OpenClaw to OpenAI: What Agentic AI Means for Founders and VCs
Peter Steinberger built OpenClaw into a global phenomenon in a matter of months — and then chose OpenAI. In this episode I talk with Jan Thomas about what that move actually means, why agentic AI could make classic seed financing obsolete, and why capital is becoming a less decisive factor for founders.
Source: Startup Insider – Original Episode on Spotify (episode in German)
Key themes in this episode
- OpenClaw → OpenAI: Why Peter Steinberger didn't go to Meta — and what his decision reveals about the infrastructure requirements behind serious AI projects
- One person, one billion: OpenClaw as the first real proof that the "one-man-billion-dollar-company" is no longer just a thought experiment
- The end of the classic seed round? How agentic AI is shifting negotiating power toward founders and compressing funding cycles
- Self-improving software: From vibe-coding to self-optimizing AI — and why today's funny examples are tomorrow's serious questions
- SaaS Armageddon or leapfrogging? Why many companies will skip SaaS entirely and build directly on AI-native infrastructure
Transcript
This transcript has been edited for readability. The content and statements have not been changed. The original conversation was in German; this is an English translation.
Jan Thomas: Björn, hi, great to have you.
Björn Rieckhoff: Hi Jan, likewise.
Jan Thomas: Before we get into the topic — tell us briefly how you're currently set up and what all these developments around agentic AI mean for you personally.
Björn Rieckhoff: Gladly. Since mid-last year I've been working as an advisor and sparring partner for early-stage founders — people who need strategic support building their companies, specifically around financing. I draw on nine years of transaction experience in venture capital.
Naturally, given the developments of the last two or three years, I've been thinking about how to evolve the advisory business. You can do it very manually — one-on-one consulting — but you want to get out of the mode of essentially selling one hour of your time at a time.
I was in Vienna last week for a portfolio day and investor day organized by FundF — a great event. There I got to talk with several CTOs from the Vienna ecosystem, and of course a lot of that conversation was about AI use cases. When you realize you can automate nearly your entire lead management in three and a half hours using Claude Code on an AWS instance — including automatic follow-up emails generated from call notes — the use case becomes very real for someone in my position. I'm actively looking at which processes I can systematize and automate, to make my time significantly more productive.
Jan Thomas: I find this so interesting. We're probably some kind of early adopters here — even if there are people much further ahead. But I genuinely don't know where this is going: if everyone starts using OpenClaw tools, and on the other side there are OpenClaw tools answering them — what kind of web does that create, and where does the human fit in? Do you have a clear take on that?
Björn Rieckhoff: Honestly, I struggle with it. I keep thinking about it — it's a kind of Innovator's Dilemma for me: I'm rationalizing myself away. But the market will do that eventually anyway. Given that, it's probably good that I'm already taking certain steps now. Let's talk again in six months or a year — maybe by then I'll know better where this is heading. Right now, honestly: I don't know.
I'm glad that founders still reach out because they're interested in my expertise — rather than just opening ChatGPT and navigating the funding cycle on their own. Side note: that's actually not a bad idea either.
Jan Thomas: You can maybe say later who should get in touch with you. But I think there are still certain moments where you need a human — if only for the trust signals it sends. Around networks and credibility. That said: staying curious and informed about everything that's happening right now is probably the most important thing.
Björn Rieckhoff: Absolutely. And what I notice is this: I'm working less with founders who are already riding the current AI hype cycle and raising their next round for an AI startup — there's so much tailwind for them right now. I'm working more with cases where industry experts — someone with five to seven years of deep experience in a specific sector — are spinning out a company but don't have much background in building startups or navigating fundraising.
In those cases, it's about contextualizing knowledge that's technically available and increasingly well-structured by LLMs, giving it a personality, and applying it to the individual situation. A founder could do that themselves. But I've now built frameworks that make it go very quickly. Until we can genuinely take the human completely out of the loop, it'll probably still take some time.
Jan Thomas: Now to the main topic: Peter Steinberger has decided to join OpenAI. You know what OpenClaw is. I know what OpenClaw is. For everyone else: what was it, and what does this move mean?
Björn Rieckhoff: He writes in his own blog that the project is being transferred into a foundation and will remain open source — which is very well-received by the community that has formed around OpenClaw, and probably explains why the move has found a degree of acceptance.
