What VCs Really Want From AI Startups in 2026

0:00On this episode of the show, I have with me Susan Liu. She is partner at Uncore Capital, and we're going to be talking about, well, finding the right companies, back to the right companies that have the right solutions within, especially the AI space. We're going to talk to Susan about her thoughts and how she sees that with early seed stage companies and valuations and adoption, and maybe where 2026 is going to take us a little bit along this AI rocket ship, I guess we can call it, but Susan, I appreciate your time. Thanks for turning the show. Yeah, thanks for having me. Absolutely. Okay, before we start, Uncore Capital, tell us a little bit about what you guys do. Yeah, so Uncore Capital, we're based in San Francisco, and we invest in seed stage companies. We invest all over, so my focus is on the B2B side, but we do B2B, we do consumer, and we do frontier tech. Our check size is, I'd say, like anywhere from like 1 to 5 million, then must have for us as a side to we're looking for a minimum of 10% ownership in order to get involved, which means we're typically leading or co-leading rounds. And then in addition to our

1:04core seed fund, we also have an opportunity fund as well that is used to continue investing in our very best companies. Absolutely. I guess the question that I always have, and you know, I think when it comes to investing companies, the earlier you get, it's just really hard. I get some difficulties, and you know, I know that there's the power law VC. I get that, no VC just wants to really lose money. I mean, that's just not how it operates. But I do find that when a seed stage, especially earlier stage VCs are operating, and they're looking for things, it is obviously a little bit more challenging. I guess from your perspective, when we're looking at early stage, especially for companies that have that quote and quote, right make up, I mean, are there some things that you've noticed just from pattern matching that you can look for and that has helped in the past? Yeah, it's a really good question. So I would say in early stage venture, you are going to lose money. I think it's unavoidable. The loss rates

2:09in earlier stage are just higher than later stages, and it's just part of the game. And like, you just have to accept that as an early stage investor. You're right, though. I think there are certain things that you could potentially look for to hopefully minimize the losses. And I think everybody has a different style of investing. For example, even if you look across our partnership, we're kind of looking for different things. So as for me, there are certain things that I like to see. And I wouldn't say it's true for all of my investments, but I think it's worked out for me so far. So how I typically look at investing, I generally specifically spend time on three things. One is team, two is market, and three, I kind of call it like product slash traction. It's a combination. So it's like a product wedge slash traction based off like the product that you're selling. So I can cover all three. And I'd say like investing, especially the early stages, is definitely more like art than science. Like this is kind of like the rough formula for me for every deal I do. Like it's kind of a different mix of each I weigh each factor a little bit

3:12differently. And you'll never have like, you know, the same type of investment. But in terms of teams I like, my focus, by the way, is on the B2B side. In particular, I really like investing in AI applications. So I'd say like typically for me, the founders that I like gravitate towards, they tend to be a little bit more seasoned. They tend to have some, you know, unique insight to certain markets. And usually this is earned insight, meaning like they've typically worked in the, you know, the space before. So for example, I invest in this company, new rule that, you know, sells sales tax software. And they're also automating the whole like sales tax process on your behalf. As like either a CPG company or as a SaaS company. And what I really liked about this CEO, the founder is that like prior to starting this company, he was actually, you know, a CPG operator. Like he was building in the past and like sale sales tax was one of the like the pain points that he encountered. So he decided to build in the space. Another one of my investments, a company

4:15called Ivo, that is building a company around AI contract review. He was an explorer. So like he knew the space pretty well too. So I generally like it when the founder has a background in the market that they're building in. And the reason why I like this, that is because I'd say like generally speaking, they kind of just like know the customers, like they understand the customers really well. So when it comes to like selling to these customers, I would say they're much better at it. So that's something that I really enjoy. Of course, like it's not just like the, you know, in addition to knowing the space really well that, you know, generally teams is more than one person. I have back solo founders, but usually like I'd say like most like founding teams are like maybe like two to three co-founders. I usually like to, I call them like these are like the industry experts. I like to pair the industry experts with a technical person. So like I say like the ideal profile for me is like somebody who's an industry expert, plus like a technical co-founder who can build the

