Introducing Starfield AI infrastructure

AI inference is local. The cloud isn’t. Starfield deploys standardized compute close to power, connectivity and demand, then connects it through one programmable network for AI inference.

What this video covers

This 30-second film is the introduction to Starfield AI infrastructure. It opens with the line that gives it its title: AI inference is local, the cloud isn't. It then says Starfield turns many small sites into one AI network.

It ends on four ideas: distributed capacity, inference that takes latency into account, placement that takes data residency into account, and delivery that can be audited. It closes on the tagline: distributed AI infrastructure at network scale.

Inference means running a model that already exists. For how the network is described in more detail, see the Fabric page.