Visteon Corporation (NASDAQ: VC) announced March 16 it will develop an edge-to-cloud AI platform for software-defined vehicles using NVIDIA technology. The platform distributes AI workloads between vehicle hardware and cloud infrastructure based on speed, privacy and connectivity requirements.

The system determines where workloads execute through runtime intelligence. Functions requiring low latency remain on the vehicle for privacy, while larger models for fleet learning and predictive maintenance use cloud computing. The platform addresses data governance, regulatory compliance and performance requirements across different markets.

The collaboration expands on Visteon's AI-ADAS Compute Module shown at the Consumer Electronics Show, which used NVIDIA DRIVE AGX Orin and NVIDIA DriveOS. The new platform integrates NVIDIA's TensorRT-Edge-LLM inference SDK, Nemotron open models and NIM microservices.

The architecture uses NVIDIA Nemotron models for speech, vision, retrieval-augmented generation and language capabilities. Automakers maintain control over data residency, model configuration and compliance strategy while Visteon provides integration services including model fine-tuning and automotive operating system software.

Applications include in-cabin assistants with safety monitoring, contextual awareness and personalized experiences. The platform aligns with NVIDIA AI Factory principles for global deployment.

Visteon, headquartered in Van Buren Township, Michigan, operates in 17 countries. The company reported $3.77 billion in annual sales for 2025 and secured $7.4 billion in new business, according to the press release.