Venture Capitalist Sarah Guo’s Surprising Bet on Unsexy AI

0:00Hey, Mims, if you had to predict the next big thing in AI, could you? Hmm. Killer robots? Homework helpers? Movies made entirely with generative AI. I'm losing track of the tens of billions of dollars that giants like OpenAI are raising to replace us. Well, predicting the next big thing in AI, that's basically the job of today's guest. She's making some big bets on founders that are looking to transform industries. And she's part of a new generation of Silicon Valley investors. That's next.

0:35One of the buzziest names right now in Silicon Valley investing is Sarah Gua. She's got a bold name resume, having worked at Goldman Sachs and then venture capital firm Greylock. In 2022, she set out to raise money to start her own fund called Conviction, which makes early stage investments in tech startups. That puts her on the front lines navigating what is a confusing time for the future of startups and artificial intelligence. High speed change is very uncomfortable, right? It's uncomfortable for human beings. It's uncomfortable for even technologists. It is uncomfortable in particular for incumbents.

1:17Founders turn to VCs for a few reasons, not just money. Though the money helps. In the best case, a VC can be a trusted advisor. They can be a vote of confidence, a signal to rivals, or other investors that there's something under the hood of a startup that has yet to prove its thesis. And as she looks at the landscape for the next big thing, she's got her eyes on software as a service, better known as SaaS in the very unsexy tech lexicon. Think Salesforce or Slack. She says everyday business tools are ripe for change. An unsaid thing about what's happening in SaaS right now is some of the interfaces that are really important today, like$400 billion of market cap important or whatever the number is, they're like data entry and update interfaces for humans.

2:09But if you have a lot less humans doing that particular task, or you can do that data entry and update automatically, like maybe the software changes a lot. From the Wall Street Journal, I'm Christopher Mims. And I'm Tim Higgins. This is Bold Names, where you'll hear from the leaders of the bold name companies featured in the pages of the Wall Street Journal. Today we ask, in a world where big name AI companies are getting all of the attention and seemingly all of the investment dollars, what are the areas of undiscovered opportunities? Sarah, welcome. Thank you so much. There's a lot of things we want to talk about.

2:47But we'd like to start off with just like a little bit about you, how you got into this and your fund in general. You announced the fund in 2022. I think you said you raised something like$100 million, which was way more than you initially intended to raise. And this came at a time when investors or in the Valley lingo, limited partners were pulling back in general. And so I'm curious, what was your pitch for why they should gamble on you and why they should gamble on your firm? At the time, I was leaving an approximately 10-year stint at Greylock. I wanted to be an entrepreneur again, and I really wanted to do focused early stage investing.

3:32My friend, he's another investor, Elad Gil, had said, like, why would you ever build a firm? Like, you know, it's very complicated. You should just use your access and go try doing investing in sort of perhaps a less traditional way. And so the original premise was I love technology. There is a lot going on in AI that I'm interested in understanding. And I want to go figure out a model that I want to work in for the rest of my career. And so that's why the thought was like, I'll go raise$50 or$75 million or some big piece of that will be my money and we'll take it from there. We ended up, as you mentioned, raising about$100 million of investable capital.

4:12And I feel very lucky in that I had, let's say, 30 entrepreneurs that I had worked with before who put money into the fund. And so that was a big piece of it. And I think the claim I made to them was quite simple, which is I'm very committed to making you top tier venture returns. I will figure it out. Let's stop there. Greylock, for those not deeply steeped in the Valley, for my mother, who I know is an avid listener, and thanks for the emails, mom. Greylock is one of the boldest of bold names in Silicon Valley. It's like you were leaving Harvard to go out on your own or something like that. I mean, this was this is like, you know, you had made it in a lot of ways.

4:56Right. And now you're going to take a gamble on yourself. My poor parents. Like I this one really hurts because they're like, oh, you're going to not go to Harvard and then you're going to not stay at Goldman Sachs. And then like you're doing this venture thing. But like then you're going to leave Greylock. What are you doing? But I I I kid. Sorry. Sorry, mom and dad. They're actually entrepreneurs and risk on in their own lives. But I I'd say working at Greylock is an incredible privilege. Like the people are very smart, has a storied history of great companies. But sometimes I think people's risk profile is just different.

5:30And my view was this technical change is going to happen once in my professional lifetime. And I can't work on anything else. Well, it's interesting you bring up your parents because I was reading your blog. You had wrote at one point that your parents, they were tech founders and that you quite literally grew up inside their tech company. I wonder, what did you see then that you draw on now as you're going around and trying to pitch yourself to founders that you can be the partner that they might need that you should be able to help invest in their companies? Yeah, it was a networking infrastructure company selling to large service providers.

