- Single Intel® Xeon® 6700-series with P-cores.
- 2U server - 420mm (16.5") short-depth rackmount server.
- 2x PCIe Gen5 x16 slots (FHFL) and 1x PCIe Gen5 x8 slot (FHFL) – supporting up to 2 double-width PCIe GPUs.
- 1x PCIe Gen5 x16 slot (FHHL) - supporting NVIDIA Bluefield® -3 or NVIDIA ConnectX® -7.
- Up to 4x Hot-Swap PCIe Gen5 E1.S and 2x PCIe Gen5 M.2 slots.
- 1+1 Redundant Titanium level power supply.
Aetina NVIDIA MGX™ Short-Depth Server
Accelerate your Enterprise AI, On-premises LLM with our Edge AI Server
The AEX-2UA1 NVIDIA MGX™ short-depth edge server, built on a modular architecture, is a single socket server integrating Intel® Xeon® 6700-series with performance-cores (P-cores) and dual high-performance double-width GPUs with NVIDIA NVLink™, delivering ultimate edge AI computing performance in a compact 2U form factor, while offering maximum flexibility and compatibility for both current and future hardware.
AEX-2UA1 Front-Access x86
Short-Depth Server
Powered by Single Intel Xeon 6700-series with P-cores and Dual Double-width GPUs to Unleash Incredible AI Power
The AEX-2UA1 short-depth edge server is powered by the latest Intel Xeon 6700-series with P-cores, providing exceptional performance for compute-intensive AI and analytics workloads. With robust support for up to dual double-width GPUs and advanced NVIDIA NVLink™ bridge, it delivers unmatched AI computing power and ultra-fast GPU interconnection. Designed to optimize AI training and inference within constrained spaces at the edge, it is the perfect solution for enterprise server rooms or non-rack locations.
Designed with Modularized Architecture
Delivering greater scalability for enterprises in the present and future
The AEX-2UA1 NVIDIA MGX™ short-depth edge server is based on modularized architecture, offering exceptional flexibility, expansion, and compatibility with current and future NVIDIA hardware. This makes it an ideal solution for enterprises seeking to optimize compute and accelerated hardware, maximizing efficiency in AI-driven solutions while reducing time to market.
Accelerate On-premises Enterprise AI Deployment
As generative AI and large language models (LLMs) permeate business and consumer lifestyles, organizations seek reliable, compact, and cost-effective edge servers to accommodate their diverse accelerated computing and unique AI workloads.
The AEX-2UA1 accelerates deployment of diverse AI-driven applications at the edge, empowering enterprises to effectively handle sensitive data with private LLM (large language model), VLM (vision language model) and diverse enterprise AI applications, particularly valuable in sectors such as finance, healthcare and 5G telco.
SuperEdge AEX-2UA1 Feature Highlight
Powered by a Single Intel Xeon 6700-series with P-cores
Suitable for tackling high demanding AI workloads with great performance
Equipped with 8 channel memory
Support Dual Double-Width GPUs to Deliver Incredible AI Power
Support up to 2 Double-width NVIDIA GPUs with NVIDIA NVLink™, boosting training and inferencing performance through ultra-fast GPU interconnect.
Advanced Network Solution
Support NVIDIA Bluefield®-3 DPUs and NVIDIA ConnectX®-7 NICs
Provide high-speed network communication, minimizing latency for data transfer
Compact 2U Short-Depth Footprint
420mm (16.5") short-depth rackmount server, optimizing space utilization
Suitable for deployments in both standard or short-depth racks or even non-rack locations, providing flexible configuration
Effortless Maintenance with Easy-To-Swap Design
The hot-swappable design of fans, E1.S storage, and power supplies allows for easy removal and quick replacements, ensuring seamless operation without downtime
Boost Enterprise AI Training and Inferencing
Tailored for edge AI applications
Accelerate enterprises data analysis and private LLM processing
Experience Incredible Aetina Short-Depth Server
Specifications
| Model Number | AEX-2UA1 |
| CPU | Single Intel Xeon 6700-series with P-cores |
| Chipset | N/A |
| GPU (Optional) | Support up to 2 Double-Width PCIe Passive GPUs |
| Memory | (8) Channel Memory |
| Storage | External/Rear Storage: (4) Hot-swappable PCIe Gen5 E1.S, support 9.5mm/15mm Internal Storage: (2) PCIe Gen5 x4 M.2, M-Key 2280/22110 |
| Front I/O | I/O Ports: (1) RJ45 1Gbe Dedicated BMC LAN port, (1) MiniDP port, (2) USB 2.0 ports Buttons: System Power, UID/BMC reset Status LED: Power, UID, Storage IO, LAN1, LAN2, Information |
| Expansion | (2) FHFL Gen 5 x 16 PCIe slot (up to 2 double-width GPU with NVLink bridge) (1) FHFL Gen 5 x 16 PCIe slot (x8 link) * (1) FHHL Gen5 x 16 PCIe slot * Not available when 2 double-width AOC is installed |
| MISC. Function | Trusted Platform Module (TPM) 2.0 Intel® Platform Firmware Resilience (Intel® PFR) |
| Power Input / Connector |
IEC 60320 C13 |
| Dimension (W x D x H) |
2U Rackmount (438mm x 420mm x 88mm) |
| Mounting | Rackmount |
| Power Supply | (default) 1600W 1+1 Redundant Titanium level Power supply (optional) 2000W 1+1 Redundant Titanium level Power supply |
| System Cooling | (4) 6056 & (2) 4056 Heavy duty fans with optimal fan speed control |
| Temperature | Operating Temperature: 10°C ~ 35°C (50°F ~ 95°F) Non-operating Temperature: -30°C to 60°C (-22°F to 140°F |
| Humidity | Operating Relative Humidity: 8% to 80% (non-condensing) Non-operating Relative Humidity: 8% to 90% (non-condensing) |