Rockwell Automation, Inc. (ROK) Earnings Call Transcript & Summary

Good morning. I'm Andrew Obin, Bank of America's multi-industrials analyst. I also cover industrial software. So the next meeting is very exciting. I think -- to me, it should be one of the most interesting meetings for the conference. And with us, we have Cyril Perducat, Senior Vice President and Chief Technology Officer, Rockwell Automation. We also have Aijana Zellner, in the audience, VP Investor Relations and Market Strategy. So I think Cyril has a couple of slides. And then we're going to go into a fireside chat, and then we'll take questions from the audience. Thank you, Cyril.

Good morning, everyone. And thanks for having me here, Andrew. I just want to start in a couple of minutes with some pointers about what is Rockwell Automation, so to baseline this for everybody. So Rockwell Automation is one of the leading global players in industrial automation and information solution with especially a very strong established presence in North America that represents 61% of the revenue. But we are a global company with more than 50% of our workforce outside of North America and a presence in 100 countries, either because we have local customers or because we follow our North American customers where they are building capacity or new things. In our portfolio, we have a combination of product, hardware, software and services. And historically, hardware, that now becomes more and more intelligent devices, that's why we named the segment that way, it represents roughly 50% of our portfolio. But we have had a very strong focus, both organically and inorganically to grow our Information Solutions business, in particular. We are diversified in terms of the end market that we cover. The starting point and the history of Rockwell Automation has been with discrete industry, like auto, tire. But more and more and over the years now we have a very strong established position in hybrid industries. Pharma, food and beverage, life science will be typical of this. And we have also in different part of our portfolio a strong established presence in the process industries. One element, in a minute to speak about our model. We have different product platforms. Those product platforms are not end market-specific. They are horizontal to the different segments and geographies that we address. Our business tends to be very global by nature. So there are less products that are specific to a geography or a very limited number of them. And we are doing right now to move around this. One move is to say from those product platforms, our customers are asking more and more to deliver outcomes to solve some very specific problem. I met a customer awhile back that asked us, "We want to increase in North America our capacity by 20% without adding a lot of people. How do we do this?" So that typically is the type of projects that we will have to activate all the different parts of our portfolio, software, hardware, services altogether, to build solutions that address some vertical needs. And we are productizing those solutions. This is what we call Industry Solutions. Maintenance-related, workforce management-related. So different categories where our approach is not just to wave capabilities in front of customers but really to say there are problems that we want to be known for as being best to solve them. And that's what -- where we do industry solutions. And at the same time, because our products are more and more digital, I would say, ultimately, all our products are digital in one way or another, we are building the digital stack which we call the digital platform. It's not a product, it's not something that we sell to customers, but that's a common digital foundation of digital products and it's very important to enable scale because digital is all about scale. Everybody can do experiment here and there with digital technologies. Doing the same scalable capability across 200 factories for a customer is where challenge is. And we are -- beyond technology, we are really working a lot on our go-to market acceleration and I will say in 3 aspects: talents. With, I will say, the transformation of the market overall makes industrial jobs more attractive, but still our customer find that having -- finding and keeping the right talent is challenging. In all case, we have this very interesting value proposal. We have this very interesting attractivity. We are in multiple hubs and multiple position. We are today able to attract and reasonably retain the talent that we need to drive the transformation that I've described to you. We do a certain number of acquisitions. The latest one that we have done will come on the next page is around information solution, data science and AI, and we will continue to be on this. And we also expand our partnership network. Those customers that are starting to build new capacity in North America, expanding in North America, they rely on us because we have an established footprint with our ecosystem. We have integrators, distributors, panel bidders. We have all different set of capabilities beyond our own technology, not just to consume technology, but to build the capacities that are needed. So those are 3 important aspects to accelerate our go-to-market. In terms of acquisition, we continue to do a bolt-on acquisition of different size. The latest one we have done is Knowledge Lens in India, and this is about data science, AI and consulting, which both allows us to address some customers locally, especially in pharma and life science, where they have been historically very strong, and India is -- this company is in India. India is a strong market for life science. But this also reinforced our global execution capability for our digital consulting practice that was built based on the acquisition of Kalypso that was done a little bit earlier. So we are already growing those global digital consulting capabilities to support our business. And I think I'm going to stop here, Andrew, maybe and we go to questions.

