Amlogic A123X & C305X2: Industrial AIoT SoCs Targeting Edge Intelligence

Key Takeaway

Posted September 11, 2026 — Amlogic unveils the A123X quad-core and C305X2 dual-core Arm Cortex-A320 SoCs, bringing Armv9 performance and power efficiency to industrial AIoT applications such as robots, dashcams, and IP cameras.

Figure 1: Amlogic A123X/C305X2 SoC architecture for industrial AIoT

1. SoC Overview

On September 11, 2026, Amlogic unveiled two new Arm Cortex-A320-based SoCs designed specifically for industrial and low-power AIoT applications: the A123X quad-core and C305X2 dual-core. These are the first silicon implementations of the Arm Cortex-A320 core, which was introduced in February 2025 as part of Arm’s v9-A architecture.

The A123X targets machine vision industrial applications including dashcams, sweeping robots, access control, and video conferencing hardware. With its quad-core Cortex-A320 configuration, it delivers significant performance uplift for image processing, computer vision algorithms, and AI inference at the edge.

The C305X2 is optimized for ultra-low-power battery devices such as solar doorbells, portable hunting cameras, and compact 2K/4K battery IP cameras. It serves as an upgrade to the C305X dual-core Cortex-A35 SoC, offering improved performance-per-watt for always-on AI applications.

Both SoCs leverage the Arm Cortex-A320’s Linux software compatibility, ensuring broad OS support and ecosystem compatibility. Amlogic will also offer turnkey reference designs, pre-validated embedded OS stacks, and unified NN SDKs for simplified development.

2. A123X Features

Performance

Quad-core Arm Cortex-A320 configuration
Scalable up to 240 MHz (industrial frequency range)
Supports Armv9-A architecture features including SVE2
Up to 600 DMIPS performance per core

Machine Vision

Targeted at dashcams, sweeping robots, access control systems
Hardware video decoding for multiple codec formats
Integrated IPU (Image Processing Unit) for camera input
Support for 4K@30fps video recording

Connectivity

WiFi 6 and Bluetooth 5.3 support
Ethernet with TSN (Time-Sensitive Networking)
PCIe 3.0 interface for external accelerators
USB 3.0 Host/Device capability

3. C305X2 Features

Low-Power Design

Dual-core Cortex-A320 configuration
Optimized for battery-operated devices
1 mW-level standby power (third-generation power management)
Fast Linux wakeup from deep standby

AI Processing

Dedicated Qualcomm QNN (Qualcomm Neural Network) accelerator
Up to 40 TOPS of AI processing performance
Sparse matrix support for efficient inference
Tensor operations for neural network workloads

Image Support

Integrated ISP for up to 3 camera inputs
Support for 2K/4K resolution at 30fps
HDR image processing
RAW image pipeline

4. Industrial AIoT Applications

Robotics

The A123X’s quad-core performance and machine vision capabilities make it suitable for industrial robots requiring real-time perception and decision-making. The SoC can handle simultaneous localization and mapping (SLAM), object detection, and path planning entirely at the edge, reducing latency and dependency on cloud connectivity.

Dashcams & Video Surveillance

For dashcam and surveillance applications, the A123X’s 4K video recording and AI-powered analytics enable features such as lane departure warning, driver monitoring system, and collision avoidance. The integrated connectivity options facilitate over-the-air firmware updates and remote fleet management.

IP Cameras

The C305X2’s low-power design and AI processing make it ideal for battery-powered or solar-powered IP cameras used in perimeter security, warehouse monitoring, and outdoor surveillance. The 40 TOPS AI accelerator enables on-device person detection, facial recognition, and anomaly detection without transmitting raw video to the cloud.

Edge Gateway

Both SoCs can function as edge gateways in industrial IoT networks, protocol translation, data aggregation, and local preprocessing before cloud transmission. The industrial temperature range and robust connectivity make them suitable for deployment in harsh environments.

5. Development Ecosystem

Software Support

Arm Cortex-A320 Linux software compatibility
Amlogic pre-validated embedded OS stacks
Unified NN SDK for AI model optimization
Turnkey reference designs for common industrial applications

Hardware Ecosystem

Evaluation boards and SMARC modules available
Integration with Amlogic’s existing peripheral IP
Long-term availability for industrial deployments
Support for industrial protocols (CAN, Ethernet TSN, PCIe)

6. Market Significance

The introduction of Amlogic’s A123X and C305X2 SoCs represents several important trends in the industrial AIoT market:

First silicon implementations of the Cortex-A320 core, validating Arm’s new v9-A architecture for embedded applications
First vendor to announce Cortex-A320 SoCs, positioning Amlogic as a leader in industrial AIoT
Quad-core and dual-core configurations targeting different performance-per-watt points
Integration of NPU acceleration directly in the SoC, reducing the need for external AI accelerators
Q4 2026 mass production scheduling, indicating volume production readiness

For industrial companies evaluating edge AI hardware, the A123X/C305X2 offers a compelling combination of performance, power efficiency, and integrated connectivity that addresses key bottlenecks in deploying AI at scale in industrial environments.

10. Sources

Jean-Luc Aufranc. “Amlogic A123X and C305X2 Arm Cortex-A320 SoCs target industrial and low-power AIoT applications.” CNX Software, 11 Sep 2026. https://www.cnx-software.com/2026/09/11/amlogic-a123x-and-c305x2-arm-cortex-a320-socs-target-industrial-and-low-power-aiot-applications/
Arm Ltd. “Arm Cortex-A320.” https://www.arm.com/solutions/microcontrollers/arm-cortex-a320
Amlogic. “Amlogic Corporate Website.” https://www.amlogic.com
CNX Software. “Amlogic Unveils New AIoT SoCs at CNX Software.” 11 Sep 2026.

Key Takeaways

Amlogic A123X/C305X2 are the first Cortex-A320-based SoCs for industrial AIoT
A123X: quad-core for machine vision (dashcams, robots, access control)
C305X2: dual-core for low-power IP cameras and battery devices
Both integrate NPU with up to 40 TOPS AI processing performance
Q4 2026 mass production, first samples available
Integrated connectivity: WiFi 6, Ethernet TSN, PCIe, USB 3.0