6:09So think like, you know, Comcast or Time Warner, those types of companies, right? Folks who deliver the TV or the Internet to New York City. This is the only childhood I had, right? So I just don't know how it might have been different otherwise. But I think norms are pretty important. And for me, it was very normal that you would take a band of your friends and go try to create a new thing for the world and be at the office at 3 a.m. trying to fix a bug for a customer and talk about sales productivity at dinner. And so I think more than any specific piece of knowledge, I would like to think it created some real empathy in me for what the entrepreneurial journey is.

6:45It is very scary, right? Even when you're winning, it's very scary. Going public is scary. Hiring is scary. Being sued is scary. Your customers are scary. And so I think that just having a very deep empathy for the combination of crazed self-belief and the terrified pragmatism of just trying to operate you against the world is something that is very real for me. I think another thing that perhaps I value a lot in investors and in partners around company building is not panicking. This sounds like an incredibly basic thing, but many bad things happen along every great company journey. And me and my partner, Mike, have been lucky to be part of some really great company journeys, right?

7:33Figma and Rippling and Notion and Facebook and all of those companies did not look perfect at any point. And I guess just being internal to a company when investors or anybody around a business panics because things are going wrong, it just makes everything worse. This feels like kind of a panicky time in your industry. I mean, Tim and I were talking about this recently. I think there's this outside perception that tech is doing great because you see these headlines about big funding rounds. but that is really limited to AI, it feels like. And as far as I know, the sort of post-pandemic slump is still with us and the IPO window opening that a lot of people were looking for hasn't happened yet.

8:22The M &A activity that people were anticipating under this administration has not picked up yet. I mean, you know that all too well, being at Figma, which did not get acquired by Adobe. Is this a scary time? High speed change is very uncomfortable, right? It's uncomfortable for human beings. It's uncomfortable for even technologists. It is uncomfortable in particular for incumbents. You know, one thought experiment I think is useful around AI is if you ask individuals, like, to understand where the feeling of, like, fear is coming from. I think that in large part, there is concern about issues that will be real, right?

9:03Job displacement and competitiveness and the like. But I'm going to give you a very biased point of view, which is the path is going to be for people to embrace the technology, and it will be very uncomfortable. Yeah, I tend to agree with that. I wonder, though, for you as a venture capitalist, it feels like it could put you in an awkward position because so many of the checks going to these big companies are gigantic, right? So you raised a$100 million fund to start. You just raised a second round at$230 million. Meanwhile, Anthropic recently raised$3.5 billion, and OpenAI got$40 billion from SoftBank.

9:43Are you able to get in on any of these rounds? Does the hype around AI make it difficult for you to do your job? The answer is, yes, building a venture firm is a multi-decade journey. So we still have plenty of work to do. But I feel very confident about our early stage access to high quality companies. And maybe I'll start by like putting a premise out there that informs our investing strategy for the fund, which is if you wanted to do like mid-stage plus rounds in OpenAI and Anthropic, and you thought that was going to be the only value creation. And actually, that is terrifying for everyone, if true.

10:19like, you know, every human being that doesn't work at one of these three companies, then this is not the fund you would raise, right? And so if you were doing that, you would raise as much money as possible to have the scale to be relevant in the$3 billion financing for one of the existing labs. So maybe two beliefs. One is between open source and the movement that keeps progressing there with companies like DeepSeek and Mistral and the like, it still feels like a very dynamic environment. to me, where it's becoming more competitive, not less on the model side. The second piece that's more relevant for us, actually, is the last mile between just having some intelligence and having impact in the real world is really long.

11:05That's like 99 % of it, actually. And so I think there are going to be a huge number of companies and opportunities for individuals to take part in that value creation. And so that's what we invest in. We have two partners between me and Mike Vernal now, but it's still really small. In the first fund, we are part of companies I think are quite interesting. Sierra, Harvey, Haygen, Cognition, Base 10, Mistral, Cartesia. We're early stage investors in all of these. And the entry price for these investments, like median, is millions, right? Not 300 billion or whatever the right alternative here is.

11:43What a world we live in where millions of dollars seems like a little bit of money, right? Don't get me wrong. I'm not saying it's cheap, right? But I believe for me and the people that work at these companies and our investors, like, I believe we can build a venture return from here. We just heard how Gwa approaches dealing with founders. But next, we want to know how she goes about judging whether an untested business is worth her time and money. You have to have some sort of North Star in terms of what you're building. but also like paranoia and pace in terms of consistently trying to understand what's going on in the environment, like the technology environment around you to make sure you are still on track and you are closest to your customer and you know what's happening with the models.

12:26Stay with us.

