GTC: What are NVIDIA's New Open Models ?

2026-03-19 · Source: AI Magazine · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Health & Medical Research · Depth: Intermediate, short

Summary

NVIDIA has significantly expanded its open model ecosystem at GTC, releasing new open models designed to advance AI across scientific, agentic, and physical domains. Key releases include Nemotron for agentic AI systems, featuring multimodal capabilities like Nemotron 3 Ultra for enterprise applications and Nemotron 3 Omni for video/document analysis. For physical AI, NVIDIA introduced Cosmos 3 for synthetic world generation and Isaac GR00T N1.7 for humanoid robotics, with GR00T N2 previewed for enhanced robot performance in unfamiliar environments. In healthcare, the BioNeMo platform now includes Proteina-Complexa for protein binder development and nvQSP, a GPU-accelerated simulation engine reporting 77x performance gains for treatment scenario testing. These models aim to empower developers and scientists to build intelligent systems beyond language.

Key takeaway

For CTOs and VP of Engineering evaluating AI infrastructure, NVIDIA's expanded open model ecosystem offers specialized tools for agentic, physical, and biomedical AI. You should consider integrating Nemotron for advanced automation, Cosmos for robotics simulation, and BioNeMo for healthcare applications to accelerate development and deployment, potentially compressing months of work into days and achieving significant performance gains.

Key insights

NVIDIA's open model expansion targets agentic, physical, and biomedical AI, moving beyond language-centric systems.

Principles

• Open models accelerate global innovation.
• AI intelligence extends beyond language.
• Synthetic data supports privacy and localization.

Method

NVIDIA's strategy involves developing specialized foundation models (e.g., Nemotron, Cosmos, BioNeMo) and simulation tools (e.g., nvQSP) to enable advanced AI capabilities across diverse sectors.

In practice

• Integrate Nemotron for enterprise-grade agentic AI.
• Utilize Cosmos 3 for physical AI simulation.
• Employ BioNeMo for accelerated drug discovery.

Topics

• NVIDIA Open Models
• Agentic AI
• Physical AI
• Biomedical AI
• Robotics

Best for: Computer Vision Engineer, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, AI Scientist

Related on AIssential

• Nvidia Launches Physical AI Models for Robots — Nvidia's new physical AI models and frameworks aim to democratize intelligent robot development and deployment.
• NVIDIA GTC 2026 Live — AI is transitioning from a specialized tool to essential, pervasive infrastructure, driving innovation across all industries.
• NVIDIA’s Open Models on Microsoft Foundry: — Microsoft Foundry integrates NVIDIA's open models and accelerated computing to unify AI development and deployment across diverse environments.
• Nvidia’s Cosmos Reason 2 aims to bring reasoning VLMs into the physical world — Nvidia is advancing physical AI with new VLMs and an open ecosystem for embodied reasoning and agentic capabilities.
• NVIDIA brings agents to life with DGX Spark and Reachy Mini — NVIDIA's DGX Spark and Reachy Mini enable building private, customizable, real-world AI agents using open models.
• What’s Next in Robotics? — Physical AI, powered by NVIDIA's ecosystem, accelerates robotics development through advanced simulation and data generation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Magazine.