NVIDIA at AI Infra Summit

The conference features eight content stages and tracks that span the entire AI technology stack, including Compute, Data Movement, Physical AI, Data and Models, and AI Data Center.

• View the full breakdown of tracks and session topics on the official AI Infra Summit 2026 site.
   

Unlike general AI events that may focus on software applications, the AI Infra Summit is uniquely dedicated to the infrastructure layer of the five-layer AI cake, focusing on the crucial intersection of infrastructure with energy, chips, models and applications to drive AI performance, scalability, and operational efficiency.

• Understand the specific focus of the infrastructure layer in the AI Infra Summit FAQ section.
   

Yes, everyone will have access to NVIDIA recorded sessions post-event, allowing them to review key technical insights, keynote presentations, and deep-dive technical talks.

• Verify access to on-demand content included with your ticket on the AI Infra Summit Eventbrite page.
• Session recordings by NVIDIA speakers will be published on this web page within 72 hours of the live event.

The NVIDIA Vera CPU is a high-performance processor designed as part of NVIDIA extreme co-designed AI infrastructure platform, meant to work as a standalone CPU in tandem with GPU acceleration for optimal AI factory performance.

• Discover the technical details and specifications for the Vera CPU on the NVIDIA Vera CPU product page and the Vera CPU whitepaper.
• Learn how the single-threaded NVIDIA Vera CPU is designed to accelerate agentic AI workloads with maximum performance at scale.

Groq 3 LPX is the interactive AI inference accelerator, an extension of the Vera Rubin platform. It extends the inference performance of Vera Rubin NVL72 with ultrafast token generation for highly responsive agentic systems. 

• Discover how NVIDIA Groq 3 LPX accelerates interactive, long-context agentic AI inference.
• Learn more about the NVIDIA Vera Rubin platform.
   

NVLink Fusion is a technology and IP for building semi-custom, rack-scale AI infrastructure. It lets hyperscalers and AI-native companies integrate custom XPUs, ASICs, or CPUs with NVIDIA’s high-bandwidth, low-latency NVLink scale-up fabric and broader AI infrastructure stack. 

• Explore the architectural benefits of the interconnect at the NVIDIA NVLink Fusion product page.
• Read about its impact on the broader world-class AI factory landscape.
   

AI factories are purpose-built, large-scale infrastructure environments that transform energy and data into intelligence for training, inference, and agentic AI workloads. They are a central focus of AI Infra Summit because delivering AI at scale requires the compute, networking, storage, software, power, cooling, and operations layers to be designed and managed as one integrated system. 

• Learn more about NVIDIA extreme co-design and how this approach is used to build efficient, flexible, and high performant AI factories in the AI Infra Summit keynote with Ian Buck.
• Explore how the NVIDIA DSX platform helps organizations design, simulate, and operate high-performance, energy-efficient AI factories.