The 120-billion-parameter Nemotron 3.5 Super VL is designed for high-stakes enterprise environments, such as financial accounting, medical insurance claims, and security operations. Nvidia claims the model outperforms competitors like Anthropic and OpenAI in specific agentic tasks, a assertion bolstered by early adoption from companies including Salesforce, Crowdstrike, and EY. By offering an open-weight architecture, Nvidia allows businesses to post-train models on proprietary data while maintaining full ownership of their infrastructure.
Nvidia pivots to local AI with new Nemotron and Bolt model release
Nvidia is pushing AI agents from the cloud to the edge with the launch of its Nemotron 3.5 Super VL and lightweight Bolt models. By targeting both enterprise data centers and local AI PCs, the company aims to offer organizations and consumers control over their data without sacrificing performance.
Simultaneously, the company is addressing the growing demand for private, local AI through its new Bolt line. These models, including the 74-billion-parameter 74B-A9B and the smaller E4B, are optimized for Nvidia RTX Spark-powered PCs and DGX Spark workstations. By executing workloads locally, Nvidia hopes to bypass the privacy concerns associated with cloud-based assistants, allowing users to handle sensitive personal or professional information without transmitting data to external servers. Super VL and the larger Bolt model are available immediately, with the smaller E4B variant scheduled for release later this month.
Comments (0)
No comments yet. Be the first!