Tech giant Nvidia, the world’s most valuable company and a child of the AI boom, is betting its future on the rise of AI agents.
The company on Monday announced a number of software and hardware updates to accelerate the development of AI agents, or AI assistants, that can perform tasks on behalf of users. Among the most important announcements is a set of tools for AI helpers based on OpenClaw, the hot agent platform that’s been making waves in Silicon Valley in recent weeks. Nvidia also announced new compute racks designed to shift the main focus of its strategy away from graphics processing units and power agents.
Wearing Nvidia’s signature black leather jacket, CEO Jensen Huang made rapid-fire announcements at the chipmaker’s annual GTC conference in San Jose. The conference attracts tens of thousands of attendees and has been dubbed the “Super Bowl” of AI.
Nvidia’s announcement is significant because so many major companies rely on Nvidia’s systems to train and power their AI services. This means new products from semiconductor giants often mirror technologies from companies across the AI industry.
Nvidia has announced a software tool to help companies create AI agents. This includes models and blueprints for creating custom dedicated assistants. We also start a set of resources for creating agents on OpenClaw that add privacy and security controls. This is extremely important considering that the popular agent has caused concern among cybersecurity experts.
Nvidia said its resources will help OpenClaw agents access systems and files without compromising security or privacy. Huang said he has been working directly with OpenClaw developer Peter Steinberger, who was recently hired by OpenAI.
Huang said OpenClaw is an “operating system for personal AI” and likened it to the importance of Mac and Windows operating systems.
“OpenClaw is number one. It’s the most popular open source project in human history, and we did it in just a few weeks,” said Huang.
Nvidia also announced updates to its new computing platform Vera Rubin. The platform consists of seven chips and is currently in full production, the company said. This includes a new central computing rack comprised of central processing units (CPUs) rather than Nvidia’s renowned graphics processing units (GPUs). CPUs are ideal for running the types of computing processes needed to power AI agents.
The company is also integrating a non-Nvidia processor into its systems, namely a new high-speed “language processing unit” (LPU) from American AI company Groq. Nvidia signed a $20 billion deal with Groq in November.
Unlike AI chatbots, which respond to questions and prompts, AI agents can autonomously complete tasks like building a website, writing a marketing pitch, or sending an email. AI agents are currently Nvidia’s biggest focus area, driven primarily by the popularity of OpenClaw and Anthropic’s Claude Code and Cowork agents.
“Every company in the world today needs to have an OpenClaw strategy, an agent system strategy. This is the new computer,” Huang said. “This is as big a problem as HTML, and it’s as big a problem as Linux.”
Nvidia is also looking to future-proof its technology in other ways. The company is launching a space module for Vera Rubin with the aim of bringing cutting-edge technology to data centers in space. Interest in the field is growing among big tech companies vying for real estate to build data centers. OpenAI CEO Sam Altman and xAI and Tesla CEO Elon Musk both talked about using space to power data centers and energy-hungry AI systems.
“NVIDIA is now focused on the future of networking in this new world of AI, beyond just computing,” Wedbush analyst Dan Ives said ahead of Monday’s NVIDIA conference.
In Monday’s speech, Huang sought to convey that the hype surrounding AI and Nvidia may continue, touting a vision of an AI-transformed future where demand for the company’s chips will grow infinitely.
Huang said computing demand “continues to increase,” adding that he expects Nvidia’s revenue to “finally” reach $1 trillion by 2027.
“There’s a reason for that,” Huang said. “We have arrived at this fundamental inflection point, an inflection point in reasoning, where AI can do productive work.”
