From AI-assisted design to digitally enabled sustainability, manufacturing is being re-engineered as a living system.
What strategies will define the next generation of globally competitive factories?
At Davos 2026, leaders described “factories that think” as cyber-physical systems built on live digital twins, AI agents, and robotics—optimized end to end from design through operations and trade. Siemens CEO Roland Busch framed the core play: physics-based digital twins that remain “alive” in production, continuously fed by real-time data to boost output and sustainability—citing “20–25% higher space productivity” and “20% less energy cost.” Siemens’ partnership with Nvidia aims to industrialize the required “technology stack” into an AI-based operating system that can be adapted across industries and applied in greenfield or brownfield settings.
Koç Holding CEO Levent Çakiroğlu stressed that impact is already measurable: a 70% productivity gain in a car plant, 13% OEE improvement in oil refining, and 95% forecasting accuracy through integrated planning and finance. Automation Anywhere’s Mihir Shukla argued for outcome-led adoption: “Chase purpose, not promises of technology,” pointing to rapid ROI (including $200M inventory savings) from autonomous agents managing decisions humans cannot scale.
Agility Robotics CEO Peggy Johnson highlighted humanoids’ emerging value in “unstructured environments,” while safety and cybersecurity remain gating factors. Panelists converged on “security first,” warning that “the greatest risk is the human,” and on the need for governance, talent, and long-term R&D commitment.
Good morning. Thank you so much for being here this morning. We're talking about factories that think I'm Jamie Heller. I'm the editor in chief of Business Insider. I am here with a terrific panel. Roland Bush, president and chief executive officer of Siemens Germany, Levent Kakeru Cakiroglu, chief executive officer of Koc Holdings Coach Holdings. Colt Holdings. Okay. From Turkey, we've got Maria Shukla, chief executive officer, chairman and co-founder of Automation Anywhere, and Peggy Johnson, chief executive officer of Agility Robots Robotics. So thank you all for being here and thanks the audience for being here. Roland, I want to start with you. Siemens and Nvidia have recently announced a partnership to build an industrial AI operating system. Could you tell us what that is and why it's important?
Okay, I try to make it short. So let's assume that you want to build a new manufacturing line for electronics. What you do is you simulate the whole thing. You make a digital twin of the product you want to manufacture of the production line, which you're going to use. So of the whole, the whole environment before you even start building it. And that allows you to optimize the product, the manufacturing line, in a very comprehensive way. This is where it starts and here we go. So this is used digital twins. And we talk digital twins. We talk really physics based digital twin. So the behave like the real thing when you when you heat it when you run power on it when you even when you run software on it. We can simulate how software runs on silicon. So once you have it, then you start sending your excavation machine. You build your plant and you run it. Output is 20, 20, 25% higher space productivity, 20% less energy cost. You have much, much faster in ramping up. You don't make mistakes.
If you use the digital twin. Exactly. You see these.
Then then comes and you want to have this digital twin alive once you are running your production because it allows you to optimize all the time. If you make a change, if there's something going wrong, you see real time what's happening on the shop floor because you have a digital twin which is still alive, is feeding by the real time data, and they can optimize if a new product comes in a new component, you know exactly what's happening. So and for that you need a technology stack. You need data. You need a very solid software simulation software. You need an operational software which runs your manufacturing side and applications which help you operate. This is what we call the operating system, the industrial operating system, AI based. And you need a lot of technology. This is where we are partnering with Nvidia. They've Omniverse, they can make a photorealistic view of this whole manufacturing line. So you look at it as if it would be a video and you can really simulate everything, including, by the way, the people working on the shop floor. So it's not only the machines but also the people. And that's very powerful. And what we want to do is we want to make this operating system as easy as possible to use for our customers so they can scale it and use it, in a greenfield application. That's what I explained. Or brownfield, if you start, you don't have to start from nowhere. You can start at any point in time.
So the.
Operating system is going to be a product that you're going to sell to your customers.
It's more if you if you consider a technology stack as a product. Yes it is. So a product means that it's not. It's you cannot plug and play it for each and every application. You really have to to adapt it to your individual needs. Obviously makes a difference. Whether you manufacture a car, you're on a semiconductor plant, or you make a yogurt or a pharmaceuticals. It works in all cases, but it needs obviously specialties. They differ. But but the principle that I explained is always the same whether you build a product or whether you simulate a molecule which finally runs in a drug manufacturing plant, the whole the whole model is the same.
