🎁 Black Friday: 50% off Unlimited and 30% off Pro annual plans
Dassault Systèmes SE (EPA:DSY)
France flag France · Delayed Price · Currency is EUR
23.45
-0.95 (-3.89%)
Nov 18, 2026, 5:35 PM CET
← View all transcripts

CMD 2026

Nov 17, 2026

Summary

AI expands the addressable market and enables usage-based monetization, while subscription becomes the growth engine. The 2029 plan targets >€8bn revenue, 35% operating margin, €2.5bn operating cash flow and €1.80 EPS.

Rouven Bergmann
CFO, Dassault Systèmes

Good morning and good afternoon. Welcome to all of you to Dassault Systèmes Capital Markets Day. Thank you for joining us here at our campus in Paris, and also a very warm welcome to every one of you joining us online. Before we start and get to the exciting moments, let me remind you about a few formalities. The financials that we are discussing today are on a non-IFRS basis, and they are in constant currency, and they contain forward-looking statements that could or may differ from actual results. Please refer to further details to our press release. Now with this, let us come to the fun part. 2026 is a year of consistent execution. This is what we told you at the beginning of the year, and we are also setting a new foundation, and our business model is becoming a growth engine.

Today at our Capital Markets Day, we will look ahead into a new industrial era, the era powered by AI, and this is creating a new cycle of growth for the company. To answer the question how we turn this exciting opportunity into concrete value, we have structured the agenda along the following lines. First, AI is a major opportunity, and our positioning is strong. Our Chairman and CEO, Pascal Daloz, will highlight how our AI-native Generation 7 solutions, grounded in science and 45 years of industrial know-how, unlock AI for the industries. Then, Adrienne Brotons, Head of Pricing, will present our AI monetization strategy, expanding from seats to usage as Virtual Companions take on more work. Second, life sciences and healthcare is back on track. With a single operating model in place and a new leadership team, we are ready to execute.

With our acquisition of ArisGlobal, the leader in safety and compliance, we now offer a unified platform across the entire drug life cycle. I am pleased to introduce our CEO of Life Sciences and Healthcare, Dr. Ashwini Zenooz. Third, our growth drivers are shaping our business model. Subscription growth is driving our annual run rate, and we are at a turning point. I will outline our financial plan to reach double-digit growth in revenue and EPS. After that, we will conclude this session with a Q&A. Now, I am excited to welcome our CEO and Chairman on stage, Pascal Daloz. Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Rouven. Thank you to all of you for being here with us today. I think one year ago, we shared our ambitions to shape the next era of the industrial innovations through 3D UNIV+RSES. I think today we are taking this ambition a step further. Why? Because the world is changing, and something fundamental is happening. AI is not simply transforming the technology. AI is transforming the industrial economy as well. This is changing what company they could invent, how they engineer, how they produce, and ultimately, how they are creating the value. Here is the questions. How do you turn intelligence into something that actually works for the real world? A product we can produce, an aircraft we can fly, a system we can operate safely, a treatment that can improve the human health.

Generating intelligence is one thing, but making it work in the real world is another one. This is precisely where Dassault Systèmes comes in. For 45 years, we have been building the bridge between the virtual and the real worlds, virtual world for real life. I think we are developing scientific expertise, deep industry knowledge, the virtual twin technology, a broad portfolio of trusted relationship with customers and partners around the world. Here, what makes so critical today. Those things are not just assets from our past. I think they are becoming our competitive advantage for the future, because we are not simply looking at the next evolution of the industrial software market. I think we are witnessing an emergence of a new market, and I call it the intelligence, the industrial intelligence.

I believe that Dassault Systèmes is really uniquely positioned to lead this market. In fact, our investment thesis rests on four convictions. The first one, I think AI is really expanding the market we could address. Second one, our 45 years of scientific and industrial expertise, I think give us a powerful competitive advantage. The third one, I think AI is unlocking significantly more value across our entire portfolio and our customer base. Fourth, I think we have a clear financial ambitions to turn this opportunity into a sustainable and profitable growth. We are already moving. We brought our AI-native solution, the Generation 7, to the market. We have introduced the Virtual Companion and the agentic capability in our 3DEXPERIENCE platform this year, in June. We have strengthened our recurring revenue model with subscription now representing more than 50%.