12:38I'm curious what you're looking for. Are you looking for the business idea or is this about founders, about picking talent and then making a bet on them? If you have to choose, I would say talent. Obviously, both things matter, right? You're looking for people who they're trying to solve a problem or they see an opportunity to serve users in a different way. And the quality of that idea is an expression of them. So at the early stage, you are really backing people. And I think one trait that is really, really important today where it may have been even a shade less important five or 10 years ago in venture investing that we are looking for is a combination of like you have to have some sort of North Star in terms of what you're building.

13:24but also like paranoia and pace in terms of consistently trying to understand what's going on in the environment, like the technology environment around you to make sure you are still on track and, you know, you are closest to your customer and you know what's happening with the models. It was a gentler time in the like middle period of SaaS investing five, 10 years ago where it was not likely that you would wake up on a Tuesday and be like, all my assumptions about how we would build this software were wrong, but that is possible today. So you said that this is a really dynamic environment still with these frontier models.

13:59If I'm making, say, like a March Madness bracket and I've got like Mistral and OpenAI and XIS Grok and Mira Morati's new startup and Amazon's Nova, and I could go on, how are you handicapping this? Or does it feel like there's gotta be some consolidation here or not? I mean, it depends on your thesis, right? I do not think that five years from now we will have an infinite number of companies training frontier models, nor should we have an infinite number of companies that dominate a consumer knowledge assistant type experience. Right. But I do think that there are multiple markets in there. And I'll just give you a few data points about this.

14:45By the way, like I'm not an investor in DeepMind, in OpenAI, in Anthropic. We co-invest with these companies. We do business with them. Our companies use them as vendors.

14:59But Ramp, the credit company, the sort of corporate card company, published some interesting data about the market share of Anthropic spend over time since Cloud 3. And that has dramatically increased. And so if you look at the API market versus the consumer market, those feel quite distinct today. I think there are points of view that instead of there being one assistant to do everything, maybe companionship and productivity are different. Maybe you have different agents doing different things. Maybe we have different form factors altogether as these things are not just informational but do more work for you.

15:43And so I would say that one of the most fun or terrifying things about being an investor or a builder in this time is the lines you drew around different categories are moving. And so I would dispute your March Madness bracket because I'm like, we're not all in the same division, right? Yeah. Sorry, I'm exposing how much I don't know about college basketball. Oh, it's okay. Sports metaphors out of me are always just total nonsense. As Tim can tell you. This is not ESPN, just to be clear. Just to make sure that I understand this. Wrong guest, yeah. It sounds like you're saying there could be one chatbot that you use at work that's assisting you with certain tasks.

16:26Maybe a different chatbot you're using for a different set of tasks at work. You have another chatbot at home. A fourth is your therapist. And then there's like five, six, seven, eight that are under the hood driving AI features within software. And you don't even realize that you're interacting with them indirectly through the API that the developer has chosen. I think that's right. I think it's possible we have this diversity of consumer experiences. Not a single company has commanded consumer attention in any part of technology history. Well, you know, it's interesting that we are coming out of this kind of era of some very big tech, right?

17:04A lot of conversations about big tech and the power of Silicon Valley and these sorts of things. And it's kind of interesting that none of these incumbents are dominating AI just yet. There's some potentials, right? Google is doing some stuff. Microsoft has taken an investment in open AI and partnering with them. But I guess some might have thought that the big tech companies would have controlled their platforms and would have been able to have distributed this technology in a way that they would just dominate. I wonder why that hasn't happened. I wonder too. Tim, we ask this question every time we look at companies because there are many claims you can make strategically about why a company shouldn't exist because Microsoft should build it or Apple should build it or Google should build it.

17:50Right. And I don't have a good answer here, except I'll actually draw an analogy back to venture. The cost to coordinate in very large organizations, it's just like it's hard and different than, for example, being creative with a risky new product or coordinating a bunch of scientists in a research effort where you have to make like very opinionated technical bets where you're spending a lot of money for a lot of money. long period of time. And it is not at all that the incumbents are not a threat to every new idea that we look at. I would just say like in this environment of very rapid change, like I do not envy the people who have a hundred thousand person ship to shift and the technology changes every week.

18:40We've been talking about one of the sexiest topics in Silicon Valley, which is AI. And now I want to talk about something even more exciting. And that is SAS software as a service for all those out there who don't know. And yes, this seems perhaps a little boring to the everyday user, but actually this is a field that's been extremely lucrative for Silicon Valley. Think hits like Salesforce, Slack, these are the tools that are really part of everyday life for many listening out there. And I'm curious how you see that kind of world changing in the startups that you're looking at. You believe heavily in SaaS and investing there.

19:21What's that like now? I think the impact will be uneven, right? And let me try to, I don't know, describe a framework with some examples of where you're going to have really big impact and where you might not. There are products where I think the core process has to happen with or without AI. And maybe I'm just not becoming creative enough about it. But for example, there's not a lot of reasoning that necessarily you want to happen when you process payroll, right? And I don't want like lots of generative creativity there. I just want to go through every time correctly. Just get it done. Just get it done.