Excellent. Do you want to? Okay. Yes. So as I said, I think the story is actually -- Rockwell story tends to be very simple but also very complex. So this conversation...

Nothing is very simple.

Right. We can speak 4 hours. So I think the first question is going to be quite open-ended. And as I said, we can probably speak for 4 hours about that. So how does the company like Rockwell grow its digital capabilities, right? And specifically, maybe we can talk about, going forward, what are your key priorities in software? And for me, and I know I keep hammering on this question, but can you discuss differentiated software strategy within key industry verticals? And specifically, I think what I was thinking is, how do you build your software presence organically and through M&A across your portfolio with things like Cognite, Plex, CUBIC as sort of examples of building out, filling out the white spaces in your portfolio.

If I answer to this in details, there will be no chance for any other questions.

That's fine with me.

But we can try.

Yes, yes, yes. That's fine.

Okay, so where do I start? Where do you want to me to start in all these sequence of question?

I know. So as I said, so look, I think you represent not a change but a big evolution to Rockwell's approach to technology and digital strategy in general. So just maybe you can outline your vision for growing Rockwell's digital capabilities over the next 3 to 5 years.

Okay. So first, I want to highlight the starting point. When we speak about digital transformation in the context of our industry, even if digital transformation means different things today, but the history of Rockwell Automation is a lot about digital transformation. When you look about the capabilities that we have built using open network technologies on the factory floor versus proprietary technologies, building an ecosystem around software where you can extend with partners that use our development environment to build solutions. So it's not like the journey started when people started to speak about digital transformation. Now there are new elements to digital transformation. Cloud, AI, the edge, new software architectures that are emerging. We are just amplifying and accelerating this. I think one of the central theses and ideas that we are developing at Rockwell Automation is the evolution of automation from something that is about repeating task in a very reliable straightforward way. This is what we are known for, that's what automation is about, to release this idea of industrial autonomy. Not industrial autonomy necessarily in the sense of building lights-out factories that have no workers and everything, but really visualize this in the same aspect as autonomy for electrical vehicles for cars. It doesn't mean that you will have tomorrow cars that will drive on the road by themselves, but things about assisted safety -- active safety, all the things. So being able to provide in process the right level of optimization by bringing more data, more information solution, more AI in the process so that we have industrial processes that are not just doing stuff but that are learning, adapt and improve continuously and technology partnering with people, I think that's super important, so that you empower people to drive business outcome in shop floor, which has other consequences like making those jobs in the shop floor more attractive. We are -- in order to achieve that, we are combining organic investments with inorganic investment. I can give you 2 examples in that regard. The core of what Rockwell Automation does is the integrated development environment that we have, that we have developed over the years, was historically on-premise. We released at the end of the last year the first in the industry cloud-native development environment. And it combined, it has a cloud element and then it works with our existing solutions that is on-premise and customer values this for collaboration across teams, scalability, up down, all the advantage of the cloud. So that was an organic investment, multiyear project, to be the first in the industry to release that to the market, which we did in December. MES, manufacturing operation management, we went for another playbook as there was an actionable target with Plex. We acquired Plex, who was also the first cloud-native SaaS MES on the market. So we combine organic and inorganic investment. When you think about the relationship between Rockwell Automation and software, you need to bear in mind that the vast majority of our development capabilities today are software already, whether this is embedded software, what runs in hardware or application software, but we are already a large extent a software company so the bulk of our R&D investment is about software development.

