ADLINK Launches First COM Express Module with Intel Core Ultra 3

Why It Matters

By packing high‑density AI compute into a compact, rugged form factor, ADLINK enables next‑generation edge solutions for robotics, medical imaging and industrial automation, reducing system complexity and time‑to‑market.

Key Takeaways

• Up to 180 TOPS AI compute in COM-HPC module.
• Intel Arc GPU offers 12 Xe cores, 120 TOPS.
• Industrial‑temperature SKUs operate –40°C to 85°C.
• Supports 128 GB DDR5 with IBECC ECC memory.
• Launch slated for Q2 2026, expanding rugged edge AI market.

Pulse Analysis

Edge AI is moving from data‑center‑centric models to distributed, on‑device intelligence, and Intel’s Core Ultra 3 architecture is a catalyst for that shift. The processor’s hybrid core design—four performance, eight efficiency, and four low‑power cores—delivers a balanced power envelope while the integrated NPU 5 and Arc GPU provide dedicated acceleration. ADLINK’s Express‑PTL leverages these blocks to reach 180 TOPS across CPU, GPU and NPU, a density previously seen only in larger servers. Coupled with up to 128 GB DDR5 and IBECC error‑correcting memory, the module offers both raw performance and reliability for mission‑critical deployments.

The technical package translates into tangible benefits for demanding workloads. Robotics platforms gain sub‑millisecond perception loops thanks to the 12‑core Arc GPU’s 120 TOPS graphics and AI throughput, while the NPU’s 50 TOPS accelerate inference for vision and language models. Industrial automation systems benefit from the –40 °C to 85 °C operating range, TSN‑enabled Ethernet, and functional safety compliance, allowing seamless integration into harsh factory floors. By consolidating CPU, GPU, and NPU on a single COM‑HPC Mini board, system integrators can cut PCB complexity, lower BOM costs, and accelerate time‑to‑prototype.

From a market perspective, ADLINK’s announcement positions it alongside competitors like NVIDIA Jetson and Qualcomm Snapdragon in the rugged edge segment, but with a distinct advantage: native Intel ecosystem support and the flexibility of COM Express standards. The upcoming COM‑HPC‑mPTL, SBC35‑PTL and MXE‑330 extend this advantage to even smaller footprints and customizable I/O, signaling a broader push toward modular, high‑performance edge AI solutions. As enterprises seek to embed intelligence at the sensor layer, the combination of Intel’s silicon roadmap and ADLINK’s form‑factor expertise is likely to accelerate adoption across autonomous robots, medical imaging devices, and next‑gen industrial IoT deployments.