NVIDIA GTC 2025

We’ll highlight our latest designs for DGX clusters, including both compute and data center configurations showcasing an example with DGX SuperPOD. We’ll discuss the architectural and design principles underlying next-gen air- and liquid-cooled deployments. We’ll also examine the network fabric, interconnect solutions, and storage options that enable maximum performance, and how these components integrate with the software stack for generative AI applications.

Often a speaker who highlights new announcements, NVIDIA’s Ian Buck will highlight cutting-edge innovations that are revolutionizing high-performance computing while prioritizing energy conservation. Discover how these technologies are enabling industries and nations to achieve new insights. Join us to explore how supercomputing is not only advancing capabilities, but also paving the way for a greener, more efficient future in tech.

Often a speaker who highlights new announcements, NVIDIA’s Charlie Boyle will lead this talk and panel. We’ll delve into the critical elements required to operationalize Gen AI at scale — including proven infrastructure with a comprehensive software stack that enables high reliability and uptime, high visibility into cluster and job status, and simplified workload management to maximize developer productivity.

Wondering how to incorporate new technology into your current DGX SuperPOD? Trying to understand exactly how usage and management of a B200 DGX SuperPOD compares to one with DGX GB200? Trying to understand something that seems simple, but somehow isn’t, on an NVIDIA DGX? This is the session for you! Connect with solution architects, product managers, and other experts in an open discussion to help get your questions answered by the people who know.

Learn about the unique architecture of these superchips and their versatile form factors that simplify data center deployment. We’ll address such questions as how the 900 GB/s NVLINK Chip-to-Chip interconnect enables the CPU and GPU to access all system-allocated memory, how you can seamlessly migrate your x86-based workloads to the Arm-based Grace CPU, and the performance gains you can expect to unlock for your applications.

The CUDA platform is the foundation of the GPU computing ecosystem. At this engineering-focused talk by one of the architects of CUDA, you’ll learn what’s new and what’s coming next for both CUDA and GPU computing as a whole.

In this session, we will dive into the GB200 NVL72 architecture and programming model, highlighting its inference performance on state-of-the-art LLM models. We will also explore optimizations techniques that enable the 72 Blackwell GPUs to work together through the NVIDIA NVLINK, functioning as one giant GPU.

In this session, leading AI cloud data centers will come together to share our experiences and insights from building and deploying colossal systems. We’ll delve into the unique challenges of networking at such a massive scale, and how we overcame them. Attendees will gain a deep understanding of the lessons learned in scaling infrastructure to support the next generation of AI, from the complexities of connecting thousands of GPUs to the innovations required to maintain performance and reliability at such an unprecedented scale.

This talk brings together Yann LeCun, a pioneer in deep learning and the chief AI scientist at Meta, and Bill Dally, a leading computer architect and chief scientist at NVIDIA, to explore the future of AI models, hardware accelerators, and the evolving computational landscape. The discussion will cover: 1.The next breakthroughs in deep learning and AI architectures 2. How hardware innovation drives AI efficiency and scalability 3. Challenges in training large-scale models and real-time AI inference

In this session, moderated by NVIDIA’s vice president of enterprise AI and automation, AI experts will share their insights and experiences in developing and deploying generative AI and agentic solutions. They’ll discuss the opportunities AI agents bring, real-world enterprise use cases, strategies for building user trust, and the challenges of scaling these technologies across organizations.