ADLINK Technology has expanded its edge AI portfolio with four new server-class systems, designed to meet the rising demand for on-site generative AI, real-time vision processing, and industrial automation. The lineup includes the ISB-W890 server board—a flexible foundation for custom workstations—and three AXE-series GPU servers optimized for high-end inference tasks in constrained or high-density environments.
As edge AI transitions from experimental pilots to production-grade deployments, the need for localized compute power has intensified. Traditional cloud-based AI solutions often introduce unacceptable latency for applications like medical imaging, robotics, or smart factory inspection. ADLINK’s latest platforms address this by integrating Intel Xeon 600 processors with scalable GPU configurations, enabling enterprises to run large language models, multimodal AI, and vision analytics directly at the edge.
- ISB-W890 Server Board
- CEB form factor for system integrators
- 7x PCIe slots (3x PCIe x16 Gen 5) for multi-GPU scalability
- M.2 NVMe and SATA III storage options
- AXE-7440GW (4U)
- Supports up to four FHFLDW GPU Gen 5 x16 cards
- Targeted for high-complexity generative AI and multimodal workloads
- Full-depth 4U chassis for high-power density
- AXE-7420GWA (4U Short-Depth)
- Short-depth (550 mm) 4U chassis with industrial ATX power
- Optimized for factory automation and AI consolidation
- Supports multiple GPU configurations for vision and inference tasks
- AXE-7220GW (2U SFF)
- Compact 2U short-depth (450 mm) chassis
- Supports two FHFLDW GPU Gen 5 x16 cards
- Ideal for space-constrained edge deployments
The ISB-W890 stands out as a modular foundation for custom workstations, offering unmatched I/O flexibility for AI researchers and system builders. Its seven PCIe slots—including three Gen 5 x16—allow for parallel GPU configurations, making it suitable for complex inference tasks like 3D reconstruction or multi-modal AI training. Meanwhile, the AXE-series servers cater to specific deployment scenarios: the AXE-7440GW delivers cloud-like generative AI performance in a 4U rack, while the AXE-7420GWA prioritizes industrial efficiency with a short-depth design. The AXE-7220GW, with its 2U footprint, is engineered for edge locations where space is limited but high-bandwidth vision processing is required.
These systems are positioned as a response to the evolving needs of industries where edge AI is no longer optional. From autonomous robotics in warehouses to real-time defect detection in manufacturing, the ability to process data locally—without relying on cloud latency—is becoming a competitive advantage. ADLINK’s new platforms also support VLA (Vision Large Action) workloads, a category that includes advanced analytics for medical imaging, autonomous systems, and smart infrastructure.
Availability for the ISB-W890 and AXE-series servers is being finalized, with configurations tailored to specific customer requirements. Interested parties are advised to contact ADLINK directly for technical details and deployment options.
