WEBVTT 00:00:00.000 --> 00:00:20.080 00:00:20.501 --> 00:00:24.041 Hello and welcome, everybody. This is Joe from Celebrate.io, 00:00:24.201 --> 00:00:27.681 your startup podcast, YouTube blog and internet radio station from Germany. 00:00:27.881 --> 00:00:32.621 Today, I bring you the second interview with Carsten, the current Business Angel 00:00:32.621 --> 00:00:34.941 of the Year for Germany. Hey, Carsten, how are you doing? 00:00:35.641 --> 00:00:39.761 Yes, I'm fine here in Zanzibar at the moment. I'm on holiday still. 00:00:40.541 --> 00:00:46.441 Yes, we do have a little game going on if you break into sweats before the end 00:00:46.441 --> 00:00:50.281 of the interview because you had to turn off the air conditioning. 00:00:50.921 --> 00:00:53.921 For this interview. So it'll be interesting. 00:00:54.881 --> 00:00:59.421 We may tell our audience that you're the Business Angel of the Year, 00:00:59.441 --> 00:01:02.981 selected by BANT, the German Business Angel Association. 00:01:03.821 --> 00:01:07.501 And they're usually elected, I do believe, in October and November. 00:01:07.761 --> 00:01:15.441 And the special thing about it is that they are awarded on the recommendation 00:01:15.441 --> 00:01:17.101 of the startups you invest in. 00:01:18.581 --> 00:01:25.901 Yes, and also the Minister of Economy of Germany is head of the jury. 00:01:26.881 --> 00:01:31.841 Ah, I see. Which will be subject to change soon. 00:01:32.441 --> 00:01:32.941 Yes. 00:01:34.461 --> 00:01:40.401 Okay, I see. We already talked about in the other interview that you're a serial 00:01:40.401 --> 00:01:43.821 entrepreneur yourself, I love that you currently run Casablanca.ai, 00:01:44.041 --> 00:01:48.201 which everybody who's watching this can see how this is working here. 00:01:48.381 --> 00:01:54.501 And I want to get straight into your investments philosophy and startup selection, 00:01:54.521 --> 00:01:59.661 because you told me when we talked about this, you get up to 10 pitches a week. 00:01:59.801 --> 00:02:03.101 How do you cut through the noise and identify opportunities? 00:02:04.459 --> 00:02:07.879 Yeah, that's a very good question, and it's not so easy because, 00:02:08.179 --> 00:02:13.079 of course, you always fear missing the one hidden gem that becomes the next 00:02:13.079 --> 00:02:17.819 Facebook or Microsoft or OpenAI or whatever. 00:02:19.979 --> 00:02:27.559 And so what do we do? We have own AI that filters through the pitch decks and 00:02:27.559 --> 00:02:29.359 takes out some red flags. 00:02:30.179 --> 00:02:34.259 So I at least annotate some red flags so that we easily identify this without 00:02:34.259 --> 00:02:36.899 going through the 45 pages of some pitch decks. 00:02:38.459 --> 00:02:43.479 So there are some things in there like nobody has had any experience in sales 00:02:43.479 --> 00:02:51.639 before or they are emphasizing technology but they don't have patterns and it 00:02:51.639 --> 00:02:53.339 seems that it's not patentable. 00:02:53.699 --> 00:02:57.959 So there are some things where we say it's probably not interesting to take 00:02:57.959 --> 00:03:00.199 any deeper look at the startup. 00:03:01.919 --> 00:03:10.539 And so this is just the first filter. It tells about half of the teams that 00:03:10.539 --> 00:03:11.959 probably they're not so interesting. 00:03:12.259 --> 00:03:16.879 So depending on the time my people have, so I have one and a half people on 00:03:16.879 --> 00:03:24.459 the investment side, Roxanne and Christoph, and so they take a deeper look only 00:03:24.459 --> 00:03:26.999 at those which were not red flagged by the AI. 00:03:28.659 --> 00:03:33.559 So the AI just principally takes the pitch deck and goes through some analysis, 00:03:33.559 --> 00:03:40.699 also puts a lot of information into a kind of a database so that we can later 00:03:40.699 --> 00:03:42.459 on analyze all the startups. 00:03:42.879 --> 00:03:46.519 And so if we miss something which turns out to be very good, 00:03:46.699 --> 00:03:52.939 we learn for the next time. So after two years or so when the startup turns out to be a high flyer. 00:03:54.259 --> 00:04:01.319 I see. I'm wondering, since when do you start utilizing AI in this way? 00:04:03.619 --> 00:04:05.399 Somewhere last year, I think. 00:04:06.719 --> 00:04:12.039 So we started programming something in, I think, April or so, 00:04:12.259 --> 00:04:13.779 and worked then in summer. 00:04:13.939 --> 00:04:17.919 And we want to do more, but right now I have more other projects. 00:04:17.919 --> 00:04:25.219 So I didn't program it, but I gave the idea what to do to somebody in my team. 00:04:25.659 --> 00:04:30.179 And not in the investment team, somewhere else in my company, in my company group. 