
4U air-cooled AMAX AceleMax AI server pairing Up to 8x dual-slot PCIe GPUs (Gen5 x16, 600W) with dual AMD processors, built for large-scale training, inference and HPC, sold and supported through GPU Vendors.
The AMAX AceleMax AXG-428A2 4U 8x PCIe GPU AI Server is an enterprise-grade server from AMAX, engineered for datacenter, virtualization, and AI/HPC workloads. It is built on AMD EPYC 9005 Series in a dual-socket configuration within a 4U chassis.
Memory support covers 24 DIMM slots of DDR5 RDIMM (6400/6000/5600 MHz), scaling up to 6 TB. Storage is handled by 8x 2.5" front hot-swap (up to 8 NVMe) + 2x M.2 Gen5 x4.
Expansion includes 14 total: 8x Gen5 x16 FHFL (GPUs), 5x Gen5 x16 FHHL, 1x Gen5 x8 FHHL with PCIe GPU support. Networking is provided by 2x 10GbE LAN (Intel X710-AT2); 1x RJ45 management port. Power is delivered by 3+1 redundant 3200W 80 PLUS Titanium, with air cooling.
Typical deployments include AI Training, AI Inference, HPC.
Configure and compare pricing for the AMAX AceleMax AXG-428A2 4U 8x PCIe GPU AI Server from verified vendors on GPU Vendors. Request a custom quote with full specifications, build-to-order options, and delivery lead times.
| CPU | AMD EPYC 9005 Series |
| Sockets | 2 |
| GPU platform | Up to 8x dual-slot PCIe GPUs (Gen5 x16, 600W) |
| GPU count | Up to 8 |
| Type | DDR5 RDIMM (6400/6000/5600 MHz) |
| DIMM slots | 24 |
| Max memory | Up to 6TB DDR5 total |
| Drive bays | 8x 2.5" front hot-swap (up to 8 NVMe) + 2x M.2 Gen5 x4 |
| Storage type | NVMe |
| Network | 2x 10GbE LAN (Intel X710-AT2); 1x RJ45 management port |
| PCIe slots | 14 total: 8x Gen5 x16 FHFL (GPUs), 5x Gen5 x16 FHHL, 1x Gen5 x8 FHHL |
| Power supplies | 3+1 redundant 3200W 80 PLUS Titanium |
| Cooling | Air cooling |
| Dimensions | 800 x 439.5 x 175 mm |
We'll email you the moment it happens. No account needed.
Free price-tracking. Unsubscribe anytime.
Comments
Smart picks based on AMAX AceleMax AXG-428A2 4U 8x PCIe GPU AI Server — upgrades and comparable options.

Scale up for bigger workloads

Scale up for bigger workloads

Scale up for bigger workloads

Scale up for bigger workloads

Comparable system

Comparable system

Comparable system

Comparable system
Earn More