5:15solution. So that's generally what I like to see from teams. Again, doesn't always have to be the case, but I think that's something that I like to see. And then when it comes to markets, you know, I think like a lot of you see this, but it's very true, meaning like I like really big massive markets. Sometimes it's not as obvious. So if it's not as obvious, like I like to see trends that show that the market is growing. I think that's really important. And I know a lot of you see stockpots, so I won't spend too much time here, but like big markets, if not big right now, hopefully one where like there's trends towards becoming a bigger market over time. And I think AI has like a really unique impact here because AI is eating up some of the labor budget because it's doing some of the work of humans. And therefore like, you know, I think in many markets, like they are growing because you know, now potentially like they could they might be able to do some of the work that historically a human could do. And of course, it's not, you know,

6:16AI is still limited in many senses. And so oftentimes you will still need a human, but oftentimes that they oftentimes they can potentially, you know, take on some of the more like mindless work. Basically the stuff that like humans don't want to do because it's super tedious. So that's on the market front. And I'd say like generally speaking, because I also do a lot of vertical software. And sometimes like the, you know, one of the questions that pop over vertical software is like, you know, when is the market too small? So after doing a bunch of analysis on this, I think like the key for me is if the mark, if you can like, if you do a market size assessment, if the market, if we think like the potential market is bigger than 300 million, it's probably big enough. But of course, again, like I'm looking for growth within these markets as well. And then after the last part, product wedge and traction, not all companies will have the traction component. But on the product wedge side, generally what I like to see here is just like a super, super strong product wedge. And

7:18generally how I validate this is I'll talk to you potential customers. So like people I know in my network that could be potential buyers of this product. And I'll ask them like, Hey, like this is the product. What do you think? And if they say like, I love it, I have to have it right now. And that's a really good sign. And that's generally look for when I am evaluating companies. But if they say something like, Oh, yeah, that's nice. It's cool. But like, I don't know if it's something that I have to have now. And that's probably a no for me. Because I like to, I like to see really strong product wedges where like it feels like it's a must have for the end buyer where they almost feel like they can't do the job without it. So I think that, you know, that signals to me that like the product wedge is very strong. And then of course, on the traction front, you know, so I would say like some sea stage startups have this other stone. Of course, like if you're earlier in your journey, you're not going to have any product traction. But I will say that like for a lot of like AI apps that I look at, there is some product traction, you know, and of course, like usually you want

8:20to see this like growing as quickly as possible. I think we all know that the Series A milestones have changed where in order to get a Series A done, like the threshold is much higher. Whereas in the past, like you just need to get to a million in terms of AR to get us, you know, to at least attempt a Series A. I'd say like now, it's like you probably have to get to like maybe like three to five million AR in a year or two get a Series A done. So a juicy companies with some traction, some revenue traction at the sea stage. And if I do see that, I like to see higher growth ideally. Yeah, I guess a question when it comes to this AI space. We are now seeing 2030 of the same companies. I've seen on the pod from just companies that come across. I'm like, oh, I've seen this business. Oh, I've seen this. And it's okay. I guess if we were to rewind the clock back 10, 15, 20 years ago, you saw less collision. I feel maybe I'm wrong. I mean, I'm just

9:22not recalling it. But I feel there's a lot more collision. I guess as of EC, someone who has especially like in these enterprise type products in the AI space, do you worry that there could be too many of a similar product or is it still a case of I got to invest in the company or the model or product that I believe in? Yeah. Yeah. I think it's not abnormal, meaning like I invest in AI applications, right? In the old world, right? Like pre AI, this was called SAS. And like, you know, if you look at SAS companies in the past, like anytime there's an interesting market, like, for example, let's take like a marketing software, yeah, MarTech. MarTech was all the rage, like, I don't know, like five to 10 years ago. And we probably we literally saw like a million companies. Like, if you look at like the old MarTech market mass, like hundreds of companies, like tackling this market because it's massive. And again, the trends are right where like I think people are making the move from on-prem to cloud. And there's just a lot of opportunity. A lot of like smart founders

10:23saw that opportunity and they wanted to build. So I say like it's not new in the sense of like, you know, for any good market, there's always a ton of competition because, you know, the reward is so great on the other end. But it does make investing trickier because like you want to make sure you're back in the right horse, right? And I think as a VC, the goal is always to back the player that will be number one or like number two or maybe like worst case number three. Generally, like, how I've seen these markets evolved is the following. Usually like, I would say like almost all the rewards go to number one, two or three. Number one will probably go public. Number two, maybe able to go public, but more likely it's M&A. And that's true for probably number three as well. And then like everybody else kind of like just doesn't work out for. That's kind of how I've seen these markets evolve. And it's basically what I expect for a lot of these like new AI markets as well. Which is why like we're also seeing Kingmaking happening in the VC world where a lot of later

