Federal Frontier AI Compute Grid

The Federal Frontier AI Compute Grid is the customer’s own grid: AI Workstations and local models on the customer’s own resources, in customer-controlled environments.

The Federal Frontier AI Compute Grid is the customer’s grid. It is made of the customer’s own resources: the customer’s hardware, or the customer’s own hyperscaler account. Federal Frontier does not host the grid or any part of it, and it is not a SaaS offering.

What joins the grid

AI Workstations and the local models they run join the grid. An AI Workstation joins by enrolling with the Frontier CLI:

frontier workstation enroll
frontier workstation run

Enrollment signs the device in through the customer’s Keycloak. Once running, the AI Workstation reports its hardware, its models, each model’s origin and a self-test result. See Enrolling an AI Workstation.

Who owns and controls it

The customer owns and controls the grid. The AI Workstations, the Frontier llm-router and the cloud tier all run in customer-controlled environments: customer hardware or the customer’s own hyperscaler account.

How chats route to it

The Frontier llm-router sends a user’s chat from the OutpostAI AI Assistant or CompassAI Ask the Ontology to that user’s own AI Workstation. If the device cannot serve the request, the router falls back to the cloud tier automatically and says so. See AI Workstation chat routing.

Model origin policy

Only models whose origin is on the allowed list can serve chats on the grid. The allowed model origins are the United States, France, Canada and the United Arab Emirates. If a device’s model is outside the allowed origin list, the request falls back to the cloud tier.

Data stays in the customer’s environment

The user’s prompts go from the router to the user’s own AI Workstation. The grid, the router and the cloud tier run in customer-controlled environments, so the data stays in the customer’s environment.