Got it. Okay, I want to come back to this, but first I want to introduce one more participant, the Minister of Foreign Trade of the UAE, Tony Ahmed, also. Thank you. Thank you for being here. Let's get right into it. We're talking about digital twins and AI systems. Oh, you're still waiting. Okay, so I'm here. Can you just bring us up to speed on the digital twins like they've been around, but they're getting better. How how important are they? What's what's exciting about them?
So you're absolutely correct. The digital twins have been around, but the technology has exponentially improved and the computing power has exponentially improved. One of the unique advantage manufacturing has is that they have been collecting data from every machine and every system a lot longer than any many other industries. So there's an enormous amount of information feeding all of those information with new, more powerful models allow you to create what Roland was talking about in Digital Twin that exhibits all the characteristics of the physical system, the and with the power of AI agents. This is not I'm going to expand the theme from not factories that think, but they think they connect and act. Right. It's it is proactive because let's say digital twins tell you that there is going to be a problem in manufacturing. You can proactively act on it. It has implications on your inventory. It has implications on your shipping planning. And so can it. Can it alert everybody and do countermeasures across the because that's where the costs are, right. So how can it proactively operate with AI agents. Can can can instantly react and bring that resilience across entire, entire supply chain and operations and factories are at the heart of it. Right. And so if you if you catch something there, you can make everything better.
You have factories, you have plants all over the world. What we're hearing, how how real is it right now or how much of it is something that's looking to the future?
We are operating with quite diverse businesses, and digitalization is a group wide imperative for us. We scale what works across home appliances, automotive and oil refining. Currently, we are operating with six ref recognized lighthouses in addition to value creation in terms of quality, productivity, efficiency and sustainability. And these factories are engines for us to transfer already proven capabilities. In that respect, what we have been talking about are the ones that we have already applied and benefited from. Take digital twins. Digital twins are emerging as productivity and flexibility engines in the world that we are operating, we have to be, operating in real life, data and decision tools in that respect, in manufacturing, digital twins technology enables us to increase productivity and flexibility at our, car manufacturing plant. The productivity increased by 70% and overall equipment efficiency increased by 13% in oil refining. By using digital twins and integrating yield, energy and blending decisions together, we have been able to increase the capacity utilization in supply chain, end to end supply chain management, combining planning and financial decisions. We have been able to increase the forecasting accuracy up to 95%. So it is real. It is working.
It's happening now.
It's happening now.
Are you ready to go, Minister? How how impactful is this for global markets and global trade? Is it is it still just a regional phenomenon or what are you seeing.
Well, it's.
First of all, the digitalization and the digital twinning, as we've been saying, is coming along and it's coming not only in factories, but it comes on all, value chain of the, the productions, lines from, the, the extractions to the manufacturing to the logistics to the distributions, including the, the customs clearances, etc.. We at the UAE, we have, very interesting experience because we've been historically investing heavily on logistics and our global network. We're connected to 250 ports around the world. And we're continuing growing this network because it's a critical now what we did at back Home in logistics, we started transforming all of our operations in the customs and ports to be a digital. The clearances for the majority of the consignments happens before they arrive. The the payments or the the deposits returns. It's an immediate, action. They don't need to wait for days. The clearances used to take a few days. Now it's just a matter of few minutes and they're they're done, etc.. After the the sophistication that we have done internally now we're taking this experience globally and wherever we manage, we do that and we associate most of the industrial zones close to the ports where we manage, because that is going to make things much faster and movements of the commodities that we produce there. Now, what we have done in the factories or the manufacturers, sector in the country, eight years ago, we, we start having, dedicated technology aspects within the ministry of, manufacturing. And, we were sending a message that technology has to take over the the huge manpower, non-skilled labor that we're bringing from abroad just to make sure that we reduce. It's not anymore about the wages. It's about digitalization. It's about improving the efficiency of how we're running the operation, how we have done it. It's it's transformative. We're applying it. We're deploying technology, and we're moving forward. Some of the examples in oil and gas, oil and gas, for sure. You cannot you don't. You produce something which is far away, kilo kilometer, below the, the the the ground. So what we the simulation assisted us in improving the efficiency of the productions, producing where the right wells and controlling the pressure of the wells. Now, moving forward, we're applying the AI. The AI tells us where to produce and etc. we don't need anymore the the simulation engineers to tell us where the AI is giving us. And we have already built that sophisticated system. Now we want even to the operation. We start applying the technology in robotics and and the the sites where we used to have a shutdown for, for 3 to 6 months now by robotics, we're just applying it while doing the operation. He's doing the reports, is doing the surveillances gives us the the the situation and apply it without having the shutdown. That used to take months. Now we're just a matter of a couple of days and the operation is back moving forward as well to some of the heavy industry aluminium, for example, we start applying the the AI in smelting refineries and that is improving the efficiency, the production, etc.. So for sure the industrial nations have a huge impact, just a matter of the will that you want to do it and you have to do it because it is coming. And we with the with the challenges that undue economic is facing, the world without digitalization will not be able to reduce the consequences of those measures.
