E335 | Mikhail Taver, Taver Capital: Deep dive into robotics and AI in the industrial sector

0:28Welcome back to the European VC podcast. that really dives into robotics and AI and industry, which is a space that's obviously very close to me. I think I've said that a million times to you guys. So I just hope you'll enjoy this episode as much as I did. Here's a few words from our beloved sponsor. This episode comes to you by the support of our partners. Portfolio IQ by Synaptic. Portfolio IQ is the most effective way to track your portfolio. It pulls data from all sources, board decks, financials, MIS sheets, forms, emails, everything. And creates a true single source of truth for your portfolio data.

1:06With Portfolio IQ, you can track all the metrics that matter, not just the standard metrics. All metrics are reviewed by a team of certified accountants. You'll get 100 % accuracy. Not sure about a number? Just click on it to see the exact cell in the document it came from. Portfolio IQ also adds a layer of standardization to make the data comparable and useful for benchmarking, valuations and reporting. All of this is as simple as forwarding an email. Zero added work for your fund or your portfolio companies. Portfolio monitoring that you'll fall in love with. Go to synaptic.com forward slash portfolio IQ.

1:42That's S-Y-N-A-P-T-I-C dot com forward slash portfolio IQ. The good folks at Synaptic will be happy to give you a no-cost trial.

1:57Tear down this wall. It's more than just an ally. This is a union of values. United and determined, we can serve as a model for other regions of the world. The nature of a problem requires a European response. Europe is a story of new beginnings. New beginnings. Let's start acting. This show is not investment advice, and the hosts of this episode may be invested in the funds and companies featured. Miguel, welcome to the European VC podcast. Hi, Andreas. Thanks for having me. Thank you. Looking forward to seeing you in Odense. That's going to be fun. Yep, yep, yep. See you there. When my peer people told me about how the conference is going to go, I mean, it reminded me of something like a Victorian sort of times detective stories, you know, when a bunch of people get looked up in a hotel on an island and then weird stuff happens.

3:00Hopefully we can make it as crazy as that. I'm not sure. Maybe we'll make it out of there. We'll try and make it something like the onion, glass onion series. I'm super excited to have you on because I'm a big believer in robotics. I've been for a long time. I think by now our audience have heard a million times me saying that I've grown up in this ecosystem a bit when it comes to the venture perspective. So for that reason, I have a pure passion for this. Maybe let's start at that question. How did you come into robotics and AI and manufacturing in industry? Well, I mean, you know, I've always been an investor in one way or another.

3:46And I've always been involved historically in sort of heavy industry deals. So I've done over 250 of those and my previous life before venture. And a lot of them involve stuff like oil and gas factories, manufacturing and so on and so forth. So I always liked it. Fast forward to 2016, we were starting a new fund. And I'm pretty bad at delegating stuff. So I knew that the team is going to be compact. So I had to have something that's got a good upside, but it's a niche thing. and I made a huge mistake by thinking that this is going to be an AI, right? I mean, we now know this is nowhere near a niche.

4:28This is pervasive. It's pretty much everywhere. It's just like electricity, as Andrew inputs it. But the first fund was focused on AI, but in all of its sort of shapes and sizes. So we got companies from sports tech, cybersecurity, food tech, industrial stuff like metals mining all over the place. So, I mean, AI was just starting there. I think it was one of the first, if not the first fund to actually have a mandate of investing in artificial intelligence. They were tech generalists, right? But not AI specialists. Well, it's been almost 10 years, 80 years since then. And I think that it's time to concentrate on the niche thing again.

5:10And when I was thinking of choosing that niche for the new farm, I thought that, you know, industry is one of the few subjects or industries or sectors where, you know, AI is there, but it's not as pervasive. It's not as has disrupted this sector yet. So there's lots of potential. We'll talk a bit about why that is probably further down the line. But I mean, I like it, which is very important, right? You have to like what you do. Otherwise, we'll get bored pretty soon. So, yeah, this is how AI and industry kind of combined. And here I am here where I discuss some of those. Yeah, yeah. Back of the year, Ian.

