Introduction

On May 20, 2025, NXP Semiconductors introduced its new i.MX 9‑A family—automotive‑grade SoCs designed to run AI inference workloads in real time for semi‑autonomous driving and smart cockpit systems. The launch aligns with rising demand for edge intelligence in vehicles, especially in mid‑tier electric and hybrid models.

“AI isn’t just for luxury—every car will soon need inference,” said Lars Reger, CTO at NXP.¹ The i.MX 9‑A integrates ARM Cortex‑A78AE CPUs with NXP’s neural processing unit (eNPU) and vision DSPs, optimized for low‑latency driver monitoring, sensor fusion, and contextual navigation.

The chip supports 4‑bit to 16‑bit quantization, over‑the‑air model updates, and functional safety (ASIL‑B/D) compliance—critical for automakers looking to balance innovation with regulation.

Why it matters now

• Automotive AI adoption is moving from pilot to production across mass‑market vehicles.
• Edge compute reduces dependence on centralized ECUs and cloud backhaul.
• Regulations are expanding to require real‑time monitoring and response systems.

Call‑out: Inference is becoming as standard as airbags

NXP’s i.MX 9‑A can perform 300 GOPS at under 5 W—enabling driver safety, voice control, and V2X response even in entry‑level vehicles.

Business implications

Tier 1 suppliers and OEMs must accelerate AI system integration to stay compliant and competitive. NXP’s move gives automakers a production‑ready platform that meets both technical and regulatory demands.

Fleet operators, insurance firms, and smart infrastructure providers may also leverage these chips for in‑vehicle telemetry, risk modeling, and adaptive behavior reinforcement.

Looking ahead

Vehicles using i.MX 9‑A SoCs are expected to debut in 2026 models from Stellantis, Hyundai, and China’s XPeng Motors. NXP also announced partnerships with BlackBerry QNX and Amazon AWS for runtime and over‑the‑air model orchestration.

Gartner forecasts that by 2028, over 70% of passenger vehicles will run real‑time AI inference on dedicated silicon—up from 28% in 2024.

The upshot: NXP’s new chips signal a new default in automotive architecture—one where AI isn’t an upgrade but a requirement. In the disruption of mobility, silicon is taking the wheel.

––––––––––––––––––––––––––––
¹ Lars Reger, NXP Tech Days Europe Keynote, May 20, 2025.