DN #18: Dirty Jobs, AI Robotics & Why Humanoids Are a Pipe Dream (w/ Jay Kapoor)

Niklas:Hi and huge welcome to you, my lovely listeners. So glad you're here. Today, you are joining me for a chat with Jay. Jay is co-founder and general partner of VSC Ventures, a seed stage VC firm that backs founders using AI to solve data, labor and safety challenges in dirty, dusty and dangerous industries. Jay, so great to have you.

VSC Ventures:It's wonderful to be on. Thank you for having me, Nicholas.

Niklas:You've been very vocal about physical AI and robotics. Can you explain why embodied AI is uniquely positioned to solve the DDD labor shortages?

VSC Ventures:I've never heard it called the DDD, but I like that. Maybe it's something I'll have to brand after this. Sure, I'm happy to start. There's a couple of places that we can take this, but maybe I'll start where my exploration of this opportunity started. And then from there, we can talk about where I see some of the opportunities in the future going. For me, when we were starting this fund about four-ish years ago, just a quick on my background, I've been investing in seed stage companies for 11 plus years. and been in around the startup ecosystem for a lot, you for a while, when I noticed, you know, where we were as a generational cultural ⁓ moment post-COVID. there was just a real ⁓ understanding of the massive disconnect between what was happening and what we'll call white collar and blue collar labor, although I think those are very ⁓ pejorative terms. But just so people understand them, there was an immense amount of progress happening for desk jobs. and not a lot happening for deskless jobs. And that wasn't because of a lack of technology or a lack of engineering or a lack of talent. So I started digging into why that was. And a couple of things I came across. ⁓ One was massive labor shortages. The industries that really power 80 trillion of global GDP were struggling to actually staff for those roles. Then you had a generational shift that was happening of the people that owned these businesses, manufacturing, transportation, construction. ⁓ You know, they were aging out and they were moving into, ⁓ you know, either either handing the business over to a family member who was younger and maybe more more tech savvy or ⁓ selling their businesses potentially to somebody younger, you know, a search firm or private equity firm. And then the last bit of it was the AI of it all. And in the last two or three years, what we have seen is that ⁓ it's very easy to say, AI is taking over labor and not just software, but the specific tasks that AI was having a lot of success with were coordination problems, were data aggregation and analysis problems, ⁓ were ⁓ proactivity from ⁓ signals that were being captured by elites out in the world. And to me, that seemed like a perfect confluence of tailwinds, labor shortages, generational attitudes changing or generational ownership changing, and then AI to see, okay, there is an opportunity here for us to invest. And when I got closer to it, what I realized was there was a ton of great technology. There's a ton of great robotics businesses, and yet they were struggling to tell their stories. And that's where we come in both as investors and partners to our companies where we realize oh there is a mismatch here between the promise of what these businesses can deliver and what they actually are conveying to the market today and I think that's where you the space that we really play in so hopefully that gives a little bit of a foundation of why I'm so excited about this category as a whole.

Niklas:Yeah, so I agree. I think for me, I'm an engineer by training. So I've always developed software and I have a degree in mechanical engineering. I have pretty broad understanding of the field. I always, when I saw the first like vacuum robots, was a step already in that direction where I found it really interesting. And it's quite obvious that we have a lot of jobs where we put people in harm's way, where it would be very helpful. actually, if machines could take over and we don't have to put people's arms away. So I find it really interesting. What I also find very interesting is that VSC doesn't just write checks, right? You have found us with storytelling beyond just the checks. Maybe you can also talk a bit about how you ended up in that place. What's the journey? Why are you so good at that?

VSC Ventures:uh... there's that's a uh... a very long story so i'll try to give the abridged version of it uh... i have always been uh... you know somebody who uh... has one foot in the media world So beyond just having some of my own content and podcasts and things like that started my career in media and entertainment I used to work for the National Football League work for Madison Square Garden You know, you're wearing a Dolphins hat. So clearly a ⁓ Miami Dolphins fan Got the chance to see how the best brands in the world Sports teams and leagues did their brand media and storytelling and then I got to the startup world almost 11 years ago now and I said how these founders they are so heads down in their tech that they sometimes struggle to convey that opportunity out. And so I started helping some of my founders when I was at previous funds, kind of learning the craft of venture. One of the things that I was helping them with was, you know, cutting a video. Now you go on Twitter, launch videos are all the rage. But 10 years ago, I was helping cut a customer testimonial video or a, you know, a founder intro video or any of these things. So helping my founders do that. ⁓ helping craft their pitches and their stories and their narratives and their copy, know, stuff that you would go to an agency for and pay tens of thousands of dollars, I was doing as just part of being a good partner to them on the cap table. And so some years ago I said, okay, well, if I was gonna go build a fund from scratch, how would I do that? What would I focus on? And to me, it wasn't about picking one vertical, right? We're not a vertical specific. We have theses that we like. This dirty, dusty, and dangerous thesis is one I've been very excited about for the last three years. I'm spending a lot of time on energy and energy abundance and availability is another thesis we spend a lot of time on. So we have these these theses and rabbit holes we go down. But the through line through everything that we do with our companies is how do we find the founders who want to be champions for a particular industry? So I have founders who

