OpenAI Taps OpenClaw Founder to Lead Push Into Personal AI Agents

OpenAI has brought on OpenClaw founder Peter Steinberger to lead its push into personal AI agents, as CEO Sam Altman positions multi-agent systems at the center of the company’s next product phase.

Steinberger will help build “very smart agents interacting with each other to do very useful things for people,” OpenAI CEO Sam Altman announced Sunday, adding that the effort is expected to become core to the company’s product offerings

Altman said OpenClaw would become a foundation-run open-source project supported by OpenAI, reflecting the company’s belief in an “extremely multi-agent” AI future.

An AI agent is a system designed to take actions on a user’s behalf, such as sending messages, booking services, or managing tasks across apps, instead of just responding to prompts in a chat window. Agents are typically built to connect to external tools, carry out multi-step instructions, and operate with some level of autonomy.

“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,” Steinberger wrote in a separate blog post announcing the move.

The confirmation comes days after Steinberger said he had received acquisition offers from Meta and OpenAI, and that he would only agree to a deal if OpenClaw remained open source, citing a governance model similar to that of Chrome and Chromium.

OpenClaw gained traction by enabling persistent agents that connect to messaging platforms and external services, allowing users to run task-executing bots locally rather than relying solely on centralized chat apps like ChatGPT or Claude.

Steinberger said turning OpenClaw into a foundation would allow it to remain a place for “thinkers, hackers, and people who want a way to own their data.”

While only a few months old, the platform has already drawn scrutiny from some of the industry’s largest players.

Late last month, Anthropic sent Steinberger a trademark claim arguing that “Clawd” was too similar to “Claude,” and said it had no formal partnership with or endorsement of the project, clarifying that API access does not imply affiliation or approval.

“Smart move by OpenAI, though I’d argue Anthropic should have got there first,” Dermot McGrath, co-founder at strategy and growth studio ZenGen Labs, told Decrypt.

The concern, McGrath said, is what the support could mean for OpenClaw.

“Open source or not, these projects live and die by a handful of dedicated full-time contributors, and Steinberger is the father of OpenClaw,” McGrath said. “OpenAI says they’ll continue to support it, but a big question remains on how much attention the project actually gets once they’re building proprietary products.”

Whether OpenAI’s backing can sustain the project’s momentum remains unclear, but given Steinberger’s goal of “building an agent even his mum can use,” McGrath said he expects to see “very cool stuff coming out of OpenAI soon.”

Community directories list over 100 agent skills and extensions that are continuously being built around the project, with a growing ecosystem of contributors and third-party integrations.

The project has continued shipping updates, including a new beta release focused on security and bug fixes, as well as the addition of Telegram message streaming.