Local AI hardware, priced daily.
Specs that actually decide what you can run at home: memory capacity, memory bandwidth, and what a gigabyte costs you. Plus guides on getting a local AI server working, with every number linked to its source.
Tracking 15 products. Prices last checked August 2026.
Best memory value
Lowest dollars per gigabyte of usable model memory. Capacity is what decides whether a model loads at all.
Mini PC
GMKtec EVO-X2 (Ryzen AI Max+ 395, 128GB)
128 GB
$3,500 $27/GB
Consumer GPU
AMD Radeon RX 7900 XTX
24 GB · 960 GB/s
$930 $39/GB
Workstation GPU
Tenstorrent Blackhole p150a
32 GB · 512 GB/s
$1,399 $44/GB
Edge kit
NVIDIA Jetson AGX Orin 64GB Developer Kit
64 GB · 204.8 GB/s
$3,299 $52/GB
Consumer GPU
NVIDIA GeForce RTX 4090
24 GB · 1008 GB/s
$3,380 $141/GB
Consumer GPU
NVIDIA GeForce RTX 5090
32 GB · 1792 GB/s
$4,800 $150/GB
Today's deals
- AMD Radeon RX 7900 XTX
7% below MSRP
$930
Setup guides
- Q4, Q5, Q8, FP16: what quantization actually costs you
The numbers in every model filename. What bits-per-weight means, what you give up going lower, and which one to pick for the box you own.
- Buying used AI hardware safely
The used market is where the value is, and where the traps are. What to check before you pay, what protection you actually have, and what to test the day it arrives.
- Setting up llama.cpp, Ollama, and vLLM
Three runtimes, three different jobs. Which one to install, the commands that actually work, and the memory knobs that keep them from falling over.
- Multi-GPU: tensor parallel, layer split, and when NVLink matters
Two GPUs do not automatically mean twice the speed. What each split mode actually does, which one to pick, and when your interconnect becomes the bottleneck.