At CES 2026 in Las Vegas, NXP Semiconductors introduced its eIQ Agentic AI Framework, marking a major stride in its secure edge AI strategy. Designed for i.MX 8 and i.MX 9 applications processors and Ara discrete NPUs, the new framework enables deterministic, real-time decision making on edge devices without cloud dependency. NXP reports that integrating this software layer with its existing secure hardware can reduce prototyping cycles by up to 40%, addressing use cases from factory safety interventions to medical alert systems with sub-millisecond response times.

NXP has already secured collaborations with leading OEMs and systems integrators. GE HealthCare demonstrated an anesthesia delivery and infant monitoring concept powered by the eIQ Agentic AI Framework at CES, aiming to streamline on-device patient-care analytics. Honeywell plans to embed the framework in its smart building controls to automate HVAC shutdowns during fire events. These alliances underscore NXP’s projection that edge AI device shipments leveraging agentic architectures will grow at a compound annual rate exceeding 50% through 2028.

To simplify deployment, NXP launched the cloud-based eIQ AI Hub alongside its enhanced eIQ AI Toolkit, offering instant access to model conversion, hardware-aware tuning and multi-model orchestration workflows. Novice developers can use automated pipelines to build on-device agentic systems in hours rather than weeks, while experts retain full toolchain integration. Security features—such as prompt-injection prevention, adversarial-input detection and a hardware root of trust with runtime isolation zones—ensure data integrity and resilience in mission-critical environments.