Okay. So you need the will. And I'm Peggy, I want to get you in here because you have a humanoid robot named digit. Yes. And, robotics has been around a long time. It's just starting to make some major breakthroughs. But it's hard, right? I mean, it sounds really good, but can you talk about some of the challenges of doing this safely and.
Sure, sure. So humanoid robots are highly complex. They're a little bit like EV vehicles. They have to take in a lot of data. They have to process it quickly. They have to make decisions, which is why we're just starting to see them clearly. AI has helped supercharge it a bit, but you still need a dynamically stable big device that can add value to a factory. So it's been it's been slow in coming, but I would say the one of the turnarounds recently with humanoids is that they can operate in unstructured environments. Many robotics are unstructured. Unstructured.
That's the that's the most recent development.
Yeah. And what I mean by that is they can step into places that are built for humans and do human work. It's largely material handling now, moving bins and boxes and things like that. Item manipulation. But that's in contrast to sort of fixed robotics, maybe conveyor belts or single arm things. Those are obviously very functional and produce a lot of output. But in a humanoid you can have something that has a lot of operational efficiency because it can do one job in the morning and something else in the afternoon, and it can go where humans go. It can go down narrow aisles. The world is built for humans, and so it can reach up high. It can go down to the floor. And that puts it in a category different from some other robotics, which are really meant to be static and in one place and do one job very well. So now that they can operate in these unstructured environments, and with the advent of AI, we can kind of supercharge the training of robots and get them out there and to do more and more things. And as their capabilities grow, they'll start to move beyond enterprise and eventually into our homes. But you mentioned safety. That's a factor. They have to be meet a very high safety bar to go from a factory into our homes.
Because right now in factories, they're mostly in like cells, right in like.
Correct? Yes. They are meant today. All humanoids have to operate inside of a work cell, and you might see videos and demos and, you know, making coffee and doing backflips and things. But when you actually put them to work, they're automated machinery and they're £200. They can lift 25kg or £50. These things are powerful. So they must stay inside of a work cell away from humans. And the bar that you have to meet to get outside of the work cell is called being functionally safe. So you want to be able to be in close proximity with humans. But to do that, you have to not harm the humans when you get close to them. So it has to sense that my robot has to sense it's approaching a human. And as it gets closer and closer, the radius gets smaller. It'll bring itself down to the ground and it'll keep carrying its payload. It has to do all that balancing. When the human passes by, it will stand up again and move on its way. We'll have one of those by the end of this year. A functionally safe robot. So then it can run around the factory.
Just just to support that. What you are talking about is the caged robots. When you typically when you see a manufacturing line for cars, the welding robots caged and they do a repetitive task as precise as possible, and they do it extremely fast, but it's always the same. It's always the same, just welding the same parts over and over again. This environment is completely different. So you don't need really too much sensing. Security is is defined by the cage. In that case, it's a completely different environment. The input tokens are sensors, different environment. And robots act accordingly. That works only with AI technologies, correct? Yeah. And that's that's what what you're talking about. And it starts really in logistics, where you see a lot of handling stuff, but that that moves more and more into different areas of manufacturing.
And the problem we're trying to solve is the some of the jobs that people don't want to do is there's.
A couple of things. One is it's very hard to hire for these manual jobs. They're dull and dirty, and they're dangerous at times because you're lifting over and over again, very repetitive. It's kind of mind numbing work. Typically, people turn over within a year, very quickly. They, you know, they, they're moving along. But the injuries are also another thing. We have an aging workforce. A lot of young kids don't want to go into environments like this. So the older employees are also getting hurt more. They've been doing this manual work for longer. There's high costs there. And so those two things are really driving the need for humanoids because it's hard. It's so hard to find humans to step into these roles. And really, humans shouldn't, you know, lift things over and over again and hurt their knees and backs. There's other jobs. There's higher order value that humans can offer versus just moving, you know, item manipulation and moving boxes.