00:04:31.359 --> 00:04:34.799 And they just programmed this. 00:04:36.759 --> 00:04:37.639 I see. 00:04:38.179 --> 00:04:42.339 It's mainly an LLM that does the analysis, but you have to do some things around 00:04:42.339 --> 00:04:48.059 to convert the PDF to set up a database and those things. 00:04:49.459 --> 00:04:53.959 You focus more on a startup's technology than its founding team. 00:04:54.619 --> 00:04:58.059 Why? And have you ever been wrong with this approach? 00:04:58.599 --> 00:05:03.059 I personally believe a lot of people are saying they're just focusing on the 00:05:03.059 --> 00:05:08.359 team because they want to seek consultancy A, B, or C. or university C, D, and E. 00:05:08.479 --> 00:05:13.139 That's basically when they get interested in invest or worked at a startup A to Z. 00:05:13.599 --> 00:05:19.239 Right. So when they focus on technology, it normally means that they look, 00:05:19.379 --> 00:05:25.199 if one of the founders was working at one of the, Batman is, I think, the new thing. 00:05:25.359 --> 00:05:28.299 So the Magnificent Eight. 00:05:29.759 --> 00:05:34.219 And if one of them worked as a software developer there, then probably the program 00:05:34.219 --> 00:05:39.479 is technically good. or the startup is technically well set up. 00:05:39.599 --> 00:05:41.779 That's what most investors estimate. 00:05:42.019 --> 00:05:46.279 The thing is that most investors come from a financial background and not from a tech background. 00:05:46.399 --> 00:05:54.159 So it's much easier for them to analyze financials or teams and take a look 00:05:54.159 --> 00:06:00.339 at the founders rather than really telling if this technology is a game changer, 00:06:00.339 --> 00:06:02.959 is really something that is – 00:06:04.837 --> 00:06:10.297 Really ahead of the competition or if it's just another startup doing something 00:06:10.297 --> 00:06:14.177 with a very good idea but not with great tech in the background. 00:06:14.437 --> 00:06:20.317 And I'm one of the few ones who are coming from tech and can thus tell something about this part. 00:06:20.417 --> 00:06:27.537 So I think this is probably one of my USPs here that I can take a deeper look 00:06:27.537 --> 00:06:29.157 into technology. Of course, my 00:06:29.157 --> 00:06:34.117 team first analyzes just these standard questions like, is the team good? 00:06:34.637 --> 00:06:37.997 Is the plan good? Is the market big enough? 00:06:38.937 --> 00:06:44.637 So generally we say we invest in a total addressable market over 1 billion. 00:06:44.877 --> 00:06:50.037 So if it's only 100 million, normally we just don't take a closer look. 00:06:50.217 --> 00:06:53.177 That's one of the things that our AI also tries to find out. 00:06:54.877 --> 00:06:58.777 And that's something that the AI and the team already challenge. 00:06:59.117 --> 00:07:04.057 And then when I talk with the founders, I'm also talking about technology. 00:07:04.217 --> 00:07:11.077 Of course, if the founding team would be like not somebody I would like to work 00:07:11.077 --> 00:07:13.277 with, then we would also not invest. 00:07:16.017 --> 00:07:22.177 And of course, technology is not everything, but then it's not that you have to get every good deal. 00:07:23.557 --> 00:07:26.457 Just you have to make sure that many of the deals you get 00:07:26.457 --> 00:07:29.757 are good but yeah so 00:07:29.757 --> 00:07:32.517 you will always miss out on some deals so 00:07:32.517 --> 00:07:35.417 i i had a opportunity to invest very very 00:07:35.417 --> 00:07:38.737 early in a startup started by two 00:07:38.737 --> 00:07:43.577 pupils they were below 18 years old and 00:07:43.577 --> 00:07:47.037 they were starting the company when they became 18 but they were preparing for 00:07:47.037 --> 00:07:53.657 it and i said you need two pivots one technically and one the idea is great 00:07:53.657 --> 00:07:59.037 but technically and also from the business model, both will not work. 00:08:06.172 --> 00:08:10.212 How did they react, given that they're not very old? 00:08:10.632 --> 00:08:12.892 No, the thing is, I said, I'm still investing. 00:08:16.292 --> 00:08:21.072 But they offered me a low valuation basis, and I said, I'm going to do this 00:08:21.072 --> 00:08:25.552 with you, and I think you will do fine, because the technical guy was very open 00:08:25.552 --> 00:08:31.132 to learning, and he was already quite great, but just some things he wanted to do would not work. 00:08:32.412 --> 00:08:37.172 And from the business model, it would also not work. I calculated this in Excel. 00:08:37.832 --> 00:08:44.112 But the CEO also had a very, he was very far for a 17 year old. So I thought I invest. 00:08:44.452 --> 00:08:52.272 And then the problem was that we had an appointment at the notary and that was in Stuttgart. 