5:46We've done a couple of dedicated episodes on that as well. We'll put it in the show notes on EU.vc. So anyone who haven't yet picked up on how AI fits into industry and the potential there can go and listen to those episodes. But Mikhail, I'd love to ask you if you should give me some concrete examples of where you've seen AI really work in industry, what would those be? sure i mean uh well let's start with that 10 years 10 years ago to well back in 2016 and 17. uh there was this very interesting lab called yandex data factory and they had a sort of cooperation and joint bench with uh a huge large iron board works uh where they optimized the uh amount of farewell you know how steel is made right you know but you basically you take iron you add some powder to it, right, in fairness, and there you are, bam, you have steel.

6:42Now, what they did was use machine learning algorithms to reduce the number of that powder, the ferro alloy, that you put into steel while keeping the quality the same. This reduced cost by, I believe it was about 5 % to 7%, and, you know, reducing cost by 5 % to 7 % in the steelworks, in an industry, that's huge monetary savings, and that directly translates into profits or market cap or whatever. but it's big money. So that was one of the first examples I've seen. I remember from my strategy studies back in the days, what's that, 15 years ago almost, like having case studies on the iron industry because the margins are so low and the slightest bit of improvement was just completely upset.

7:26Exactly. That's one of the reasons I actually like and I think that investing in industry is good because you can add tremendous value by doing well, I wouldn't call them negligible, but rather incremental changes. That's good. That's very, very ripe for AI. Another good example where I'll be blatantly advertising my own portfolio from the current fund, but yeah, we have a company called Earth AI. They use AI to find mineral deposits and they do it faster and they do it cheaper. and I mean we're talking like a hundred times faster I mean the whole mining industry made I think yeah I believe it was 45 discoveries during the year 2023.

8:12FAA made three in half a year so I mean you know what's there not to like right another good example would be I mean there's loads of predictive prescriptive maintenance stuff around I mean it becomes increasingly common basically that's where you put a wire to you on a heavy machine and it listens to how the machine vibrates and tells you look this is going to be broken in about half a year please replace it before you have to stop the whole thing and restart it which is often well expensive so I mean there's the three examples there's plenty more but I mean we can talk about that for ages but let me know could you share a bit on the usual VC metrics of success when we're looking at Yandex and EarthAI?

9:02Anything about their fundraiser? I was lucky enough. This is a funny thing. I always say that I always read my emails. Always do. And always read my messenger. So outreach does work in my case. EarthAI had a summer intern, a lady who contacted me via a LinkedIn messenger just out of the blue and we invested. And about half a year down the line, they went into Y Combinator, which was 2019, right? That was a good year. They went there. Bam. Your valuation raises a lot, right? They're not doing their new round. They keep making discoveries. So it's just, you know, I love, I'm enjoying the rate. am i right in saying that uh earth ai are the uh guys also featured by packing mccormick on not boring uh yes these are the ones i mean yeah if you've seen that long read yes i mean it's a bloody book or something right this guy is amazing and he's an inspiration to a content creator like myself uh so we'll make sure to put that in the show notes as well so everyone can go in and learn a bit more.

10:21Yes, stock up on coffee and popcorn or whatever. It's a long journey. Whatever is your thing. Yeah. Okay. But then tell me, so you just laid it out very well. There's not really happening disruption in these sectors. Maybe you could talk a bit about some of the impediments to disruption happening in these sectors. I know that one of your core pieces is the communication lacunas between startups and corporations that's really keeping this from moving. Yeah, well, I mean, the whole industry has not been that much disrupted, probably partly because it was always too large for these things to come into.

11:02It was more of a private equity play, right? And PR boys play it safe. They don't really like things being disrupted. They want, you know, stable as a bond thing. but yeah there's lots of communication lacunas and that's what I found out when I invested in FAA and that was a that's why I thought that was a great opportunity right I mean I've heard a lot from corps when you you know especially in the heavy industry so I have a lot of examples where they just don't understand each other I mean let's start with the communication gaps between startups and corporations, which is one. The startups don't normally understand the real scope and scale of a corporation, right?

11:48Let's say you made it through all these investment committees, procurement departments and all that, and you strike a deal. You go, okay, I'm producing oil and gas pipe monitoring drones. They love it. They go, well, we need, say, 10 ,000 of those by end of this year and then they go oops i can only make two with my screwdriver and garage and you know i've not spoken to factories i don't really know how to go about this that's it they just don't understand each other uh startups don't understand corporate sales cycles i mean in their weirdest sort of very conservative i guess if uh uh scenarios they think that you know they is going to take a whole year to sell something no it takes two years three years five years and This is normal for industry, right?