Minister, I want to come back to you. You said there's so much potential. If you have the will. What what what do you mean? Is it money? Is it political will? Is it, public sentiment open to this? What is the challenge?
See, the the most important thing is the, you know, the issue, the challenge, and you work toward it. Sometimes you have the money, but, you haven't diagnosis the right, solution to the problem that you're having. I'll give you an example. As a nation, we welcome everyone, and we are attracting the talents. And one of the challenge that we have, we're heavily dependent on the unskilled laborers for the constructions, which is exactly what, my colleague was talking about. It's it's becoming aging. It's too difficult to attract new young people to start working on the constructions. And robotics is the main solutions. Are we going to stop there? Absolutely no. The construction is part of the development of the nation. And we're going to have we're having the largest development projects in the region. So we're heavily investing in robotics factories to ensure that it does this, this, this job, instead of continuing bringing or looking for for the such, labors finance is critical, for sure, but many countries are having the finance, but they don't have the, the, the right solutions or the right diagnosis diagnosis to, to, sort out the issues. What we have done in the country, we came up with very holistic approach when it comes to robotics and AI, starting with the government appointments, the policy developments, capacity, buildings. But the main game changer that we have done is the R&D investments, which is usually a very costly matter. And many, many, stakeholders are avoiding that because they said someone else will do it. I'll just copy paste it. But when it comes to AI, it's too late. And digitalization and technology, it's too late to wait for someone else to do it because you're going to be out of the game.
Can't you just hire Siemens?
We're working with them, but you have to start working on R&D because you ensure to customize things to your own ecosystem, your own environments, your own conditions, etc. and then you start deploying and taking things forward. And the the finance is combined with that because you need to have, the entrepreneurs supported to bring those ideas and R&D, thoughts into practicalities and commercial phase, which we're doing through different stakeholders. We have we're applying in the country the two models, the the deployments and the investments, model where we invest in international technology and bring it to the country. But also we have the consumer market where we apply anyone, anyone's technology in the country who would like to pilot an experience. They're, they're they're, projects.
So you're you're opening your doors to.
Be able to pilot.
We're becoming a regulatory labs to everyone.
A laboratory.
And at the same time, we're doing our own as well, investments which which match our own interests, etc.. So coming back to your question, the will is there and should be there, but you have to know what you want and you have to wait for a long term. You cannot just do it in overnight and think it's going to happen. Because if you don't have that patience and consistency, you will not be able to achieve whatever you want.
Within your company. How much are you doing of your own investments, your own development, versus working with partners and other providers.
Who, of course we need to work and collaborate with technology partners. But in addition to that, we have been building up our capacities, capabilities in digital space. In that respect, I can give an example of platform 360, that we have developed in-house. And, it has been deployed in more than 40 sites. It scales, execution and decision quality across the businesses as complexity increases, the need, to speed, accelerate the response speed and consistency become a real constraint. In that respect, this platform standardizes the diagnosis and decision across, businesses, and it is one of the most important aspects, from our point of view, to scale from pilot to the enterprise level.
Is the ability to evaluate and make these decisions. That's the technology that what is the technology you're saying that is so important to scale across?
We have been, deploying artificial intelligence, machine learning, IoT, all the digital twins technologies that we have talked about. But it's important to have a single architecture, to utilize, what we have as a digital capacity across the businesses. One of the aspects of, intelligent factories is to be able to scale what is achieved in pilot studies, and it is not about deploying more technologies, it is more about strengthening the governance, operating systems, talent aspect and the culture and leadership within the organization to make sure that we benefit from our capabilities across the board.
I want to go to the audience for questions, but first I just want to come to you, which is, we are talking about R&D and the importance and trying new things. And yet companies have a business to run. They have numbers to make. And how do they what are you hearing from your clients? Like what's they probably feel a lot of pressure to get with the program and get going. But they also have their their daily obligations. And also, is it better to be starting anew with fresh technology or just trying to like AI ify what you already have?