00:08:52.692 --> 00:08:58.192 And I had an appointment in the evening with a politician in Berlin, 00:08:58.452 --> 00:09:00.252 a member of the parliament. 00:09:00.732 --> 00:09:07.332 And the notary called and said, we need to shift this by one hour because my 00:09:07.332 --> 00:09:08.952 previous case takes longer. 00:09:09.532 --> 00:09:13.532 And I said, then I cannot get my flight and I cannot go on the meeting with a politician. 00:09:14.072 --> 00:09:17.712 So I will have to cancel today. Let's find another appointment next week. 00:09:17.792 --> 00:09:20.632 And we didn't find it next week. We didn't find it the week after. 00:09:20.872 --> 00:09:23.692 And after about four weeks, they came back to me. 00:09:25.972 --> 00:09:30.512 We had an appointment in four weeks or so. but they said now our valuation is 00:09:30.512 --> 00:09:35.572 more than twice what we had before because we have found some other investors who wanted to invest. 00:09:37.072 --> 00:09:42.972 And then I said for two pivots, I'm not going to do it, not going to do it at that valuation. 00:09:44.552 --> 00:09:51.092 And we still discussed a bit but did not agree. And what happened is they started 00:09:51.092 --> 00:09:55.652 this and they did the first pivot very soon And then they got, 00:09:55.812 --> 00:10:03.032 I think, 2 million on a 10 million basis after less than six months. 00:10:04.217 --> 00:10:09.057 And that was quite a lot already because we were far lower in our valuation before. 00:10:10.077 --> 00:10:13.157 And I'm not so sure how much they finance. 00:10:13.277 --> 00:10:18.217 I think they have taken up something like 20 million already or so and are doing very well. 00:10:18.917 --> 00:10:23.677 So I did wrong there because I thought with the two pivots necessary, 00:10:23.957 --> 00:10:29.837 the new valuation is not appropriate because the risk is still high. 00:10:29.837 --> 00:10:36.337 Two pivots necessary and the valuation is not ultra low as it was before just not very high yet 00:10:38.497 --> 00:10:43.717 And I was definitely wrong there the other 00:10:43.717 --> 00:10:46.697 thing is yes I sometimes I just 00:10:46.697 --> 00:10:49.637 sometimes make mistakes in investing of course and that happens to 00:10:49.637 --> 00:10:52.997 everyone and there can be two sides of mistakes so one 00:10:52.997 --> 00:10:55.797 is you have missed the chance and the other thing 00:10:55.797 --> 00:11:00.037 is you invest in a company where where it doesn't pay off and the company does 00:11:00.037 --> 00:11:06.357 not do well and I also had this in the past I invested in a company that just 00:11:06.357 --> 00:11:14.297 didn't want to grow they were happy after they were slightly profitable and still at a very low level 00:11:15.397 --> 00:11:20.617 and yes we can run the company now we don't need any more investors so what's 00:11:20.617 --> 00:11:26.877 the problem and that that's also not so good so so even if the company doesn't go broke, 00:11:27.117 --> 00:11:31.077 it might still not, by far, not pay off for the investor. 00:11:33.137 --> 00:11:39.057 Talking about investor approach here, what is for you a hard no or a red flag 00:11:39.057 --> 00:11:41.257 when you are evaluating the startup pitch? 00:11:42.717 --> 00:11:46.737 So one is if I can't work with the company, that's a hard no. 00:11:47.317 --> 00:11:54.497 Also, if sometimes they think they have invented something totally new and I find out that, 00:11:55.277 --> 00:12:01.537 that what the company invented is already there or very easy to reproduce and 00:12:01.537 --> 00:12:03.417 you can't get a patent for it. 00:12:05.450 --> 00:12:10.230 So that's normally a red flag for me. So, for example, I had somebody with an 00:12:10.230 --> 00:12:16.010 invention in chemistry, and they would not tell me what it actually was. 00:12:16.190 --> 00:12:19.610 So actually, they had a discovery in chemistry. That's what they told me. 00:12:21.330 --> 00:12:26.970 And when they finally told me after I signed a hard NDA and lots of things, 00:12:28.170 --> 00:12:34.210 I found out in less than five minutes that the effect was already known. 00:12:34.210 --> 00:12:37.490 But under a different name with 00:12:37.490 --> 00:12:40.330 a slightly different approach but it didn't make any 00:12:40.330 --> 00:12:44.090 difference so it was definitely existing so that 00:12:44.090 --> 00:12:49.970 that's a very hard no so what else is a hard no so I'm not so much on this single 00:12:49.970 --> 00:12:54.630 founder doesn't work so single founder can work well I don't have a big problem 00:12:54.630 --> 00:12:59.130 with that I know it for myself so I started multiple companies normally as a 00:12:59.130 --> 00:13:03.770 single founder in the beginning and then took up some people later, but, um, 00:13:04.710 --> 00:13:08.150 uh, fact find that I actually grew to a hundred people without any co-founder. 