Niklas:you

VSC Ventures:are building AI agents and technologies for electricians. And what I loved about meeting this founder was he wants to be a voice of the customer, like publicly. When you think about a problem happening in the electrical industry, you should think about conduit and you should think about Ian. And he's building his relationships and presence in that network. We have another founder who's building technology for law enforcement. And he wants to be the voice of the crime analyst or the police chief, the people who are actually

Niklas:you

VSC Ventures:working really hard to keep communities safer and he's empowering them with the tools and technology to do so. So when I'm looking for founders, I'm looking for somebody who wants to be that. And they don't necessarily have to have it all figured out, but they have to have the want, right? You cannot, as they say, take a horse to water and make it drink, right? And so for me, the opportunity is find somebody who wants to be the voice of that customer and wants to be a champion for that industry and help them through our skill set and brand media and storytelling to do that. And I do that

Niklas:you

VSC Ventures:through partnership with my business partner Vijay, we started this fund together four years ago. He started a PR agency almost 23 years ago, has helped thousands of companies through that. And when we got together, we said, great, if we can invest in these founders and give them access to all the work that the PR agency would do as part of the investment, then they're getting a massive amount of value for having us on their cap table. And that's a really unique thing. I don't really know of any firms in the PR and kind a marketing world that offer a full agency strength of service. There are folks that have expertise, I don't doubt that. But to have a 30 person agency behind you working on your company, helping launch you, helping tell your story, ⁓ I mean the impact of that cannot be understated. And don't take that from us. You'll go on LinkedIn and you'll see our founders say things like, man, having these guys on the cap table, worth its weight in gold, right? That's the kind of stuff that I love to hear, that the founders get value out of our partnership beyond just our capital.

Niklas:And I think this lately, it has surfaced a lot more because especially in the software space, software has become easier to build. And ⁓ now people just hit the point where you have to do growth a lot earlier, I think, in their journey than they did before. this whole, I think at least on X, see a lot as a trend that you say like this CEO. has to be at least semi-famous in the industry and the thought leader for it to work. Because now this organic distribution becomes more and more a key ⁓ thing for success of a company and also for scaling it. How do you think about it? Has it always been like the change in your mind?

VSC Ventures:Yeah. I think it's become more important. So I think it has always been valuable for the founder and CEO to be a champion and spokesperson for their company and their industry. You know, you go back some past examples, Aaron Levy from Box, Brian Chesky from Airbnb, ⁓ Ryan Peterson from Flexport. These folks are very active on social. They are very active in the press and media. Just to you an example of Ryan, if a ship gets stuck somewhere, in some straight beat the Suez Canal or the Strait of Hormuz or whatever it is, Ryan Peterson is on CNBC, Bloomberg, Twitter, and anywhere else that will have him talking about it. And the value that accrues to his brand is being seen as a expert and thought leader and thoughtful and sometimes ⁓ can ⁓ push back on people publicly, right? And kind of come with that expertise. it stands out in people's minds. To the point where something was happening with the Strait of Hormuz recently, and my first thought was, I wonder what's happening with shipping. Let me go check Ryan's Twitter account, and of course there he was talking about it, right? And I'm using Ryan as a very easy example, but you can find industry leaders across our industry doing that, whether it's semiconductors or it's data centers or whatever. So what I tend to say is you don't have to start like that. I think today, and this is maybe a different path to go down, don't need to have a decade of experience in an industry to be a champion for that industry. That, I think, is new. And I think AI has a piece of that, right? ⁓ Understanding, gaining depth of understanding about an industry is faster than ever. I still think you have to have first party research and communication with customers, and I do that as part of my diligence process to actually call and validate when a founder's telling me something. ⁓ Is it actually true, right? Once I actually speak to somebody who's in the industry to understand it, but I think the old idea of founder market fit and having somebody who comes from a decade in the industry being the only way that they can build a business in that industry, I don't think is true anymore. I think if you want to be a champion for the industry, if you want to immerse yourself in it, build the relationships, build the network, speak for them when there's important things happening in that category, in that industry, there's never been a better time to build your own distribution, build your own media, be founder led. So the importance of it is growing I would say but so is the ability for anyone to succeed at it.