11:24stage investors are piling in into like the number one or number two player. And while you're seeing like all these funding rounds like happen back to back. Meaning like somebody who will raise a series A and then like three or six months later, like you'll see that they raise a series B. They probably have had some like good progress in terms of company building, but probably not a ton. But like VC's kind of understand this dynamic. And they want to make sure they're in the number one player before like you know the rounds get too expensive. So I mean like this stuff is definitely happening. And again like with AI like we've never seen growth like happen this quickly before. And so like you know the time for a company to get to a hundred million in terms of ER is much faster nowadays and in the past like in the old I guess I would say like relative like to the old SAS world. So you know for me personally like I like markets where there's a lot of competition because you know that it's well I guess you never know for sure but you have a hunch that it's a probably a really good market where you guys having like a massive impact. An example this is law right. You kind of know like illegal tech like you know something's happening there because

12:28there's just like a lot of companies, a lot of funding, a lot of interesting companies being built. But I would say like you know the key for founders is just make sure you like well ideally you become number one or number two. And then if not that you just kind of have to have a really good story about like why you should exist in this market. You know I like that a lot and I guess the question when you're looking at you know what we've been talking about SAS you're talking to Martek. At some point I guess you know when we're looking at these products and we're looking at the market we're early. Is there a thought that potentially investing where there's less dollars going as not I mean we hear that a lot investing just in stock market and the reason I'm asking you is we're seeing the the hard rush into AI all the dollars are going into AI. Obviously we're hoping that that's going to come true that's that's that's it looks like it should there's a lot of positives but is there a risk of not looking at other areas or is it a case in the VC world where it's

13:33listen this is where the dollars are going because the competition for the next thing is too great to miss out on it. I don't want to call FOMO but maybe there's a touch you can't miss that big swing. Are you talking about like companies are like investing companies are not AI focused? Yeah I mean investing companies are not AI focused maybe slightly different focus or maybe yeah I'll take this two ways so one is look I think a C-Sage investors we we aim to be ahead of the trend yes like if I'm backing a company and it's already a hot market well I guess I still invest in some some like markets are already hot but like my goal at least is in early stage of investor is to get into that market before it's hot. So like for example when I invested in new role which you just talked about which is the sales tax company like sales tax was not hot at all people were like this is such a boring sector of the market like why would you like invest in that company here but then I did my work and I was like oh my gosh like this is like a multi billion dollar market and all the players are like essentially legacy players

14:34what they've been around for like a really long time like Avallara vertex etc these are old companies I think there's room for a next generation player here so that's why I ended up making a bet in this space but it was not definitely not a sexy market when I invested it's a sexy market now because people realize the opportunity the goal for me is always to get in before like the market gets super sexy and then you know as a release to like the AI component like investing in markets that are like not AI driven I definitely think there's still opportunity in non AI markets like for example like I do a lot of vertical software investing and like a lot of you know I'd say like AI is present in some vertical markets but there's still markets where like AI still not a thing because like you know maybe they're maybe they're just starting to top technology like there's other markets where like you know it's mostly like pen and paper and there's just no like software yet and that happens or maybe like they're still moving from on-prem to cloud solutions like you know it's not like sure you can add AI to it but like it could just be as simple as like

15:36going to the cloud right and so I'd say like for me like if I were to like slower investment that was non AI focus it would be in vertical markets most likely and I would be looking for systems of record core systems of record like that's what like you know essentially it's some sort of software platform where it's how you run your business and like you know it's like the main place you go to to like run your business like that that's what I would be looking for absolutely and I definitely think there's an opportunity there it's just finding the right markets yeah yeah I mean I guess that's yeah you we've been kind of talking about finding that right market obviously lots of move into AI I guess when the one component of this is there's a lot of product being built we also see a lot of studies coming out I think the the MIT one was one about the percentage of failures from companies themselves implementing solutions and obviously there's a lot of companies are piloting these AI solutions that are popping up when it comes to