00:13:08.730 --> 00:13:14.790 So, and then, then I hired a COO in order to just get me a bit away from management 00:13:14.790 --> 00:13:16.690 and let me go back to innovation again. 00:13:19.197 --> 00:13:25.717 Because I don't like management. And this guy really loved management. 00:13:25.977 --> 00:13:31.797 He was in a 10 times larger company before at sea level and joined us and just 00:13:31.797 --> 00:13:37.637 restructured, reorganized how we worked in terms of how we managed the company. 00:13:39.857 --> 00:13:45.237 When we talked before, you told me you have currently a portfolio of 15 investments. 00:13:45.857 --> 00:13:52.837 First, my personal understanding is that this is usually a big portfolio for 00:13:52.837 --> 00:13:54.557 German business angels, right? 00:13:54.837 --> 00:13:58.857 Yeah, for business angels, yes. So there are some people. 00:13:58.977 --> 00:14:05.277 I have found it also the frequent angel circle, which is a sub-circle or subgroup 00:14:05.277 --> 00:14:09.117 in band in the German business angel network Deutschland. 00:14:10.837 --> 00:14:17.017 And we accept people who have done 10 invests in their past so you don't need to hold 10 invests now 00:14:18.097 --> 00:14:23.817 even if you have only one or two or so right now and you have already exited 00:14:23.817 --> 00:14:31.417 or some have gone broke or so and you have done 10 investments in the past you 00:14:31.417 --> 00:14:36.057 can join very very few members of Bund have had 10 investments 00:14:36.777 --> 00:14:40.977 So, I don't know how many there are in the Business Engine Network Deutschland 00:14:40.977 --> 00:14:48.957 who could qualify but it's probably 100 or not so much more and 34 are in our frequent angel circle. 00:14:49.237 --> 00:14:52.617 Actually the one with the most investments has done more than 100 investments. 00:14:53.037 --> 00:14:58.297 We already have one with 75 who was also nominated for Business Engine of the 00:14:58.297 --> 00:15:04.037 Year and we have one more with more than 50 investments and all of those investments. 00:15:04.037 --> 00:15:08.097 And so we're talking about investments of 50 K euros and up. 00:15:08.197 --> 00:15:12.057 We're not talking about people who are just, um, by a bit by a business end 00:15:12.057 --> 00:15:17.097 of the club or so diversify their money and invest 10,000 euros each in 50 startups. 00:15:17.417 --> 00:15:19.597 So, um, so it's, um. 00:15:21.151 --> 00:15:26.091 Um, so, uh, and we have one who has done also about 20 investments, 00:15:26.251 --> 00:15:28.911 but his typical, uh, investment is 1 million. 00:15:30.111 --> 00:15:35.431 Um, so he doesn't start with 1 million. He normally starts with 250, 250 K. 00:15:36.031 --> 00:15:42.491 Um, but, um, but over time he grows in the companies and sometimes he also starts with a million. 00:15:43.031 --> 00:15:46.551 Um, so, uh, there are these people. Um, 00:15:48.011 --> 00:15:52.531 and, um, so in the frequent angel circle, I'm rather one of the smaller investors, 00:15:52.531 --> 00:15:57.211 but as compared to the typical business angel, I would say, yes, 00:15:57.291 --> 00:15:58.511 this is quite a portfolio. 00:15:59.491 --> 00:16:03.831 So I've done 21 investments, 15 are still active. 00:16:04.171 --> 00:16:12.851 Still active. Exactly. And I was wondering, which one are you most excited about right now and why? 00:16:12.851 --> 00:16:15.631 Ah that's that's actually quite a number 00:16:15.631 --> 00:16:18.911 of really interesting investments so um so i'm i'm 00:16:18.911 --> 00:16:22.091 very positive for valent that's a london investment 00:16:22.091 --> 00:16:30.691 i made so outside germany in in the uk um valent v-a-l-e-n-t um they um they 00:16:30.691 --> 00:16:37.091 detect social media attacks so you know sometimes a company has a negative social 00:16:37.091 --> 00:16:39.311 media, and that's okay if it's natural. 00:16:39.651 --> 00:16:47.351 But sometimes a competitor actively attacks the company on social media. 00:16:47.531 --> 00:16:49.531 It also happens in politics. 00:16:52.082 --> 00:17:02.442 And Valent has a number of customers who are big and who have been under or 00:17:02.442 --> 00:17:09.302 who were under social media attack from competitors or political opposing people. 00:17:09.862 --> 00:17:15.422 So that means normally there is some truth in the story. 00:17:15.702 --> 00:17:20.242 So something has gone wrong with something. So like disappointed one customer 00:17:20.242 --> 00:17:26.182 or so, and an attack normally means that somebody takes this story and makes 00:17:26.182 --> 00:17:28.942 it huge as opposed to the positive stories. 