So I think the the we have thousands of customers in manufacturing vertical alone, running AI agents in millions. The possibilities and the benefits are enormous. And some of them are within three months. So you can unlock a vast amount of capital. Let me give you some examples. Take an example of Sumitomo rubbers. They started using us in the planning and then in inventory management. And eventually as they started connecting AI agents, one of their biggest benefit came in container planning and shipping, because if you optimize that effectively, there was a huge dollar amount. Take another customer. Cargill. Just, just just in order management and order to cash processes you can unlock. Let's say if you do a straight through order processing, increase that percentage by by 20 or 30% more, you can unlock a vast amount of revenue stream. So by by targeting the specific areas, I'll pick one more on inventory management. We have a European customer, and we were trying to see how to create an autonomous inventory management. And we initially we were trying to mimic what human beings were doing. And it occurred to us, why are we doing it? This is a lot more powerful. So we created AI agents that dynamically manage inventory in 32 warehouses and 685,000 items four times a day. There is no amount of human workforce that could do that. So, and it saved them $200 million in inventory cost. So each of these things could unlock a vast amount of capital that is being vested in this current system. You take all of that, you take the will. You take the multi-year planning and reinvest it into into a sustained competitive advantage. This is a this is a phenomenal time to be doing this.
So it's really not optional. And just like you could just keep like, getting just better margins. It's a simple stuff. Do you have any questions in the in the room here. Over here please just let us know if you don't mind standing up your name, where you're from and your question. Thank you.
Hello everyone. My name is Anita Tejuosho. I'm a global shaper from Nigeria, and I work for Nigeria's Presidential Initiative for Unlocking the Healthcare Value Chain. So that's trying to boost local healthcare manufacturing. And one of the ways that we do that is through technology transfer. And so my question is one, I want to learn more about the work you all are doing in Africa, not just as a market to sell to, but as a partner to build industry and also to know what what defines a great partner when it comes to technology transfer. What are you looking for when you want to look towards sustainable growth of industry in Africa specifically?
Do you want to take that or.
I can give it and can give it a try. Maybe I give one example and what what we are doing in Africa, big time is and it starts in northern Africa, in Egypt. Here we, we have, we have the largest project, actually, Siemens ever did in the history, which is about 2000km of railway lines, from the rail to the signalling system, electrification and the trains, the rolling stock, the stations connecting Red Sea and Mediterranean Sea connecting the north and the south down to Sudan. So in this changes the whole the whole economy of this system, because we're connecting 90 million people and there's a lot of technology which is going in. This includes, of course, by nature, technology transfer, because finally, you operate the system on the ground. You have people how to know how to how to make a kind of a overhaul of your trains run signal and systems. Actually, it's European technology which comes to the country. So there's along with that project, it's not only boosting the economy because it obviously changes the logistics systems completely. The way how you connect people, including obviously call it tourist line, which is connecting Hurghada in Luxor. But but it really brings technology to the people. We do that, by the way, in, in training people, Siemens invests 430 million just training our own people on new technology, AI, technology. We do that every year, but we are opening this platform also to our customers and our partners to take advantage of the technology, which we have, because it requires a lot of knowledge. Maybe one word on Africa. This is not about AI technology. This is a channel one. When you when you develop, when economies develop, they develop in four ways. Energy, logistics, infrastructure. So communication mobility then comes industry and then comes higher education healthcare. It goes concurrently. But you see these waves and and not before you have the first two. You really are able to build up substantial industrial technology. And that's maybe one of the biggest problems of many, many countries in Africa that they are in wave one and two. Egypt is moving straight forward to obviously the next one. So and that and then last point here, the advantage is now what I encourage is if you anyhow building that up now in that world we are living in, just try to leapfrog and use AI technologies all the way from the beginning. It's much, much easier if you do that right away from the beginning rather than have a brownfield where you start migrating. Keep it short. It's a long.
Story, and you mentioned talent and training. One of the things we were able to do is give free licenses and capabilities to the region. And in one particular case, we trained 700 women. And the great part is that they don't have to unlearn and they can. And this technology are so natural to learn that it doesn't take four year degree to learn them. Right. So there is an instant mobility. And out of that 700 women, 525 found a job in first week. This is remarkable, right. And this is a genetic AI jobs. So the I think there are multiple dimension to it, but talent is a very important one.
We have operations in Egypt and in South Africa, factories. And in addition to the products, with the highest energy and water efficiency, we also transfer the capabilities that I have mentioned here to those factories as well. And those, definitely create value in the countries that we operate in. And that includes investment in our talent in order to make sure that the technologies are being actively used and further developed by our colleagues, maybe out of, intelligent factories, I can give an example, which is close to my heart. Our colleagues have developed, refrigerator, which is powered by solar. It is quite vital in rural areas where, the access to electricity is very limited.
