At CES 2026 in Las Vegas, NXP Semiconductors unveiled its eIQ Agentic AI Framework, a software platform that brings real-time, autonomous decision–making directly to edge devices. Built to run on the i.MX 8 and i.MX 9 application processors as well as Ara Discrete NPUs, the framework enables developers to coordinate multiple AI models—vision, audio, time series and control—in parallel with deterministic low-latency performance. NXP estimates that the optimized scheduling engine, which dynamically allocates workloads across CPU, NPU and integrated accelerators, can reduce inference latency by up to 40% compared to generic runtimes. The toolkit includes built-in security measures—secure boot, runtime isolation, hardware root of trust—and safeguards against prompt injection and model spoofing, addressing both safety-critical and data-sensitive edge use cases in industrial automation, smart buildings and medical monitoring.

To accelerate prototyping and deployment, NXP also rolled out the eIQ AI Hub, a cloud-hosted portal offering immediate access to its full eIQ AI Toolkit suite—Time Series Studio, GenAI Flow and the new Agentic AI Framework. Through the hub, developers can spin up cloud-connected hardware boards for real-world performance metrics, then seamlessly transition to on-premise deployment. NXP reports that early adopters have shaved integration times by 30–50%, thanks to automated conversion pipelines and preconfigured hardware profiles. A modular licensing model lets teams customize feature sets, from beginner-friendly workflows to advanced, hardware-aware tuning pipelines, supporting both rapid proof-of-concepts and high-performance production builds.

Building on its automotive roadmap, NXP introduced the S32N7 series of super-integration processors—32 variants built on a 5 nm foundation—designed to centralize core vehicle functions such as propulsion, gateway, body and safety domains. According to NXP, consolidating dozens of discrete ECUs into a single hub can lower total cost of ownership by as much as 20% through reductions in wiring and hardware modules. The superset variant, S32N79, is already sampling with key customers, including Bosch, which plans to integrate the chip into its next-generation vehicle-integration platform. By delivering high-performance networking, hardware isolation and AI acceleration on one SoC, NXP aims to give automakers a scalable path for over-the-air feature updates and the deployment of virtual sensors across their fleets.