00:17:30.162 --> 00:17:35.942 And this happens. This is not just something that could happen, but it does happen. 00:17:36.162 --> 00:17:43.182 There are malevolent people who actively attack their competitors in social media. 00:17:44.362 --> 00:17:51.182 And Valent detects this with AI. And they have a very strong AI team, I can tell this. 00:17:51.782 --> 00:17:56.662 Actually, when one startup starts to attack another one on social media, 00:17:56.922 --> 00:17:59.762 I would be wondering, don't you have anything else to do? 00:18:01.542 --> 00:18:04.802 Yes, so there are a number of countermeasures you can do. 00:18:05.962 --> 00:18:10.722 And that's a project that Valen can do that cannot be done automatically yet. 00:18:11.222 --> 00:18:12.702 So some things are legal. 00:18:13.222 --> 00:18:16.822 So you can, if you find out, if the first thing is you try to find out who is 00:18:16.822 --> 00:18:22.162 behind it, But if they're really good, they somehow hide it behind a Russian group or something. 00:18:24.222 --> 00:18:28.662 But sometimes they can find out who is attacking or who is behind it. 00:18:30.504 --> 00:18:34.884 And then there are other measures like how do you work with the press because 00:18:34.884 --> 00:18:39.184 normally then the press takes up this and you have to do some countermeasures there. 00:18:41.084 --> 00:18:45.124 And how do you deal with the social media and so on. I'm not an expert for that. 00:18:45.244 --> 00:18:51.724 So you have to interview Veiland the next time in order to talk about this topic. 00:18:51.904 --> 00:18:58.364 But I think it's a great company and a former investigative BBC journalist is 00:18:58.364 --> 00:18:59.384 part of the founding team. 00:18:59.964 --> 00:19:04.464 But they're also really great in AI. So I admire what they do. 00:19:04.984 --> 00:19:09.464 So that's one of the startups. I can tell about more if you want to talk about more startups. 00:19:10.044 --> 00:19:15.444 I do have still a lot of questions for you, and we're already running 20 minutes of recording. 00:19:17.024 --> 00:19:21.424 Yeah, and keep in mind that the AC is not running right now. 00:19:22.884 --> 00:19:29.284 I was wondering what one startup investment, that, can you name one that completely 00:19:29.284 --> 00:19:31.504 surprised you with how fast they scaled? 00:19:32.844 --> 00:19:40.504 I have one where I was wondering how, I was astonished how quickly everything went well. 00:19:41.104 --> 00:19:46.424 I have one startup, I cannot name it if I tell more about it because we are 00:19:46.424 --> 00:19:51.064 under a, not too hard, but still I'm the NDA with a buyer. 00:19:52.024 --> 00:19:56.244 So, but I invested in a hardware company. which 00:19:56.244 --> 00:20:02.424 had a lot of innovation and I was fascinated by the innovation and the hardware 00:20:02.424 --> 00:20:07.544 company was they showed me their very first prototype of one part of the of 00:20:07.544 --> 00:20:14.884 the hardware they were building and so what happened was 00:20:16.893 --> 00:20:23.873 They took in investors, they tried to bootstrap and then found out that it will 00:20:23.873 --> 00:20:26.453 not be sufficient. They had some money, but not enough. 00:20:27.333 --> 00:20:31.613 And then they took in investors and I was one of the people that they asked 00:20:31.613 --> 00:20:37.633 because of AI, because they thought AI will play a big role also in their future product. 00:20:38.333 --> 00:20:43.493 And so they had the very, very first prototype of something where you need hundreds 00:20:43.493 --> 00:20:45.393 of thousands of in big companies. 00:20:46.553 --> 00:20:51.473 And it did not work stand alone. You needed some other components also and they just had this one thing. 00:20:51.633 --> 00:20:54.993 But I was fascinated by the height of the innovation they built. 00:20:56.413 --> 00:21:02.993 Then it took only 18 months until they sold to a big company. 00:21:03.153 --> 00:21:08.853 They were still like, I don't know, minimum one and a half years pre-seed, 00:21:08.953 --> 00:21:15.493 so pre-revenue. So definitely what they had built then was a fully working prototype, 00:21:15.493 --> 00:21:16.993 but in a very small scale. 00:21:18.353 --> 00:21:25.933 So like when you say I'm automating, let's say, a car manufacturing factory, 00:21:27.173 --> 00:21:34.053 and you just demonstrate one robot which screws in the screws or something at one place. 00:21:34.053 --> 00:21:39.533 So they could show a smaller part of everything, but you could see that this 00:21:39.533 --> 00:21:42.693 is going to change the world in this special aspect. 