G893-ZD1-AAX3 Gigabyte Dual EPYC 9004 8U 6NG893ZD1DR000ACX3 Server Barebone
Do you have a major deal consisting of various components? Then you're at the right place!
We offer you the best prices for your purchase:
- - If you already have an account with us, please log in. If you don't have an account, you can apply for one here.
- - simply collect your required items in the desired quantity in CART, and click the 'GET QUOTATION' button. and you're done!
- - You will receive your offer within the next 24 hours on working days.
Gigabyte G Series
Do you have questions about this product 6NG893ZD1DR000ACX3?
G893-ZD1-AAX3 Gigabyte
High-Density 8U HGX GPU Server platform for AI and HPC featuring NVLink/NVSwitch fabric with the G893-ZD1-AAX3 Gigabyte
The G893-ZD1-AAX3 Gigabyte is an 8U rack-scale HGX platform engineered for GPU-centric AI training and accelerated HPC, pairing 8x NVIDIA HGX H200 GPUs on a unified baseboard with a dual-socket AMD EPYC SP5 CPU foundation (9005/9004 generations). This architecture is designed to keep large model pipelines fed, using high core density and high TDP headroom to sustain data preparation, orchestration, and CPU-bound preprocessing without constraining GPU utilization.
To support large-batch training and high-throughput inference, the G893-ZD1-AAX3 Gigabyte integrates a 24-slot DDR5 ECC RDIMM memory topology (up to 6TB) for high-capacity host memory and stable long-duration runs. PCIe 5.0 expansion enables scalable I/O and accelerator-adjacent connectivity, while 8x hot-swap 2.5" NVMe bays provide low-latency local datasets and checkpoint targets; dual M.2 boot options isolate the OS from high-IOPS data volumes. Integrated 2x 10GbE and an AST2600 BMC with IPMI remote management simplify provisioning, fleet monitoring, and service workflows in production clusters.
Why Choose the G893-ZD1-AAX3 Gigabyte?
The G893-ZD1-AAX3 Gigabyte delivers a pragmatic density-to-performance profile by consolidating 8x HGX H200 with NVLink/NVSwitch into a serviceable 8U envelope, reducing node count, cabling complexity, and inter-node latency for scale-up workloads. Dual SP5 CPUs, PCIe 5.0 lanes, and redundant hot-swap power are aligned for enterprise availability, predictable scaling, and repeatable deployment patterns across AI and HPC racks.
Gigabyte designs this platform around datacenter-operational requirements: remote lifecycle control, component serviceability, and power delivery tuned for sustained accelerator loads, supporting modern rack standards and thermal practices.
Key Applications:
- Large language model training
- Multimodal foundation model training
- High-throughput batch inference
- GPU-accelerated simulation and computational science
- Visualization and rendering
- GPU-accelerated VDI and virtualization consolidation
At Happyware - Server Hardware you benefit from non-binding offers tailored precisely to your AI cluster requirements for the G893-ZD1-AAX3 Gigabyte, with an extended warranty option up to 6 years to protect long-lived infrastructure investments.
Technical details of G893-ZD1-AAX3 Gigabyte
How long is the Warranty and how you can get technical support for G893-ZD1-AAX3 Gigabyte
| Warranty: | 36 months (part) |
|---|---|
| RMA Url: | https://www.gigabyte.com/Support/Consumer/Warranty |
| Description: | Items must be returned first. The manufacturer does not send replacement items in advance. Delivery times are not guaranteed, and on-site service is not offered as standard. |
| Support Portal: | https://support.happyware.com |
|---|
Customer reviews about G893-ZD1-AAX3 Gigabyte
Login
- Express shopping
- Save your data and settings
- Order overview and shipping information
- Manage your newsletter subscription
Similar/Alternative Products
Deals for G893-ZD1-AAX3 Gigabyte
We don’t have any deals for this product yet.