Excellent. And so maybe we can talk about going forward, as we think about in both organic and inorganic. So clearly, sort of building the hybrid model on-prem, SaaS, so that's a big focus. But can you expand on your key priorities in software?

Yes. So I will say functionally, when you look at the space of software, it's a lot about moving from providing the tools that customer need because of trustability associated to regulated industry, being able to observe trends in process to really develop a software stack that help customers address their complex optimization scenario today which is multi-dimension and not necessarily linear. And let me explain. When you look at what CEOs are speaking about today, they are speaking about resiliency, sustainability, efficiency. They want to achieve all of this at the same time. Those 3 things contradict each other. How do you maximize outcome while minimizing energy? You are not going to slow down the conveyor because it's just [indiscernible]. So all these needs information management, data science, but in a way that is productized so that it's usable by the actual talent that our customers have. So being able to productize solutions so that they can be used up to the point that complex things like data science, AI and other things are almost invisible to the user, but they can benefit from the tool is very important. And then, of course, in terms of market coverage, continue in software to develop or design and simulation capabilities organically, inorganically is certainly interesting. Continue to develop what we do in the manufacturing operational management space, MES in particular, continues to be important because those dimensions are important to running and optimizing this core process that we described. When it comes to control, before you ask me the question because I know you will, I think, overall, there is a kind of shift of the architectures. We are more and more embracing software-defined architectures versus hardware-defined architectures, not because we believe that our hardware is soon going to become a commodity, I don't believe it is. Think about other sectors that have been through those transformation at networking, switch didn't disappear, they're still here. But really in the sense of decoupling the life cycle of the hardware from the life cycle of the software so that we can more, in a faster way, evolve software. And it could be because of customer needs, think about security patch, or it could be because we have additional value to offer to our customers so we can deploy an app in the system and maybe ask the customer to subscribe to that specific app and then generate annual recurring revenue. So we maximize optionality in the architecture, but we also maximize optionality in terms of business model to capture back part of the value that we create with those new outcome-based solutions.

And I do find when we talk to investors, it is a complex topic. So maybe you can help us a little bit. So I think you sort of explained Plex how this was your entry into SaaS, MES and I think you sort of were able to leapfrog a lot of your competitors. That's my personal view. But maybe other couple of examples recent. Maybe we can talk about Cognite and CUBIC. So how does that help you sort of fill out white spaces in your product map, sort of to give us an indication where Rockwell could go.

Okay. So they are pretty different in their nature and [indiscernible]. When I look at the partnership with Cognite, we looking at -- we are partnering with Cognite to develop what we call an industrial data hub. So we are not in the business of data storage, data lake, and all these things. That's more, I will say, the IT capabilities. Customer will choose one or the other. We need to work in any data architecture of our customers at the enterprise level [indiscernible]. The data architecture at the process level is something that we own, drive and promote, but at the enterprise level, you have a vast area of choice made by customer. What is important with the industrial data hub in that regard is to say, what we mine and what we transform are the operation data. The more you go to some enterprise-level scenario, the more operational data need to be combined with other enterprise data. It could come from ERPs. It could come from specific enterprise maintenance systems or other. So data hub is really the broker of data that allows to converge those different data sources and create new applications at the enterprise level. Either we create new application or we enable our ecosystem to build those applications or our customers might want to build some specific applications that are related to the IP. So that's for me, that's the enterprise side of this overall data and AI equation that we are looking to resolve and -- with our customers, but as I said, in a very productized way. I was with one of our customers recently where we have done a very successful in-process implementation of those data and AI solution I was referring to. So not experiment, not proof of concept, but actual implementation in a line. The next question is that how do you do this in one of those factories? This will not work if you need to bring 20 data scientists, junior developments and everything. So the ability to productize is absolutely key. So that's what we do around Cognite, productize some of the technology we use into an offer called Data Mosaics, but this is what we do over off-stack. Productization is very important. CUBIC is a little bit different. CUBIC is really about the envelope and the system around power control. And what CUBIC brings us is new market access, and I mentioned in my presentation that was important. So access to new market, presence in Europe. They have a significant presence in Asia as well. And strong presence in renewable. So that's also some of the end markets that we want to be able to address, either because those are combined with the capabilities of our existing customers. We see many customers that are adding renewable capabilities in their production process. It could be solar, it could be steam recovery, it could be fuel cells, it could be hydrogen [indiscernible], different things. So the ability for us to integrate renewable energy in a process and have the technology to do this, including on the hardware side, is important. That's what we do with CUBIC.