00:21:43.813 --> 00:21:49.773 And then they already sold. So that went very fast, and it was still far beyond 00:21:49.773 --> 00:21:52.793 before they could sell the product to any customer. 00:21:53.393 --> 00:21:57.393 And we thought that we have, in the current situation, we thought that we have 00:21:57.393 --> 00:22:03.953 to have the first customer. and we tried to get some customer who signed LOI 00:22:03.953 --> 00:22:06.333 or so, which was also very difficult in the current situation. 00:22:06.693 --> 00:22:13.093 But actually they had some inquiries from big companies who wanted to invest 00:22:13.093 --> 00:22:14.273 in them or rather buy them. 00:22:18.193 --> 00:22:23.053 And we went fast with one of them. And so after only 18 months, 00:22:23.253 --> 00:22:31.173 I could sell my shares again and was very happy with the multiple I got. 00:22:33.433 --> 00:22:37.273 You learned a lot when you do so many startup investments. 00:22:37.573 --> 00:22:42.353 I was wondering, what was one of the counterintuitive lessons you learned? 00:22:42.353 --> 00:22:47.813 You said you're still open to invest in single-founder companies and not teams. 00:22:48.693 --> 00:22:50.013 What else did you learn? 00:22:51.733 --> 00:22:57.933 Oh, that's a difficult question. Maybe you have to skip that if something was 00:22:57.933 --> 00:23:01.393 counterintuitive and I still had a gut feeling, you mean, kind of, right? 00:23:01.613 --> 00:23:08.553 So it's not like normally you do this, but in this case, I'm going for it. Yeah. 00:23:12.662 --> 00:23:16.602 But nothing comes to my mind at this moment. Sorry. So for counterintuitive 00:23:16.602 --> 00:23:19.902 investment, or counter-rule investment. 00:23:20.182 --> 00:23:25.082 So, I mean, normally what I said about this, so what I said about this hardware 00:23:25.082 --> 00:23:28.682 company right now was also normally when you invest in hardware, 00:23:28.842 --> 00:23:31.522 you think it takes 10 years until they have an exit. 00:23:31.782 --> 00:23:37.742 And in this case, it was not necessary because it was so clear that this is 00:23:37.742 --> 00:23:39.342 very disruptive for the industry. 00:23:39.662 --> 00:23:43.222 And they were so far ahead of anybody in this field. 00:23:46.362 --> 00:23:56.462 I see. When I talk to many successful investors, they do have personal investment thesis or rules. 00:23:56.722 --> 00:24:03.342 Do you follow some of them? Do you have some and you follow them before making a deal? 00:24:04.682 --> 00:24:11.202 So we have a list of a few things we check, But these are the things that Roxanne 00:24:11.202 --> 00:24:16.782 checks when she talks, or some of the AI checks, some Roxanne checks. 00:24:17.162 --> 00:24:24.582 And so the general thing is that we are always ready to learn and maybe still 00:24:24.582 --> 00:24:31.622 invest if something is not according to our list, 00:24:31.682 --> 00:24:35.322 but if some other parts are overwhelmingly good. 00:24:37.716 --> 00:24:42.836 So in general, for example, we invest pre-seed. 00:24:42.976 --> 00:24:46.056 So that means if the company is already running and just wants to grow, 00:24:46.376 --> 00:24:48.956 normally we don't invest. Still, I did this once. 00:24:50.316 --> 00:24:53.996 So I invested in a company called Hybrid LiDAR. 00:24:55.436 --> 00:25:00.996 And so they're not really running, but they have first customers who have done 00:25:00.996 --> 00:25:04.556 tests, runs with them and paid for that and so on. 00:25:05.876 --> 00:25:13.556 And but this is already quite a big company and but I thought this innovation 00:25:13.556 --> 00:25:15.636 is also very high so I invested also. 00:25:15.876 --> 00:25:24.336 I do a new type of LiDAR of the distance measuring which is used for self-driving 00:25:24.336 --> 00:25:30.796 cars except by Tesla who is only using cameras but it's also a great application 00:25:30.796 --> 00:25:34.016 for manufacturing and robots and everything. 00:25:34.276 --> 00:25:40.856 And they have a much higher resolution at a less costly LiDAR device. 00:25:41.276 --> 00:25:45.256 And they're currently going to the first prototype. 00:25:45.836 --> 00:25:52.636 So not prototype, the first production run. So for the first customer. 00:25:53.396 --> 00:25:57.176 I'm preparing for that right now. But they're already quite big. 00:25:57.316 --> 00:25:59.836 So normally I don't invest on valuations over 5 million. 00:25:59.836 --> 00:26:03.976 Every investor is currently getting giddy when they hear AI, 00:26:04.276 --> 00:26:08.176 which was in the past something like blockchain, quick delivery, and so on and forth. 00:26:09.456 --> 00:26:16.336 Except for those current technologies, if you could only invest in one emerging 00:26:16.336 --> 00:26:19.916 technologies in the next five years, what would it be? 00:26:23.942 --> 00:26:30.162 So, half a year ago, I would have said deep reinforcement learning. 