Mark Zuckerberg and Sam Altman are both trying to buy the same open-source project — and the guy who built it can’t decide. OpenClaw, the viral AI agent that exploded onto the scene in late January 2026, has become the most fought-over piece of open-source software since Android. Here’s what’s happening and why it matters.
What Is OpenClaw?
OpenClaw is an open-source autonomous AI agent that lives on your device and actually does things for you. Not “here’s a suggestion” — it manages your email, organizes your calendar, handles files, and executes tasks across apps, all through WhatsApp, Telegram, or Slack. You message it like a human assistant, and it works in the background.
What makes it different from ChatGPT or Claude is the word autonomous. OpenClaw doesn’t wait for you to ask a question. It monitors, plans, and acts on its own. It supports multiple AI models — Claude, GPT-5, Grok, Meta’s Llama — and runs locally on your machine, which means your data stays with you.
The project has racked up over 180,000 GitHub stars and millions of installs since going viral. It evolved through several names — Clawdbot, then Moltbot, then OpenClaw — each iteration more capable than the last.
The Man Behind the Lobster
Peter Steinberger isn’t a first-time founder having a lucky moment. The Vienna-based developer previously built PSPDFKit, a PDF framework used by companies worldwide, and sold it to Insight Partners for over $100 million. After that exit, he spent three years exploring — and tried 43 different projects before OpenClaw clicked.
Now he’s sitting on the hottest open-source project in AI, losing $10,000 to $20,000 a month running it, and fielding calls from two of the most powerful CEOs in tech.
Zuckerberg vs. Altman: The Bidding War
Both Meta and OpenAI have put concrete offers on the table. Steinberger has spoken openly about his interactions with both leaders, including a detailed conversation on the Lex Fridman podcast.
Mark Zuckerberg reached out via WhatsApp — no calendar invite, no assistant scheduling the call. When Steinberger picked up, Zuckerberg asked for 10 minutes because he “needed to finish coding.” Their first real conversation included a 10-minute argument about whether Claude Code or Codex was better. Zuckerberg later called Steinberger “eccentric, but brilliant.”
What impressed Steinberger most was that Zuckerberg was actually using the product — tinkering with it, building with it. Not just evaluating a deck.
Sam Altman took a different approach. Steinberger described their conversations as “really cool” and called Altman “very thoughtful, brilliant.” OpenAI is already contributing tokens to the project and had constructive discussions about the project’s direction. But Steinberger noted that OpenAI’s engagement felt less personally invested in the product itself compared to Zuckerberg’s hands-on approach.
And there’s a third player lurking: Steinberger revealed at a Vienna event that he’s also spoken with Microsoft CEO Satya Nadella.
The Open-Source Condition
Here’s where it gets interesting. Steinberger has one non-negotiable condition: OpenClaw must remain open source, regardless of who acquires it. He’s pointed to the Chrome/Chromium model as his template — a commercial product built on top of an open-source core.
This is a massive constraint for any acquirer. Meta might be better positioned here given its track record with open-source AI (Llama, PyTorch). OpenAI, despite the name, has moved steadily toward closed-source models. Whether Altman would agree to keep the core open is one of the key questions.
Steinberger has also made clear that money isn’t his primary motivation — he already had a nine-figure exit. “The beauty is if it doesn’t work out, I can just do my own thing again,” he said. He could also start his own company with the funding available to him, though he’s said that path appeals to him less because it would pull him away from actual development.
Why This Matters Beyond the Deal
The OpenClaw acquisition battle is really a proxy war for the future of AI agents. McKinsey projects that AI agent applications could generate $2 to $4 trillion in annual value. Industry analysts predict agentic AI could automate up to 50% of knowledge work by 2028.
Meta has already placed a big bet in this space, acquiring competing AI agent platform Manus for approximately $2 billion in January 2026. Adding OpenClaw would give Meta a dominant position in both open-source and commercial AI agents.
But OpenClaw’s explosive growth has also exposed serious risks. Security researchers found over 42,900 OpenClaw agents exposed due to misconfigurations, with 15,200 vulnerable to hackers. When you give an AI agent access to your entire computer — email, files, passwords — the stakes of a security breach are existential.
The Bottom Line
As of today, no deal has been announced. Steinberger has 3,000 pull requests to process and a project to run. But the fact that the CEOs of Meta, OpenAI, and Microsoft are all personally courting a single open-source developer in Vienna tells you everything about how high the stakes are in the AI agent race.
The question isn’t just who will acquire OpenClaw. It’s whether the open-source AI agent model can survive contact with Big Tech — or whether it’ll be absorbed, repackaged, and locked down like so many projects before it.
Sources: Trending Topics · OfficeChai · OpenTools · H2S Media · CNBC
Leave a Reply