Compute

Northbridge runs its own GPU machines: local AI models for development, an isolated coding sandbox, and compute workloads when the cards would otherwise sit idle. It is experimental infrastructure that we run for ourselves. We don’t sell compute time today, and nothing on this page is an offer of any kind.

What we work on: running models locally (llama.cpp on AMD and NVIDIA cards), power-aware scheduling, heat and safety guards, and measuring real costs honestly.

Early-access ideas that grow out of this are listed under products in development.