12:35I mean, there's nothing wrong with that. It's just the way things work. I literally yesterday was on a call with a fellow creator now. He used to be a venture-backed startup founder. And he said, like, and now he's then helping other founders navigate the hardware space. Because what he said, I've been working with all the big guys in windmill production, uh vestas siemens every one of them and every single time they met a milestone where they had thought that they would get a contract it would be yeah but you know then the next and it's like these these stories for hardware founders that aren't used to navigating it you know at some point the money runs out and and exactly exactly and i've seen so much of it including my own portfolio that you know this is just something that you you have to educate your own portfolio i mean i might be to educate any startup but you know just plan for a different safe cycle this is a different beast industry hardware is different another thing is startups don't understand how corporations work on the inside i mean it's not just for hardware and not just for industry but it's very common pretty much everywhere i mean how budgeting works right i mean yes you made it to the budget, yes, you stroke the country, but you made it this year, this means you'll be probably, likely put in the next year's budget, which is going to be, you know, done towards the end of November.

14:07And so you'll get your first sale in about a year and a half, and your first revenues might come even later. So that's another thing that, you know, startups don't understand about corporations, or corporations don't understand about startups, but, you know, it's the startup's a responsibility to adjust to that. Yeah. I'm curious. One of the things we love talking about on this podcast is GP LP dynamics. And I'd love to ask you, so as an example, my good friends at Joint Capital, they have a lot of industry in their LP base. And part of their value add to founders is, of course, that even in their due diligence process, in their vetting process, they connect them to their industry LPs.

14:50and then sometimes the investments end up happening, sometimes it doesn't, but they've already been connected. So great value out there. But at the same time, this is also a way to very quickly get potential, both acquirers in the end, but also just important customers into the process. Do you do something similar? Well, I don't have industrial piece in the first fund, but I'm planning to have some in the second one. but I'm not a fan of having LPs that sort of get that much involved in the fund's operations. I mean, I'm happy to ask for help if I need to, but if it's too much of it, if it's too proactive, then I always wonder, why invest in the fund that pay management fee?

15:35If you know better than I do, then you save yourself some money and invest yourself. But I mean, I'm always using my network. They don't have to be my LPs, you know uh i still have a huge sort of uh notebook with all the necessary contacts and i'm happily using it to help my portfolio style steps plus so of course i use that to do due diligence because i mean you know i i'm not going to claim i'm an expert in you know oil and gas or metals or ai for that matter of anything but i i know who to ask for expertise and that's the main I think that's how I advise my LPs and portfolio companies. And this is a great example of the different models that are in venture, right?

16:20And what I just say to all the emerging managers I talk to that are thinking about taking on corporate LPs, it's always like it just changes what you do with the service you deliver as a VC. because if you take strategic investors in, they will be coming in with a different mindset than the pure financial investor you just described, right? Well, I just want a return on my investment. If I wanted anything, I would be involved and not pay management fee. But they do it the opposite way around because it's somewhat of a long-term commitment to have almost consulting or at least network partnership around how we develop and innovate and so on.

16:59So it's like so interesting to just see the two different models and you're describing it as because you're enjoying it probably say not I would love these guys and you say no I never I still come from the sort of large company background so I know how that works and I mean I see that on some boards of the portfolio companies where I see where you have a sort of young startup or sort of fund that mainly invested in the early stage plus myself and some other colleagues that come from large corporations and industry is just so different. I mean, that's good. You have a balanced sort of view on the board.

17:35That's important. But it's so noticeable. I was surprised to see that. Yeah, it is incredible. But maybe this is a good pivot or a good segue into talking about then the communication lacunas, to use that word again, between startups and VCs. Because that's also, now we just spoke about it between startups and corporates. But there's an equally big problem in the fact that if you're a startup in the hardware industry space, then most VCs will not be exactly susceptible and ready to take in what you're coming with. Oh, you're going to have fun with that. I mean, I certainly did when I was monitoring the fundraising.

