Nvidia turns the AI factory into a product, and agents into the workload
For most of the past half-decade, the narrative around Nvidia Corp. has largely centered on the graphics processing unit. The questions have been who has allocation, how much faster the next generation is and whether anyone can close the gap. At GTC Berlin early today, Chief Executive Jensen Huang spent surprisingly little of his keynote […] The post Nvidia turns the AI factory into a product,…
Over the past five years, Nvidia Corp. has been primarily associated with graphics processing units. However, during the GTC Berlin keynote, Chief Executive Jensen Huang shifted the focus towards the challenges of deploying AI, emphasizing speed, power, security, and the ability to utilize enterprise data. Huang emphasized that AI has evolved from a tool to becoming part of manufacturing infrastructure.
He introduced DSX POD, a reference design for AI factories, the OpenShell Reference Architecture for safe agent deployment, and expanded its Agent Toolkit with a new family of structured-data models called Kumo. Nvidia also deepened its involvement in telecom with Nokia and European operators. Huang used a back-of-the-envelope calculation to illustrate the economics of AI infrastructure, stating that a one-gigawatt AI factory costs around $50 billion and generates $40 billion in annual revenue.
Delays in deployment can result in significant revenue loss, amounting to $10 billion. Nvidia's DSX POD aims to address the challenge of hardware idleness by embedding the manufacturing process within the factory itself, thus reducing deployment time from six months to 14 days. The DSX platform, which includes features like warm-water cooling, power optimization, and automated load management, is designed to improve efficiency and reduce resource consumption.
The keynote also highlighted the increasing importance of agents in AI systems, with Huang outlining the five components of an agent: model, harness, sandbox, knowledge, and tools. The focus on agents underscores the need for robust security measures to prevent autonomous systems from taking unintended actions. Nvidia's ecosystem partners are actively aligning with its solutions, reflecting a growing industry trend towards treating data centers as partners in grid operations.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
