SPARSE

COMMUNITY DATA CENTRES

AI

Local AI where locality matters

CEAC is designed primarily for inference: running useful AI workloads closer to the people, organisations, devices and data that need them.

Local language AI

Chat, retrieval-augmented generation, document processing, embeddings and knowledge search.

Speech and vision

Speech recognition, text-to-speech, image analysis, CCTV and sensor inference.

Community services

Education, local organisations, community services and privacy-sensitive workloads.

Escalation

Workloads that exceed the local node can move across the CEAC network or to remote cloud and specialist compute.

The Generation-1 design uses four NVIDIA A10 accelerators, while the wider CEAC architecture is intended to support replaceable compute as accelerator technology changes.

Explore the Generation-1 computeHow distributed execution works