KION, OHB and Siemens: How data ecosystems will shape the next era of industrial AI

Very good. And uh as you can see on the screen, so well done is my name is Peter and I'm looking after technology at Seammens and strategy and a lot more and many more.

Well, uh, for one, these are great partners and, uh, we going to get there in in a second. It's all about trust and, uh, we know each other for a longer time. And so we together want to master the big challenges all of us are facing. And no matter if you are Keon, if you are H OB, if you are Seammens, we all have the same challenges. Obviously geopolitics is a big issue right now. Everybody's facing with uncertainties in the supply chain. Nobody knows what's going to happen. Then we have a demographic change. We don't find enough people. Skilled labor is really there's a massive shortage around it. There's the quest for being more resilient and more resourceful when it comes down to also um sustainability and energy efficiency. So we all face the same challenges and what we said clearly is um what we have to address those challenges is with technology. We're going to solve it by being not working harder but by being smarter by having access to all the data all the knowledge that is out there and this is I think this is what unites us. We want to be successful in the respective industries and uh we want to do this smarter.

Yeah. So first off, we love partners that have really big challenges and big dreams and aspirations as Christina and Rob both indicated. So building something in space really revolutionizing the supply chain. What you need for that is intelligence. It's being smarter. It's the data. And this is why the session of this year about how do we build data ecosystems. The big difference to all the discussions we have on Gen AI and about large language models is that while we have uh large language models that are more for consumers, they are based on language as the name would indicate large language models and that language usually you find on the internet and so those language models are being trained on all the available material that is been that has been published on the internet in industry. Can you imagine? OP now going out and saying this is my latest design for my satellite constellation up there and publishes it on the internet with all the 3D designs and all the manufacturing data and all the documentation. Certainly not. Rob wouldn't do that either for his his solutions of course. And so the big difference in industry is the data. The data challenge of how do you now collect data that's coming from production, from engineering, from design and now build the next generation frontier models that we want to apply in the physical world. And the only way to do this is by building open data ecosystems. And that's why it's so great to have also now Rob and Christina sitting here where we have these partnerships now where we clearly said okay let's exchange data where everybody brings their domain expertise and the right set of data together where we as seammens then can help to bring this together in a in a way where we encapsulate that knowledge and that domain now how into models that then can be scaled universally and that's the trick because none of of us can do it alone.

Semens can't. Yes, we do produce. We produce trains, substations and so on, but we don't have enough data. OP doesn't have enough data and also Keon doesn't have enough data. But together, we're getting a step there. And in addition, of course, we can apply all the great AI experts that we have, build a model around it, make sure no IP is being violated. This is the key thing. And then start to scale it. That's the only way how to do it.

Yeah, there's this great quote that says the future's already there. It's just unevenly distributed, right? And I think uh Christina raw both of them they will bring the future here and that make it uneven. So that indeed they have in their respective industries an unfair advantage because they moved early. Um and what we showing here at Hanofa Messa is uh the first uh that that you can see where this is going to take us. It's the IGEN engineering agent which helps actually industrial automation to scale faster because you have to program them and today you don't find the programmers anymore as much and so we can significantly accelerate that in terms of quality and time and so this is where industrial AI comes already uh into play where we can show it's trustworthy it's secure it's reliable and now we just optimize this one step of programming a machine. But then we said, well, but then it's peace meal. Then you have all these many many many steps if you think about from designing a satellite or designing a forklift or an AMR. But what we want to do is the design, the engineering, the production, the operations, the maintenance thereof, what we call the digital threads, so that you actually can optimize it with one shot and really make that work. This is what Rob was already alluding to is what we call the industrial foundation model. Mhm. So the industrial foundation model thinks about end to end. It thinks about okay if I have to engineer and design a piece I already can do it simultaneously where I think about material properties where I think about the layout and the design of it where I think about how to manufacture it so that I can optimize it in one shot and this is why it's so critical because you need to have all these data from engineering and for manufacturing and for operations and you only can do this in these partnerships and this is why I wholeheartedly believe we In Europe, we do have a big advantage in the world. And that is because of the domain knowledge that we have. And so if you bring the main knowledge together with great partners uh like Keon and OB build the industrial foundation model together on a trustful way, we can scale it much much faster than anybody else.