NXP Semiconductors (NXPI) COMPUTEX 2026 Keynote summary

Keynote themes and industry context

• AI is rapidly moving from the cloud to the edge, transforming sectors like automotive, industrial, and smart infrastructure.

• Physical AI's challenge is mastering reflexes, not just reasoning or language, as highlighted by Moravec's paradox.

• Edge devices require real-time, low-latency, low-power, and highly secure solutions to be effective.

• The blueprint for elite physical AI is inspired by biological intelligence, emphasizing distributed, layered control.

NXP's neural axis architecture and applications

• NXP's neural axis architecture mimics biological intelligence with three layers: reasoning, coordination, and reflex.

• This architecture is applied in drones, self-driving vehicles, and humanoid robots, each requiring ultra-fast, local reflexes for safety and reliability.

• Separation of layers ensures mission-critical functions are handled independently, reducing risk and increasing trust.

• Humanoid robots benefit from distributed control at joints and limbs, enabling rapid, autonomous responses.

Bridging perception and understanding in robotics

• Robots need not only to perceive but also to understand the physical world, requiring world models and Vision-Language-Action (VLA) models.

• VLAs bridge what robots see, are told, and how they act, enabling safer and more adaptive behavior.

• NXP's eIQ Toolkit enables efficient deployment of advanced AI models at the edge, optimizing for power and latency constraints.