Back to Playbook
Chapter 2

Product & Platform Architecture

The four integrated layers that power Palantir's enterprise data operating system

Platform Components

Foundry

Enterprise Data OS

Gotham

Defense & Intelligence

Apollo

Continuous Deployment

AIP

AI Decision Platform

Reference Architecture Overview

Palantir's platform architecture consists of four integrated layers that work together to transform raw data into operational decisions:

1

Data Integration Layer

Connects to any data source (databases, APIs, files, streams). Handles data ingestion, cleansing, and transformation with complete lineage tracking.

2

Ontology Layer

Maps real-world concepts to data (customers, products, transactions). Defines relationships and provides semantic understanding for non-technical users.

3

Application Layer

Low-code/no-code application builders, pre-built applications for common use cases, and custom development framework.

4

Deployment & Control Layer (Apollo)

Continuous deployment across any environment. Security, access control, monitoring, and governance.

Foundry: The Enterprise Data Operating System

Data Integration

  • ✓ 400+ pre-built connectors
  • ✓ Real-time and batch processing
  • ✓ Schema inference & profiling

Pipeline Builder

  • ✓ Visual data transformation
  • ✓ SQL and Python support
  • ✓ Incremental processing

Ontology Management

  • ✓ Object type definitions
  • ✓ Property mappings
  • ✓ Relationship constraints

Workshop Builder

  • ✓ Drag-and-drop interface
  • ✓ Pre-built widgets
  • ✓ Custom React components

Example Use Cases

  • • Supply chain optimization
  • • Customer 360 views
  • • Risk and compliance monitoring
  • • Operational dashboards

Gotham: Intelligence and Defense Platform

Graph Analysis

  • ✓ Entity resolution
  • ✓ Link analysis & pattern detection
  • ✓ Temporal & geospatial analysis

Investigation Tools

  • ✓ Case management workflows
  • ✓ Evidence collection & tagging
  • ✓ Collaboration and handoffs

Operational Planning

  • ✓ Scenario modeling
  • ✓ Resource allocation optimization
  • ✓ Mission planning & execution

Security Features

  • ✓ Classification handling (TS/SCI)
  • ✓ Attribute-based access control
  • ✓ Audit logging & compliance

Example Use Cases

  • • Counter-terrorism investigations
  • • Criminal network analysis
  • • Battlefield intelligence fusion
  • • Border security operations

Apollo: Continuous Deployment Infrastructure

Core Capabilities

Environment Management
  • • Multi-cloud deployment (AWS, Azure, GCP)
  • • On-premise installation
  • • Air-gapped and classified networks
Deployment Orchestration
  • • Blue-green deployments
  • • Canary releases
  • • Automatic rollbacks
Security & Compliance
  • • End-to-end encryption
  • • Certificate management
  • • Compliance scanning
Monitoring & Operations
  • • Health checks and alerts
  • • Performance metrics
  • • Log aggregation

🎯 Key Differentiator

Apollo allows Palantir to deploy and update software in environments where traditional SaaS cannot operate—submarines, forward operating bases, secure facilities—while maintaining central control and governance.

AIP: Artificial Intelligence Platform

AI Agent Framework

  • ✓ LLM integration (GPT, Claude)
  • ✓ Agent orchestration & chaining
  • ✓ Tool use and function calling

Model Management

  • ✓ Model versioning & lineage
  • ✓ A/B testing framework
  • ✓ Performance monitoring

Governance & Safety

  • ✓ Output validation & constraints
  • ✓ Bias detection & mitigation
  • ✓ Explainability features

Platform Integration

  • ✓ Connects to Foundry Ontology
  • ✓ Leverages existing pipelines
  • ✓ Surfaces insights in applications

Example Use Cases

  • • Automated report generation
  • • Predictive maintenance recommendations
  • • Fraud pattern detection
  • • Decision support systems

Ontology System: Mapping Reality to Data

The Ontology is Palantir's secret sauce—it maps real-world entities and relationships to underlying data sources, allowing non-technical users to work with familiar business concepts.

Object Types

Define real-world entities

  • • Specify properties and data types
  • • Set validation rules
  • • Define constraints

Relationships

Connect objects together

  • • Specify cardinality
  • • Enable traversal and querying
  • • Maintain referential integrity

Actions

Define operations on objects

  • • Implement business logic
  • • Maintain data consistency
  • • Trigger workflows

Synchronization

Keep ontology in sync

  • • Handle schema evolution
  • • Manage data quality
  • • Track lineage

Key Benefits

  • ✓ Business users work with familiar concepts instead of tables and schemas
  • ✓ Consistent data model across all applications
  • ✓ Simplified application development with semantic understanding
  • ✓ Enhanced data governance and compliance

Deployment Options

SaaS (Foundry Cloud)

  • • Fully managed by Palantir
  • • Multi-tenant architecture
  • • Regular updates and patches

Private Cloud

  • • Dedicated infrastructure
  • • Customer or co-managed
  • • Isolated tenant

On-Premise

  • • Behind customer firewall
  • • Full customer control
  • • Apollo-managed updates

Air-Gapped

  • • No internet connectivity
  • • Manual update process
  • • Maximum security