Venture with Grace
00:50:02 · Jay Zhao
[00:50:02] S1: Yeah, so 1,000,000,000 then bought. MaintainX obviously exited outcome. Spectro Cloud, they just reached the 1,000,000,000 mark, and we'll have one company that we'll be raising that will be at that mark. So the hit rate for Fund one, we think it's pretty good in terms of unicorn to the amount of companies that we have. And the portfolio health wise, for one, we only have one write off out of the pre CNBC companies. And most of the companies, it's kind of in the middle, they have either maintained at the same valuation or they have a slight markup, one to three x in the middle. And then the three x in the bottom is breakout companies. That's the one that really drives the majority of return of the Venture fund. I think a lot of LPs were surprised by the model write off. And I think it kind of rooted to some of the founder philosophy that we have. I think one way to remedy micro hype cycle is to make sure you have enough runway as long as you can before you find a proper market fit. I think one of the mistakes that I have seen for companies and founders who raise huge round in the high top of the market is that you spend money too freely. And then what happens is that you never find product market fit. So you have to wind down the business, return half capital to LPs, to investors, or you just never go anywhere, you shut down at the end. There was multiple cases like during the 2021 era to 2023. So we tend to like the founders are super scrappy and super, know, very capital about the capital spend. One example, so to answer yes, it could be a problem. So we backed the company out of 1.2. We invested 500 ks in the company, have a meaningful ownership of the company. Very smart team, very young team as well. And one day I got the investor update and they were like, the monthly burn rate is $7,000 I was like, is there an error? Usually it's $7,000 They're like, No, it's $7,000 and it turned out that a three person team, they still live in the Airbnb place, share the same living room and stuff and only have two bedrooms, three people, and just with a triple low burn rate, but their top line AR is like $1,500,000 Wait,
00:52:35 · Grace Gong, Jay Zhao
[00:52:35] S0: can I also put in a small check-in-law, [00:52:37] S1: this company is doing? [00:52:38] S0: How do you invent them to partner with? [00:52:39] S1: Yeah, we can talk about it. These companies, I think the unique challenge for them is that I actually want them to go out and raise more capital. [00:52:50] S0: Just like play more aggressively about [00:52:52] S1: Play more aggressively. But I think one of the potentially tricky things is that when they get to next month, they will get to a profitable and then grow the business more sustainably with their own kind of block making mechanism. But then you can do that for a period of time. But if you do that too long, then you don't become a scalable business. So that's another topic, another area that maybe require a different time and stuff. But yeah, so what we try to do is that we don't try to do too end of the extreme, we try to be in the middle, so you have the best of both worlds. Yeah, so even for the breakout companies, they always manage their spend pretty carefully, and when they hit a certain milestone, the Series A term sheet, you often times got preempted and then you know, I guess in a lot of ways, you may never have to worry about fundraising. Once you get the fundamental, the top line level is right, you know, the capital market is quite efficient right now. [00:53:47] S0: I love that. So I want to be mindful of the time, maybe we do one real category question and then one final one thing is like, what is some AI trend that you feel like underhyped and then what is something that you feel overhyped?
00:53:47 · Grace Gong, Jay Zhao
[00:53:47] S0: I love that. So I want to be mindful of the time, maybe we do one real category question and then one final one thing is like, what is some AI trend that you feel like underhyped and then what is something that you feel overhyped? [00:54:01] S1: So think the underhyped part is still with agent infrastructure. I mean, I'm pretty bullish about the space, but it's very dry. [00:54:12] S0: Is there any particular subsector that you felt like it's kind of interesting to you? Because I feel like nowadays, every company is an AAM front company. [00:54:20] S1: From [00:54:22] S0: that to data center, literally anything and everything is infra.
