Home/Products/On-prem & edge servers
InInventec

On-Prem & Edge AI Servers

Servers for the fleet back-office: train and fine-tune models, aggregate robot data, and run inference on-site — in short-depth edge chassis or GPU-dense 2U nodes.

Product imagery coming soon

Inventec · 1U short-depth

EdgePro 1U

Intel Xeon 6 edge AI server

An ultra-short 438mm chassis that fits where full racks don't — warehouses, depots and telecom closets — with modular GPU, storage and networking upgrades.

438 mmultra-short depth
Xeon 6processor
ModularGPU / storage / NIC
Form factor1U, ultra-short 438mm depth
ProcessorIntel Xeon 6
ExpansionModular GPU, storage and networking upgrades
WorkloadsEdge AI, data pre-processing, fleet data aggregation
Ideal forWarehouse back-officeDepot / yard serversData pre-processing
K888G7

Representative image — final chassis may differ.

Inventec · 2U

K888G7

Dual Intel Xeon 6 GPU server

Four double-width GPUs and up to 26 U.2 drives in 2U — on-prem training, simulation and model serving without going to the cloud.

4× GPUPCIe Gen5 x16
26× U.2NVMe bays
176PCIe lanes
Form factor2U
ProcessorDual Intel Xeon 6
MemoryDDR5 RDIMM, 6400 MT/s @ 1DPC
GPUUp to 4× PCIe Gen5 x16 double-width FHFL
StorageUp to 26 U.2 SSDs
CoolingOptional liquid cooling
Ideal forOn-prem trainingModel servingSimulation
K885G7

Representative image — final chassis may differ.

Inventec · 2U

K885G7

Dual AMD EPYC 9005 server

Up to 192 cores per socket with 24 DDR5 DIMMs and two full-size GPUs — for CPU-heavy pipelines, data processing and mixed workloads.

192cores per socket
24DDR5 DIMMs
2× GPUdouble-width
Form factor2U
ProcessorDual 5th Gen AMD EPYC 9005 (Turin), up to 500W
Memory24 DDR5 DIMM slots
GPU2× PCIe Gen5 x16 double-width FHFL
I/O80 PCIe Gen5 lanes
Ideal forData processingHPCMixed workloads

Need a different configuration?

Memory, storage, wireless, OS images and mounting can all be configured to your spec — we'll confirm lead time and pricing.