NVIDIA has announced Vera, a system designed to optimize both throughput and agent speed in AI inference workloads. According to the company, Vera achieves high data processing rates while maintaining fast response times for individual agent operations. The announcement appears on the official NVIDIA blog highlighting the dual performance capabilities.