Venture with Grace
00:02:08 · Grace Gong, Jay Zhao
[00:02:08] S0: What was your company doing before? [00:02:11] S1: That was after an e commerce company back in the days and that was before Shopify, before 3PL became more popular. And I have to hand code a lot of the things myself in building the site. And what's even interesting, I don't think any founder should do that nowadays, is that I have to go to the warehouse and then move things around. But in terms of inventory stuff, So there was a lot of kind of just grinding going on. But at the same time, I understand how hard it is to start a company, raise capital, because after three years, I just burn out. Like, it's it's really tough to start a business, during the voting stage. But what it taught me is that when we do what we do at Leonis right now, we're almost like a startup, but it does give us a lot of empathy trying to connect with them and understand what they go through because it's never a straight line up to the right. But we want to be there, know, at the Stuttgart and Dallas.
00:03:07 · Grace Gong, Jay Zhao
[00:03:07] S0: For sure. Want to talk about, since I feel like when you were building your founder investor networks in the Bay Area, like how did you kind of like that plugged into like the Airbnb of the world? And for like in the AI era, I feel like obviously researchers like a really huge like, you know, group of people that like, you know, it's like a super crucial to like the AI space. Maybe we'll talk about like the network building, AI researcher land as well. Yeah. Since you guys had like, I believe as an LP at the very early age and how you spotted the first AI wave and how you created the relationship with these really important people in [00:03:54] S1: Yeah, we're kind of lucky in some ways because Steve Jobs is [00:03:57] S0: Are you guys still in touch with him though? He's very public these days.
00:04:02 · Jay Zhao
[00:04:02] S1: I we are. We're in touch with all the LPs through our investment activities, also the AGM LP meetings and stuff. But obviously everybody is busy nowadays, especially over the past three years. The reason I mentioned the context is that back in 2020, AI was just emerging. OpenAI was a non profit at the time. So how to get started with Leonis? It was somewhat coincidental because at the time I didn't want to join a platform VC firm anymore, and I didn't want to start a VC firm because... Just because people were starting VC firms at the time. I think the origin of Leonis, you talked to my partner Jenny, what we see eye to eye is how can we start and build an entry franchise that's highly research driven and conviction driven. Because a lot of times VC will make investments and in hindsight we see that trend early on but it's very much of a lagging indicator to those investment cases. What we try to do is that because we have a lot of LPs we're early engineers and researchers at labs. We're trying to understand how can we identify these companies one year or two years ahead of the AI capability curve. I think if we do well, we can not only just avoid the hype and high valuation and give a return to our LPs, but at the same time really help the founders. Often times founders start something, it's less obvious, it should be less obvious, like when they start at C companies. Those type of founders, they typically want to have VCs who are thoughtful and understand what they do. And that's really how we build out the platform and also people around the Leonis investment team side.
00:05:46 · Grace Gong
[00:05:46] S0: I guess when you are identifying the early trends, how do you think like what is something that's like, know, one year ahead of the trends compared to like something that's like super early? Because like I think that the blow up like of ChatGPT like was like kind of unpredictable. It was not technically like planned it for these like AI companies. So I'm curious, how do you think about like, what is something that's like gonna be like the huge hype like the next year? And when you guys were first building these like research circles and is there any... I know like Jenny was like early employee and when you guys are selecting the people that are you're connecting or working with, what were some early insights that you have gained that shaped the thesis of the fund?
00:06:39 · Jay Zhao
[00:06:39] S1: We do a lot of research posting on our website and on Substack and try to share our opinion and our viewpoints about the tech stack. But at the same time, we call our belief like the world view. We believe certain things will happen over the next two or three years and we want to be loud about our worldview and so that we can attract the right founders to it. So for example right now we're wrapping up our fund three, we're gonna close fund three by the end of this year and heading to next year for the next vintage of 2027 to 2030, there are three things that we spend a lot of time thinking about. One one single world view is that we believe in the future, people will have more AI agents working for the person and for companies, both on the person side and the personal side and the company side. And the implication of that is that the whole infrastructure built out for AI agents need to be rebuilt and that has huge implications for agent payment, e commerce side has huge implications for like for example we invest in this company called Agent Phone. They are almost like a Twilio for agents. So it definitely needs to build from the ground up in terms of agent identity, persistent identity and compliance and so on and so forth. So we think that's a very interesting opportunity. You don't see the volume as big right now, but we think one or two years later it's going to build up. It's going to be explosive growth. The other area that we spent a lot of time thinking about, and frankly, Seeds Fund two, two years ago, when we invested in a company like Bigger AI, that was an investment that we did in 2024. And at the time, I think a lot of VCs and companies were mostly talking about chatbot and copilot at the time, two years ago. And we put out a thesis and research post to our OPs. And we think that as AI becomes more capable in terms of executing long time horizon tasks, the agent should be able to complete tasks, action, like end to end, right? So I think we put out our worldview to our LPs, to our operator network, and when we met the team at Victor, Frederic and Peter, they were at Meta, you know, working on two or three paper at the time. So we spent a lot of time thinking about can we align from that world view and that's how we are the right companies.
00:09:09 · Jay Zhao, Grace Gong
So that was consistent to be a thesis for All Funds three, but we package in a way that's easier for frankly for our OP to understand. It's called system of actions. So basically the idea is that the last generation of companies like Anaplan, was a system of record. Salesforce is a system of record, monday.com is a system of record. And they want to be a system of record because they want to be at the premium place to charge this kind of SaaS premium pricing predictable annual revenue. But going forward, I think those companies face big trouble because just a lot more vulnerable. The next phase of companies, software companies, really have to be the system of action where agents can perform action in a reliable way. That's where we really spent a lot of time in the other piece of token economics, agent name wrap, and all it's kind of a stirrup from both thinkings. [00:10:06] S0: I wanna talk about, I guess when you guys were, I wanna shout out to the audience. Hi, LinkedIn user. We'll address your question in a minute. Hi, John. So LinkedIn user is asking how do companies position themselves and their value are ahead of the curve with new legislation, which impact research AI companies in healthcare, for example, when it is not something most VCs are paying attention to.