00:26:30.522 --> 00:26:31.722 So, that's one part of AI. 00:26:32.822 --> 00:26:37.442 But it's AI that was underestimated for a long time. 00:26:37.622 --> 00:26:42.322 In my prediction, in the beginning of 2024, I said deep RL, deep reinforcement 00:26:42.322 --> 00:26:48.802 learning, will do a lot in the coming year, in 2024. 00:26:48.802 --> 00:26:52.342 And actually it only did now with DeepSeq, 00:26:53.282 --> 00:26:57.602 everybody could see that deep reinforcement learning is still a very important 00:26:57.602 --> 00:27:03.382 factor because this was basically the foundation of why they could be much faster 00:27:03.382 --> 00:27:06.022 and much better than other approaches in the past. 00:27:07.622 --> 00:27:12.382 So it's a combination of this transformer technology of LLMs with this deep 00:27:12.382 --> 00:27:13.182 reinforcement learning. 00:27:13.322 --> 00:27:18.582 So there was some deep RL also in the previous LLMs, but it had only a minor role. 00:27:18.802 --> 00:27:23.822 And now it became part of the core of the innovation of DeepSeek. 00:27:24.902 --> 00:27:30.942 So that's what I've said then. And right now, I think what is totally underestimated 00:27:30.942 --> 00:27:34.022 is robots, even though they already have some hype. 00:27:34.182 --> 00:27:39.342 But I think people cannot see how big the disruption will be in five years. 00:27:39.802 --> 00:27:44.882 I actually talked to Jan Lippert, another interview that we just published. 00:27:45.502 --> 00:27:49.802 He is Frankfurt-born, even though he was raised in the US. 00:27:50.202 --> 00:27:55.662 He is working currently on a company that combines robotics with large language 00:27:55.662 --> 00:27:57.622 models, which is also pretty interesting. 00:27:57.802 --> 00:28:01.362 And it's something I haven't heard a lot about before. 00:28:02.733 --> 00:28:06.753 So definitely you need to combine them if you want to have practical applications 00:28:06.753 --> 00:28:09.713 in smaller scale companies. 00:28:09.973 --> 00:28:15.033 So in bigger scale companies, when you need like 200 robots doing the same thing 00:28:15.033 --> 00:28:19.713 in a factory or so, you can train the robots by other measures. 00:28:20.413 --> 00:28:25.813 But if you want them to just take over any kind of small job that helps you, 00:28:26.033 --> 00:28:30.373 you need to also be able to give them commands in that national language. 00:28:30.373 --> 00:28:36.053 Like when you have them, let's say, at the building site, and it's a helper for the mazen. 00:28:36.933 --> 00:28:43.493 So then you want to tell them, like, just hand me over five more of those stones. 00:28:44.273 --> 00:28:47.413 Hand me over this tool and so on. 00:28:47.933 --> 00:28:54.093 Can you just go downstairs and carry up two more sacks of the cement or something? Yes. 00:28:56.073 --> 00:29:03.613 And you want to command them in that way, and you don't want to show them every 00:29:03.613 --> 00:29:07.093 step, which is the normal way of learning for the new robots, 00:29:07.213 --> 00:29:09.573 so they observe and they copy. 00:29:09.833 --> 00:29:14.293 But also for their internal reasoning, for understanding what you're doing, 00:29:14.473 --> 00:29:18.833 I think they're also using maybe not large language models, 00:29:19.053 --> 00:29:24.433 but the same kind of foundation models with some transformers inside or so who 00:29:24.433 --> 00:29:28.953 do represent what the person is doing. 00:29:29.573 --> 00:29:35.433 So you can reason what is it that it's doing because you don't go like, 00:29:35.613 --> 00:29:40.793 okay, now the hand moves 20 centimeters in that direction, then the fingers grab. 00:29:41.153 --> 00:29:46.293 Rather, you want to do something like he's grabbing the tool, 00:29:46.713 --> 00:29:51.373 he's taking the tool, putting it to the screw, turning the tool. 00:29:51.373 --> 00:29:53.913 So you want it to understand what is happening. 00:29:54.153 --> 00:29:59.493 And for that, it already needs some knowledge representation internally. 00:29:59.493 --> 00:30:02.973 And that's easier with language, as it seems, than with most other things. 00:30:03.613 --> 00:30:12.673 I was smiling when you started to kind of combine this robots with learning, 00:30:12.673 --> 00:30:15.593 with what came to mind are the apprenticeship models. 00:30:15.733 --> 00:30:19.713 Because if you're an apprentice in Germany, they will play some practical jokes 00:30:19.713 --> 00:30:22.393 on you in the start. But I also was a victim of that. 00:30:22.513 --> 00:30:26.853 And I was wondering what jokes the people will play with the robots. 