We have taken actions to improve the performance of our businesses, and you saw the improvement in life sciences and the turnaround we did in the consumer industry. Now, we are moving for preparing AI to capture the business opportunity. Today, in my presentations, in my keynotes, I really want to address three questions. How big is the industrial opportunity? Why we are uniquely positioned to capture it? How do we turn this into a financial results? Let's start with the first topic, the transformation of the industry. For decades, industrial competitiveness was really built around three capabilities: design better products, manufacture them much more efficiently, operate them effectively. These capabilities are still essential. However, I think AI is changing the game. More importantly, this is changing what companies they can actually create. Look at the products emerging today.

Air taxis, humanoid robots, intelligent medical devices, adaptive surgical systems. They are not traditional products with AI added afterwards. They are AI-native products, meaning the intelligence is becoming part of the product itself. Now is the challenge. Software, electronics, mechanics, sensors, AI must be engineered together from day one. This is changing the entire engineering processes. You need to understand how those technologies, they will interact. You need to simulate their behavior. You need to validate the entire systems before industrial production start. An AI-native product, in fact, is requesting an AI-native industry. This is precisely where our technology becomes essential. We help our customers moving from imagining intelligence products to engineering systems that actually work. This is what we do in the virtual world first, and then in the real world next.

This transformation, I think, is creating a major opportunity for our businesses. Historically, our market has been driven by design engineering, simulations, manufacturing, life cycles, management, so you know them. Those markets remain essential for Dassault Systèmes. AI is opening up entire new industrial activities. Think about it. Physical AI, agentic engineering, industrial orchestrations, certification of intelligent systems, intellectual property protections, just to name a few. They are not simply adding AI features to the existing software. I think we are expanding the industrial world with our platform, that our platform can support. At the end, more users, more activities, more workflows. This is a fundamental change, because we are not selling anymore the software to the same market. I think we are addressing a much larger market by enabling entirely new industrial work. That's the opportunity behind industrial intelligence.

Let me show what it means for our customers. Take XPENG. Their ambition is to go far beyond electrical vehicles. They are developing entirely new category of intelligent products, advanced mobility, eVTOLs, humanoid robots, just to name a few. Those products are requesting a very different approach to engineers, to do the engineering, with new disciplines, new technology, new systems to validate. I think together with them over the time, we have built a modular environment to help them to extend innovation across multiple product categories. We are not just helping XPENG to design better vehicles. I think we are helping them to invent entirely new type of products. We call them the new category of products. Recently, I had the opportunity to discuss with XPENG Founder and CEO, Mr. He Xiaopeng. He is an interesting character.

He was previously the one who built the Alibaba mobile platform before making a bold move into the auto industries with a clear ambition. He wants to bring AI to mobility. More than a long explanation, let's hear directly from him about his visions and how our collaboration is working over the time. Launch the video. Xiaopeng, thank you for taking the time. When we look at XPENG today, we see something bigger than the evolution of the car. In fact, you are bringing intelligence into very different physical systems, from electrical vehicles to flying mobilities and robotics. When AI becomes part of a product itself, what can we invent?

He Xiaopeng
Founder and CEO, XPENG

Actually, I personally believe that in the next decade, all large vehicles will become one robot. This robot may not be human, but it will still be a robot. I think today's AI uses a sort of non-human thinking logic. That's why our autonomous vehicles can definitely operate globally, and it's why we believe robots can speak multiple languages and emulate different people, different roles, and perform different tasks.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

The industry is changing.

He Xiaopeng
Founder and CEO, XPENG

Yeah.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

With our 3DEXPERIENCE platform, agentic platform, and the 3D UNIV+RSES, I think we are connecting scientific models, industrial knowledge, virtual world to help company like you to understand the complexity and validate the decisions before committing, in fact, to the physical world. I think this ambition is not simply to better design products. It makes an entirely new category of products possible. Now, how do you turn that potential into reality?

He Xiaopeng
Founder and CEO, XPENG

I think we've gone through different phases, but what I saw at the very beginning was that we actually rely on people and experience to program. As a company that highly values technological leadership, we use Dassault Systèmes, whether it's our hardware development or autonomous driving development, to accelerate, reduce costs, and increase efficiency by simulating the physical world and the digital world first. Our supply chain has experienced tremendous acceleration over the past three or four years, especially in the physical world. I'm very happy working together with Dassault Systèmes to improve the acceleration from the physical world to the real world. There's room for continued collaboration between our two companies. This is something I am really looking forward to.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Exactly. I think the more intelligent the product becomes, the more important it is to master the complexity behind. I think this bring us to the engineering foundation requires to invent what's coming next. In fact, XPENG, you are exploring very different product categories, and each of them, they have their own architecture constraints, the engineering challenges, integration hardware and software, managing complex systems, validating the performances. How do you expand into these new categories without reinventing the entire development processes every time?