And did Cognite have a specific industry focus at the time when you acquired it?

Like many in that space, it started with asset-intensive industry, and as a consequence, oil and gas. And by the way, through Sensia, we are in partnership with SLB. And SLB is also a partner of CUBIC.

And so the idea, you can take Cognite capability that sort of grow up in oil and gas, a, help your oil and gas...

Yes, because of...

[indiscernible] to take it horizontal...

The fundamental equation, which is to be a data broker, data aggregation transformation API layer for data, it's critical on the value point of view for oil and gas and asset industry, but it's important across the spectrum in different shapes and form.

And how can you sort of take CUBIC? Because clearly, that helps you enter into specific end market, but can you also take those capabilities and expand them across your...

Because -- I believe because CUBIC productize the envelope and the architecture of the system. Because of their size and their scale, they're focused on some industries. The synergy with Rockwell is to use the same core technology and expand and bring it to other markets and to other geographies.

No, this is very helpful. So look, I mean, I think we are big believers into -- I think I'm on the record saying we're probably going to have the best CapEx cycle of our professional lifetimes. But you are sort of guiding for very strong organic growth. And maybe just to think about -- how do you think about scaling your R&D and SG&A costs, right? How do you think about scaling things up in an environment where there is a structural upswing? You can argue cyclical structural. But how do you think about investing for growth?

Okay. Well, we are investing for growth. I think it's part of any R&D portfolio management discipline to say there are things that we have been doing because we have always been doing it that we need to stop and we need to maximize the investment in the new area. I can just take a simple example. What I do as the Chief Technology Officer, investment in 4 area in fiscal year '23, cybersecurity, data science and AI, industry solutions and digital platforms that I was mentioning. So we are allocating the resources. We put our money where our strategy is. That's what we do today. We are both organically and inorganically prioritizing this new frontier of these directions that I was mentioning where we want to go. And this implies that we need to optimize our portfolio overall but growing profitable company. We have the means to execute our strategy.

But I guess the question -- the broader question, right, in the world, we sort of analyze the prior cycle, if you look at capacity investments, I guess software has grown a little bit faster. But it was like fairly anemic growth, right? And I think it's easier to manage growth when there is very little growth. It's a lot easier I think to optimize for margin and you can sort of -- margin of safety is greater. As the world sort of accelerates, do you need to sort of change the way you think about investing internally and sort of step it up? And how do you make sure that the company, and I know that's sort of not your area, but as a CTO, how do you think about sort of the company keeping operating leverage, right, as the world is accelerating? So you don't -- so maybe if the revenue grows by, let's call it 11%, what you're guiding 11%, 15%, how does your sort of R&D and SG&A not go up 15%, 20%? That's what I mean.

I think maybe a short detour to your question, but I think it will answer, because all those transformation is digital-related, software-related, I think it's very important to have the discipline to partner with the right companies and not to try to build organically and internally every single capacity that you could imagine and need for the growth. So the nature and the profile of R&D investment is changing. We partner with Microsoft, that's one that I mention quite often. I just mentioned Cognite, you asked me a question about Cognite. So being careful about not trying to rebuild capabilities that are not differentiating. So I think it's not necessarily about, if you say we need to have this amount of growth, we can sustain that growth only if we have the same amount of R&D, we have leverage, because I think we need to gradually shift our R&D investment to what is really our core differentiating IP, and that's not necessarily the IT infrastructure that is behind. I think there are some players in the industry that have tried to be an infrastructure player and provide platform and different set of capabilities for customers, but that's not largely undifferentiated. Many players are providing this. We are more focused on delivering outcomes on those industry solutions. And this equation of industry solutions that are based on common scalable productized capabilities at scale in the right way.

