I sat down with my Monster this morning and put on the @openclaw Clawcast Episode 2 with @hrudolph , and somewhere in the start of it Peter Steinberger (@steipetefinally put a match to the "OpenAI secretly owns and controls OpenClaw" story that's been rotting the conversation for months.Official Openclaw Youtube
Friends Of The Crustacean Discord
If you've spent any time around AI agents since the winter, you know the one I mean.
It started in February 2026. Peter, the guy in the lobster shirt who kicked off the biggest open-source agent boom any of us had seen, said he was joining OpenAI to push harder on personal agents.
Sam Altman posted his support. OpenClaw would move to an independent nonprofit foundation. OpenAI would sponsor tokens and help cover costs. A few other companies would pitch in too.
And the timeline lost its mind.
Headlines shouted "OpenAI Buys OpenClaw." Influencers ran their victory laps or their doom threads. Memes everywhere. The tidy little version stuck fast: Peter sold out, the project's a puppet now, and every bug or rough edge or volunteer-made screen is proof the corporation swallowed it whole.
Months later, each release brought it roaring back. The app shipped and the reply was instant: "worst slop I've seen, obviously because OpenAI owns them."
Then @hrudolph Hannes Rudolph, about a week into the community manager job, said the thing that everyone wanted to ask out loud. Sitting in front of 400-plus live listeners on the new Clawcast, he asked the thing nobody else would.
Word for word:
"I've only started here just over a week ago, and one thing that came to light right away… people said 'Oh, OpenClaw's not stable'… But honestly, (jokingly) "I thought we were owned by OpenAI." That's what everybody in those threads is saying. I mean, aren't we?"
Peter, who'd just moved continents, wrestled US paperwork, and kept dozens of open-source projects breathing, didn't blink.
His answer ought to be pinned somewhere:
"No?… honestly by now I don't know why… It almost feels like malicious that people drive this narrative." "I went to great lengths to like make sure OpenAI does not own it, does not even have rights to say anything about the direction of what we do.
The goal was always that OpenClaw is Switzerland, and we wanna make it great for any model." "The support that OpenAI gives us is they give us tokens… it's kinda hard to hire people on tokens alone, so I'm also glad that we have other companies that support us with money so that we… can hire some people."
He pointed out they picked the harder nonprofit foundation path on purpose, Pushed out iOS and Android apps, sat through Apple and Google review to do it by the book, and put every release through hundreds of test machines with no venture money behind them.
He got to the app backlash that lit this latest fire too:
"We had apps since January… most people just didn't know… instead of putting our efforts together, everyone was doing their own little side thing… I was like, people, we just have to get the app out… We don't have resources like a VC-funded company."
And it worked, he said. It pulled contributors back together, bugs got fixed overnight, and it gave people the local-first thing that actually counts:
"your messages don't have to go through WhatsApp or Telegram… you also have an option to fully end-to-end control who sees your messages."
The people who fed this thing? Let them speak for themselves:
All of it standing against Peter's own tweets
https://x.com/steipete/status/2036268009854296257
("OpenClaw is owned by the independent OpenClaw Foundation, which I lead"), Sam's post, the foundation docs, and now a whole podcast.
Here's what I think after listening:
Does anyone actually grasp how much the support for OpenClaw mattered?
This wasn't "another agent tool."
It changed how agents work at every level, from side experiments to something people touch every day. It figured out multi-channel orchestration across 16-plus platforms, local-first privacy, session memory that holds up under load, agents talking to agents, and the whole composable stack of parts.
It made the industry solve the ugly stuff first, rate limits, security surfaces, twenty million config combos, the anti-bot walls on messaging platforms, so everyone downstream could move quicker.
It spun up a fresh corner of enterprise growth: orchestration layers, monitoring, security wrappers, custom integrations, and all of it feeds every LLM and inference provider directly. Smarter agents mean far more compute demand. Good for builders, good for users, good for the whole stack.
I think OpenAI offering Peter a seat was one of the sharpest talent calls going. (also, I'm no expert)
You don't let a founder with that much hands-on agent mileage wander off. Landing someone like him is top of the league. Sweat equity plays.
Backing the project with tokens and chipping in on dev costs - why on earth wouldn't they? It feeds both sides. The tide that lifts every boat, and this tide was already floating the whole agent economy.
But jumping from "job offer plus sponsorship plus founder works there" to "they own it, they call every shot, Peter's compromised, the whole thing is captured"?
Peter has given the world zero sign that OpenClaw bends toward OpenAI.
Once and for all: can we grow up and drop the bullshit?
The support for OpenClaw was one of the most important accelerations of the entire agent era.
It laid the road.
-Gabe