He Xiaopeng
Founder and CEO, XPENG

The goal is for us to become explorers of the future of mobility, from the underlying hardware to the system architecture, then to the software, all the way to the design that the customer can actually see and interact with. It is about how to plan and design it in a holistic, long-term, modular way, and how to validate it. XPENG believes that 70%-80% of a car's and a robot's modules can be shared. However, I also want to point out that between creating a car, an eVTOL, and a robot for the first time, the differences are very significant. However, as you progress, you will find that their modular capability will greatly increase. Dassault Systèmes solutions have been very helpful to us in this regard.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

This is a fundamental shift, and inventing a new product is not only part of the challenge, because moving from an intelligent vehicle to a humanoid robot also means mastering very different manufacturing requirements. How do you ensure that the intelligence you build into the products can be translated into manufacturing processes that are reliable and repeatable and scalable?

He Xiaopeng
Founder and CEO, XPENG

Our entire production and supply chain management is conducted very strictly according to automotive logic. Our flying cars are really quite different from cars. We basically leverage automotive manufacturing experience to innovate. During this process, we will encounter a lot of mistakes and setbacks from R&D and design to manufacturing. But I believe if we can learn by trial and error, and if XPENG can collaborate with Dassault Systèmes to integrate this process capability into our system, we will have data that improves our virtual twins and simulation. Then the second time, we will turn it into a system, an infrastructure that is reusable and scalable.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

The opportunity for me is really to scale innovation across the entire industrial systems. I think it is really whether AI can change the way engineering gets things done.

He Xiaopeng
Founder and CEO, XPENG

Companies that manufacture hardware will face significant changes. For example, today, if we put our camera hardware into AI, this AI can then redesign a similar camera, whether from an engineering perspective, a materials perspective, or a mechanical perspective. AI will definitely change us.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I think a general-purpose AI can generate an answer, but engineering requires much more. It needs to understand the systems, respecting the physics constraints, connecting decisions to their consequences. I think XPENG is an amazing showcase of ambitions to invent a new category of products, and we believe the virtual world combined with industrial knowledge and AI can help to turn this ambition into reality.

He Xiaopeng
Founder and CEO, XPENG

I do not think there is any other CEO of a hardware design firm like Dassault Systèmes who is able to think so far ahead. This is what I really like.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Mr. He Xiaopeng.

He Xiaopeng
Founder and CEO, XPENG

Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

So, thank you, He Xiaopeng. This is exactly the kind of transformations we are talking about. Customers are no longer, in fact, asking us to simply improve the existing products. They are asking us to help them to develop things. They were previously extremely difficult to do and sometimes even impossible to engineer. The question now is, why are we well-positioned? Because this is not a market we are entering today. It is a market we have been preparing for over 45 years. We started with 3D design, to design very complex assembly, very complex products. With PLM, we connected design and manufacturing to be ready at the same time with your production system when your design is finished. With 3DEXPERIENCE, we brought these capabilities at the enterprise level, connecting people, processes, and knowledge together. And now with 3D UNIV+RSES, we are connecting AI-native virtual twins. That is what we do.

But our advantages, I think, goes far beyond our technology. It is a combination, again, of scientific expertise, industrial knowledge, integrated workflows, a broad portfolio, and trusted relationship. I think a competitor can build an AI model, a competitor can access to computing power, but reproducing 45 years of scientific expertise, industrial processes, customer trust, this is a very different challenge. Anyone can generate intelligence, but the reality, very few can make it reliable. And this distinction is extremely critical in industry such as aerospace, energy, and healthcare. In those industry, AI must be explainable, traceable, validated, trusted. In this context, I think Airbus is a perfect example. And today, I am particularly pleased and honored to welcome Guillaume Faury, our CEO of Airbus. Guillaume leads one of the world's most iconic industrial companies at the forefront of aerospace innovation and transformations. He is also a longstanding partner of Dassault Systèmes.

