TBPN - FULL INTERVIEW: Why I Spent $70M to buy AI.com

Episode Date: February 11, 2026

This is our full interview with Kris Marszalek, recorded live on TBPN.We discuss why he bought AI.com the moment it became available, the long-term vision for a consumer “chief of staff” personal assistant for billions of people, how he locked in a Super Bowl spot before the product existed and pulled the trigger just weeks before airing, what happened when the ad drove massive traffic and signups, and why he turned down a $500M+ offer for the domain right after closing because he’s going all-in on building the product.TBPN is a live tech talk show hosted by John Coogan and Jordi Hays, streaming weekdays from 11–2 PT on X and YouTube, with full episodes posted to podcast platforms immediately after.Described by The New York Times as “Silicon Valley’s newest obsession,” TBPN has recently featured Mark Zuckerberg, Sam Altman, Mark Cuban, and Satya Nadella.

Chris, good to meet you. What's happening? Nice to meet you guys. How are you? I love the background. The logo's looking great. Is that real or is that AI? It's totally made up.

I was like it feels pretty quick to have the actual design. Manufacturing. It takes time. Well, take us through the thesis for AI.com. Yeah, wait. Before we get into the Super Bowl, like, let's maybe rewind to maybe probably a year ago. You see a domain on the market. I think they had been trying to sell it for a while, I imagine, and you came in as a buyer.

But walk us through that whole journey. Yeah, I bought another domain, and the agent who was brokering this told me about AI.com being in a process of being sold, if you will. So I immediately recognized the importance of it and just jumped on it. Got on the phone the same day. got the deal done, we shook hands. There were some ups and down through the process, but we managed to get this done. You've had some good success buying iconic domains.

What did you pay for crypto.com? Have you ever disclosed that? I don't think we ever did, but we paid $12 million. I would argue that there was a more, difficult decision, if you will. We were a small company back then. $12,000 was about a third of our capital. Wow. And it was bang in the middle of the 2018 bear market. So people were discussing whether crypto is going to survive or not. Yeah, the person selling it to you was probably like,

probably like, this guy is an idiot. Of course, of course you ended up looking at the Staples Center. I'd drive by it all the time. Okay. So yeah, you, you see this domain is on the market. Do you have any idea what you wanted to do with it at the time of acquisition or do you just think that hey this AI thing's probably pretty important? Maybe I should own AI.com. Look, we were building products ever since Chad GPT launched and playing with consumer facing side but also internal tools. So constant experimentation. And the vision from day one was, you know, we want to build a consumer product.

We believe that you have 4 billion people having personal assistance that should play the role of kind of a chief of staff for your entire life with great context, being proactive, getting things done for you rather than just chatting. So the vision didn't change from day one. And then, yeah, I guess fast forward, how did you, how did you process, before we get into the kind of Super Bowl and that whole strategy, how did you, how did you process kind of the open claw launch? It sounds like you guys are integrating the, are you building? Some of those patterns? Yeah, at least some of those patterns. Are you kind of forking the project? Talk about that. So as we started building this product about mid 2025, it was never just, it just didn't click. You didn't have this.

You couldn't pass the uncanny Valley. And I think the pivotal moments where we started seeing this change is Opus 4.5 release. It started working much, much better. And we obviously saw CloudBod going live and adding it some element. of their architecture to it, I think it gets it done in terms of how it feels as a product. You just need to solve a whole litany of issues to make it consumer friendly. Like how do you set up with being technical, you know, the security issues around your data? Those are serious, serious issues when you want to bring something to the mass market.

So I think we're combining everything that we've built over the last, say, eight months into something that we can roll out to the end users and we are starting doing this tomorrow. Wow. Tomorrow. Okay. Before we get into the product and kind of more of the vision, let's fast forward again to the Super Bowl specifically. How did that all come together?

It felt like it was coming together quickly, but we know we ran a much, much smaller ad. We ran a regional ad. You do have to lock these things in ahead of time, but walk us through the process of kind of preparing and then experience. experiencing the Super Bowl Sunday? So I bought the domain in April. The deal closed, and we got the domain successfully. So I'm like, okay, we need to launch this, and it deserves a global stage.