Niklas:really interesting. then going back to your fund, you chose a very specific field. I call it DDD again. ⁓ Why exactly did you focus on that? I mean, you could have picked anything, right? So why is this one?

VSC Ventures:Yeah. Yeah, yeah. Well, it's one of the theses that I got really excited by about three and a half years ago, right as you know, chat GPT moment was happening, everybody was thinking about where is AI going to play in AI applications, software applications, law, health care, you know, all these different things. And my point of view, even back then, and I, you know, myself on the back and say it's only gotten stronger is that the more software driven a technology is, the easier it is for an AI lab to take its market share. And you're seeing it start to prove out. Claude has a law function, Claude has a design function. We've had multi-billion dollar businesses in law and design, and ⁓ both OpenAI and Claude and Gemini are coming for it. So to me, I started looking at where is the durability, where are the motes? and dirty, dusty, and dangerous problems to me. One is because I think a lot of investors didn't fully understand it. So I think the opportunity wasn't being priced in, was the sort of simplest way to think about it. The second was, I think these founders sometimes come from backgrounds that VCs don't understand. If all you've done is spent your time bouncing around companies in the Valley, do you really understand the problems of an electrician or the problems of a laundromat owner or the problems of a police officer or the problems of a solar panel installation company. No, I doubt you do. So to me, finding the founders that understood those problems really deeply and understood the technology to actually solve for it, that was the most important thing. And if they came from... backgrounds that the Valley didn't understand, then all the better for me because that was an opportunity to kind of compete and come into rounds before they were fully inflated in their pricing. As I'm sure you've been following, the median seed round is ⁓ up 2X from where it was three years ago. So that's what we as a fund have to think about. So to me, there was a confluence of I'm really excited about this thesis, I'm really excited about the founders that are building in it,

Niklas:you

VSC Ventures:there's a pricing mismatch between the value I think these companies can deliver, the multi hundred billion dollar businesses that can be built in this, and the fact that it's not being priced today because everybody's excited over here by AI software. So that's why I got excited.

Niklas:Yeah, so I generally agree. It's often easier to turn to things that don't look so attractive on the surface, that don't attract as many people. So you have less competition and then it's probably a more profitable playing field. So this is how I would think about it. If you're going in the most crowded space, competition will be crazy. And then the question is, if there's one player who can make a sizable return, it's just really hard. What I would also be interested in, think currently, least from what I'm hearing, it's not only seed rounds are up, it's also expectations on growth are up a lot. Like, what is your feeling? What does it mean for founders currently if they are raising? Is it different like two years ago?

VSC Ventures:It is a little different than two years ago. I think the progress that you need to show while you are in a fundraise is equally as important as the progress you show going into a fundraise. So by that, I mean a typical fundraise conversation process can last, let's call it two to maybe six weeks, right? Is typically what we see at a founder who started a fundraise, let's say in September, September 2nd, I remember the date. So September 2nd, he started to kick off the raise and he had a term sheet by October the 14th, right? And that's, I would say, typical. You see some rounds where you hear, ⁓ they launched the round and two weeks later it was filled. What I'll tell people is, like, don't believe all the press, because those are usually conversations that maybe were having, you know, happening in the background, and right when they kicked off a round, everybody was already warmed up and ready to go. But all that to say, ⁓ in that six-week timeline, There is a lot that you as a founder can show and prove about how fast you are moving. Speed of execution is the most important thing I hear from my co-investors today. So if you talk to a founder week one or two and they maybe give some feedback where you say, ⁓ yeah, you know, this is actually important for us. It was maybe a little bit later in the roadmap and the investor says, no, no, from what I'm seeing, I think this is the thing that's gonna differentiate you and it's gonna really matter. If you're able to ship that using all the tooling at your disposal, AI or otherwise, Within that time frame. I mean that impresses the hell out of any investor you're able to come in there and say ⁓ You gave me feedback on September 1st and I shipped something by September 15th and ⁓ it's functional and look I'm already gonna start putting in a production and you know Here's why I think it's gonna have a great impact on our business the ability to ship and ship well and do it on a timeline that impresses investors Can really accelerate a fundraise? So I think that is one of the most tactical The expectations on product, because you can create products so much faster, have gone up. And as a function of that, the expectation on how quickly you're able to accelerate have gone up. Now, there are still some businesses that will set function, right? Anytime you are doing hard manufacturing, anytime you are building... a factory line, first of a kind. Anytime you are doing some level of hard science or you have to go and create your own ⁓ training data in order to train your robotic models. I recognize it's going to take time. So that really requires having an honest conversation with your investors about what the timelines are. when you're in a fundraise, you're not going to ship a product every week to show people your product velocity. That's not the kind of business you are. So you have to show other things. Like, hey, we use these AI tools to capture data better or solve some other problem our customer was having on the way to solving the main value problem we have. how quickly you can show ROI, that is something that every investor is looking to see. And the timeline for that is just compressed.