So Guillaume, thank you for being with us today. It is really a pleasure to have you here.

Guillaume Faury
CEO, Airbus

Hello, Pascal.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I was saying in my introduction that we have a longstanding relationship. You remember when we started to work together?

Guillaume Faury
CEO, Airbus

Oh, yes, I do.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Now maybe, I have few questions for you, if it works for you, Guillaume. The first one is, Airbus is facing a significant long-term growth opportunity. Seeing you have been vocal about it, you are telling to the investment community that is really the case. But capturing this growth opportunity is also requiring from you a fundamental transformation of your industry and from Airbus. What will drive the next growth in aerospace and for Airbus, and what do you need to change to make it happen?

Guillaume Faury
CEO, Airbus

Yeah. Well, thank you, Pascal. As I said, very happy to be with you today. I know you have your investors with you, so it is important. I am very happy to be here because we enjoy a long-lasting, successful, and trustful partnership and relationship with Dassault Systèmes, and we could not be what we are today without this partnership. We have decided in 2019 to expand, moving into the future with a partnership on 3DEXPERIENCE that is absolutely crucial to our future successes. Indeed, we are in an industry and in sectors, civil aviation, aerospace, space itself, helicopters, and others, where we enjoy growth, actually a very fast growth. We have more than 10 years of backlog for our commercial aircraft, and we need to ramp up.

We are also at a point where we are preparing the next generation of products, and that is where the partnership is really important to us. We are in a situation where the complexity, the size, the complexity of our products is increasing. We are looking at the product not in isolation, but we are looking at the product, the production system that will manufacture all those sophisticated products, and the world of operations when those products will be operated by multiple clients over a long period of time. So we need to bring the intelligence of human beings collaborating together to design and manufacture those products, to service them. We need to bring the capabilities of data to use artificial intelligence in combination with humans. This in a rather complex environment over several countries, inside Airbus and with our partners outside of Airbus.

Nobody can manage that complexity in the sense of what human can access, so we need the ability to orchestrate all of this. Actually, the partnership we have with you through your product, 3DEXPERIENCE, but also your people, your teams, the expertise you were mentioning before, is a unique capability that we see in the world to prepare, to organize these capabilities, and to be able to design the products in their sophistication, in their environment, in a physical environment. We do not impose to the laws of physics what they tell us. It comes with a good understanding of the world, and this is what 3DEXPERIENCE is capable of bringing to us. So actually, the partnership that we have with you, that we have with your teams, is absolutely essential to our future successes.

A bit like what was said with your previous interview with her, we are also looking at the partnership with you as a key enabler of managing that sophistication, that complexity, and driving it in a world where we need reliability of data and what we are doing. We need sovereignty. We need cybersecurity. So these are quite hard requirements when it comes to managing digital, but that is what we are developing with your teams. That is what we are testing already today, and that is the environment in which we want to develop our future products, humans and agents collaborating on a unique set of data, from design to manufacturing to operations of the product. That is basically why we are excited to be working with you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Guillaume. Now, if I understand well, growth means for you increasing productions, securing the supply chain, accelerating the decarbonization, right? Preparing the next generation of aircrafts. You mentioned it, in an industry where safety comes first, what capabilities simply cannot be compromised, and what role the virtual twin and AI can play in this?

Guillaume Faury
CEO, Airbus

Well, actually, the virtual twin is at the center. It's something we can no longer work without it. We don't, as you sort of already alluded to, we don't deliver an aircraft or a satellite almost complete. We don't deliver an aircraft that is almost safe. So the level of details and understanding of what we're doing, the level of mastery, the ability to go into the very details of objects and production systems, which are very large, distributed all around the world, with a high degree of complexity, where you have software and hardware working together, systems connecting with others. We have connected products moving forward.

This can only be made through bringing reality, physical reality, into a digital world where we can assess, we can test, we can compare, we can multiply the number of configurations we are evaluating, and we can get the support of AI to increase the number of possibilities. That's what we're doing through the digital twin. That's what we're doing with the 3DEXPERIENCE agentic platform. There's no product or no production system tomorrow that will exist without having had before its digital twin, its digital brother, on which we will have applied a lot of tests, a lot of experience, a lot of simulation, and create the ability to bring much more sophisticated products much faster in a more reliable way. Obviously, as I said earlier or before, we need reliability, we need cybersecurity, so the boundaries and the constraints are quite multiple.