And in May, we were one of the first companies to actually buy the spot. Oh, no way. Wow. At that time, we had just the domain and idea what we want to do with it, but the product doesn't exist. and you know I I know that we only have one shot to get this done correctly and I didn't want to release the product until I felt that it's there

for the end user you know these things you need to be able to develop an emotional connection with the product in order for this to be sticky and retentive so I only made the decision that we actually going to pull the trigger on this a couple of weeks ago.

And that's why the ad felt like it was quickly put together. It was quickly put together. Wait, so you bought the Super Bowl spot. But you always fall back to crypto.com if you wanted to? Yeah. What was the idea? Like, hey, if we don't run it for AI.com, we have the crypto.com ad ready. We'll just run that.

We have the crypto business. we have a prediction market's business. You know, we could do, there's always some level of optionality, right? But this is the moment to run an AI ad, as you guys have seen. Yeah, it is. And timing is really important in life, you know, skill, timing, and a lot combination of these things. It's a good reference. Okay, so you put the ad together in effectively two weeks, and then you run it,

and what happens then? because I think you got the attention from buying the world's most expensive domain ever. And you got the attention of like, hey, there's this new AI product I've never heard of with a crazy domain running a Super Bowl ad should pay attention to it. And then you got a whole other kind of amount of attention from people being like, wait, I just got an ad for AI.com. And I landed on the website and the website's down. So what kind of happened? did you you spent the 70 million on the domain, the 8 million on these spot, and then you didn't have enough to host it?

Or what happened? I'm assuming a lot of people got through, but certainly a lot of people got stuck as well. I think we are happy with the outcome. We had about 300,000 people signing up. Wow. Wow, let's go. Can we get the gong?

Can we get the gong for 300,000 signups? There's been a lot of big numbers. That's a big number for one day, for one day. But how do you even prepare for that amount of traffic? Like, how do you, what was going on in the war room? Guys, you know, we're on a platform that is used to spikes, right? And we've got a great DevOps team. And we've got all the stuff that you usually would expect, auto-scaling and whatnot.

So there were intermittent problems for some. people, but largely it held up. So I think fundamentally it's the name and the fact that there is a certain element of curiosity there and we designed it. It was a very simple call to action. Go and sign up. So I think it worked. Yeah. Talk to me about consumer AI.

ChatGPT is broken through Gemini and Google. They've been leveraging, you know, the Google platform and the network to on-end. board consumers, nano banana was a big moment. Sora and the meta vibes app sort of made a splash, but haven't been super sticky, but where do you see the gap in breaking through with consumers in a new way, or just doing what consumers already expect, but better, cheaper, faster?

Like, where is the consumer AI opportunity now that we're three years into the CHETGPT boom? I think fundamentally you're able to actually get stuff done, right now. So that's a big differentiating factor for the user experience. And we don't really know where this experiment is going to take us, taken us, given how the domain is resonating with people, you know, we can introduce social network elements to it. I think the fact that every single person on Earth is going to have an assistant of this sort

unlock new type of interactions and make our lives just better through serendipity, advice, staying on top of things and being proactive, really understanding us. So I'm pretty excited about the wandering aspect of it. We try to keep an open mind and not really be set on one thing. We now have 300,000 people waiting for us to give them the product. Are we going to very quickly iterate on it? I'm a huge believer in moving quickly and listening to actual customers, and we will see where the journey takes us.

I take a very long-term view. You know, what can we do with this in 10 years? I think it worked in the crypto space, and the opportunity here, the size is much, much bigger. Yeah, so much of what happening crypto was sort of permissionless, you know, bankless, this open source, these networks, anyone can set up a note. and part of this latest open claw clawed bot, you know, hype cycle is driven by the fact that you like you get a Mac Mini, it can talk to iMessage, it can talk to WhatsApp, it can talk to Telegram, it can go sort of wherever you go as a person and that feels unique because

maybe Open AI can't go over to WhatsApp because Mark Zuckerberg doesn't want to let him. And so I'm wondering about how you're thinking about the trade-off between certain things that are only possible with an open source AI system that is sort of acting as an imposter as a human versus you're a company. If you want to integrate with WhatsApp, you might have to give meta a call. So how do you think about delivering the vision of like a truly universal AI agent that can do things with the realities of the business community? There are a lot of business and UX challenges here. Yeah. So we'll have to resolve.

them one by one. Sure. And our view is, we want to stand on the side of the consumer and help them make these, you know, technical choices, make sure that their data is safe, make sure that they can do what they want to do without putting themselves at risk and solving these issues with access to data, with user experience. It cannot feel like a chore. You know, today you need to be really technical to get value out of it.

