Observed arrival · 2026-11-04
Rackprint’s part-by-part AI server builder
A free 3D configurator estimates whether AI models will fit on a selected server and calculates its power use, generation speed, and indicative price.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- People comparing hardware for running AI models
- Worth noticing
- Its speed estimate uses memory bandwidth at 55% efficiency, and the site labels prices, power, and speeds as estimates rather than guarantees.
Field notes
The page exposes its calculation method rather than presenting the outputs without context: weight estimates include a 1.05 multiplier, while the KV cache estimate uses model architecture and FP16. For mixture-of-experts models, the speed estimate uses active parameters; the page also gives a separate allowance for GPU-usable unified memory on Mac. Rackprint advises checking compatibility with a reseller before buying.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
◎NicheUnusually specific use
One card from the complete issue