18:15I mean, VCs nowadays, they used to invest in startups that do SaaS or solutions or apps and stuff like that. And they think through that prism and they try to interpret hardware industrial startups through that prism. You know, you have a geologist coming to tier 1 BC going, right, I'm a geologist. I extract minerals from earth. They go, okay, what's your LTV? What's your customer acquisition cost? I'm a geologist. Sure. Yeah, I understand that. What's your LTV? And this is just a dialogue that goes nowhere. And I mean, it makes it hard to fundraise. It makes investors miss very good investment opportunities just because they speak a different language.

18:56And this is probably, I hope this is an opportunity for me because I mean, I speak both. I tend to speak both. Plus, you often have VCs, again, that they don't have industry experience, industrial experience, right? So like I said, they either decide to not invest, which is bad for them, probably bad for the startup. But even worse happens when they do invest and they start demanding the tempo, the pace of, say, acquiring traction that you normally see in a SaaS from a sort of deep tech industrial startup that takes five years to develop the product, let alone find the product part. That's even worse.

19:36And now imagine these kind of investors sitting on boards and actually making those decisions and implementing them. That's a recipe for a disaster, I think. So that's a lot of misunderstandings. This is going to change. I mean, like I said, the VC people are just coming into the industrial space. I think they started to notice that probably a couple of years, maybe a bit less than that, when Elon Musk started talking that, you know, oh, we need lithium, otherwise we're not going to get batteries for our EVs. This is how we started to understand that, you know, it is important to see what's on these days and you can make money on that.

20:13But now you have tier one coming in. You have, I've just seen the news that Iron Spring Ventures, That's a Texas fund doubling down and making a second fund up to invest in industrial. So, yeah, it's heating up, right? It starts to get VC attention. So it's very good, but not yet. I'd love to ask you two questions. One is, how do you work with these journalists? Because the journalists can be incredibly valuable, both in terms of they bring network and oftentimes the generalists start the larger funds. So they also bring the capital to be able to follow on and take a company to the next stage.

20:56But they also bring oftentimes, because again, they're the tier ones, they also bring the VC accreditation of a big brand that comes into a company. So what do you do to kind of help ensure that everyone's aligned when they come in following you in an investment? Do you do it with work with the founders? How do you make sure that you're bringing on a partner that understands the game they're getting into? I always work with the founders very closely. I'm a hands-on type of person, which is why I normally tend to have no more than 20 companies in my portfolio. There's as many as I can manage with my style.

21:39I'm not a great type of person. But yeah, it takes persuasion. You know, you read some psychology books or go to some psychology, group psychology, group decision-making classes. That's useful for boards. That's useful for negotiations. Just talk, you know, work with people. Investments, especially in VC, I always say that it's a lot more than finance and financial mathematics. It's a lot more about people and psychology and stuff like that. So just work with them. I don't really know how to say that. No, no, but you're absolutely right on that. Now, before we close, I would love to ask you about regulation, because regulation have always been something that limits technological progress, and it definitely has an industry as well.

22:29So I'd love to ask you a bit about that, because I know you have some big and important opinions. Well, I mean, yeah, I always said that, you know, people keep asking me as a tech investor why don't we have flying cars now and I always reply that it's not the technology that does not enable us to have flying cars or whatever the technological progress it's the people that limit that it's the regulator so I don't know well we're in Europe so let's take EU AI Act the infamous or famous wherever you put it, depending on how you look at it. Is it infamous or is it famous? Well, I don't know. I would say it's infamous from my view, right?

23:20Because, I mean, what's the concept of unacceptable risks? I mean, I always quote that example. Is nuke acceptable? Nuclear power, right? I mean, bomb is not acceptable. It's a weapon, right? Mass destruction. But nuke energy, is clear energy, is good. And, you know, with the European approach, it's what outside, right? You can't really wiggle yourself from that. It's forbidden. Verbotten. You can't do that. That's no good. It's going to be incredibly interesting to follow what the EU will do now that we can get AI on our Apple phones. That's the difference. That's not the AI act. That's the DMA.

24:02That's the Digital Markets Act. that's why Apple does not surf all over. It's probably going to be both of them, right? Because the AI ad, no one knows, right? But they're just looking at their focus and what the fuck what is it for them? 10 % of their market? They'll say, okay, we can probably spot it. Plus, you have this DMA stuff that, you know, they're now beating Microsoft for bonding teams. I mean, why beat anyone for bonding teams? I mean, you have to feel certain for them. So yeah, you're not going to have new features from Apple. So yeah, I would say that the same thing. But the funny thing is that the US started to move in that direction as well.

