Skip to main content
WerkHub AI Business
Most popularMost Popular

Business

Mittelstand, professional services, healthcare. Form factor: 4U rack-mount server.

With Digitalbonus subsidy (50%)
from 18.500 €
37.000 € without subsidy
Digitalbonus Bayern: 50% grant for Bavarian SMEs (<50 employees). Other funding programs may apply.
GPU
2× NVIDIA RTX PRO 6000 Blackwell Max-Q
VRAM
192 GB GDDR7 ECC
Users
30–100
Concurrency
20–40 simultaneous users

Full Hardware Specifications

GPU

Model2× NVIDIA RTX PRO 6000 Blackwell Max-Q
VRAM192 GB GDDR7 ECC
Bandwidth3600 GB/s
AI Performance8000 TOPS (FP4)
FP16220 TFLOPS
TDP300W per card
InterfacePCIe 5.0 x16
ArchitectureBlackwell

Processor

ModelAMD Threadripper PRO 7955WX
Cores16 / 32T
SocketsTR5
PCIe128 lanes

Memory

Capacity256 GB
TypeDDR5 ECC
Configuration8×32 GB

Storage

Capacity4 TB
TypeNVMe Gen4

Platform

Form Factor4U rack-mount server
MotherboardASUS WRX90E-SAGE (TRX50 / WRX90)
Power Supply1600W 80+ Titanium
NetworkIntel X550-T2 (Dual 10GbE)

AI Performance

Max Model (FP16)70B (tensor parallel)
Max Model (Quantized)120B+ (Q4) / multiple concurrent 70B
Concurrency20–40 simultaneous users
Power Consumption1000W

Pre-loaded AI Models

Every appliance ships with a curated selection of production-ready open-source models. Swap or add models anytime.

Llama 4 Scout
Llama 4 Scout
17B · FP16
Gemma 4 31B
Gemma 4 31B
31B · FP16
Qwen 3.6 35B
Qwen 3.6 35B
35B · FP16
GLM-4.7 30B
GLM-4.7 30B
30B · FP16
Mistral Small 3.1
Mistral Small 3.1
24B · FP16
Devstral 2 24B
Devstral 2 24B
24B · FP16
Devstral 2 123B
Devstral 2 123B
123B · Q4
Qwen3-Coder-Next 80B
Qwen3-Coder-Next 80B
80B · Q8
Nemotron 3 Super 120B
Nemotron 3 Super 120B
120B · Q4

AI-Powered Applications

Run agentic coding tools and automation platforms locally — with full data privacy. No API keys, no cloud dependency.

Claude Code
Claude Code
Codex
Codex
OpenCode
OpenCode
OpenClaw
OpenClaw

What's included with this tier

Every WerkHub AI appliance ships ready to run. No extra setup fees, no surprise add-ons — software, installation, and support are all in the price.

Software stack

Full AI software stack pre-installed (vLLM, Open WebUI, WerkHub Console, all pre-loaded models and applications, multi-model configuration)

Installation & onboarding

Delivery + on-site rack integration + network setup + full-day staff training

Support

Next-business-day on-site support, 12 months included. Remote monitoring + guaranteed SLA response.

Alternative GPU Configurations

This appliance is also available with AMD or Intel GPUs at different price points. Contact us for details.

AMDProven
GPU4× Radeon AI PRO R9700
VRAM128 GB GDDR6
Bandwidth2560 GB/s
Price25.000 €

4× blower-style GPUs. Proven multi-GPU vLLM on ROCm 7. 128 GB total — 64 GB less than NVIDIA option but higher aggregate throughput.

IntelNew
GPU4× Arc Pro B70
VRAM128 GB GDDR6
Bandwidth2432 GB/s
Price23.000 €

Cheapest 128 GB configuration. Needs multi-GPU inference validation. OpenVINO stack.

Ready to get started?

Request a quote for this configuration or book a consultation to discuss your specific needs.

Request a Quote