Request for Startups: 10 AI Ideas We’re Ready to Fund

Hello and welcome back to the Gentek Podcast. My name is George Van Darien, general partner at Untapped Ventures. We have a really exciting episode in store for you as we are at the halfway point of 2025. Things are moving so fast in AI. So wanted to publish our inaugural Request for Startups episode where I could share a little bit more about where all of our research in the fast moving world of AI has led us and what are the next 10 investments we want to make. We're beyond the days of just the AI assistants and copilots. We're not looking for something that's a tool, a feature. We're really looking for full scale platforms. It really divides into two areas, agentic AI and convergence AI. What we're really looking for is SaaS killers. We have a tremendous opportunity for a new batch of AI founders to take a first principles approach to what does the customer in this segment really want, what are they looking for? And from a first principle standpoint, build a whole new experience where the UI is different. It's not the same old SaaS UI that we've become used to with all these database screens, but it's really able to help run my business for me, do as much as it can. And me and my team only need to do the 2%, 5%, 10% of what it can't do on its own. And I think we're going to see a number of SaaS incumbents turnover as the new AI versions are going to replace them. Welcome to the AgentIQ podcast where we dive deep with the pioneer AI founders reimagining business and investors betting on them. We bring you the war stories, insights and perspectives you won't hear anywhere else. I'm your host, George Van Darren, a general partner of Untapped Ventures, a VC fund that invests in early stage AI startups. Let's go. Hello and welcome back to the Gentec podcast. My name is George Van Darien, General partner at Untapped Ventures. We have a really exciting episode in store for you as we are at the halfway point of 2025. Things are moving so fast in AI. So wanted to publish our inaugural Request for Startups episode where I could share a little bit more about where all of our research in the fast moving world of AI has led us and what are the next 10 investments we want to make.

Right.

And so I'm really glad that you're here with us and let's go ahead and dive in. So what I'm going to do is I'm going to Provide some context and some background and then we'll jump into those areas. So with that said, let's dive right in. So I think first of all the things to understand as we begin is what's our thesis? Right, so our thesis centers around, we're betting on agentic AI and convergence AI and I'll define what that means. We want to be able to lead, co lead and even follow in some cases precede and seed rounds with 500k to $2 million checks for extraordinary founders that are building in these spaces that we're going to talk about ideally leveraging some kind of an inflection or an insight that unlocks a massive market.

Right.

Something that could become really big, create a multi billion dollar outcome and be a multi fund returner. And I think it's really important for, you know, founders to understand what VCs are looking for, especially early stage VCs like us. Very similar in terms of what our goals are.

Right.

In order for us to be able to have a multi fund returner, that means that it's got to have a billion dollar plus or larger outcome. So IPO for 10 to 50 billion or get acquired in the billions. So what does that take? That takes a big enough market that's going to create $100 million of ARR business.

Right.

So what does that mean? Kind of what it really boils down to and obviously we have a 20 point investment criteria, but it really boils down to I think the two most important ingredients are team and tam.

Right.

