NVIDIA GTC AI Conference

Built on open source foundations, security by design, sustainability at scale, and digital sovereignty, Mistral AI’s approach helps unlock bare-metal GPU performance, build trusted, future-ready AI systems, and advance collaboration by giving innovators access to the same infrastructure powering our models.

In this session, experts will walk through practical, end-to-end workflows for physical AI development using NVIDIA Isaac™ and Omniverse™ technologies.

We'll walk through a practical decision guide for open models from picking the right-sized models based on the agentic tasks to software for customizing, evaluating, and deploying the models at scale.

We’ll dive into debating the concrete architectural trade-offs required for low latency and high throughput. This includes unique training challenges, new model architectures, inference and runtime optimization, low-precision and NVFP4, and scaling topologies.

This talk outlines the singular mission required to deliver scientifically useful accelerated quantum supercomputers by 2028: the complete convergence of AI with quantum hardware, systems, and algorithms.

The session will highlight challenges in deploying scalable, safety-critical edge AI and reveal how these global innovators are architecting their next-generation solutions using NVIDIA’s accelerated computing and end-to-end edge platform.

This session explores how you can build the right context-engineering architecture that helps LLMs deliver high-signal AI code reviews.