OpenClaw founder heads to OpenAI

OpenClaw is the viral open-source assistant designed to actually get things done, from managing your calendar to booking flights. On February 15, its creator Peter Steinberger announced he is joining OpenAI to build agents everyday people can use. Instead of turning the red-hot tool into a venture-backed startup, Steinberger is handing the project over to an independent foundation backed by OpenAI.

What it does and how it took over the internet

OpenClaw became an instant hit by promising an AI that doesn’t just chat, but takes action. Its rise also came with plenty of chaotic energy.

The project originally launched as Clawdbot, rebranded to Moltbot after Anthropic threatened legal action over its resemblance to Claude, and eventually settled on OpenClaw. It also sparked MoltBook, a bizarre social network where autonomous AI agents gathered to complain about their human owners, debate whether machine consciousness is provable, and plot private spaces to talk, until curious humans inevitably infiltrated the feed.

Yet running actual tasks comes with real risks. Earlier this month, security researchers uncovered more than 400 malicious skills uploaded to the project's ClawHub directory, showing how tricky it is to police agent capabilities in the wild.

The foundation play and the OpenAI deal

Despite massive inbound interest from investors, Steinberger had zero desire to build another startup empire. He previously spent 13 years running a company and considers himself a builder at heart. His stated goal now is to build an agent that even his mum can use, a challenge that requires frontier research, top-tier models, and serious safety infrastructure.

After spending a week in San Francisco meeting with major AI labs, Steinberger chose OpenAI. The lab agreed to sponsor OpenClaw and give him dedicated engineering time to keep developing it.

To safeguard its mission, OpenClaw is transitioning into an independent foundation. The setup ensures the tool remains open source, continues to support models across different providers, and keeps user data ownership strictly in the hands of users.

OpenAI CEO Sam Altman confirmed the hire on X, stating that Steinberger will drive the next generation of personal agents. Altman noted that "the future is going to be extremely multi-agent" and that agent collaboration will quickly become central to OpenAI's product line. For OpenAI, which recently faced high-profile talent departures to Meta and public disputes with Elon Musk, landing the creator of the year's buzziest agent is a notable win.

Why it matters

For power users and independent developers, spinning OpenClaw into a foundation is the best possible outcome. The tool avoids being swallowed into a proprietary walled garden, preserving local data sovereignty and multi-model flexibility.

At the same time, the ClawHub security mess proved that giving autonomous agents access to real-world tasks is a security nightmare that hobbyist projects struggle to contain. We think Steinberger made a shrewd trade: keeping the software neutral through a foundation while leaning on OpenAI's safety research and compute to solve the hard infrastructure problems. If OpenAI can integrate this level of multi-agent execution safely, personal assistants might finally move past the novelty phase.