And so is this an amazing team that is the best team in the world to execute on this idea. And if everything goes perfectly well, is this TAM big enough that this is going to become a not just 100 million ARR company, but a multi billion dollar ARR company and therefore a multi billion dollar company. And if you look at some examples from our portfolio, we have that right. For example, we invested in a company called Xtropic that's building a new kind of AI chip. And you can imagine if that works out then that's going to be in hot demand. A new AI chip that's much more efficient and we could easily see a hundred billion dollar tam. Similarly, we invested in liquid AI, a new kind of foundational model out of MIT and many others. So I think if we boil it down, there's two kinds of startups that we invest in. The ones that are more in the deep tech world and right, so there's, there's a scientific or at least a big technical breakthrough and then secondarily there's the application layer, right? So these are more typical software SaaS, but in the AI space and within there, what we look for, we typically find that we have six kinds of founders we're talking to, right? So we're always chatting with the founders that are currently at the AI labs and leaving and wanting to, you know, start something new, right? So OpenAI, Anthropic, DeepMind, Google Brain. And so we're very, very interested. So if you're at one of those places and you're leaving, would love to kind of chit chat, brainstorm and be able to spitball ideas with you. And secondly, very similarly, we have FAANG operators, FAANG and MAG7, right, that are looking to leave Meta and the rest of Google, let's call it Tesla and others. Third would be your lean AI builders. And lean AI builders could vary. Right now we have solo founders that could be using vibe coding and building something new. We have vertical domain experts that come from an industry and are very deeply ingrained with the problem and solving for that for that vertical. And then from there we have the academia and research, right? And from there you have these PhDs and postdocs that are working on those big scientific and technical breakthroughs and, and then kind of an X factor category is what I call the outsiders, right? The immigrant, the international founders, the just the ones that are pure hustlers and use their aptitude and their drive and their hunger. They may not be your typical sf, Stanford, Berkeley, mit, but somehow they just figure out how to make amazing things happen. And there's a lot of data that supports each of these profiles and why they're successful in AI today. And then like I said earlier, we're not just looking for seed and precede. A lot of AI VCS invest that seed and precede, but we even come in a couple steps before that, which is stealth and pre stealth, right? So you may be very early, you're not ready to raise your first round, but you're thinking about something, you're in that ideation stage, problem discovery. You're in that process of maybe trying to find the right co founders and we're building a community for that. And in the show notes and at the end I'll talk about, you know, so we have two paths. If you're ready for funding, great, pitch us. And if you're not, we have a new stealth program that we're, that we're rolling out that is meant for people that are looking to start something and we're, we're not afraid to be that first check in, you know, even pre revenue, pre product, pre everything. Okay, so hopefully that provides you some good context on who we are, what we're looking for as an AI, native VC. Okay, so let's talk about the 10 themes and ideas. Right, so basically it really divides into two areas, agentic AI and convergence AI. And so when I talk about agentic AI, what I mean is where we're at in the overall AI lifecycle is, you know, we're beyond the days of just the AI assistants and copilots. I think those areas have been really filled up. We're not looking for something that's a tool, a feature. We're really looking for full scale platforms. And then we'll talk about convergence in a moment where AI is converging with all the other exponential technologies and, and really unlocking new possibilities. So let's start off with core AI and there, you know, all the interest is around agentic AI. So what's happening, Some of the micro trends or macro trends I should say that we're seeing are you have vibe coding, right? With vibe coding you have these new tools like Cursor and Windsurf and also Lovable and Bolt that makes a building easier. It's the tide that rises all boats. But on the flip side it, it makes it harder to build something truly different and special. So I think in many ways it's become easier to build, but harder to build a better venture scale company.

Right.

And so what we're really looking for are founders that are leveraging all these tools and have a bunch of agents on their team.

Right.

And you have this, the lean AI leaderboard now with all the fast scaling AI startups. And I think it's really important to be leveraging those AI agents within your business. So if we stay on the topic of agentic AI, I think if you look at Gartner's Hype Cycle 2024, the innovation triggers, a lot of those technologies in the first area are the ones that are the newer areas that we're interested in that are just starting to get developed. And so, you know, I think if we dive in, ultimately what are we looking for?

Right.

So I would say one, we moved from SaaS to SaaS. So what I mean by that is we've moved to software as a service, to service a software, right? This is the new acronym SAS Service A software. So software used to help a human deliver a service, now the software is the expert and delivers the service. Right. So we're very interested in companies that are Building entire full stack AI companies as YC calls it. And for example, we made an investment in Harper which incidentally went through YC recently and they've built a full stack insurance company that is completely agentic.

Right.

So we love that, we love where someone is building not a tool for insurance, but just saying, hey, I think we could build a better insurance company, a full insurance company that competes against all the many insurance companies but does so in a better way using the current tooling. The next thing that is related is outcome based pricing. I don't think that's necessarily a business model but, but outcome based pricing is another trend where, you know, why, why pay a license fee when you could pay for the result?

Right.

And I've always been a big believer that the best vendors that any company works with are willing to give 100% money back guarantee and they're willing to price based on the outcome as opposed to time and materials or usage.

Right.

And so I think there's a tremendous opportunity now with the Gentec for companies to be pricing based on that ultimate decision. So whether it's on whatever, the KPI is a closed lead, downtime avoided, you know, emissions reduced and I think the plumbing is ready.

Right.

You have real time data streams, you have contract standards, you have AI evaluators that could judge fairly. And I think the outcome here is that, you know, you can have a higher price and better margins by having something that could deliver the end outcome for the customer. So if we combine the first two services, software and outcome based pricing, what we're really looking for is SaaS killers.

Right.