And a refrigerator powered by solar. Right. You better have a open roof.
Here because, you know, South Africa, Egypt are more, well structured when it comes to investments. But we start as well as a nation investing in those LDCs within Africa. But it's not the answers will not be only applied to Africa, but everywhere. There are a few things which any partnerships has to be to take in consideration the long term. There is no disturbance on this investments. That's a critical matter. Second one is energy, which you you mentioned. You have to ensure that there is an enough supply of energy to, to to power those manufacturers, technology, supplier areas, etc.. Third one, availability of land. Sometimes lands are not owned by governments, owned by people. So it's going to be a very challenging to to have access to the land to do those factories land, land accessibility. And the fourth one, logistics, which is even if you do the the investments, you have to ensure that there is connectivity between wherever you are to the ports or to the closest, area, fifth one, which is as well, I see it as an important one. And to guarantee at least the, the, movements of the project is offtaker, is it going to be the government guaranteeing that or the private sector? And the last one is the availability of talent, which is, an issue nowadays because everyone is trying to pull the talent when it comes to technology, to their, to their parts. So it's an ecosystem. You cannot just take one, one part of it and keep the others. If you don't have all, all of them, you don't move on. And some of the countries outside Africa as well, they have the the lands. And they gave us the, the, accessibility to power. But the logistics is nightmare. So you cannot move on on that. So you even if you have the the main ones, you have to look at the whole thing because otherwise the visibility of the projects will not be there.
So a lot of challenges that great question. Thank you so much. Do we have another question in the room.
And behind you as well?
Thank you so much. Please stand up and say your name.
Hi, I'm from India and, I lead a home textile manufacturing company, which is one of the largest in the world. And I manufacture for 60 countries across the world. My supply chain is very complex. It starts from the farmer to spinning to weaving to processing to cut. And so my question to, the panel here is when do you start looking at Digital Twin? And do you look at it in a holistic perspective. Because I deploy 20,000 people, I've actually put in, you know, data, you know, where machines can speak to one another. How do you use it? And how do you look at using digital twin? And when you talk about, you know, business, you're talking about return on capital employed as well. How do you look at it? So I think for me the question is when do you start deploying digital twin in a complex supply chain where your customers are sitting in United States of America, UK, Europe, Japan or Australia, and you are looking at end to end.
I start I run a technology company and I would tell you I wouldn't start with technology at all. Digital twin the right answer is maybe not. I would look at, there are only three outcomes you can have. You can either increase revenue, reduce manage cost, or reduce risk. Right. And so in your business where is the biggest benefit maybe part of running a resilient supply chain. You can reduce cost effectively and generate $50 million of benefit. I would do that first and use that money to fund everything else, create more oxygen for the, so I would prioritize based on business outcome. I, we often tell customer that chase purpose, not promises of technology. And I run a technology company. But that's otherwise you will always have a gap between AI's promise and impact of AI. And the diffusion of AI. Technology is a means to an end. Amazing technology, by the way, but means to.
An end.
Do you have a.
You have a. Let me add, I mean, I know a little bit of textile industry, but but you rephrase it and you have it starts with your supply obviously. Then you, then you manufacture I assume you have a lot of manufacturing machines, which are, which are complicated. And you want to let them run 24 over seven, obviously. And, and then and then comes the demand. The let me for the in the middle part. If it comes to machines, this is where I believe you could really start working on. And you don't need a digital twin of your, of your textile machines. But what you need in the first place is collecting the data, how they operate. So start with the data layer. So you collect the data and then you can. And this is the beauty about it. Once you have a good call it operating software which controls your machines, you can throw AI on it. And they tell you this machine is going to fail in a week from now. So you and they tell it even for that reason. So you can order a spare part. This machine doesn't stay a week. It stays maybe just an hour because you know exactly what's happening. This is possible with AI technology right now and again. And this this makes your CapEx run 24 over seven makes a big difference. I'm not talking about getting more in autonomous operations, which is in some cases easy and difficult. I'm pretty sure you have your suppliers for your machines. And that's something where I mean, we have this technology we could, could, could support also to do that, that piece. The other one is the supply chain. If you really want to see how do you get this whole thing running. Where can you cost, can you cost out. Then you are more in the ERP system. Which you which you obviously run as well. And the best is if you start tearing down data silos, I'm pretty sure you have data silos and you supply new machines and you sell and start start tearing these silos down. Critical data warehouse, data lake, whatever. Then you can once you have that, you can do exactly what you said. You can connect the data AI hubs, you and, and this is super powerful because it opens you. It's a view on your operations, which you never had before.