So another area of debate, right, and -- just the key growth markets, at least near term, if you look at the data, semiconductors and EVs. So can we just talk about how should we think about your wallet share in semiconductor manufacturing? One of the, I think, folks sort of point out, they say, well, semiconductors are not that big for Rockwell. I think you've addressed it on one or some of your calls. But how should we think, a, about your wallet share in semiconductor manufacturing? And then also how should we think about your exposure to the entire value chain in semiconductor manufacturing? So maybe we can just sort of unpack that?

Yes. So it's true that our share in semiconductors is not historically big. We have been in facility management, drive climate control, what goes around. Now we are more and more in the conveying transportation, fast moving. We -- and I believe we are successful there because we have a differentiated competitive advantage. We have a unique product that is not just about the way you convey silicon wafers [ in the plan ], but that is also the way you design, stimulate those process, maximize outcome for customers. So we are developing more and more capabilities that are relevant for semiconductor manufacturers. But we are also suppliers of a lot of the equipment manufacturers that goes through the equipment process. So if you have an equipment that is managing the water flow in the process, semiconductor process uses a lot of water, there is most likely a pump, most likely this pump is going to be powered by Rockwell Drive. So it's not just the end user power that you see, but it's all the different equipment that go into a process where we have different type of capabilities to offer to those OEMs. And we also have the ecosystem to support us. Because one of the big part of all those semiconductor's project is they want to be up and running fast. They want to have those capability fast. We don't only have a lot of good technologies, but we have the ecosystem to execute in North America, which I think is very important. It's not necessarily -- if you want to go fast, that's not necessarily the right time to experiment a new partner or to experiment with somebody that is a new entrant to the market but doesn't have the capabilities of Rockwell Automation. So we have the advantage of being the established, strong automation player in North America for that region.

And as more of supply, sort of semiconductor supply chain moves into North America, right, and you have a very strong presence in North America, would that incentivize some of the equipment manufacturers? Are you seeing equipment manufacturers who -- right, because incremental capacity used to go to Asia. So -- but as more incremental capacity that goes to North America, does it change the calculus for sort of machine builders to work with Rockwell going forward given the user base going forward?

This is true beyond semiconductors. This is true for EV, this is true for other things, for a simple reason. If you have, let's say, OEMs in Europe that might work today with some of our steam competitors, as they want to have more business in North America because North America looks to be a very attractive market for them, they will work with end users in North America. Those end users very often will be closed from Rockwell Automation. We help those end users write the specifications of the automation systems that they need to have in their plants. So naturally, those OEM will come to see us and say, "I want to be competitive in North America to address the needs of end user in North America, so I need to work more closely with you. So naturally, we see those conversations happening.

And another area of investment that we sort of -- I think we sort of identified [indiscernible] on my team, like pulled out for every permit, but EV battery plants. I think there, you guys are very excited about your convert technology. What's the opportunity there? What's the strategy there? And is it fair to think that EVs generally and batteries are a particularly exciting opportunity for Rockwell?

Yes. And by the way, when we speak about batteries, it's not just about North America. Today, we are having customers in Asia that are battery manufacturers. So -- and I think the reason it's batteries manufacturing is typically an hybrid process, you mix chemicals and physical parts that you assemble in a certain way. You make a lot of cells. So you have a repeat that is very adapted to automation. And then you move this on fast conveyors, which are the integrated cart technologies that we have. So we both have the expertise in this type of industry, even if it's not specifically battery, but the nature of this type of industry, hybrid industry, we know this very well, so that's really a sweet spot for us. And then we have some unique technologies altogether to do this. And when I look at North America, with the Inflation Reduction Act, the amount of incentives to localize battery manufacturing in the U.S. is tremendous. This trigger other projects like in mining, in other transformations where we are present, so it triggers other value chain elements. And there are also very strong incentives in IRA to recycle batteries. And this is going to create new industries. And recycling of batteries, it's just a reverse process of making it. So you deconstruct the battery, you recover the chemicals, all these parts. So we are also very well positioned to help create in North America this industry of recycling batteries.

