SCBC 2026 | Arjun Sethi Keynote w/ Jan VanEck on Crypto’s Future

Yeah, it's—well, I don't. That's probably the first part of the answer. Look, I think there's just so much stuff that's happening every single day from all walks of life all over the world, and it's happening at a faster and faster pace. And so, I think maybe I just wake up with a pretty healthy paranoia of what I need to spend my time on. It's probably not too dissimilar to when I was younger—there was just enough busy work to make sure that you had an understanding of what your workflow would look like, even if it's not for work, it's just to learn. And yeah, honestly, I don't really have a framework. It's just that I'll hear something, I want to learn about it, and I want to make sure I'm not behind in any capacity. And so if my kids are doing something and I don't understand it, then I'm definitely behind. So my kids probably use more workflow tools than I do and I just try to learn from them in some way.

Well, I'll take a step back. So at Kraken today, in the parent company Payward, we barely use emails. It's mostly for external—Okay, thanks for that dig—but it's all internal, right? We have shared Slacks with other institutions, and most of the communication is asynchronous, and we're also a fully remote org. So the workflow was actually tailored to use as many tools as possible. But I think lately—and this is all new over the last two to three months—people have started talking about what's your AI workflow and then what are the agents you use on top of it. And I think that there's just so many new use cases of what you can do actually daily. Like I now have a team which is ridiculous, but when I first started it, I just had agents for each personality type. So how do I respond to my parents? How do I respond to my kids? How do I respond to my personal dating life? How do I respond to each and every company that I'm a part of? It's actually just separate orchestrations and then it comes to me at different levels. That's today and then next week it's probably going to be something different. So sometimes I start from scratch to re-architect these pieces.

If I was to go into the architecture of how we built it, it'd be long, but the way you can think about it is that each and every personality type that I have—I sound like a sociopath—but each one has their own virtual private servers and that's where everything is set up today. And then I have orchestration to be able to communicate with them, and then they communicate with each other before they communicate with me today. It's kind of buggy, but it works pretty well.

I started failing, then obviously I asked for help along the way.

Yeah, I think there's this concept that I believe in which is regression to the mean. You get used to a certain workflow and you just don't want to change it. And the more you get used to it, the more you regress depending on what you're doing. And so what we build is financial infrastructure. You can say tokenization, you can say you're building out agents, but you're not really doing it unless it's happening bottoms-up and top-down at the same time. And a lot of the times I think that you see executives or people that are not what I call close to the field just saying, "Use AI," but then they can't spell it, and they say, "Use machine learning," they can't spell it. They say, "Protocols," again they can't spell it either. And so that's kind of like the big fear, not just for me, but any of the executives that are at the company. And so a lot of it is, "Okay, well, what can you do yourself? And how do you actually inject that type of DNA as quickly as possible?" And it's not just for—AI is kind of the flavor of the month that people are talking about, but what are new workflow tools that people are using? We do KYC, AML, there's our compliance team or legal team—what are they doing? So you kind of have to be close to all of these pieces in order to run some of these companies and institutions.

It's dependent on workflow. The engineering teams are probably going to use tokens per second more. Our data team actually is probably the highest usage of it because they're building out systems for the rest of the org. And then it's going to be lower for lower frequency items that might not need as much interpretation. And then how do we use frontier models to our own models internally? And then how are we using it for each piece? We're just consistently iterating. But the reason we're doing that is I'm extremely paranoid that we will fail if we don't do that as an organization and we don't understand how these tools work. Because I also feel like two or five years down the line, these are also just distribution channels. So I think of Claude as another operating system. I think of OpenAI's systems as other operating systems. Who knows what's going to be moving forward? And that's how people are ingesting their time, their workflow. And so that's different than traditional internet, different than mobile. And so my customers are going to be somewhere, and I need to understand how to reorient our infrastructure, our tools, our experience for our customers, because we have small to medium retail, we've got mid-market, we've got enterprise. Then we have institutions, we've got hedge funds, market makers and takers. They're all in different locations all over the world. And so what is someone like—how do you explain an agent to someone in Argentina? I'm not sure you can, but you can probably explain to them that you've got some sort of system or a bot or a friend running on your behalf to do something, which might be just saving your money.

I don't actually make bets or do trades.

Yeah, there's this concept from some sci-fi books, which is the future's already here, it's just not evenly distributed. And I think that concept applies. If you're really rich today, you can build all these things. You can have a wealth manager. You can have someone trade on your behalf. And I think it's more about how do you bring that education, those tool sets and systems to everybody. And so I think the heartbeat of crypto—at least why I got into it—was that there was this bifurcation after the global financial crisis for the folks that were alive during that time, that it's not that the system was rigged, but the system could be unfair to a certain level of folks. And that's how the Bitcoin white paper started, and people started talking about how do you free your capital and how are you empowered to do your own. And so that started off as what I call a part of the self-directed movement. But it's not about self-directed movement. It's more like, what are the tools that you can use to make your own decisions without having all these intermediaries that extract along the way? And I think in fintech and banking and a lot of these companies, what they do is they extract along the way. And a lot of what our goal was—or at least community-wise—was how do you bring that access that everyone else has, like better yields, better trading strategies? How do you bring what Citadel and Jane Street do to the average person if they choose to want to make those decisions? And so that's actually how Payward was born, is that a lot of those systems we built for what we call professional traders, and then those professional traders helped us build for consumers and vice versa.