I mean.
That might have just paid for your ticket to. To Switzerland. Some great advice there. We have one more question I think back here.
Thank you. My name is Julia. I'm also a global shaper from Italy this time. Good. I work in human robot collaboration at the Italian Institute of Technology. My question is that, I mean, technologies are fundamental, bringing them to the factories, but I think that increases the it's the exposition to cybersecurity risks. So just imagine if a malware stops production. That's a great damage. Right. So I was wondering, how are you addressing this problem if you are. Thank you.
You want to get in here, Peggy? Because these robots are pretty risky.
They are. I mean, when you think about what your mobile phone can do with the sensors on that, this is a different level because of the sensors that are on robots, obviously there they have lidar, they have cameras, they have hearing devices. They are machines inputting data. It's a lot of data that's very, very sensitive. So from the very beginning of our development of the robots, we knew that that was going to be an issue for any customer. So we keep as much data on board the customer or offload it locally to the customers, sites inside their factory. And we don't own any of the data. We, with their permission, we can help process it for them. For instance, off in the cloud if the robot needs to make a decision or something. But the data is the customer's. And that was sort of the principle we went in with, so that every the way we built the robot was to keep all of that data protected, to keep the comlinks protected, and, ensured that, you know, others bad actors couldn't get to it because clearly that's very competitive data as well.
I would also I would also say that as part of design, the you want to design it so that where do you use artificial intelligence and where there are dynamic reasoning and where you lock it in so that it does not deviate? I use an extreme example of a cruise missile and it can dynamically decide many things. It's kind of a robot, but it has a fixed destination that does not change. That's not dynamic because you knew it before you launched where it is going. Yeah, there are some things you do not want it to be dynamic. And in this exuberance of AI, sometimes we replace natural intelligence with artificial stupidity. And that's that's the wrong answer, right?
Did you want to pop in here because. Yeah.
So.
So hard to beat that.
I mean the before before digital technologies, robots and whatever came in very often you see customer our customers running a manufacturing site disconnected, disconnected from the rest of the world except one, one router, which is which we throw every cybersecurity on it. What you what you have. So it's an island that doesn't work anymore. I mean, that goes end to end. You start somewhere in the sensor, a thermostat, which you can hack eventually all the way to the cloud. So the attack angles are now 360 degree, much, much more than before. So you have to start from the from the from the beginning, how you design your cybersecurity landscape. And what we do is we deploy technology, which is snorkeling on each of of the bus systems, which we run in a manufacturing site looking for patterns which are unusual. This goes all the way to the cloud with all the cloud services, which you get from from your hyperscalers, from your suppliers. So it has to be an end to end kind of solution, and each bits and piece has to come together. Most importantly is you. When you you cannot avoid somebody attacking you, right. Cannot even avoid somebody going behind your firewalls. But what you can avoid is once they are behind that you stop this thing because you detect something. And this is it takes normally I mean, minutes or maybe half an hour or hours before really damage can happen and you stop it before that. So, there's much more to talk about it, but it's a super complex and, and, and we have a huge ecosystem of partners, new technology startups. They come up with crazy ideas and huge technologies. We are very much tuned to do that and use it as as much as we can.
I think you almost have to design security first.
Exactly right.
It can be afterthought. It cannot be afterthought. This is your question is so important.
From the government point of view. We started early on cybersecurity. And the other thing which we did, we built the right protocol and we start making sure that people are aware. You cannot just work in isolation as a cybersecurity threat. You have to know to make sure that the employees, anyone who is in the factories or in the government, employees, they are aware of the protocol and what are the steps that has to be taken, and for sure the protection measures, etc.. But the most important thing is to start from the beginning. Cybersecurity has to start hand in hand with the whole development, otherwise it's going to be too late and ensure that the awareness is going and the development is continuing. It's not does not stop somewhere.
And the most important reason for an attack and a successful one is the human.
Yeah.
That's that's where the greatest risk is. And that's why it's so important what you're saying to have well-educated humans on this. Yeah. Okay. Lots of risk but huge potential. Thank you all for an amazing panel. Thank you for being here. Have a great day.
Thank you, thank you.
It was great. Thank you all so much.