Nvidia has announced a new suite of AI technologies focused on video analysis and agent-based automation, as revealed during CEO Jensen Huang’s CES 2025 keynote address. The company has launched several blueprints that enable developers to create AI agents capable of analyzing video content, processing documents, and automating various enterprise tasks.
The new technology includes the Nemotron family of AI models, available in three sizes – Nano, Super, and Ultra – designed for different computing requirements. These models are built on Meta’s Llama architecture and are optimized for tasks such as instruction following, chat functionality, and coding. The video analysis capability addresses a significant market need, as current estimates indicate that less than 1% of footage from 1.5 billion enterprise-level cameras worldwide is actively monitored.
Nvidia has partnered with several companies, including CrewAI, LangChain, and LlamaIndex, to develop specialized blueprints for AI agent orchestration. These blueprints integrate with Nvidia’s AI Enterprise software platform and include features such as the PDF-to-podcast conversion tool and video search and summarization capabilities. The company states that the technology can be deployed through various cloud platforms, including AWS, Google Cloud, and Microsoft Azure.
Sources: VentureBeat, VentureBeat, VentureBeat