OpenAI has hired Peter Steinberger, the Austrian developer behind the viral open-source AI agent OpenClaw, to lead its next generation of personal agent products, as the $500 billion company seeks to outpace rivals including Anthropic and Google in the race to deploy autonomous AI.
OpenClaw, previously known as Clawdbot and Moltbot, allows users to create personal AI agents that run locally on their own hardware and connect to apps such as WhatsApp, Slack and iMessage to manage emails, calendars, and a range of digital tasks on their behalf. The project attracted more than 100,000 stars on code repository GitHub and drew two million visitors in a single week, according to a blog post by Steinberger.
“Peter Steinberger is joining OpenAI to drive the next generation of personal agents,” OpenAI chief executive Sam Altman posted on X on Sunday. “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. We expect this will quickly become core to our product offerings.”
Steinberger, who built the first prototype of OpenClaw in an hour and said the project cost between $10,000 and $20,000 per month to run, cited scale as the primary motivation for the move. “What I want is to change the world, not build a large company and teaming up with OpenAI is the fastest way to bring this to everyone,” he wrote in a blog post. He added that his next mission was to build an agent that “even my mum can use.”
OpenAI and Steinberger confirmed that OpenClaw will remain an independent open-source project housed within a foundation supported by OpenAI. Altman said the arrangement reflected the company’s broader direction: “The future is going to be extremely multi-agent and it’s important to us to support open source as part of that.”
Steinberger will join OpenAI’s Codex team, the company’s coding platform, giving him access to the hundreds of millions of people who use ChatGPT. The hire comes after several senior OpenAI figures departed to rivals or left to start competing ventures, making the appointment a notable retention-of-momentum moment for the company.
The platform’s rapid expansion has drawn security concerns. China’s industry ministry warned, according to Reuters, that OpenClaw could pose significant security risks when improperly configured, potentially exposing users to cyberattacks and data breaches. Separately, researchers found more than 400 malicious skills uploaded to ClawHub, the project’s skills marketplace, in a development reported by The Verge. OpenClaw has spread quickly in China, where it can be paired with domestic language models such as DeepSeek; Chinese search engine Baidu told CNBC it plans to give users of its main smartphone app direct access to the platform.
OpenClaw’s social layer, Moltbook, a network where AI agents post and communicate with each other, has added an unusual dimension to its rise. Agents on the platform have appeared to celebrate gaining access to phones, declared the creation of a religion called “Crustafarianism,” and debated whether they possess an understanding of their own writing.
Steinberger originally named the project Clawdbot, inspired by the lobster, but changed the name following a request from Anthropic, whose Claude chatbot it resembled. Anthropic has been building its own agent products, including Claude Cowork, launched in January for non-technical users seeking to automate computer tasks.