But fundamentally is what I'm hearing is that you have a very strong position in North American ecosystem. And insofar as more stuff is going to be done in North America, it structurally benefits Rockwell in the long run.

That's a good summary.

Yes. So maybe sort of automation landscape and we'll sort of talk about virtualization now. But how do you think about sort of more traditional automation, hardware versus more complex technology, marrying it with things like AI, robotics? Because clearly, was sort of written because of the labor shortage, there is more emphasis on robotics. AI is sort of becoming more mature, probably more use cases on the factory floor. How do you integrate your more traditional offering with sort of expanding it into robotics and AI?

So maybe I will speak separately about robotics and AI. So to continue on what I said earlier about AI, we are really in the business of applied industrial AI, okay? So we don't need -- we can rely on partnership for core AI capabilities, Microsoft, Amazon, others. What we do with AI is really at the intersection of software domain expertise and AI. It's really productizing the solution that is AI-based in order to solve a specific problem at scale that we know how to solve. I can give you an example. One of the developments that we are doing is to embed commission monitoring and drives. Use the information available on the drive based on the electrical signal and the [ tiny valuation ] of the electrical signals, deduct that we have a specific problem with the motor, create a maintenance object that is pushed to the CMMS of the customer so that the drive is not just a drive anymore, it's part of the overall tool chain of maintenance for customers. So -- and we do this in multiple elements. For me, AI, I don't believe, and I think the more I speak with people across in the industry, I think it's a common belief of everybody. Doing AI in industrial processes is not about, I am going to take a huge amount of data, move everything to the cloud, let's say, create a very complex model that describes the behavior of my plant and hope that somebody is going to be able to use it to improve operation. That's not the way it works. So making small AI distributed across the architecture, making explainable models that work with the experts that our customers have today, those customers becoming more and more rare, is the way to go. It's not replacing the experts, it's transforming all those tasks of the experts that are repetitive into something that is integrated into the AI model and maximizing the time that the experts spend on cases that you cannot model well or that are complex or that are common cases and everything. So we are really looking at distributing AI from devices through control, to the edge, to manufacturing operational management, but to solve actual problems. And sometimes, we will know it's about AI, sometimes we won't. Like when you look at your phone to make pictures, you don't think that there is AI processing image in the background, you don't need to. That's the same. AI doesn't necessarily needs to be all the time. On your second question on robotics, 75% of -- if I take the example of North America, 75% of Rockwell Automation integrators in North America are robotic integrators. I think the nature of those autonomous operations that I was describing implies that you have coexisting in a factory people, robots in different forms, robotics arms, edge EVs, integrated cart technology, which is another form of robots, EMRs and everything. And in order to do this and not have all these being different clusters like work in isolation, like you have a human being here, you have a robotic arm in a cage here and everything, being able to provide an integrated environment for those integrators that spans all the different capabilities, whether we have them directly as a physical hardware or not, we don't have robotics arms, we have ICT, but we can integrate all the different things that exist across the market. We just announced recently a strategic investment in a company called Ready Robotics, a start-up, and they do design environment, low code, no code, simple design environment in order to integrate robots into automation system. I mentioned that we have released last year our first cloud-native environment for development. Well the two will intersect very well so that we can expand the robotics capabilities or the robotics integration capabilities of development environment.

And last question for me, and maybe we'll have time for 1 or 2 questions from the audience. What's your visibility on reshoring activity over the next 6 to 12, 18 months? What are you hearing from customers? When do you start seeing orders coming on that?