That's what we are finding in the 3DEXPERIENCE relationship we have with you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

What is striking me, in fact, is that every critical engineering decision must be traceable across the entire lifecycle and ecosystem. At the same time, the technology is moving incredibly fast. How do you balance sovereignty, security, and trust with the need to innovate? Again, where do you see the greatest potential for 3DEXPERIENCE at Airbus in the coming years?

Guillaume Faury
CEO, Airbus

Well, actually, what we have decided to do at Airbus moving forward is to put all our activities on 3DEXPERIENCE. We are doing it on Airbus commercial developments and the production systems, on helicopters. We have a good cooperation on helicopters already now for many years, also in the field of defense and space. We go across the board. We go step by step because we need to mature our ability to work with 3DEXPERIENCE. We need also to mature the platform itself as we are moving forward, given the constraints we mentioned before, the traceability, the reliability, the cybersecurity, the sovereignty. Ultimately, for the products that will enter into the market in the next decade, they will be born, they will be native on 3DEXPERIENCE. They will be maintained, they will be further developed on 3DEXPERIENCE.

What is new as well is the fact that we no longer look at products in isolation, but we look at the product in their production environment, in their industrial environment, and how they are operated along the years in a single source of truth, of data, that can be accessed by everybody in the same way, and going beyond the borders of the company. We really want to see this going to our main partners. That is what you have started to do, and in some cases, we have started doing it together. But we need this going further to most of our main partners and suppliers so that we can expand the boundaries of the potential of this way of working.

Because working in 3DEXPERIENCE gives speed, gives reliability, gives control of what we are doing, and gives this potential for use of AI, meaning doing much more than what we would be doing in a standard way. Yes, we see that this is an enabler of bringing products to the market. As you said it earlier, we have few challenges, but one is to accelerate the ramp-up. It is a lot about supply chain management and industrial performance, but it is also the preparation of the next generation of products in this new environment.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Guillaume. I know you have a busy schedule, and thank you for making the time with us. You have, I think, your board meeting just after. One final word you want to say to the folks here in Vélizy with us, the investors?

Guillaume Faury
CEO, Airbus

Well, I know you are with your investors. I am a big believer of what you are doing, so that is maybe the only thing I can say. We have a long-term partnership that we have renewed recently. We are super happy with the way you try to understand our business and bridge with what you can bring. It is really teamwork that I can see when we are working with you. You mentioned the difficulty to replace the understanding of an industry for a newcomer. I can testify that indeed, we are working with experts and leaders in your organization that really know what we are doing, and that is very helpful because we know we have a platform that is designed and further developed to fulfill our needs for the future. So, I am very happy with this partnership, and I am looking forward to keep working with you, Pascal.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Guillaume. I think the perspective you gave to us illustrates something fundamental about the transformation of our industry. As you say, Airbus is preparing the next generation of aerospace, facing unprecedented challenges in engineering, manufacturing, sustainability, industrial performance, costing, just to name few. But what makes this partnership so important is, as you say, we are working together on precisely those challenges. So, Guillaume, thank you very much again for your leadership, for your trust, and for your longstanding partnerships. Ladies and gentlemen, please join me in thanking Guillaume.

Guillaume Faury
CEO, Airbus

Thanks, Pascal. All the best for you. Have a good day. Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Guillaume. Now, let's look at what makes our approach to AI so different. Today, the industry is investing billions in foundational models and computing power. Rightly so. But intelligence alone is not enough. You need much more for the industry. You need to understand the physics, the materials, the behavior of the complex systems. Guillaume was mentioning this. If you take an aircraft wing, AI can describe it, probably generate a partial design, maybe suggest some improvements, but can it tell you whether this wing is safe, how the material will behave under stress, whether it can be actually produced? And there are very different questions that the system cannot answer. In the industry, almost right is wrong. I repeat, almost right is wrong.

When you are designing an aircraft, developing a medical device, or generating a power plant, I think almost right is simply not good enough. You need much more. You need science, you need engineering knowledge, you need the validations. And this is the difference between generating an answer and delivering an outcome you can trust. And this is where AI for industry is focused on making intelligence much more powerful. These ambitions to make intelligence work are our ambitions. We are not in the race to build another foundational model. We are building the industrial intelligence layer that makes AI reliable, useful in the real-world applications. This is why we are developing industry world models. They bring scientific and industrial knowledge into AI, representing how physics and the biological world works actually. Let me simplify maybe, or let me simplify the architecture.