16:40adoption when it comes to usage I mean we're going to have to see some traction on the enterprise side to I guess justify the spend at some point when you start looking at that you start looking in ROI have you started seeing anything that starts looking like real like litmus tests so what people are looking for to to get out of an AI model so they know hey this is worth the X-pen we're spending yeah um I think your questions around like just like how I think about ROI for like AI um spend right yeah um so I mean like I think you kind of nailed it which is like I think a lot of companies that are purchasing AI solutions are looking for the ROI component and that's really important like look AI sexy like there's all this promise but like not all AI products can deliver meaning like have meaningful impact on your business and I just say generally like businesses are looking for probably like two things one is that either drive more revenue or like save cost summer yeah if you cannot prove one of these two I think like it's going to be a tough sell and like

17:42companies are not going to renew um there's a lot of ex you know we know we all kind of know this there's a lot of experimentation happening right now where like your CEO probably be like hey like we need some more AI in our company right and then like you as like maybe the business leader will have to go and like figure this out and you're probably trying a bunch of things and I think that's okay like you know I think it's good that people experiment um but if you're selling to these companies like you have to make sure the ROI is super clear as either like a revenue driver or or like really prove out the cost savings um and like help your um customers be like measure this because like that's basically how they're going to figure out whether or not they're going to renew with you or not um my guess is that um in the next year or so we're just going to see a lot of churn because um I think a lot of companies again like a lot of companies have been experimenting over the last year um but again like not not every solution is proving out the ROI and if they're not they're just going to cut budget for this like it's just part of the game absolutely I guess

18:42as you're kind of looking forward we're still early we've we've now been within the AI space for a couple of years all the attention it's got and I I guess do you expect more of the same so it's 2026 going to just be a continuation of 2025 I know we we dubbed one year the year the POC I mean what are your thoughts on 2026 moving forward yeah um I would say I am I'm trying to figure this I like which stance I want to take but I say like I'm cautiously optimistic in the sense that like I think AI can still have a massive impact I mean like we're kind of just scratching the surface in terms of what AI can do and like it's impact on many markets um and the thing that like makes me excited is that like I am seeing and I have a pretty massive impact in multiple markets like legal which we talked about like the ROI they're super clear um health care is another space where like boy like there's so much like um manual work happening in those markets that I'm glad AI is

19:45automating a bunch of it because like some of it just seems a little unnecessary at the moment in terms of like manual work session human labor that's kind of going into um like certain like health care processes um but at the same time like again like we kind of talked about like um you know people being very expensive experimental and trying a lot of solutions like I just you know I do feel like the churns coming at some point I think this might be the year um we're like we're finally gonna see a lot of churn for some companies it's not all like you know at the end of the day it's not just all revenue growth like you like you have to think about retention um and I think you know some early AI companies are just not focusing on that enough and when it happens when the churn comes it's gonna be a rude awakening um ideally it doesn't happen to our AI companies but like it's definitely gonna happen we're just gonna see that and then like as a company you're just gonna have to start thinking a little bit more about like hey like how essential is our solution um and then like you kind of go back to like first principles

20:46which is like you know make sure you have a good customer success department I can like you know flag um you know companies that might churn and then you know work on those um accounts to make sure they don't churn etc so um you know I again I do think like AI is gonna continue to have a really big impact but we are gonna start seeing like some churn happen um with some AI companies which will be interesting absolutely I guess yeah just maybe a question from that when you're looking at I guess we talked a little bit about just uh we've even talked about valuations but we're looking at valuations I mean AI itself I know meant these bringing about some kind of potential efficiency for many of the discussions we're seeing right I mean that could mean somebody has more time you need less people I mean there's a lot of different I mean there's few different things we've heard but when it comes to revenue um outcome based is is apparently

21:47becoming a thing where it's like hey I'm saving you x percent that's that's I want to charge based on that is that is that starting or has that caused you know the VC world to start looking at valuations different or is it still yeah we're just not there to actually get that far yeah yeah I say we're not there yet meaning like I actually don't know of that many companies that are charging based off of outcomes just yet I think like there's a lot of talk about it because it would be a very interesting business model but we're not quite there yet I would say um yeah I mean like look when it comes to like valuations for AI companies it is off the charts in the reason why and it's kind of justified to be off the charts if I'm being honest with you like I've never seen valuations this high ever since I've been a VC and I've been doing this for 13 years um and I the reason why it's justifies because um you know these companies are growing so much faster than I've ever seen before um and I you know of course like revenue growth is a big factor in terms of what like valuations should be in addition to that it's not just software spend anymore like it

