By Isabelle Bousquette
SAN JOSE, Calif. -- The energy was electric inside a tent at Nvidia's annual GTC event, where developers and tech executives clamored around high-top tables to experiment with OpenClaw.
The open-source AI agent orchestration system, which went viral in recent weeks, became a key topic of Nvidia's event. On Monday, CEO Jensen Huang announced NemoClaw, a software development tool kit aimed at making these OpenClaw agents enterprise-ready.
"The importance is profound. It's the most popular open-source project in the history of humanity," he said.
At Nvidia's on-site Build-a-Claw experience, technologists were building "claws" for everything from road trip planning to job applications, research and content marketing. Nvidia executives were standing by to help.
Claws are autonomous agents and can plan and execute tasks on their own, and, critically, spin up their own subagents to tackle specialized tasks, access files and themselves delegate tasks to other subagents.
They represent a big leap beyond question-and-answer-style AI chatbots as well as recent iterations of AI agents, which typically have narrow use cases and run for a set amount of time -- although claws also come with a new set of security concerns. For claws to work as a true personal assistant, they need access to all of a user's data.
So what are people using them for today?
Daniel Buchmueller, co-founder and chief technology officer of defense logistics startup Gallatin AI, said he was looking at building a claw that could ingest all the daily newsletters he received, and then rewrite the information as a single newsletter, customized to his personal preferences, possibly delivered to his inbox at market close every day.
This type of asynchronous, always-on capability separates claws from conventional AI chatbots where work happens in real time through back-and-forth exchanges. "Your computer doesn't need to sleep at night. It can do work for you," he said about the advantage of claws.
But that conceptual shift can be challenging for even the savviest of technologists.
"Everybody is at step zero," said Luke Wignall, Nvidia's director of technical product marketing. When users approach claws with a simple prompt that can be answered with a single response, he's always pushing them to think about the next step.
For example, Wignall relayed a story of a college student who was testing a use case where a claw recommended jobs she could apply for. He suggested she go further, optimizing her résumé based on certain positions, then connecting the claw to her LinkedIn account.
"Using claws is not a one and done," said Wignall, adding that they can learn about a user and their preferences over time, gauge the effectiveness of what they are bringing back to the user and then use that to improve.
The proactive nature of claws is part of what makes them so compelling to Hein Kolk, who works at the Dutch research institute TNO. Kolk said he has used individual agents for research tasks in the past, but often finds he is spending so much time creating and operating them that it isn't always a timesaving.
Now, he is looking at a potential claw that he could spin up once and that would regularly scan the latest research, papers and literature and report back highlights to him once a week.
Ash Bao, head of marketing at AI-native consulting firm Gruve, said she has more than 10 ideas cooking with her marketing team on how to make use of claws.
One involves building a claw to call colleagues when they are at conferences to interview them about trending and important themes. Typically, she would make those calls herself, but outsourcing it to a claw means the colleague can take the call whenever they are free and she can look over the insights when she starts her workday.
Still, Bao, Buchmueller and many others at the Build-a-Claw experience were reluctant to start using claws in actual business use cases, in part because of the security vulnerabilities associated with giving claws their logins, passwords and proprietary data files.
That's part of what Nvidia's open-source contribution, NemoClaw, will help tackle. Kari Briski, vice president of generative AI software for enterprise at Nvidia, said it is a good idea to keep these claws locked down on separate devices while the technology is so new.
ServiceNow incorporated elements of NemoClaw architecture into AI specialist agents that it released internally last month, according to Amit Zavery, president, chief operating officer and chief product officer. Some customers are already using the agents, which also employ governance and orchestration from ServiceNow's AI Control Tower product. A general release is planned around May, according to Zavery. The agents have been achieving 80% to 90% resolution rates. "We will see a lot more of those things," he said.
Bao said she is still formulating a plan with her CTO around how to move forward with usage, but agreed that use cases like marketing, which rely more on publicly available data across the web and social media, will be among the first worth exploring.
Nader Khalil, the director of developer tech at Nvidia who was central in the building of NemoClaw, said another critical application could be better training of autonomous-driving vehicles. But on the personal side, he also set up a claw that monitors cheap last-minute flights to visit his family in San Diego.
Has he bought any flights on it yet? "I haven't yet," he said from the front of the bustling Build-a-Claw tent. He was too busy preparing for this conference.
Write to Isabelle Bousquette at isabelle.bousquette@wsj.com
(END) Dow Jones Newswires
March 18, 2026 16:54 ET (20:54 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments