Video Systems accelerates Industrial AI with Lenovo, AMD, and Infostar.
A new AI-ready infrastructure reduces model training times by over 50% and lays the groundwork for Video Systems' next step toward Physical AI through SENTIO.The evolution of Machine Vision and Artificial Intelligence applications requires increasingly greater computing power, especially when multiple research and development teams must work simultaneously on training, validation, and deployment of AI models.To support this growth, Video Systems, together with partner Infostar, has built a new infrastructure based on Lenovo and AMD technology, designed to offer performance, scalability, reliability, and energy efficiency to support the most advanced AI workloads. lenovo_videosystems_casestudyA New AI-Ready PlatformThe solution includes two Lenovo ThinkSystem SR665 V3 servers, equipped with AMD EPYC 9124 Server CPUs and NVIDIA L40S GPUs.The infrastructure supports the design, training, validation, and deployment of AI-based Machine Vision models developed on the Ingenium® platform. The implementation was carried out with the joint support of the Infostar and Lenovo teams, allowing workload migration without interrupting ongoing operations. lenovo_videosystems_casestudyThe platform also benefits from Lenovo Premier Support for Infrastructure, which guarantees 24/7/365 direct access to technical specialists, helping to maximize infrastructure availability and reduce issue resolution times. lenovo_videosystems_casestudyMeasurable Results in AI DevelopmentThe new infrastructure has allowed Video Systems to achieve significant improvements in research and development activities:over 50% reduction in AI model training times;30% more availability for developers thanks to virtual GPUs;reduction in energy consumption for TFLOP/W. lenovo_videosystems_casestudyVirtualization of GPUs also allows hardware resources to be divided into multiple independent instances, enabling different AI workloads to be executed simultaneously.The platform is used not only for training Machine Vision models but also for Digital Twin simulations, predictive maintenance analysis, and parallel development of multiple projects, accelerating the path from R&D to industrial application. lenovo_videosystems_casestudyFrom Industrial AI to Physical AIUpgrading the infrastructure also represents a technological foundation for the next phase of Video Systems' development: Physical AI.Through SENTIO, Video Systems is advancing research in reasoning-driven robotics, aiming to connect perception, reasoning, and action within physical and industrial environments. lenovo_videosystems_casestudy“Our next step is to build on decades of industrial AI experience and advanced reasoning-driven robotics through SENTIO. We need reliable, scalable computing that gives our teams room to develop and test new ideas.”Alessandro Liani, Founder & CEO, Video Systems Srl lenovo_videosystems_casestudyThe new platform enables multiple teams to work simultaneously on AI models, creating a stronger foundation for the development of intelligent robotics through SENTIO and for exploring new applications in the Aerospace & Defence sectors as well. lenovo_videosystems_casestudyA Technological Partnership to Support GrowthThe collaboration with Lenovo, AMD, and Infostar allows Video Systems to have an infrastructure designed not only for current needs but also to support future expansion of AI activities.“Lenovo and AMD technology offered superior performance, and the energy efficiency to help us control operational costs. Knowing that Video Systems is set to continue growing rapidly in the near future, the scalable nature of the solution is crucial to helping us meet demand effectively.”Alessandro Liani, Founder & CEO, Video Systems Srl lenovo_videosystems_casestudyFor Video Systems, this investment represents a further step in the evolution of its technologies for industrial quality control, Machine Vision, Artificial Intelligence, and Physical AI, with the aim of accelerating innovation and bringing new solutions to market more quickly.