00:30:26.993 --> 00:30:28.553 Robots, yes, certainly they will. 00:30:28.833 --> 00:30:32.953 People will want to play tricks on the robots and then laugh at them. 00:30:33.233 --> 00:30:37.673 Yeah, I see. I'm getting a little bit more on the bird's eye view. 00:30:38.073 --> 00:30:43.073 What do you think is the biggest advantage that European startup founders have 00:30:43.073 --> 00:30:44.753 over Silicon Valley entrepreneurs? 00:30:46.766 --> 00:30:52.586 So, you mean the founders in Germany or the small investors in Germany, the business angels? 00:30:53.106 --> 00:30:58.346 I would rather go for the founders because if they do have advantages, 00:30:58.346 --> 00:31:01.906 also the investors who invest in them have advantages, right? 00:31:03.086 --> 00:31:07.986 So I cannot tell exactly about the US scenery, but what I would say is that 00:31:07.986 --> 00:31:13.846 probably the founders in the US will get money more easily. but then they have 00:31:13.846 --> 00:31:16.326 to deliver more on KPIs more quickly. 00:31:16.846 --> 00:31:24.326 And when you're in Europe, probably founders who have more of a longer time 00:31:24.326 --> 00:31:30.486 until they will reach results probably have a bigger chance of doing this with 00:31:30.486 --> 00:31:34.986 German investors and European investors, so business agents especially, 00:31:35.426 --> 00:31:42.906 than in the U.S., I think, because we're not so hard results-driven as most of the U.S. 00:31:43.566 --> 00:31:48.086 Investors I've talked to so far, but there might be others also in the U.S. 00:31:49.466 --> 00:31:54.766 So maybe some more disruptive innovations, which take a bit longer, 00:31:55.166 --> 00:32:02.846 can emerge in Europe more easily because the founders will not have so much 00:32:02.846 --> 00:32:05.186 pressure to get results quickly. 00:32:06.986 --> 00:32:12.226 I see. Before we go into a very small ad break, I was wondering, 00:32:12.386 --> 00:32:18.486 how do you manage to run Casablanca plus all the companies, all the startup investments you have? 00:32:19.086 --> 00:32:24.386 The secret is people. So you have to have the right team. You have to have the right people. 00:32:24.646 --> 00:32:27.846 And I'm very happy about my PA, Wenke. 00:32:28.926 --> 00:32:36.486 Also very happy about the other people in the team who actually manage a lot on their own. 00:32:37.546 --> 00:32:41.226 And I don't have to say okay to everything before they do it. 00:32:41.386 --> 00:32:47.166 They just do it and ask for forgiveness rather than for permission before. 00:32:47.746 --> 00:32:51.706 And normally everything's right. And, of course, in Casablanca I have a COO 00:32:51.706 --> 00:32:57.766 who runs the daily business and he's also an experienced entrepreneur. 00:32:57.966 --> 00:33:01.446 So this is his third company. He was also at Accenture. 00:33:01.806 --> 00:33:09.366 And so it's Marcus Vollmer. He also was a CEO of some other German smaller companies, 00:33:09.366 --> 00:33:11.686 which were not startups. 00:33:11.906 --> 00:33:13.966 So he knows how normal business runs. 00:33:14.406 --> 00:33:22.146 And so I have a very experienced person at my side, and he's running everything 00:33:22.146 --> 00:33:24.966 in sales and normal daily business in Catablanca. 00:33:25.406 --> 00:33:29.546 And the same is true for Omicron Data Solutions. I have Albert there, 00:33:29.726 --> 00:33:35.846 who is also very experienced and a great operative Geschäftsführer. 00:33:36.966 --> 00:33:42.746 CEO. Okay. We'll be back after a small ad break. 00:33:49.555 --> 00:33:52.215 Gerring, thank you for staying with us, guys. 00:33:52.455 --> 00:33:56.035 We're talking with Carsten Krauss, Business Angel of the year 2024, 00:33:56.615 --> 00:34:00.575 current Business Angel, because the award is only the award ceremony. 00:34:00.575 --> 00:34:03.415 It takes only place in late fall. 00:34:03.675 --> 00:34:05.895 So he's the current Business Angel. 00:34:07.535 --> 00:34:12.035 Talking about this winning Business Angel of the year 2024 is a huge honor. 00:34:12.395 --> 00:34:15.215 What does this award mean to you? 00:34:16.375 --> 00:34:20.135 I was quite astonished when I finally got it. So in the beginning, 00:34:20.195 --> 00:34:24.655 I thought, yes, I'm doing very well and probably I have a good chance. 00:34:24.875 --> 00:34:28.555 So like 30% or so. And then I saw who got into the short list. 00:34:28.995 --> 00:34:32.295 And out of, I think, 83, they selected seven. 00:34:32.535 --> 00:34:35.515 I was one of them. And the others were so great. 00:34:35.695 --> 00:34:40.475 So one of them had 75 investments and so on. And then I thought my chances are 00:34:40.475 --> 00:34:42.315 like below 10% to get elected. 00:34:42.895 --> 00:34:47.595 And when they finally, I didn't hear anything before. 00:34:47.595 --> 00:34:54.875