Reference builds
Six configs, one per profile - from the cheapest new-parts build to an 8-GPU cluster node. Every component and total is computed from the same dataset the builder uses.
updated 16 Sept 2026Budget AI Lab
The cheapest all-new-parts build that still runs a useful local model. For testing the waters before committing to a bigger rig.
Open the build ->
Home AI Lab
One fast GPU in a quiet tower. For an individual running models locally - no rack, no server-room noise.
Open the build ->
Prosumer Mini Rack
Two GPUs, ECC memory and real storage in a small 4U rack. For the advanced individual or indie hacker.
Open the build ->
Small Business Private AI
A local server that a team can rely on: redundancy, backup, UPS and 96 GB of ECC VRAM.
Open the build ->
Company AI Server
A proper GPU server: 4-8 cards, redundant power, BMC management and fast networking, for a company that has outgrown a workstation.
Open the build ->
Private AI Cluster
Multiple GPU nodes, shared storage and a high-speed fabric. Priced per node - plan for three or more.
Open the build ->