LiveRamp (NYSE: RAMP), the leader in data collaboration, today announced native support for NVIDIA AI infrastructure, upgrading its clean room architecture to handle the world's most compute-intensive AI workloads. AI partners and brands can now securely and seamlessly train and deploy sophisticated models using LiveRamp clean rooms or via the LiveRamp Marketplace at up to 15x speed, without exposing data or model weights. By moving to GPU-optimized clean room infrastructure, LiveRamp now supports the most advanced model training and inference.

By implementing GPU-optimized infrastructure, LiveRamp eliminates the technical hurdle of AI partners needing to rearchitect models for CPU-based environments. AI partners can now plug-and-play existing code within LiveRamp's clean rooms. Model providers and brands using Marketplace or Clean Rooms can stitch together disparate datasets with unprecedented scale and ease for AI-driven marketing. This ensures compute-intensive models are trained on the most complete and accurate foundation of insights.