Niklas:Yes. this is, I think, very nice to hear or interesting to hear. What I'm also interested in is when you look at the industries that you're investing in, which industries do you feel are primarily underfunded? So where would you see, where do you believe founders should look more if they are looking for an opportunity now? Where should they look if you look at the space you're

VSC Ventures:So if you were to ask me a year ago, I would have said technology for tradespeople was underfunded or underappreciated. And I think in the last year, it's actually flipped the other way. There is way more capital in these categories building for, you know, frontline workers, electricians, recycling professionals, you know, whatever. ⁓ And the number of ideas are fewer and just less interesting. You know, people are building observability tools, they're agents on dashboards, or they're building fleet management and to To me, it's somewhat uninspired. Where I think things are underfunded, specifically as it pertains to robotics, ⁓ is actual implementation. I'll give you a stat. There are close to a quarter million factories in America today. We are the second largest manufacturer in the world behind China. And 90 % of these factories still do not have a single robot in 2026. Now compare that to China or compare that to really any major manufacturer in the world. It is unacceptable. But the reason for that is not because people don't want robots. It's because they haven't figured out how to actually implement them and show ROI in a way. And yeah, we used to have systems integrators that used to do this, know, people consultants effectively, whose job it was to go and understand your plant and your factory and figure out what robot to get you and actually install it and do it there. And systems integrators are going away. That is the biggest trend that I've seen the last couple of years is those folks, unless you're in very highly specialized industries, ⁓ pharmaceutical or defense, where you have very specific specifications. the role of systems integrator is going away. And you're actually seeing that onus fall on the startup itself. So what I'm thinking about is not who's building the robots. I think we got plenty of people building robots. And that's great. You can go raise a lot of money building robots. Who is actually doing the implementation? And who's succeeding at it? And that's who I'm excited about because we will not, there is no world in which we solve this problem through human labor. In America, cannot exist because we have a population that does not want more immigrant labor to solve these problems. We have a generational ⁓ mismatch of expectations with people who do not want these jobs. Who wants to spend 14 hours in a warehouse or in a recycling plant picking glass bottles off a conveyor belt? No, that is a problem for automation. So I'm excited about who's actually implementing the automation and whether they are vertical or they are horizontal. That's what excites me right now. ⁓ And then the second one I'll give you is anything which kind of builds off the first, ⁓ anything around the data center ecosystem. If you believe as I do that the next three years will continue to accelerate data center build outs, ⁓ they are super short staff right now. The capital is there, the demand is there, getting human beings to actually build these things. is the bottleneck and it's why you're seeing SoftBank investing, I think it was $40 billion in data center automation. It's why you're seeing a lot of our portfolio companies getting a ton of demand from the data center ecosystem. So anybody who is serving the data center ecosystem right now, surprisingly, still underfunded.

Niklas:Yeah. So I'm, I believe strongly in the base idea, like get more electricity. We need a lot more solar. think that it's like the more resources we have, especially like renewable resources, I would say that are quite easy to, to harvest without changing our environment much. Like solar is a very good example. think also the idea of putting data centers in space, which I think Elon is pursuing a lot which goes in the same direction. So go closer to the source of energy, ⁓ harness it directly. You don't? Yeah, we can talk about it. ⁓ Because he said...

VSC Ventures:I'm not a believer in that, we can talk about that another time. No. Yeah, I mean, I'll tell you in a couple, no, please, go ahead. I didn't mean to interrupt. Please go ahead.

