Aetina Expands NVIDIA MGX SuperEdge Portfolio with Certified AEX-2UA1 and New AEX-2UA2 Servers
The two compact 2U short-depth GPU servers bring flexible, high-performance computing to enterprise edge and near-edge environments, supporting generative AI, AI agents and on-premises inference workloads.
www.aetina.com
Aetina Corporation, a leading provider of Edge AI solutions, today announced the expansion of its SuperEdge portfolio for enterprise AI with the NVIDIA Certified System (NCS) AEX-2UA1 and the new AEX-2UA2. Designed in compact 2U short-depth form factors, the two NVIDIA GPU servers extend high-performance accelerated computing into enterprise edge and near-edge environments where space, data sovereignty, latency, and deployment flexibility are critical.
Built on NVIDIA MGX modular reference architecture, the AEX-2UA1 has achieved NVIDIA Certified System (NCS) certification, while the new AEX-2UA2 expands the SuperEdge family with an alternative rear-access design and broader AI infrastructure flexibility. As enterprises move AI from pilot projects into production, infrastructure must handle larger models, concurrent workloads, sensitive data, and real-time interaction—without relying solely on centralized cloud or conventional full-depth data center servers.
“Enterprise AI is moving from experimentation toward production, where AI agents and generative models need reliable infrastructure that operates close to enterprise data,” said Jame Su, Software Division Director of Aetina. “With the AEX-2UA1 achieving NVIDIA Certified System status and the addition of the AEX-2UA2, we’re giving enterprises more flexibility to match GPU performance and deployment format to their AI workloads, helping them move from development to production with confidence.”
A Strategic Infrastructure Partner for Enterprise AI
Aetina is a select system vendor globally acknowledged by NVIDIA to design MGX servers, combining architecture expertise with its long-standing strengths in industrial and edge AI system design. This positions the AEX-2UA1 and AEX-2UA2 as infrastructure building blocks for Enterprise AI, enabling organizations to:
- Standardize AI platforms across edge, near-edge, and enterprise environments
- Deploy validated hardware aligned with the NVIDIA AI software stack and roadmap
- Reduce integration risk and accelerate production rollouts of AI applications
AEX-2UA1: NVIDIA-Certified Multi-GPU MGX Server
The AEX-2UA1 has achieved NVIDIA Certified System status when configured with 2x NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and is further qualified by NVIDIA for NVIDIA L40S GPU and the NVIDIA H200 NVL GPU. This provides enterprises and system integrators with validated GPU options to balance performance, memory capacity, and budget across different AI workloads.
Powered by a single-socket Intel® Xeon® 6700-series processor with P-cores, the AEX-2UA1 supports dual double-width GPUs up to 600W with NVIDIA NVLink for configurations with H200 NVL, enabling high-bandwidth GPU-to-GPU interconnect for demanding training and inference workloads. It also provides a dedicated PCIe Gen5 x16 slot for an NVIDIA BlueField-3 DPU or ConnectX-7 NIC, plus up to four hot-swappable PCIe Gen5 E1.S bays and two PCIe Gen5 M.2 slots with NVMe RAID 0/1/5/10 support.
Measuring 438 × 420 × 88 mm, the compact 2U short-depth chassis brings high-density GPU computing into space-constrained deployments such as enterprise IT rooms, smart manufacturing sites, telecom environments, retail facilities, and AI innovation centers.
AEX-2UA2: Expanding SuperEdge Deployment Flexibility
The new AEX-2UA2 expands the SuperEdge portfolio with an alternative rear-access 2U short-depth architecture. It pairs an Intel® Xeon® 6700P processor with a broader range of high-performance GPU configurations, allowing system integrators and enterprise customers to select an infrastructure design that better fits existing racks, cabling, maintenance workflows, and installation environments.
Generative and Agentic AI in Action: Powering Private AI Knowledge Assistants and Executive Decision Intelligence
The expanded SuperEdge family runs the full range of enterprise AI workloads—from private AI knowledge assistants to executive decision intelligence—entirely on-premises. GPU configuration scales to the task: NVIDIA L40S handles cost-efficient semantic search and knowledge Q&A across departments; NVIDIA L40S and RTX PRO 6000 Blackwell run mid-size (7B–13B parameter) models for AI agents that draft summaries, automate workflows, and answer compliance questions in real time; NVIDIA H200 NVL GPUs and RTX PRO 6000 Blackwell support large (70B+ parameter) models for executive decision intelligence across vast internal knowledge bases; and dual NVIDIA H200 NVL GPUs with NVIDIA NVLink let enterprises fine-tune models on proprietary data on-premises. With inference processed locally, executives gain real-time visibility into operations, employees get instant answers instead of searching disconnected repositories, and enterprises can adopt generative AI at scale while meeting data privacy, governance, and compliance requirements.
One SuperEdge Portfolio. More Possibilities for Enterprise AI
The NVIDIA-Certified AEX-2UA1 and new AEX-2UA2 give organizations the flexibility to align GPU performance, system architecture, and deployment format with evolving AI workloads—from private generative AI and multimodal inference to AI agents and digital humans.
Learn more information at : www.aetina.com
