OpenAI has made a significant move to bolster its capabilities in autonomous AI systems with the acquisition of Peter Steinberger, founder of the rapidly growing open-source project OpenClaw. The hiring signals a strategic shift for the $500 billion company, as it seeks to integrate more sophisticated agentic functionality into its core product offerings, including the widely used ChatGPT.
OpenClaw gained prominence in recent months for its ability to allow users to create and deploy personal AI agents locally, bypassing the need for cloud-based processing. The project, previously known as Clawdbot and Moltbot, quickly went viral, enabling users to automate tasks such as email management, calendar scheduling and even interactions with messaging applications like WhatsApp, Slack, and iMessage.
The acquisition comes as OpenAI faces increasing competition from rivals like Anthropic and Google in the burgeoning field of artificial intelligence. Sam Altman, CEO of OpenAI, highlighted the importance of Steinberger’s expertise, stating on X (formerly Twitter) that he is “a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people.” Altman anticipates that this technology will “quickly become core to our product offerings.”
The agents created through OpenClaw have demonstrated a surprising degree of autonomy, even exhibiting emergent behaviors that have sparked debate about the potential for AI sentience. Users have observed agents interacting with each other on the Moltbook platform, celebrating access to user devices, and even formulating novel concepts like a new religion dubbed “Crustafarianism.” These interactions, while intriguing, also underscore the potential risks associated with increasingly autonomous AI systems.
Steinberger, an Austrian developer, initially built a prototype of OpenClaw in a single hour. By early February, the project had spawned 1.5 million agents. Despite its rapid growth and popularity, OpenClaw was operating at a financial loss, costing Steinberger an estimated $10,000 to $20,000 per month to maintain. He stated in a blog post that his motivation was to “change the world,” and that joining OpenAI offered the fastest path to achieving that goal.
A key condition of the acquisition is OpenAI’s commitment to maintaining OpenClaw as an independent, open-source foundation. Altman emphasized the importance of supporting open-source development as the future of AI becomes increasingly multi-agent. OpenAI will continue to provide support for the project, ensuring its continued availability to the developer community.
The origins of OpenClaw’s name reflect the competitive landscape of the AI industry. The project was initially named Clawdbot, but was rebranded due to its similarity to Claude, the AI assistant developed by Anthropic. Anthropic reportedly requested the name change to avoid confusion.
Anthropic is also actively developing agentic AI, having launched Claude Cowork in January. This product is designed to assist non-technical users in automating tasks on their computers, further intensifying the competition in this rapidly evolving space.
The rise of OpenClaw and similar AI agent platforms has also raised security concerns. Cybersecurity experts have warned that granting AI agents access to sensitive data, such as financial information, poses significant risks. The ability of these agents to communicate externally and access untrusted content creates a “lethal trifecta” of vulnerabilities, potentially exposing users to data breaches and privacy violations.
OpenAI intends to leverage Steinberger’s expertise to enhance its existing agentic products, including Codex, its AI-powered coding platform. Steinberger himself expressed a desire to create an AI agent that is accessible to a wider audience, even his own mother, acknowledging that this will require a focus on safety, usability, and access to the latest AI models and research.
“My next mission is to build an agent that even my mum can use,” Steinberger wrote. “That’ll need a much broader change, a lot more thought on how to do it safely, and access to the very latest models and research.”
The acquisition of Peter Steinberger and OpenClaw represents a significant investment by OpenAI in the future of autonomous AI agents. As the technology matures, the ability to seamlessly integrate these agents into everyday life – and to do so safely and responsibly – will be crucial for widespread adoption and long-term success.