00:54:30 · Jay Zhao
[00:54:30] S1: I would say more like a genetic infrastructure, meaning that you have to build out a certain infrastructure for agents. The company agent phone, they are like Twilio for agents. There's actually a lot of plumbing stuff to do on the back end to make sure agents can use a phone number in a consistent way and also if you are a company or developers you can acquire that number in the full compliance way. So it's almost like early days of Stripe, right? Stripe is a simple API there was a lot of plumbing and integration they needed to do on the back end for the new e commerce economy back then. Same with Marqeta, the fintech API company, they had to ride a wave with the on demand economy and Marqeta becomes successful from the near death experience at the time, it was because they catch that wave and power DoorDash into the car and the breeze at the time. So think we're at... So anyways, think, obviously we spend a lot of time thinking about the space, but when these companies talk to other VCs, I feel like their experience is not that great, just in terms of finding that capital market fit. So I think that space is under hyped. But to answer your question, we're not thinking about infrastructure like data center or the chips or semiconductor. We're not actually expert in that. So that part, some investors will play it out. But it's mostly how can we view the new infrastructure for agents to operate in a consistent way. One example is that if you ask agent to go through your CRM like Salesforce, some of this stuff is actually not good, very bad. Adios looks slightly better, but the data format is actually structured in a way for humans to consume. For agents, it's actually very hard for them to go through whole things. Either you go through too many tokens or agents become lazy, they will report back to you and say, Hey, I looked through a larger record, but actually they did not. So that required from a fundamental level, that's the agent infrastructure that needs to be rebuilt. So that's on the hype one. The over hype one, I would say physical AI is pretty overhyped as a category. But there is pocket opportunities there as well.
00:56:40 · Grace Gong, Jay Zhao
[00:56:40] S0: Yeah, I really like, so last week I also chatted with Boris and Kevin's company Bedrock Robotics. I feel like those are like, so those are like a lot of the companies that I feel like coming out of that space is that actually become huge is like people with industrial experience. I feel like physical AI is a space that like you literally need like industrial experience to actually understand like what's going on. By the way hi everybody. I really like everyone's like, yeah. So, yeah. [00:57:13] S1: A different comment. Yeah, [00:57:15] S0: yeah, it's like people's random comments. I appreciate everyone's yeah, like I feel like, yeah, like there's like a certain aspect of like, I feel like AI infra you need like the AI needed founders like YC, the high school dropout, blah blah blah, but like I feel like for physical AI it's like one of the space that you really need to actually have Yeah, work like on the capital, it's like very intensive and it's like also life or death, you don't want your car to be out on road and then five people die from it. Yeah, sure. Okay, so I wanna end with a quick one minute feral for you. Sure. What's your favorite book? [00:58:10] S1: Yeah, that's a good question. [00:58:13] S0: No one reads anything these days. [00:58:15] S1: Yeah, no, there is one book that has impacted a lot of my thinking in this channel. I need to find the book name, it's called Twilight. It's actually a history book. [00:58:25] S0: Oh, okay, so not Twilight the Empire.
00:58:25 · Grace Gong, Jay Zhao
[00:58:25] S0: Oh, okay, so not Twilight the Empire. [00:58:28] S1: The Empire. It's actually the top of the China Opium world at the time. It describes how I wrote that history through a history book like everybody else, right? The writer is a really good writer and he actually went through the whole geopolitics dynamics and how UK was viewing China at the time is not as straightforward as the history of world portrayed. There was a lot of internal debate within China, within The UK at the time. Just how the whole thing unfolded is quite mind blowing to me because it kind of talks about human psychology and how history unfolds in certain ways that's sometimes completely out of random. Anyway, so that's the book I recommend to a lot of people to read because [00:59:17] S0: I love reading history stuff, it's kind of interesting. Who made the biggest impact in your career? [00:59:26] S1: I would say my first mentor when I was starting my own company, he's like, I don't know, probably three years older than me And he helped me a lot when I was just a young founder with nothing. A lot of people don't know about this, I actually come to this country by myself, which is quite unusual. Yeah. Oh, turn? [00:59:51] S0: Yeah, I think, but in my era, I feel like there's a ton of people who just come to The US by themselves.