Venture with Grace
00:26:53 · Jay Zhao, Grace Gong
[00:26:53] S1: by That's the off the true. I think the part that's super interesting is that out of the five scores, it's actually two and three historically produced the best venture outcome. [00:27:07] S0: What is two and three?
00:27:07 · Grace Gong, Jay Zhao
[00:27:07] S0: What is two and three? [00:27:10] S1: Like two or Oh, three and four is actually the best. One and two is tough. One and two meaning that from our AI score ranking, One and two meetings that it's not like it's impossible for them to become successful, it's just that the headwind is too strong. We would still take the meeting, we have a lot of pitch meetings right now, so we actually just use AI, what we call alpha, alpha rank the meeting priorities, so we're just kind of trying to figure out where we can spend the time most effectively. So three and four star founders are actually the best founders. So maybe translating into some qualitative characteristics, These are the founders, either repeat founders, past startup experience, but not a huge asset, but they learn the lesson. They learn what it takes to start companies. Maybe it's a second round, third round of doing that. I think that's additive to the startup component. Then two or three founders are, you know, so one thing that's kind of interesting is that within the AI era there's a lot of maybe this is a hot take, but there's a lot of professors who start companies. According to alpha analysis, actually it's tough. It's tough to have a professor as a founder. I don't know why. I think probably if I have to break it down, it's like you just need the full time commitment into the creation phase of the company. And I think professors are oftentimes engaged with a lot of activities on their own.
00:27:10 · Jay Zhao, Grace Gong
[00:27:10] S1: Like two or Oh, three and four is actually the best. One and two is tough. One and two meaning that from our AI score ranking, One and two meetings that it's not like it's impossible for them to become successful, it's just that the headwind is too strong. We would still take the meeting, we have a lot of pitch meetings right now, so we actually just use AI, what we call alpha, alpha rank the meeting priorities, so we're just kind of trying to figure out where we can spend the time most effectively. So three and four star founders are actually the best founders. So maybe translating into some qualitative characteristics, These are the founders, either repeat founders, past startup experience, but not a huge asset, but they learn the lesson. They learn what it takes to start companies. Maybe it's a second round, third round of doing that. I think that's additive to the startup component. Then two or three founders are, you know, so one thing that's kind of interesting is that within the AI era there's a lot of maybe this is a hot take, but there's a lot of professors who start companies. According to alpha analysis, actually it's tough. It's tough to have a professor as a founder. I don't know why. I think probably if I have to break it down, it's like you just need the full time commitment into the creation phase of the company. And I think professors are oftentimes engaged with a lot of activities on their own. [00:28:43] S0: I would say they're kind of not really in touch with current the realistic commercialized equipment.
00:28:43 · Grace Gong, Jay Zhao
[00:28:43] S0: I would say they're kind of not really in touch with current the realistic commercialized equipment. [00:28:51] S1: Yeah, that's right. [00:28:52] S0: I would assume they are probably perfect if they are doing... Like, who does a lab that's fit into her [00:29:01] S1: strength. [00:29:01] S0: And I got to acquire because it's kind of like, talent acquisition for these companies. Yeah. I would like to finish.
00:29:14 · Jay Zhao
[00:29:14] S1: Yeah, it's definitely talent acquisition. It's at the lab level. It's the talent acquisition, it's definitely a talent density play. So I think in that lens it makes sense. But for application stuff and for software infrastructure stuff, algorithm suggests that basically the three or four star founders, they tend to have the best shot in this delivery mature scale return, both for them and for investors. So to answer your question, the reason that matters is because oftentimes we got invited into hot deals at high valuation. It's the same lens that we use when it comes to growth stage companies. We actually have a multi stage fund that we launched last year, and the goal is to translate our early stage research and role view into selectively pick a few companies at growth stage to invest and back. And the idea is that if you invest at high valuation or at a growth stage, then you really have to believe two things. One is even though the valuation is high, which in and by itself is not a bad thing, but you really have to believe that there is a 3x or 10x potential growth from there, right? And then that's one component we have to get to a conclusion. And then the second part, which is the engine behind that is the moat question matters a lot. In the hype environment, will get concentrated to pattern recognized. Founders will get concentrated to narrative driven companies, but nobody really digs underneath in terms of... I think part of the reason is that AI is very new, so a lot of business models are emerging as we see it. But I think there are some cases where we can't get to a conviction level that the moat actually will compound over time, therefore we can still get to 10x return. So that's how we evaluate the high valuation deals and those type of companies. Obviously it's a very high level abstract answer, but because the fund is being stealth, we have some positions in some growth stage companies we can talk about maybe offline later. When
00:31:29 · Grace Gong
[00:31:29] S0: you're thinking about going back to the portfolio construction piece, so obviously the fund is significantly bigger. I guess do you guys just make the check size bigger when you're doing the investment? How do you think about follow on as For