AI Infra Summit: NVIDIA Vera Rubin and DSX Platform Advancements Showcase Energy Efficiencies of Optimizing Tokens Per Watt for AI Factories
Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA, Tuesday spoke on AI factory efficiency at the AI Infra Summit, the Santa Clara Convention Center event that has morphed into a Coachella of infrastructure tech. Before a packed audience — with more than 8,000 attendees this year, up from 3,500 last year — […]
At the AI Infra Summit held at the Santa Clara Convention Center, NVIDIA vice president Ian Buck discussed advancements in AI infrastructure efficiency. With over 8,000 attendees, the event showcased new collaborations and results from NVIDIA partners. Amazon's Annapurna Labs and d-Matrix are working on custom high-bandwidth memory and Fusion platforms, respectively.
NVIDIA and its partners presented Emerald AI's flexible-load program, which achieved a 23% improvement in performance per watt using NVIDIA DSX MaxLPS. Pinterest utilizes the NVIDIA Blackwell platform and Dynamo inference software for conversational AI in visual discovery. NVIDIA's full-stack AI factory platform includes Vera Rubin systems, Dynamo inference software, NeMo libraries, and networking solutions.
The focus has shifted from peak performance to validated agentic tokens per megawatt. NVIDIA's DSX MaxLPS delivers up to 1.4x more tokens per megawatt, while NVLink unites large-scale accelerated computing. This results in AI infrastructure that generates more tokens, improves efficiency, and maximizes value from every megawatt of power.
Written by urgent.news from NVIDIA Blog's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