22:50could also be labor spend services spend etc so you're eating into other budgets and so now again markets are much larger so um does it justify a higher valuation I'd argue yes which should you invest no matter the valuation probably know like it does take a little bit more analysis there to kind of figure out like the right price and like you know you know what the potential cap of that should be um but I think every like VC has kind of like their own internal um goals in terms of like you know how much they want to return for the fund and all that stuff um but yeah like just given all that I feel like yes valuations are high but you know there sure markets where it justifies it absolutely do you think that when we're and I I'm definitely we're we're way too early for outcome based anything at this point but when you start looking at the companies that you are investing in in this in this you know in enterprise AI space

23:51are there things that as that you're seeing in terms of successful I guess people who've moved from seed stage to the next round of funding are you seeing some common things from an AI strategy perspective and the reason I ask is it's fundamentally a different product you mentioned Martek we're not delivering a CRM or something that's going to help us here in process we're delivering some potential block box secret sauce for telling people is going to make their world better um I guess some people are going to position that better than others but are you seeing successful founders is there something they do better with that or is it just purely product dealer because it's either going to work or not based on a customer yeah um I would say like um while the end product is different meaning like it's not just software anymore it's you know it could be an agent a co-pilot whatever um I'd say like the the fundamental I would say the fundamental of a business when it comes to like raising a series A is still roughly the same

24:56but like the bar is just higher um I mean we kind of just like touch on this like the in terms of like where you have to be like AR wise has gone up from like one to like maybe three to five million in terms of AR um because again like companies that grew just grew in so much faster by the end day I think like series A investors are still looking at like the the same things yeah meaning like you still have to solve a real valuable product or a problem I'm sorry um where uh you know maybe again maybe like the product is different we're like it's not just like a CRM anymore now it's like an agent but the agent still has to deliver value I think that's really important um you're still backing a great founding team meaning like you kind of have to believe like they can take it all the way and like take the company public like that's still the same um still has to be a massive market um but again knowing that the markets could be potentially bigger now um and then you know yeah I think you always have to factor in like what else is going on the ecosystem like you want to make sure you're backing like a top player in that space with a unique angle um and like the it's

26:01up to the founder to like really sell that story so I'd say like um at its core like the series A components are still roughly the same the bar is just higher I mean look it's I think it's still I mean it's really hard to get a series A done these days just because like there's a lot of really interesting AI companies that are just growing quickly um but if you like again like if you if you're if it's a good company like they should have no problem raising that series A and I'd say like it's series A investors are also very keen to have more like AI in their portfolio so um you know I think it's definitely still possible it's just again the bars higher than ever before yes and actually yeah that's actually a really good point because as you were kind of talking through that I was like it might actually benefit investors to go heavier seed stage in some cases because there is a lot of unknown and if you want to invest in it in more unknown and more opportunity obviously getting earlier is where you go but to the flip side you have to prove out

27:04your business your product and the product in this case is it's not as easy to touch and feel it's got to be kind of not a different level and then you get your series A so as you're talking I was like oh I could see the bar being different it's not like oh these features are coming oh yeah is it actually really work can it solve this problem you've claimed it stated and that's a whole different it can be bar them before yeah um I think I think that's right um and look I think in a day like series A investors still kind of evaluate company the same way which is like how do you figure out there's real ROI such a product is valuable you talk to customers like that still happens even when I prep my founders for their next race like a series A for example I'm like hey like make sure you have three to five companies or three to five really happy customers that are willing to speak on your behalf that will tell your VC like why you love them and why you're a great company and why you're even spending the money you're spending on this company um I think like having the customer tell the VC is like the most powerful thing you could do um when it comes to raise it your series A absolutely love it Susan thank you for taking the time thank you for sharing

28:07with us um if somebody does have a follow up question uh you covered a lot of different areas for us but is there a good way of touching base with you yeah um you can find me on LinkedIn um and then if a founder uh wants to send us a pitch or emails pitches at uncorcapital.com all right perfect then again since this is a thank you for the time I appreciate it thank you absolutely all right that's it for the episode we back again different guests different topic until then two things I'd love it if you could share this episode with somebody else who could uh find value in it I think Susan covered a lot for us within the startup space investing the startup space especially seed stage she uh actually dove in and told us what she looks for some of what are the key components to successful startup really uh haven't changed but there are some wrinkles that are even a little bit more challenging in this AI space so share this with someone who is a founder wants to be a founder if you want to pitch we got a link for you this season uh provided also like subscribe comment let me know how this shows going for you until next time thank you and goodbye