And so what does that mean? That means that our belief is that a lot of the traditional SaaS, you know, from the generic horizontal ones like your CRM and marketing automation companies like Salesforce and HubSpot, all the way to your vertical SaaS companies like for example Service Titan for Home Services and many other vertical SaaS. I think we have a tremendous opportunity for a new batch of AI founders to take a first principles approach to what does the customer in this segment really want? What are they looking for? And from a first principle standpoint, build a whole new experience where the UI is different. It's not the same old SaaS UI that we've become used to with all these database screens and, and all that, but it's really able to help run my business for me, do as much as it can. And me and my team only need to do the 2%, 5%, 10% of what it can't do on its own. And I think we're going to see a number of SaaS incumbents turn over as the new AI versions are going to replace them. So if you're building in that space, very, very interested to chat with you. I, I think we're going to see a lot of turnover here in the coming years. So the next area that I want to touch on is, you know, a combination of, I would say multi agent simulations, digital twins, decision intelligence, causal AI. You know, it's this whole idea of moving from LLMs that are just predicting the next word to systems that actually can tell us the why behind things. Right? The causation, not just the correlation. And then also run what ifs, right, what if scenarios with Monte Carlo simulations and be able to provide whether that's on a city or a supply chain or a company or an entire economy or the entire world. You know, I can run scenarios and have it give me intelligent advice before a single dollar is spent or single life is at risk. So there's a lot of opportunity there and really interested to chat with anyone that's building in that space. Next thing that I want to chat about is agentic AI infrastructure, right? And so with agentic AI infrastructure, as agentic AI blossoms just like any other AI technology and any other technology, we need that, we need that infrastructure layer and there's a lot of infrastructure players already. And I think there's going to be more opportunity to really be able to bring this to a point where it can be easy for people to build and manage agents, you know? You know, I think there, there's an opportunity for master agents. Maybe I have one agent and that works with all these other agentic tools or different ways to go at this problem. But I think overall we're going to need to make it easier for people and companies to have agents that work that could create agents easily, that could plug into agents. The protocols that are coming out lately are a big step in the right process. But as I talk to enterprises, they're still struggling to figure this out. The next thing that I'll touch on is ui.

Right?

We're going to need a new turning point in ui and I think just in the last week or two we got a glimpse of this with Cluli, right? Clulee, with their very exuberant founder, has shown us this ambient glass overlay as possibly a new interface to AI and I think that's really exciting. But there might be others, there might be wearable projectors, gazing, gesture control and, and just in general innovation on what does this agentic UI look like? What does AI interfaces look like? So very interested on that point. Earlier we talked about Vibe coding and so one thing we're really interested in investing in is Vibe everything else.

Right?

So Vibe coding to create an app is great, but there should be, you know, now we're starting to see Vibe marketing, right, Where I could just tell it a campaign and it could just build it out. I think there should be Vibe marketing, Vibe recruiting, Vibe Finance, Vibe Market, Vibe Vibe anything I need to do where I could just tell it and it could just build it out. Not necessarily an app that I plan to publish, but just a tool. It could be a dashboard, it could be a mini app, it could be a script, it could just tell it and it could just build it for me within my tool set. That's almost like vertical Vibe coding or vertical vibing in different areas. And then the last thing I'll say in this area before we move on to convergence is, let's face it, I mean, the reality is 40% of roles are going to be reshaped by AI. And so millions of people are going to need to find new skills, new paths to learn quickly, to prove mastery. We're going to need an AI that's going to help people navigate the process, transition, learn new skills, find new jobs, start new businesses, find their passions. So, you know, we're really interested in founders cracking a new kind of AI for reskilling and upskilling people that have lost their jobs due to AI. Okay, so we just finished exploring the AI native domain within the AI stack and our excitement around agentic AI. Now the other exciting area is what we call convergence AI. AI is the catalyzer for all the other emerging technologies and exponential technologies. If you take a look at Cathie Wood Ark Invest's Big Ideas 2025 report, you could see there that AI is the leading catalyzer for the other technolog, including everything from robotics, precision medicine, multi omics, blockchain, drug discovery and design, material science. Right. Neurotech, robotics, quantum computing. AI is the one that is the main lane that connects and branches out to all these other areas. So we're looking for founders that are using two or maybe even more of these exponential technologies in convergence. Right, so let's take a look and go deeper on some of them. Let's talk about AI and robotics and AKA embodied AI. Humanoid and mobile robots are stepping out of demo videos, doing tricks into warehouses, restaurants and construction sites.

