NXP CTO Lars Reger on Physical AI, Robots, Drones, and Trust

NXP's NavQ95 is not a product launch—it is a strategic repositioning. As robots and drones migrate from labs to factory floors, the binding constraint shifts from raw compute to trust: can an edge node execute a tamper-proof perceive-infer-act loop in milliseconds? On the tech stack, NavQ95 fuses inertial sensing, secure MCUs, and lightweight inference into a single module, elevating NXP from component vendor to decision-stack vendor and structurally outflanking STMicroelectronics and Infineon, whose pure-silicon models cannot replicate that integration moat. Compliance is the quieter battleground. The EU AI Act's traceability mandates and data-sovereignty fragmentation across Southeast Asian supply chains are converting security from a cost line into a market-access prerequisite. NXP's two-decade automotive-grade safety pedigree now translates directly into embodied-intelligence credibility. Over the next 18 months, two trajectories stand out. First, the trust layer in physical AI will be priced as a standalone SKU, echoing the TPM-chip premium logic of the mid-2000s. Second, geopolitical decoupling will spawn regionalized security-stack variants of the same architecture. NXP's modular design is already positioned to capture that fragmentation premium.