by giving technical and non-technical staff alike a shared, consistent picture of the organization.

Where it shows up

The Ontology isn't a separate product — it's built into both Foundry (Palantir's commercial platform) and Gotham (its government/defense platform). A concrete example: Rio Tinto's team built out its own Ontology in Foundry over several years, which the company says now lets it deploy AI use cases faster across areas like risk identification and supply-chain management — see our separate page on the Palantir–Rio Tinto partnership for more.