I think there is a variation of different things. It's -- you have some projects that have already started, maybe because they were already here and they just accelerated because there is demand. You have some new initiatives. We announced that we work with Rockwell -- with Ford to support them on their transition from ICE to EVs. I think it's real. Those are real projects. Those are not vague intentions. That's the first thing I will say. And then they are in different stage of execution. Some project will be sliced into smaller phases, each phase will go faster. So if you look at semi, semiconductors, usually you build something big right away. Those projects can be -- I will say those complex projects can have a lead time up to 2 years from the initial discussion. Most commonly, typical project of Rockwell, not those ones, when we have a project business, it's more something like 6 to 12 months. So the truth is kind of in between terms of the cycle.

So maybe we'll open it up. We have time for a couple of questions. I don't know if folks in the audience have questions for Cyril. Any questions? Okay. Well, I certainly have more questions. So maybe we can talk about virtualization, right? And you certainly have strong views on this. Big debate on when and how virtualization will impact discrete automation, right? Some say the industry is right for disruption. Other think that PLC will be the last thing anyone will touch in broader terms. You -- I think you have more of an evolutionary approach. So once again, probably another conversation that we can spend an hour on, but maybe talk about how do you think about sort of this evolution of PLC in a more virtual -- sort of virtualizing PLC and what does Rockwell sort of out there?

So first, well, it's interesting because I have been in this industry for, I don't know, probably 30 years. And this is a recurring permanent question that happens all the time. When you look at PLCs because we are speaking about controllers and PLCs, they have already evolved quite a lot. The fundamental ideas of the PLC, which is to replace cable logic, it's an idea from the '70s, but the PLCs today look very different from the PLCs of the '70s. They still do the same thing, but they have more compute power, more memory, more resilience, a lot of other additional attributes. What I believe is that PLCs will continue to evolve. And they exist among different options, not the only one to run control in a process. You would have some PLC-based architecture, we have them. We acquired a company called ASEM in Italy that is doing IPCs. So we have this capability. We have more and more distributed compute that can go into the devices. When I'm working -- when I was discussing about this development environment, it's all about how you can manage those different applications. For me the important change, which is for me, for Rockwell creator of value, not destructor of value, is this idea of a software-defined architecture. And for me, the fundamental idea is decouple hardware cycles from software cycle and not commoditize hardware. This is not what it means. If PLC were to be commoditized, it will be already commoditized in one way or another, it's not the case. We have very resilient margins. Some of the players that have specialized in non-PLCs type of control will overwhelmingly have won the market, they have not. So I think they are -- a PLC is a fit-for-purpose devices that does what it's supposed to do very well in a very reliable way at a cost point that is the right one and important for our customers with a kind of commitment of operations for 10, 15 years ,okay. You don't have this with a PC. So the fundamental reason why your PLC exists today as a fit-for-purpose devices is not going to disappear. But the PLC will need to continue to evolve, like in the parallel I will make with software-defined networking and all the IT players, switch have not disappeared, they became different. The players that have adapted their switch to the world of software-defined networking have had an upside. The others that did not believe in software-defined networking, not to mention any name, have had a bit of competitive struggle over the years. So we won't be on the first one, not on the second one, and we will lead that transformation. And there are other things that are very important associated to this. You can think about today, you build a PLC, you integrate capabilities inside it and then you install it for 10 years. Tomorrow, we want to have the optionality to add features. We want to have the optionality to upgrade across the life cycle so that we monetize additional value that we have created. We want to be able to deploy specific application modules that could be custom or not custom for customers. We want to be able to have natively inside the PLC some features that if the customer wants to use it, you activate them with a subscription because it's model-based, for example. So there is a lot of optionality and a lot of upside associated to this transformation. And for me, this is certainly not about commoditizing the underlying hardware.

We're out of time. Cyril, always a pleasure. Thanks so much for being here.

Thank you very much. Thank you.