24:40It was that SB 1027 from California. I think that, you know, they have some weird points. I don't know why they follow the suit. But I mean, starting from stuff like, you know, making modal developers responsible for their use, which is totally different in spirit to this very famous Texting 230, right? Where the platforms are not responsible for the users generating that on the platforms. You have those thresholds that rely on computational power necessary to train the model. I mean, it's like progress does not exist. I mean, chances are you're going to have to adjust that tech every month or at least every half a year or it's going to become outdated and just not work on one hand.

25:28On the other hand, you know, some say that, you know, regulation should be totally even tougher because you have this very well-known example of one of the co-founders of OpenAI, Ilya Suskever, who left because he felt that, you know, it wasn't safe enough. It wasn't regulated enough, but you can't rely on corporations to regulate. You probably need government to step in, but do it a bit more intelligently, right? I mean, I will say that it's best. You cannot regulate R &D. That's the best approach. I mean, you let the scientists develop whatever they develop. You regulate the end use. Because otherwise, you'll just stop the progress in your own block or your own country.

26:12And the scientists will just flock elsewhere. I mean, they like flu. They go wherever there's less trouble to work. So they just lose them to Israel, like in the case, or China or whatever. And we're not even touching on costs of complying with all those acts. I mean, you're a small startup, you're just bootstrapping, you know, you're paying yourself, I don't know, a thousand-year salary, but you have to hire a lawyer to make sure that you're not breaking any laws. That's crazy. That's not right. Yeah, you're completely right on that. Yes, so that's a big topic and something that I want to talk more about on the podcast.

26:50We're going to have some people that are experts in tech regulation over the next coming months. So keep an eye on that, everyone. Now, I'd love to ask you about another contrarian belief that you have, because I asked you beforehand, what do you have of contrarian beliefs? And you gave me a good one. So I'll ask you to just tell us about that. Yeah, I mean, if you use sort of a common logic, rational economic behavior, then, you know, with times are tough, companies are likely to, you know, start looking for ways to save, to cut costs. And this is where automation comes in. This is where AI comes in and this is where technologies come in.

27:32Right? That's how it should be according to the books. But it doesn't. And I mean, this is not just applied to technical projects. I mean, you know, if your sales fall, chances are you're going to start, you know, sacking your marketing team, which is the counterintuitive thing to do. But that's how companies do. And I've seen a lot of even my portfolio companies that, you know, oh, good, there's a crisis. You know, companies need to start saving. We'll come in and offer them a solution that doesn't require people. No, it doesn't work. The first thing that gets slaughtered is the innovations budgets because people want to play it safe, especially during crisis.

28:08and this is where psychology contradicts purely rational economics. Yeah, and you're absolutely true and it has a bunch of ramifications for what products should be made slash what products have a good timing in the market. Advice to young people in the industry? Don't skip school. Honestly, I mean, you know, there's a lot of legends of successful founders being dropped out from university? Yes, but no. I mean, there's a lot of foundational knowledge. You have to study. You have to study a lot of theory. You never know how that affects how you think. You don't even feel it, but it's still going to affect how you do things.

28:57And another thing I would say is don't replicate other people's actions. There's a lot of popular stuff like mentors or just watching someone, how they do it and replicating their actions and probably even achieving the same results for some time. But you have no idea what the educational background for that person is or why they do what they do. So at some point, you'll just get a very different result because you don't know how many factors they consider. And this is, well, it's a very basic example, right? You see someone falling, I don't know, in a hell of the ground and being saved by throwing a rope down there, right?

29:37You say, ah, that's a good technique. And then you use it on someone who got stuck in a tree. Oops, the person's dead. I've never heard that analogy before. But it is quite fitting. That said, I'm a big fan of mentors. And I always advise people to find someone who you really admire and try and get close to them, try and see what they do. Of course, sift it, make sure that you take it in and take it for what it is and build your own path. But I'm a big fan of mentors. Am I to understand that you're feeling that we're maybe over-indexing on mentors a bit in today's society? Well, I don't know. I mean, like I said, the room is different.