Niklas:No, because I can just say what I read, think today actually, because Elon said in three years, ⁓ it will be cheaper to deploy AI in space than it will be on Earth.

VSC Ventures:Yeah. Yeah, I think getting them up there is probably possible. I think getting them powered is probably possible. Getting them maintained is going to be a massive problem. And so that's where I think the bottleneck is. If we're able to solve the maintenance of these in space, then sure, maybe we're on the right timeline. Launch capability is there. ⁓ Obviously, energy abundance in solar is there. Can you get the right components to do it? Fine. But then you have to maintain these things. And these are very expensive assets we're talking about. These GPUs are very expensive. And so it's the reason why we made a bet on wave energy and solar energy, as opposed to, you know, data centers in space. Because to me, it sounds really sexy. And it gets SpaceX stock, you know, the private shares pumped. and people are excited about it, ⁓ which is fine. That's his right to do. But I think if you are expecting it to happen in the next decade, ⁓ I would take the over on that. I think it's going to take a lot longer. ⁓ There are other places that we can put data centers terrestrially today ⁓ that are going to have a much bigger impact than the few that we might have in space ⁓ as a pilot test.

Niklas:I think you can differentiate, right? So first of all, I don't think that it's a replacement of one of the other. I think you will still build as much solar as you can on planet Earth and get as much energy and the seasons I'm 100 % on. In space, I would question if you really need the maintenance. So if you look at, for example, the ⁓ Starlink satellites, they just let them.

VSC Ventures:Sure, yeah, yeah. Hmm.

Niklas:crashed, right? When they are finished, that's kind of what happens, right? So you let them control the burnt down. So the question is, how large would you build these satellites? I mean, you can build them small. I haven't done a deep dive, but my first take would be build small satellites and don't maintain them. Just let them crash down. That would be my take on it. So you don't have the maintenance costs. You just need to care about redundancy of sorts.

VSC Ventures:Yeah.

Niklas:Obviously you cannot lose your data, but beyond that, the advantages. then if you look at it from like a pure data processing standpoint, if you move data through styling or something anyway, then at least from a data transfer standpoint and connection standpoint, it's not worse. And you go closer to the source of energy. So the very fundamental thing, as long as you don't have to maintain it and you take that out, which I agree, which is nearly impossible, you have advantages. And you don't have to secure them the same way you need to secure a data center on Earth, because at least the human effect is taken out of it because it's so unreachable. So yeah.

VSC Ventures:There are no protests outside of space data centers. No, it's a pretty big if I'd be very curious when you actually do the math on what it costs to let these things just naturally decay. When you're putting GPU racks up there and how much those cost, ⁓ what would be the willingness to let them naturally decay as opposed to maintain them? But either way, can circle back on this and once you have the math, I'd be so interested to see it.

Niklas:Yeah, definitely. I'll take a look at it and ⁓ time will actually tell. We'll see in a few years how it plays out. Once again, I don't think it's a replacement of one or another. So I really strongly believe that we need to harness as much energy from the sun as we can on this planet to advance humanity. That's absolutely necessary. And if you look back to dirty, dusty and dangerous industries, if you look at

VSC Ventures:Yeah.

Niklas:A startup that really excited you that automated something. What's something you think is really great that you saw lately?

