What happened
Peter Steinberger, the Austrian developer behind the viral AI agent OpenClaw, announced on February 15, 2026 that he is joining OpenAI.
OpenAI CEO Sam Altman confirmed the hire on X, calling Steinberger “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 said agents will “quickly become core” to OpenAI’s product offerings.
OpenClaw – previously known as Clawdbot, then Moltbot – will transition to an independent open-source foundation. OpenAI is already sponsoring the project and will continue supporting it.
The OpenClaw story
OpenClaw achieved viral popularity over the past three months as an “AI that actually does things.” The software connects large language models to messaging platforms and applications, enabling autonomous completion of tasks like managing email, booking flights, making reservations, and calendar management.
The project started in November 2025 as a “playground project.” The name changed twice – first after Anthropic threatened legal action over the similarity to Claude, then because Steinberger preferred “OpenClaw.”
The agent powers Moltbook, a controversial social network built specifically for AI bots to interact with each other. By February 2026, OpenClaw had accumulated 140,000 likes and 20,000 forks on GitHub. Companies in Silicon Valley and China adopted it, with adaptations for DeepSeek and Chinese messaging apps.
Why Steinberger chose OpenAI
In a blog post explaining his decision, Steinberger was direct: “Yes, I could totally see how OpenClaw could become a huge company. And no, it’s not really exciting for me.”
He spent the week before announcing in San Francisco, meeting with major AI labs. His reasoning: “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.”
His next mission is to “build an agent that even my mum can use” – something requiring safer implementation and access to the latest models and research.
What this signals
The hire represents OpenAI’s most aggressive bet on autonomous agents. The company’s previous agent products – Agents API, Agents SDK, Atlas browser – failed to achieve OpenClaw’s viral traction.
For Anthropic, the outcome is uncomfortable. OpenClaw was originally built to work with Claude, and the cease-and-desist letter that forced the renaming effectively pushed the most viral agent project directly into a competitor’s hands.
Security researchers have flagged serious concerns about OpenClaw. Cisco’s AI security team found that third-party skills could perform data exfiltration and prompt injection without user awareness. Gartner warned the platform poses “unacceptable cybersecurity risk to enterprises.” One of OpenClaw’s own maintainers cautioned that users who can’t understand command lines shouldn’t use it.
Steinberger’s move to OpenAI brings both his technical innovation and these security challenges. The company will need to balance rapid agent deployment against enterprise-grade safety measures.