So there's plenty of work and it's difficult and that's part of the opportunity. Yeah, one of the, some unrequested feedback for me, I had kind of heard that AI.com was like potentially something like some type of, you know, leveraging some open claw technology. And then I got hit with the Google login. And I was like, I don't have time to read through kind of the full terms, the service, privacy policy, and really understand. So I would love to see, I mean, maybe there's plans for it, but I'd love to see just like being able to create an account on the platform to play around with it just because I was looking at my Gmail, which my life is on, my work email, which obviously has its own privacy concerns, all that stuff. You said you're rolling out the product tomorrow. What's the first thing that you want people to do with it? I think this is the big part of the product figuring out how do you onboard people to it and get them to do a.

couple of things so that you can see value very quickly and connect with it. I think today it's pretty hard to get the feeling for what it can do without really connecting your email and calendar. Yeah. We will see. I feel that there's going to be a lot of experimentation there and we will look for user feedback and truth in data in numbers of what really works and what doesn't. To a certain degree you need to gamify it until such point where where users are deep enough that it actually gives them the feeling of like, wow, this is special. This is different. Yeah.

Yeah, that gamification is so important. You see the Studio Ghibli moment where, you know, the latest images in chat Ghibit launches, and you don't even have to think. You just have something in your camera roll, and you're going to type Studio Ghibli, and you're going to get the value prop. And then a couple months later, you're still going to be going there for slide inspiration and stock photography and all the other things that you can do. But there's a killer app on day one that you come in and you and you get joy out of

Did you ever talk to the original owner of AI.com? I already sat on it for 30 years Because his initials, his first and last name, his first name starts with I so he bought the domain. I'm shocked that he held onto it for so long. You would think like IBM Watson in like 2010. Oh totally. We'll give you a million back for it. But he held on. Look, we spoke on the day when we closed the transaction because he had a, it was a little bit of a bidding war. He had a very serious bidder on the other side and it required connecting in order to get it done. That's good. Your deal, Scott.

And by the way, right after we closed the deal, I got the approach from the other side offering 500 million plus, not for $500 million. plus no for 500 million for the domain. I think I could have pushed it to a billion if I wanted to, but I didn't want to. So I think you guys need to understand I am pot committed. I love it. I love it. I love it. I love it.

I love that you're just thinking, you're viewing this. Like, obviously you're taking it very seriously, but you're also taking the approach of like it's very early days in, you know, what will be a long journey for the project, but also the industry. and you're just going to listen to your users and figure it out. But the conviction to turn down what would have been turning 70 million into 500 or a billion in 24 hours is admirable. Will you train a foundation model? I think I'm more focused on getting this to scale and getting the data fly will going so that we can deliver for our users.

Our users don't really care about which model runs in the background as long as the data. the job gets done and their data is safe. But once you get to a certain scale, who knows? Anything's on the table. I like it. Well, I'm sure you'll be back on. I'm excited to see this rollout and roll- I'm signing up tomorrow.

I might be using a dummy Google account, but I will be signing up and testing this out. I'm excited. And then I'll slowly forward myself data from my real account to give you a little bit more, a little bit more to see what it can do. But I'm excited for the launch tomorrow. And congratulations.

I mean, a fantastic career, but also this particular project, really, really fun execution and a wonderful story. So thank you for coming to share with us. You're basically the mayor of Los Angeles through the crypto.com arena. Do you come through much? You're a Lakers guy? I've been in D.C. last week, and then I stopped over in Silicon Valley.

I have never been to the arena. Never been to the arena. Wow. You got to come sometime. Catch a game. Maybe we should catch a game. Yeah. They also do monster truck rallies there. Underrated Crypto.com arena experience, especially if you have kids. This is a big monster truck rally. I'm a big monster truck guy. I don't really follow basketball that much, but I will be watching Gravedigger live from the crypto.com arena. Anyway, thank you so much for taking the time. Have a great to be you. We'll talk to you soon.