VSC Ventures:⁓ well, let me tell you about one of our investments that is doing this in industrial cleaning. ⁓ so let's, let's talk about that, right? You don't talk about dirty jobs. ⁓ it is impossible inside of facilities to get reliable labor to do the basics of, you know, sweeping, mopping and facility maintenance. And I mean that anywhere from gyms to malls to college campuses to hospitals, ⁓ this is somebody's job and it's hard to find these people. I'll underscore that for a second, Nicholas. There are so many people in the AI industry that are doing themselves a disservice talking about AI taking jobs that people actually want. And we're not spending enough time talking about the jobs that people just don't want to do, that you cannot hire labor for, right? And I'll come back to industrial cleaning in a second. We made an investment a few years back in recycling. They built a robot that sits across conveyor belts and recycling plants and picks materials based on the computer vision. And so as it understands, ⁓ this is a glass bottle, great, we sort it over here. It's a plastic cup, we sort it over here. That used to be a human's job. states it still is. Where human beings will stand over that line for 14 hours a day, manually sorting glass, paper, plastic, that job has a 500 % annual turnover. That means by the time you have hired somebody and they've figured out where their spot in the parking lot is, they've already quit and taken a job at an Amazon warehouse or somewhere else. Right. So that's one example. ⁓ Another one we have in laundromats, ⁓ laundromats post covid have struggled to hire the staff and talent they need to turn around the volume of clothing that people want actually clean and dry, clean and laundered and all this stuff. So we invested in a laundry robotics company. that's creating a five minute personal dry cleaner. And now they've actually put them across cities as their own laundromat. So instead of being a laundromat, the company's called Presso, you have a Presso mat where you can just walk in with a couple of coats, let's say, and each garment within five minutes, you can get perfectly dry clean. Compare that to going to a dry cleaner and it takes a week for them to turn it around, if that, right? So to me, I'm really excited about opportunities where these are that human beings do not want and should not be doing. So coming back to industrial cleaning, we started with a business called Autopilot. Cleaning is just the first vertical they're in. They've already expanded out to a few others. But when he came to me, ⁓ the concept ⁓ that he started with is I have all of these sports facilities, just simply tennis courts and pickleball courts, that need to be cleaned at a high turnover. Because you want to get, as soon as one match is over, you want to clean the court, the next match going. ⁓ You could imagine really anywhere at basketball courts or gyms, anywhere that you have people with aggressive foot traffic and turnover, you need cleaning to be done. And they just can't hire enough people. And so the robots are there and the opportunity is there. And here was a founder that was actually putting those two together and figuring out how to build, structure, finance, and deploy in a way that I've seen so many robotics companies and they build amazing technology and the customer deployment is an afterthought. And that's what got me really excited in the dirty dusty industrial cleaning. Something as simple as that, where you have massive employee turnover and low retention and a human may cost $20 an hour. And you know, they show up for a few hours in the day and a robot costs $5 an hour and it can operate 24 seven potentially. And so that's kind of the stuff where I get really excited.

Niklas:It's exciting. And what I'm always wondering is we have this more general humanoid robots, right? That Tesla is now heavily going into. And if you look at it, what do you think? Where we see, especially also in this area, like these more generalistic robots taking over that are built after humans. Or do you believe in specialization, at least for the foreseeable future? What do you think is going to happen?

VSC Ventures:Yeah. I think humanoids are a pipe dream. I think they have function inside of environments with high variability and environments that are built specifically for humans. And those environments typically are the home, not factories, not facilities, not. you know, manufacturing environments, logistics environments, those are typically standardized. and they don't change very often. When you build a warehouse, you put the racks where the racks go and the racks are pretty much gonna stay there. Same thing inside of a loading dock for ⁓ your trucks. Same thing inside of a hospital potentially. So to me, we are putting a lot of money and emphasis on a form factor that is designed to mimic humans. ⁓ The thing I always think about is if you were going to create something that was better or more efficient than a human, why would you give it human weaknesses? Why do you give it an ACL? Why do you give it an elbow? Why does it need two arms? Why wouldn't it have four? Why wouldn't it have six? Why does it need to have the same form factor as a human being? ⁓ Is it just for the cultural acceptance that we will feel more comfortable with a humanoid than a six-legged wheeled spider bot? I don't know. But to me, ⁓ the opportunity inside of the industrial use case, which I think is way bigger and near term than the, you know, consumer use case, which is robots in the home, I think is going to look specialized. And I think it's going to potentially start with a baseline of existing technology, you know, be it a humanoid, be it ⁓ six degrees of freedom, you know, two arm robot, whatever it is. ⁓ But it's not going to be the Neo or the figure or whatever. I think those are great. And we saw with Boston Dynamics that you can get some really great content out of them. But to actually deploy them at the scale that these companies believe they're going to be deployed, ⁓ mark me down as a skeptic. I think near term, we should be focused on actually solving the problems that these customers have. And that starts with purpose-built robotics.