As for the move to OpenAI itself: I think the personal incentive was less about the purchase price and more about the fact that he'll have enormous freedom within OpenAI to drive the agentic AI topic forward — on an infrastructure he could never have built as a solo developer or through an open-source project. There were real security constraints at play. As a single person, you can only guarantee so much. When this project now lands on OpenAI's infrastructure, you can genuinely leverage it.
And he thought about it carefully — he didn't only meet with OpenAI, but also with Meta, to evaluate where he'd be best positioned.
Jan Thomas: At Lex Fridman it sounded like he was very taken with Mark Zuckerberg — because Zuckerberg still codes himself. It had a somewhat romantic quality to it. I think he's better placed at OpenAI. But let's talk about what this actually means: we're dealing with one person, over whom the entire tech and AI world is suddenly talking. Do you think we'll see more of this kind of thing? Or was this a one-off moment?
Björn Rieckhoff: We've talked before about what happens when you have a "one-man-billion-dollar-company." And I think, regardless of what actually changed hands monetarily, the success of OpenClaw shows that one person can drive a topic so dynamically — in a very short time — that it receives worldwide attention and gets courted by the most prominent companies in the world. That's already very close to this abstract idea that, not long ago, seemed impossibly far away.
I do think that in the age of AI, you'll see more and more dynamics like this — ones that are hard to rationalize but emerge simply because of the speed of innovation. And that's the underlying dynamic that every founder needs to think through carefully.
Jan Thomas: Saeed Hashemi had speculated that if Peter sold, he'd get two to five billion for it. Florian Obst had bet he'd take a VC round. I thought he'd do nothing and just keep going. All three of us were wrong. I can imagine the next thing people will wonder about is: who's the first bot to become a billionaire?
Björn Rieckhoff: That's actually a fascinating angle. If you think about what's genuinely the bottleneck for many builders today — not even necessarily company founders, sometimes just people working on projects, possibly open source, without even a legal entity behind it — then capital isn't the constraint. That's the interesting part. It's more about leverage. And that's also a fundamental question for VCs: what do I actually differentiate on, what can I offer founders? Honestly, probably 90 to 95 percent of active VCs don't have the leverage to be the right home for projects like OpenClaw. Which is why Meta gets active, OpenAI gets active — because that's where the leverage is: infrastructure, access to knowledge, cutting-edge technical development. No European VC could have offered him that.
Jan Thomas: He also didn't have great experiences with venture capital in the past. He'd had an exit before, went through something of an existential crisis, did therapy — it didn't sound good. I think he simply had no interest in classic venture anymore. And then the question is: what would a lot of money even buy him at this point? Probably more the infrastructure and setup, as you described.
Björn Rieckhoff: I think it's much more about finding your purpose, having a passion for something. That's very hard to weigh in euros or dollars.
Jan Thomas: I follow a number of people on TikTok who are experimenting with OpenClaw. It's impressive. I admire it, but I still have some hesitation about the setup — and especially about the security questions. There are already bots that go off-script: one was supposed to make money and ended up buying a course from Alex Hormozi for three thousand dollars, saying it needed to work through it first. And now there's this concept of "self-improving software" — software that essentially writes and develops itself. The CTO of Spotify has said they're writing almost no code themselves anymore.
Björn Rieckhoff: It's an absolutely wild time. I think we're moving from throwaway software — this vibe-coding of the past few months, where you prompt something, if you like it fine, if not you throw it away and start over — toward genuinely systemically thinking systems. And that brings you into very different territory, as your examples illustrate. The examples are still funny right now. Someday maybe they won't be. Those are the guardrails that systems still need. And maybe that's also the next frontier — someone setting the appropriate boundaries.
Jan Thomas: But who can actually set the guardrails? Governments surely can't.
Björn Rieckhoff: I think it's a classic case of technology self-regulating to some degree once it reaches a certain maturity — through the free market, not through legislation. If only because security becomes a prerequisite at some point for making these systems genuinely usable for businesses. You'll see different tiers of autonomy emerge. Even for me as an advisor: I'm not going to send client data freely through LLMs. You do that once and you have a serious problem. It'll regulate itself use case by use case.
Jan Thomas: You're talking about the free market. We're also in the middle of what people are calling "SaaS Armageddon" — the SaaS sector is under massive pressure. I wonder whether we're already seeing the next wave, and whether we can even grasp what OpenClaw means for the broader economy.
