Integrated Edge AI Processor with Secure Wireless Connectivity

NXP Semiconductors has expanded its i.MX 9 family with the i.MX 93W application processor, a system designed to integrate edge computing, AI acceleration and secure wireless connectivity in a single package. The processor targets embedded systems in industrial automation, smart buildings, healthcare devices and IoT gateways where local AI processing and low-latency connectivity are required.

Edge AI integration for physical environments
Edge computing architectures increasingly rely on localized AI processing to control physical systems with minimal latency. Applications such as smart building management, industrial automation and connected healthcare devices require embedded platforms capable of processing sensor data locally while maintaining reliable connectivity.

The i.MX 93W application processor addresses these requirements by combining application processing, an integrated neural processing unit (NPU), and multi-protocol wireless connectivity within a single system-in-package architecture. According to NXP Semiconductors, the integration can replace up to 60 discrete components typically required in comparable embedded designs. Reducing component count decreases board space, simplifies supply chains and lowers design complexity in embedded electronics.

The processor is designed for physical AI deployments where multiple AI agents operate collaboratively at the edge. For example, in smart building environments, AI agents can coordinate building subsystems including lighting, heating, ventilation and air conditioning (HVAC), occupancy monitoring and access control to optimize energy usage and building operations in real time.

Integrated AI acceleration and wireless connectivity
The processor integrates a dual-core Arm Cortex-A55 CPU and an Arm Ethos neural processing unit capable of delivering up to 1.8 eTOPS (effective tera operations per second) for machine-learning inference workloads. This architecture supports edge AI applications such as sensor analytics, computer vision preprocessing and intelligent device orchestration.

Wireless connectivity is provided through an integrated tri-radio subsystem that supports Wi-Fi 6, Bluetooth Low Energy and IEEE 802.15.4 protocols. The latter enables communication frameworks such as Thread and Matter used in interoperable IoT networks. Integrating these wireless standards within the processor package eliminates the need for external radio modules and reduces RF design complexity.

For embedded developers, this integration addresses common challenges associated with radio coexistence, antenna tuning and wireless certification, which can otherwise extend development cycles.

Security architecture aligned with regulatory requirements
Security mechanisms are integrated directly into the processor through the EdgeLock Secure Enclave architecture. This hardware root-of-trust enables functions such as secure boot, secure firmware updates, device attestation and protected device access.

The security architecture also supports compliance with regulatory frameworks including the European Cyber Resilience Act (CRA). Secure provisioning and lifecycle management can be implemented through the EdgeLock 2GO key management service, allowing original equipment manufacturers to deploy cryptographic keys and certificates during manufacturing or device operation.

Pre-certified reference designs for faster deployment
To reduce development complexity, NXP provides pre-certified reference designs based on the processor. These designs support both single-antenna and dual-antenna configurations and have already undergone wireless certification in multiple regions.

Pre-certified RF architectures help developers avoid extensive radio testing and regulatory approval processes, reducing the time required to bring embedded products to market.

Application Areas for Edge AI Systems

Key Capability
Combines AI acceleration, wireless connectivity, and embedded security in a single system-in-package architecture, enabling compact and energy-efficient designs for distributed edge computing environments.

Prepared with AI assistance and edited by Sucithra mani

www.nxp.com