Aetina Launches Compact 2U GPU Servers for AI
The new short-depth systems target enterprise AI workloads at edge and near-edge sites where space and data sovereignty are constraints.
Updated on Nov. 6, 2026 in Data Centers
Aetina has expanded its SuperEdge portfolio with two compact 2U short-depth GPU servers designed for enterprise AI. The AEX-2UA1 has received NVIDIA Certified System status, while the AEX-2UA2 offers a rear-access design.
Why it matters
The systems are intended for edge and near-edge deployments where space, latency, installation flexibility and data sovereignty are considerations. Local inference is positioned as a way to address data privacy, governance and compliance requirements.
The AEX-2UA1 measures 438 × 420 × 88 mm and supports two double-width GPUs of up to 600W each, plus up to four hot-swappable PCIe Gen5 E1.S bays and two PCIe Gen5 M.2 slots. Pricing, availability and independent performance benchmarks were not reported.
The players
Aetina
Aetina is a Taiwan-based provider of edge AI computing hardware, known for its SuperEdge product line of embedded and edge servers.
NVIDIA
NVIDIA is a U.S. chipmaker whose GPUs and MGX modular reference architecture underpin AI computing systems across the industry.
The details
The AEX-2UA1 pairs a single-socket Intel Xeon 6700-series processor with a dedicated PCIe Gen5 x16 slot for an NVIDIA BlueField-3 DPU or ConnectX-7 NIC. The AEX-2UA2 pairs an Intel Xeon 6700P processor with a range of GPU configurations, and both systems are built on the NVIDIA MGX modular reference architecture.
Timeline
The AEX-2UA1 received NVIDIA Certified System status in the announcement dated November 6, 2026.
The Tech Race
The AEX-2UA1 and AEX-2UA2 are built on NVIDIA's MGX modular reference architecture, extending that design approach into compact short-depth form factors. The move places Aetina among vendors competing to supply edge-deployable AI servers using NVIDIA's GPU platforms.
For organizations running AI inference close to where data is generated, the systems offer a smaller footprint that can fit space-constrained sites. Buyers handling sensitive data may find local processing simplifies compliance, though pricing and availability were not disclosed.
The takeaway
Compact GPU servers with NVIDIA certification are narrowing the gap between data-center AI capability and edge deployment. Organizations evaluating local inference should weigh footprint, data governance needs and vendor certification status before committing to hardware.
Further reading
Explore more on Data Centers for the infrastructure trends behind compact edge AI servers.
Source note: This article includes information reported by Sourcesecurity.
