Some love it, some hate it. But nobody can ignore it. OpenClaw, the open source platform for building personal AI agents, has had one of the most extraordinary rises in the history of software. A few weeks ago I set it up on a server and built my own AI assistant, August. It has been quietly running in the background ever since, handling tasks and learning as it goes. That felt like enough of a story on its own. Then Jensen Huang took the stage at Nvidia’s GTC conference this week and name-dropped it as one of the most important software releases ever. So let’s unpack what just happened and why it matters.
From a side project to a global phenomenon
OpenClaw started as a relatively small open source project built by developer Peter Steinberger. It is a platform that lets you build your own AI agent, connect it to tools and services, and have it run tasks on your behalf, continuously, in the background. Think of it less like a chatbot and more like a digital employee that actually does things rather than just answering questions.
The adoption numbers are hard to believe. Linux, the open source operating system that powers most of the internet, took roughly 30 years to reach the download milestones it holds today. OpenClaw matched that in three weeks. It is now the fastest growing open source project in the history of software by both stars and downloads on GitHub. That is not a typo.
What Jensen Huang actually said
At GTC 2026, Nvidia’s annual developer conference held in San Jose this week, Huang did not just mention OpenClaw in passing. He built a significant portion of his keynote around it. He compared it directly to Linux, Kubernetes and HTML, three technologies that quietly became the invisible backbone of the modern internet.
“Every company in the world today needs to have an OpenClaw strategy,” he said. “This is as big of a deal as HTML. As big of a deal as Linux.”
He also framed it in a way that anyone can understand. Mac and Windows are the operating systems for your personal computer. According to Huang, OpenClaw is the operating system for your personal AI.
That is a bold claim. But given the adoption curve, it is hard to argue with the direction of travel.
What Nvidia actually launched
Image Credit: The News Stack
The headline announcement was NemoClaw, Nvidia’s own platform built on top of OpenClaw. It adds a security and privacy layer that the original OpenClaw currently lacks, wrapping agents inside a controlled sandbox so they cannot access data or systems outside of defined boundaries. It was developed in collaboration with Steinberger himself, who was recently hired by OpenAI but continues to maintain OpenClaw as an open source project.
NemoClaw installs via a single command and can run locally on Nvidia RTX PCs, dedicated workstations or Nvidia’s own DGX supercomputers. It is hardware agnostic, meaning it does not require Nvidia chips to run, which signals that Nvidia is prioritising reach over hardware lock-in.
The honest bit: this is not for you yet
Here is where I have to be straight with you. If you are a regular person who heard Jensen Huang’s speech and wanted to set up NemoClaw this weekend, you would hit a wall fairly quickly. The setup still requires command line access, an Nvidia API key and a reasonable level of comfort with developer tools. The enterprise tier adds infrastructure costs on top of that.
NemoClaw is currently aimed at developers and large organisations. It is in early alpha. The rough edges are real and Nvidia has said as much publicly.
The original OpenClaw remains the more accessible starting point for individuals. That is what I used to build August, and it runs well on modest hardware with a manageable setup process.
Why this moment still matters
None of that changes the significance of what happened this week. When the CEO of the world’s most valuable technology company dedicates keynote time to an open source project and calls it foundational infrastructure, the industry listens. Enterprises start building strategies around it. Developers start contributing to it. Managed, user-friendly versions start appearing.
We have seen this pattern before with Linux, with Kubernetes, with every technology that started rough and became essential. The difference this time is the pace. What took those technologies years is happening in weeks.
AI agents are not a distant future concept anymore. They are already here, already running, and now they have the attention and resources of one of the most powerful companies in tech behind them.
I built mine a few weeks ago on a quiet Tuesday evening. It did not feel like history at the time. It is starting to feel a bit more like it now.