Sovereign AI for the Supply Chain, Built on Infrastructure You Control

-

Dell AI Factory with NVIDIA provides an on-premises technology stack of secure compute, storage and networking for Palantir’s operational AI, helping ensure customers can build sovereign AI systems on their terms.

• Organizations can scale AI while keeping sensitive data, models and operations under their control, including for critical supply chain workflows.

Every company across every industry has a unique supply chain that can be affected by complex factors spanning availability, transportation, weather and more. Supply chains that leverage AI must securely connect data, decisions and actions across the operation while remaining sovereign, preserving control over proprietary data, models and operational knowledge.

The Palantir Sovereign AI OS Reference Architecture with NVIDIA and Dell brings those requirements together to enable on-premises AI deployments, including workflows like the new Palantir AI stack for supply chain announced today with NVIDIA. First built for NVIDIA’s own supply chain, which requires managing over 1.3 million components for each NVIDIA Vera Rubin rack, this collaboration can optimize supply chain operations for industries such as manufacturing, healthcare, energy, aerospace and technology. The sovereign-by-design foundation protects proprietary data, knowledge, intellectual property and models on-premises within an organization’s own security domain.

The bottleneck is not the model. It is the operating environment.

Enterprise data is distributed across ERP systems, manufacturing and logistics platforms, documents, sensors and data lakes. The challenge is turning it into a reliable representation of the business while keeping an organization’s most precious proprietary data secure.

Palantir Foundry and Ontology provide the semantic backbone, deployed on the Dell AI Factory with NVIDIA. Ontology models the assets, processes, people and relationships that shape an operation, giving applications and AI agents a governed business interface instead of dozens of direct connections to back-end systems.

Inside the Dell AI Factory with NVIDIA, Dell’s storage engine, including Dell ObjectScale and Dell PowerFlex, provides scalable options for enterprise data. ObjectScale supplies scalable S3 object storage, while PowerFlex provides distributed NVMe block storage. A rail-optimized leaf-spine network design separates GPU, storage, services and management traffic.

GPU-optimized Dell PowerEdge platforms, including the PowerEdge XE9780 with NVIDIA Blackwell Ultra GPUs, provide the on-premises accelerated compute foundation for model customization, retrieval-augmented generation, inference and agentic workloads. CPU-based platforms, including the PowerEdge R670, support control-plane services, Ontology workloads and data processing.

The NVIDIA Spectrum-X Ethernet networking platform and NVIDIA ConnectX SuperNICs support the GPU fabric, while Dell PowerSwitch networking isolates services and management traffic.

This creates a governed path from enterprise sources to Palantir Ontology. Data can be discovered, enriched, transformed, queried and retrieved where it already lives, while Ontology objects preserve lineage and business meaning for developers and agents.

Security and control from silicon to software

Supply chain data can reveal supplier relationships, production constraints, product roadmaps and competitive advantages. For many enterprises and sovereign organizations, it cannot move freely into a public cloud environment.

Dell AI Factory with NVIDIA infrastructure incorporates hardware security capabilities, secure supply chain practices and platform-level controls and adds governance across data preparation, search and activation. Palantir Rubix provides a zero-trust Kubernetes runtime with multi-tenant isolation and policy enforcement. Apollo supports centralized deployment, monitoring and lifecycle management across enterprise data centers, sovereign facilities and air-gapped environments.

From supply chain data to supply chain decisions

Before an AI agent can make a recommendation, the underlying data has to be discoverable, current and trustworthy. Running Palantir on the Dell AI Factory with NVIDIA enables data orchestration, ingestion and data preparation, federated analytics and powers semantic search. Palantir Ontology then connects those data products to the real-world entities and relationships that govern supply chain decisions.

Palantir AI Platform (AIP) connects large language models, optimization engines and business rules to Ontology. Right-sized, NVIDIA Nemotron open models, like Nemotron Lightning, can be post-trained with organization-specific data using Palantir Foundry and Palantir AIP, providing the accuracy of larger frontier models but at a fraction of the cost. NVIDIA cuOpt enables teams to model constraints, evaluate scenarios and understand the operational impact of allocation decisions.

The result is a system that identifies risks, recommends actions and keeps experts in control.

For example, in a material allocation workflow, the platform connects supply, inventory, production and demand signals, represents their relationships in Palantir Ontology, evaluates alternatives and surfaces the choices most likely to protect production. With NVIDIA NeMo libraries and Palantir Autopilot, operational feedback can improve specialized models over time while preserving governance and traceability.

That is operational, sovereign AI for critical supply chain operations: connected to the business, deployed where it is needed and designed to improve through use.

The Palantir AI Stack for supply chain deployed on the Dell AI Factory with NVIDIA combines Dell PowerEdge compute, Dell storage, NVIDIA accelerated computing and networking, NVIDIA AI software and open models and Palantir Foundry, Ontology, AIP, Rubix and Apollo.

The goal is a sovereign environment that turns intelligence into action securely, built on infrastructure that enables scale.

Contact Dell Technologies to discuss a sovereign AI supply chain solution powered by the Dell AI Factory with NVIDIA.