19:59And so I think as an example of like, if I want a database of my employees and to process payroll, like I will still want that five years from now. On the other hand, I think when you look at things like Salesforce, right, because you mentioned it, Salesforce has so many products, but one of the core things is you have the customer relationship management database where you have a list of opportunities and you take notes on them and then you assign them to sales reps and then you do forecasting, right? Wait a second. I just want to say to you, finally, we have a great description of what Salesforce is about.

20:37I think a lot of people don't. They see the name on the building, but they have no idea what it is. Yeah. What does Salesforce do? It's a list of your customers and then you update it. But you can imagine ways in which that could be really different, right? I'm sure you guys and your listeners have at some point chatted into a website and then there's some multiplexed agent serving 10 different people at once being like, oh, can you give me your name again for the 14th time, right? Absolutely. And it's like, what is your customer ID? I'm like, I don't know. Just find me. But I think in terms of like how the customer relationship database could work in the future, you could populate it automatically.

21:19You could do many interactions automatically. You could upsell customers automatically. And so I think an unsaid thing about what's happening in SaaS right now is some of the interfaces that are really important today, like$400 billion of market cap important or whatever the number is, they're like data entry and update interfaces for humans. But if you have a lot less humans doing that particular task or you can do that data entry and update automatically, like maybe the software changes a lot. The future might be bright for software, but what does Gua think about a new era of political expression in Silicon Valley?

22:02There has been some social division as folks are trying to figure out where one's allegiances lie. And the Overton window of like what you could express as a technology person and venture capitalist has opened dramatically over the last five years. That's next.

22:37When I think about some of these younger founders that I talked to, there's really this incredible before and after, generationally, a lot of it because of the pandemic and how just central that was to a lot of people's experience. And I'm just curious, have you witnessed this? Do you feel like there are significant differences in the younger generation of founders now who you are talking to in any way, like in terms of their priorities or the technologies they're focused on or anything else? Absolutely. But it's really hard for me to tell if that's just because I'm becoming an old. You know what I mean?

23:12I'm like, were they always like this or is it just like I'm getting very ornery? We work with several sets of dropouts. I can describe them now. So there's a company called Sola that does business process automation. So, for example, like, you know, in a freight brokerage company, you have to move a lot of paper around between different systems to take a load from a shipper to its destination. And, like, it would be nice if more of that happened automatically. The founders are MIT dropouts. So they're in their very early 20s. I'm trying to remember if they could drink when we first backed the company, and I think the answer is no, or at least one of them could not, which amused me.

23:53But I'd say, you know, we're not backing every random set of dropouts, right? Like, happens very rarely. You can't have an Ivy League dropout on your table and automatically get funding from you? No. Not anymore. Well, I don't know. It's a pretty heated time, but no, not for us. And I think the ambition and capability and strategic thinking, like, they happen at all different ages. And, like, this is funny because this is an enterprise software company, right? We're selling business software to enterprises to improve their operations, like, efficiency, which is not – As 20-somethings are wanting to do.

24:37it's uncommon versus like, oh, let me start like a consumer social thing for my friends on campus, right? And so I think the surprise to me from a behavior perspective might just be me growing old, which is like the communication can be quite casual. But I think it's just, I think it's actually just stylistic because these people are like amazing professionals. So they're like, I got a million dollar contract, bro. Something like that, Except I would need to urban dictionary the words. Well, I don't even know if we're talking about age young, but we're talking about almost a new generation of entrepreneurs who, you go back 10 years ago, were excited about the potential of the app economy, a potential of what mobile was doing.

25:25So we saw a lot of app-based businesses, whether it was Uber or Instagram or whatnot, kind of coming out of that. But now, you know, I feel like we're seeing a lot of AI, but we're also seeing a lot of maybe hard science kind of businesses that are inspired by the likes of SpaceX and the defense company Andrel. And this idea of changing the world in like the literal sense of the world. Are you seeing those kinds of pitches in your life? Software can change the world, Tim. Absolutely. I take that back. But if you are, I believe that so genuinely, so I'm willing to say it. But if you're asking, like, do we see companies that are doing hard sciences beyond computer engineering?

26:10Absolutely. Right. So we work with a company in the consumer robotics space. They are young PhDs and PhD dropouts. And yet I'm so comfortable saying that they are world experts in what they do. And it's in the pedigree, right? Like Tesla and DeepMind, whatever else. But it's characteristic of the fact that one thing that might be really impressive to me as I become more of an old is, you know, people start their journey into becoming a world expert at a very early age, right? Where I'm like at 16, I'm like hostessing at Outback Steakhouse. And I still had what I would consider to be like a very specific childhood where I went to the startup office every day.