Niklas:Yes, So my background is mechanical engineering and I also did a PhD in production management. I understand the manufacturing environment quite well and there are robotics you have to say for completeness in manufacturing for long time. It's basically it's one arm that can move at a certain degree of freedom usually, right, that you can steer and that will have a tool or set of tools that you can exchange and they are integrated in a manufacturing plan. I would first of all agree that within an environment that you can define yourself and you build a manufacturing plant for purpose. ⁓ It probably makes sense to use a purpose-built robot. You don't need the flexibility of something walking around. If you look at an assembly line, for example, where you have a human, you would still have to do the math on is the human weed actually cheaper, including maintenance, versus somebody you hire. That's what I told people when we did. productivity analysis back in the day. when we did courses and coachings for people in manufacturing, I always told them, maybe you shouldn't worry so much about automation today. Because actually, is also, and that's also what you said earlier, like deploying robots ⁓ also carries a cost, which is maintenance. You have to deploy them. They are less flexible. There's usually a trade off discussion on this, so that's not trivial to solve. And if you look at humanoid robots, I think the market is huge. So that's clear, but it might not be in manufacturing in the first place. Maybe it's somewhere in warehouses, maybe somewhere at home, as you said, or outside. Maybe it's something else. But I would also say, especially in manufacturing, it's more like purpose-built robots.

VSC Ventures:there's an immense amount of trust that you're going to have to have trying to get a humanoid into somebody's home. I have a toddler, he runs around and bumps into things that are you know, walls and furniture. Now imagine running into a hundred thousand dollar appliance or whatever it costs. Let's say they bring it down to a reasonable cost. ⁓ But that could topple over and injure him. you are going to have to develop an immense amount of trust that these things can be in the home. And I think that's that's quite far away. So I understand what they're trying to do, which is to say, Well, we are quite far away from having it in the home. But we're not that far away from having it in the factory. We're not that far from having it in the loading dock. And the argument that I would make there is that is the exact wrong place for it to go. That is exactly where purpose-built robots make sense. And so, yeah, there's just a fundamental mismatch between the vision that these companies are selling and the use cases that they are going after, which is what feeds my skepticism.

Niklas:And you can fully automate a warehouse today. So this is also something that there are usually reasons why you don't do it, but you can fully automate it. There are robots for warehouses, for manufacturing. These things exist. It's not a completely empty market. So yeah, really interesting. If you looked at startups currently coming up and the labor trends, you think a startup wants to go and go in a place where there's an immense labor shortage and solve a problem there. What do you see where the largest labor shortages currently are and what will become worse over time?

VSC Ventures:I think the labor shortages are myriad. ⁓ I think the shortages where you have long lead times on training humans and where there is a lack of structured data and knowledge that can be passed on for automation ⁓ is where these labor shortage are going to get worse. So the example I'll give you because we've already invested in this category ⁓ is electricians. ⁓ We are losing four electricians for every one new one we train. So even if we start fixing that problem today, everybody listening to this drops what they're doing and goes to, you know, vocational school to become a certified electrician. It will take them a few years to actually become functional and useful as a, you know, electrician on the field, right? Can actually be deployed. The second bit of it is the data that you need in order to inform decisions. Where do I move this wire? What does it actually feed, you when a data center goes down ⁓ for even an hour, that is half a million to a million dollars in lost revenue. every single time a data center goes down. And we are overloading these things and overclocking these things and pumping more and more energy through these chips so we can drive more compute. And we don't have the electricians to fix it. Don't take it from me. Jensen Huang on a podcast about a month ago said the biggest bottleneck that we have at any given time is the numbers of plumbers and electricians. I've already made an investment in electricians. I'm looking at plumbers next. Just to try to figure out when there's a lack of data and there's so much tribal knowledge that goes into actually feeding how you make these decisions on the field. How can we as a population serve this problem? So it kind of ties the two themes that I was talking about earlier on, which is dirty, dusty, and dangerous, right? What are the jobs that human beings just don't want to do or aren't equipped to do matched with where is the tailwind today? And that is the tailwind for intelligence. And that comes from data centers. So everything being built around it, insofar as it intersects with dirty, dusty, and dangerous, it's an exciting place for me to dig in.

Niklas:So, and maybe as a final word, if people now are building in the space, are listening to the podcast, they want to reach out to you. How do they best reach you?

VSC Ventures:I am always on Twitter and LinkedIn, you can find me. I also host a weekly talk show called climb, where I interview ⁓ investors, founders, industry experts that are building companies in this space. We've done you know, three episodes now on AI safety, security, and, and privacy. We have done conversations with experts that are building, you know, the copper wiring interconnection at the data centers, all the way to the application layers that all of that compute serves. So if you're interested in understanding what is happening in AI and why it matters to you as a founder or an investor, come check us out on Climb and you can find us on YouTube at ClimbVSC.

Niklas:Awesome, Jay. So good to have you. Thanks for being on the podcast.

VSC Ventures:This was so fun, Nicholas. Thank you for having me.