HPC & STORAGE SERVERS
LLM-3U-AM5
Your models. Your rack. Your rules.
Purpose-built rackmount inference for private LLMs — dual 10GbE, PCIe Gen 5 x16 pairs, zero cloud egress.
2xPCIe x16 Gen 5
192 GBDDR5
10 GbESFP+ x2
HPC & STORAGE SERVERS
Your models. Your rack. Your rules.
Purpose-built rackmount inference for private LLMs — dual 10GbE, PCIe Gen 5 x16 pairs, zero cloud egress.
EVERY ANGLE
CONFIGURE
Choose a configuration and add it to your quote bag. We quote your exact build — no list pricing, no obligation.
UNDER THE FINS
ON-DEVICE AI
On-device inference with AMD Ryzen + dual GPU. Memory is the ceiling — here's where LLM-3U-AM5 lands.
Runnable at 4-bit quantization within the platform's memory budget. Larger models fit on higher-memory Premio AI computers below.
THE LINEUP
NVIDIA Jetson AGX Thor
NVIDIA Jetson AGX Orin
Intel Core Ultra + RTX GPU
NVIDIA Jetson Orin NX
Intel Core (machine vision)
Tell us the workload — we’ll configure the LLM-3U-AM5 around it.