Sure. So there's two co-founders at Payward, and that's the parent company of Kraken—Jesse Powell and Thanh Lu. So I met Jesse probably at one of these things or one of these events. I don't think he really cared about me because I was on the venture side or moving into the venture side at that time. But I had invested into Payward through an SPV, just to gain access. I'd invested in a bunch of other exchanges as well, but Kraken was one of the first of them.

I think it was 2013 timeframe, 2014.

Yeah, it's like spray and pray at the time. Strategy that worked for a little bit. But later on, when I started my own firm in 2018, I spent more time with the other members in the ecosystem. I came back to Kraken, and I think it was more that they had a missionary-style attitude towards what they were building. Sometimes to their detriment, which is like here's how customers actually receive experience. But they kind of stuck to it. And then over time, I had started building and incubating my own projects in banking and some protocols in DeFi as well. And they started scaling at a faster pace, and so he kind of asked me to come on board and said, "Hey, I was already on the board over a period of time since I started doubling down on the company, if I just come on board full-time." So it took me about a year to decide to come on board full-time.

So when we started Tribe, it was based off of a concept of how do we identify what you know versus who you know. So a lot of the world is built off of a network. You and I are friends and maybe if I decide to fail upward, I'll come to you and ask you for seed or Series A money, and then you might be inclined because you're like, "Arjan's smart enough, so let me give him some capital." And so unfortunately, actually a large portion of early-stage investing is very intuition-oriented, and in many cases is unfair. And so my team and I—because we were operators, we came from leveraging data to make decisions. And when you make those decisions for a consumer product or a social network product or a mobile product, you are enabling it with cohorts of information. So we said, "Okay, well can we use that same type of framework, build our own framework for private companies where we can get access to anyone in the world capital?" Because we think that can be built anywhere, and especially with software making it easier to start, and now even from vibe coding to production coding it's even faster. Could you give that capital to somebody at 50k, 250k, a million? It ended up being the sweet spot was what I call enterprise value between let's call it 15 million to about 250 was where we could identify at least some signals.

No, just enterprise value.

So we have this concept where we said, okay, how do we qualify product market fit for someone who engages with a product and how do they distribute that product? How does it grow? Regardless if that's enterprise software, if it's a consumer product. And that's what we did our whole life. And so that's how we started that company and that was the value proposition, which is that we would either identify these companies, we'd identify markets, or we'd start them ourselves. And most of the capital was principal capital or at least at the time 20% of it. So we said, we're going to do it our way and we don't want to be beholden to institutions to say we should do it your way. Or them telling us you don't fit in this bucket list. So that's how we started it. And so it wasn't that we were an incubation house, it wasn't that we're just a VC firm, it was more I just want to do whatever I want to do with our time and our capital, and that's how we concentrated it into a very small set of companies, and then we started some, and then in some cases we ended up running some.

Look, I think maybe I've just had so many failures throughout life. Choosing who you work with—forget governance for a second—it's just really important. Are they on the same page as you? Do they have the same value sets as you? When times get tough, are they going to leave your side or do you leave their side? That part of the matchmaking ends up being the most important. They don't need to be the smartest people. Just more that they're driven towards a certain set of values or virtues that you care about. And my first set of companies that I started with ended up being resume building, where someone was like, "Hey, this person's really good. She should join your board or they should just be an executive." And they always turned out to fail. Or in some cases, I failed because of that dynamic. That could be just specific to me—that I'm terrible in how I run my companies. But that was the early stages of it—which is, okay, well, who are the folks that should be on your board? Is it like your early-stage investor, or is it people that really matter around the table? So every company that I'm a part of where I'm on the board, a lot of what I tell the founders is, hey, you don't need to have a general from the defense company. You can have someone who's been in the field of battle who actually buys these things. Or if it's an AI company, who are the folks that are actually on the ground building the orchestration layer? Not necessarily that they used to work at OpenAI or Google. In many cases we had the student at Boston University that was just really good at building new systems, and he's on the board of one of our foundations, and I don't know, he's like 26 years old. I think it's easy to teach people how to be a good board member, governance, and what the structure looks like, and then ideally they have some value sets. But I think it's very hard to teach people functionally what you're really good at and how do they add that dynamic to a boardroom.

I've always been averse to the check mark, which is like, "Hey, this person served on an audit committee." Well, you want someone who served on an audit committee that really understands your company and what you're trying to build versus someone who served—I'm just thinking Wells Fargo because that was my first bank account—but it's like someone who was at Wells Fargo in the '70s, why should they be on your board? They don't really know anything about what you're doing moving forward.