Right.

The hardware exists and now what we need is the high level brains that can Understand language, vision, physical goals simultaneously. And so there's a tremendous opportunity here with taking it to the next level. I think Tesla and Figure are leading the way, but there's plenty of opportunity in this emerging field. The next area is AI with life sciences. We've seen advances with synthetic biology, with CRISPR and gene therapy, and lots of advances in longevity in this concept of anti aging and slowing down aging and reversing aging in a day and age where sequencing a genome costs less than a nice dinner.

Right.

And we have now open source protein models that put big pharma grade tools on our laptops. So that means that tiny teams can spot new biomarkers, design proteins and personalized therapies. And the regulators are generally supportive. The budgets are big. So you know, great for a computational biologist that's ready to commercialize some new science. So if I sum it up, it's really AI and multi omics, right? And I think that we have a lot of great new drugs to discover, diseases to solve, healthspan to increase and it's an exciting area. Similarly we have AI and material science, right? We have battery, concrete and semiconductor revolutions that will be unlocked by new molecules. We can graph neural nets that can explore trillion compound space. Now automated labs that could synthesize and test candidates overnight. Climate funds are begging for investable breakthroughs. So I think there's a lot of opportunity here to create new materials. And whether that's new materials that go inside the body, that's new materials for concrete, new kind of wood, concrete, plastic, we have the physical world gives us a lot of problems to solve using AI material science. So now let's move on to neurotech. So AI with neurotech in just a few years we saw brain computing interfaces move from a lab novelty to FDA cleared devices in under five years and again Elon leading the way there with neuralink. We have high density, low noise electrodes plus multimodal language models enabling real time decoding of speech and intent. And so there's a lot of applications here from restoring mobility to next gen gaming. You know, this is an area that I'm personally very passionate about. The name Untapped Ventures comes from helping us reach our untapped human potential. And a lot of that has to do with our brain, right? So the next area I want to touch on is AI with defense and I would put space and aerospace in there. You know, I think with everything we see based on the current geopolitics, including the recent war with Israel and Iran, the cold war, the AI Cold War with China, everything happening in Russia. There is a lot going on. So we need to stay ahead with better AI for national defense and our military. Space is the next frontier. Google organized the world's information or the Earth's information. You know, there's a need to organize more of the space and cosmos information. So I think it's a, it's an exciting area. And then you have just aerospace in general, there's a new, new types of autonomous aircraft and aircraft innovations happening. And again, I think that AI is a supporting layer that's going to help us have new breakthroughs in these massive industries that were kind of sleepy for many decades, but they are growing at a pretty, pretty good CAGR clip. Very similar to that is the field of energy, where we had areas like nuclear and solar and fusion and fission. There's a lot of activity because the GPUs are very thirsty and the grid is creaking and there's issues about what's going to happen with energy and any kinds of innovations here where AI is helping us get improvements in energy lower, improve performance lower the cost lower. The carbon footprint is going to be huge. So those are some areas that we're excited about. So hopefully that was helpful. I know there was a lot in this episode and hopefully your head is spinning but, but only in good ways. And even if it is a little bit in a bad way, maybe it'll settle in, it'll sink in and you'll get hit with the right idea in the shower in a couple days. But either way, you know, I just want to say thanks for listening. If you're early and in stealth and just kind of thinking about things and if any of what I said of any of these areas resonates with you and you want to brainstorm ideas and have a sounding board, me and my team would be happy to do that with you. Just grab some time with us. There's some links in the show notes. If you're further along and you're ready to raise a round or you're already raising one, please pitch us for funding. Would love to take a look at your full pitch and be able to evaluate that and see if we can be the right partner for you. And either way, please share this episode with the smartest builder that you know.

Right.

The world needs their talent on the field right now. There's a tremendous opportunity and so we want everybody to be on the field, not on the sidelines. And if we can help in that journey, we'd love to do that. Our team will have as much of the research that led into this, which was countless hours of research. We'll include as much of the links in the show notes as we can for every market study that that was referenced. So once again, I'm George Bandarian. Thanks for listening to Agentic. Go build something audacious and let's uplift humanity together and let's help humanity reach its untapped potential.