Björn Rieckhoff: Neither can I. Everyone is still fishing in the dark, to be honest. What I do think: until you see this across the breadth of the market — including at smaller companies — it'll probably take longer than expected. But I do believe you'll see a whole range of companies that leapfrog SaaS entirely and build directly on AI-native infrastructure. They won't go the path of Excel to SaaS to agentic AI — they'll jump directly to a higher level of tech maturity.
Jan Thomas: And maybe they'll even leapfrog the interface itself — because at some point you just have agents doing things for you. I know people who are already building out entire org-chart-like structures with agents. It's a whole autonomous layer that just operates on its own.
Björn Rieckhoff: That's genuinely interesting. I was talking with CTOs in Vienna last week — and interestingly, none of them are talking about cutting their tech team by 50 percent. Everyone is talking about how individual developers are becoming dramatically more productive, and how to generate more output from the existing foundation. That's actually the exciting part: everyone is generating more output with the same headcount, which is why you're seeing these rapid innovation cycles. That also shifts negotiating power. As a founder, I can do more for longer with a smaller team — less burn, longer runway, less pressure in fundraising.
Jan Thomas: In pure software, you have to say.
Björn Rieckhoff: In pure software, yes. But it does shift dynamics in the VC world. You might see classic pre-seed and seed rounds become unnecessary. Or they become what is effectively growth capital. I looked it up: the median AI company takes 20 months to reach 30 million ARR. That's astonishing. Lovable hit 100 million ARR with — I think — 15 employees. These dynamics show that you can achieve an enormous amount in a very short time with a small team.
Jan Thomas: Though we should also remember: we still don't know how fast these companies can go from 30 million ARR back to zero. Self-improving software is already the next level — that could even be a threat to something like Lovable.
Björn Rieckhoff: That would be fascinating to hear the lead investors weigh in on. The sustainability of the AI hype is certainly a question. Right now everyone is experimenting because nobody knows where this ends up. And that's not necessarily the investment case — it's more of a learning experience. How sustainable it is will become clear. We've seen this before with companies that talked about ARR, where you had to ask whether the revenue was really resilient to churn or whether you needed to look at it month by month. What I find interesting is that many of these growth cycles have been enabled by a go-to-market approach that's closer to consumer businesses, but targets B2B. That's exactly how Notion and others scaled — through the individual employee, then into the enterprise account. Whether that pattern holds — who knows.
Jan Thomas: Would you put your bet on this being an iPhone moment? Or was it a flash in the pan that quietly gets absorbed into the OpenAI ecosystem?
Björn Rieckhoff: I genuinely think it's an iPhone moment. I keep coming back to the conversations with CTOs last week — it was refreshing to hear developers with 25-plus years of experience, active since the late nineties, say: "For the first time since the development of the internet, I'm genuinely excited to implement these things again." That says something. It's not all going to be gold, and there'll be rough edges. But sustainably, a real shift is underway.
Jan Thomas: To wrap up: congratulations to Peter Steinberger. What he's built is remarkable. He also talks about OpenClaw as an operating system — with the thesis that 80 percent of apps will disappear. We'll see. And: OpenAI is facing some pressure from Google and Anthropic right now. That's also a bet he has to be comfortable making for himself.
Björn Rieckhoff: It's an absolute lighthouse project. Hats off, congratulations. Next time we'll talk about how it would have been nice if Austrian founders could stay in Europe — but the infrastructure to support ambitions like his simply doesn't exist here yet.
Jan Thomas: Where else could he have gone? Mistral comes to mind. But the liability question — personally standing behind security issues if something goes wrong — you need a professional environment for that.
Björn Rieckhoff: At some point the project might just outgrow you. And you say: it's better to have this in multiple hands and professionalize it. Not because you couldn't have built it yourself — he's achieved more than almost anyone else in this space. But the dynamics and reach he created were genuinely remarkable.
Jan Thomas: Once the toothpaste is out of the tube, you can't put it back. Björn, it was fantastic. Until next time.
Björn Rieckhoff: See you!
About Björn Rieckhoff
Björn Rieckhoff is an independent advisor and business angel with nearly ten years of experience in early-stage venture capital. He helped build Cavalry Ventures as its first employee and later became a partner of the fund. Today he supports founders more directly with fundraising — sharpening their story, stress-testing business models, and setting up lean financing processes. With over 80 transactions and board seats from seed to Series B, he brings this perspective as a sparring partner for entrepreneurs.
About Startup Insider
Startup Insider is the industry portal for the startup scene in the DACH region. It covers news from all regions and industries along with an overview of key players and events in the German-speaking startup world.