The foundational model generates and understands the content. The world models help understand the environments. The physical AIs interact with the physical world. But the industry world model brings scientific and industrial knowledge, and the virtual twin provides the environment to simulate, test, and validate. Here is the new chain, if you want. Generate, understand, simulate, validate, and then act. And this is our approach. And we are already seeing dividends of this potential. Take our design simulation benchmark. A standard language model approach reaches approximately 75% reliability. With our AI approach, we are exceeding 92% reliability. The important point is not just the difference between those two numbers, it's what they demonstrate. When AI is grounded in a scientific and engineering knowledge, it becomes significantly more reliable, and that's why the industry needs us. Now, let's make this concrete. Imagine introducing an intelligent robot into a factory.

It needs to work alongside people, move autonomously, adapt to this environment, and react to unexpected situations. Would you simply switch it on and hope everything was going well? Of course not. You need to understand how it will behave. What happens if someone walks in front of it? If it encounters situations it has never seen before? I think training a robot on thousands of scenarios in the real factory will be far too much expensive, difficult, and sometimes extremely dangerous. So what do you do? You learn and test virtually before operating physically. You create a virtual twin of the robots and its factory. You simulate the interactions, you explore the behavior, you train the system, you validate it before the deployment. And the result is what? Faster learning, lower risk, better performance.

So learn and certify in the virtual world before you operate in the physical world. That's how we make physical AI ready for the industry. This brings me to the new positioning, the virtual twin factory. Think of it as the place where AI meets industrial reality. It brings together intelligence, science, engineering, industrial knowledge. This is where ideas are explored, these systems are designed, simulated, and trained, and decisions are validated and certified before implementations. In another word, it's where we move from what AI can generate to what industry can actually build and operate. The virtual twin factory is much more than a technology environment. It's a new infrastructure, it's a new industrial infrastructure, I should say, for the AI economy. Again, we have spent 45 years building these foundations.

Bringing AI into industry at scale also is requesting enormous computing power, and that's why we have our collaboration with NVIDIA, and that's why this collaboration is also important. NVIDIA builds the AI factories, they provide the computing infrastructure to create and develop intelligence at scale. We are building the virtual twin factories. We provide science, engineering, and industrial knowledge to connect the intelligence to the reality. Let me put it simply. AI factory creates the intelligence, the virtual twin factory makes it work for the industry. The two are complementing each other. The two is essential part of tomorrow industrial infrastructure, and together, I think we can make AI ready for the industry. What is also very interesting with this partnership, it goes both way, in both directions. NVIDIA is also using our virtual twin technology to design and simulate its own AI architecture.

Here you can see, this is the latest architecture they have built, and you see they have built a complete virtual twin, which is extremely important when you need also to deploy the system at scale and in a time-wise. This is not simply a technological partnership. I think, as I say, we are complementing our area of expertise working together. I had recently the opportunity to discuss this transformation with Jensen Huang, the NVIDIA founder, and at that time, he was a CEO. There have been some recent development at NVIDIA, and obviously, those matters, these are matters for NVIDIA and its board, and I will not comment on it.

But Jensen has been one of the leading voices shaping the AI revolutions, and I think his perspective on the emergence of the new industrial era, the role of physical AI, is particularly relevant for what we want to build together. So let hear what he has to say. Thank you, Jensen, for taking time today.

Jensen Huang
CEO, NVIDIA

Happy to see you, Pascal. Every time in France.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

More than 20 years ago, we started by putting accelerated graphics into the hand of the engineers. Today, we are entering another inflection point. AI is moving beyond generating content. What fundamentally changes when AI moves from answer to engineering to the physical world?

Jensen Huang
CEO, NVIDIA

This is really the golden ages for Dassault. When we met, it was during the time when proprietary workstations was making a transition to open platform workstations. It expanded the opportunity for Dassault and for us tremendously, and we became great partners. We have the Dassault Virtual Companions, and so these agentic systems has to be grounded on physics, grounded on science. For a long time, we've been limited by the number of designers, and now every designer will have 100, 1,000 Virtual Companions, and all these Virtual Companions are going to be using the Dassault tools.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

For 40 years, Dassault Systèmes has built the scientific and engineering foundations to represent, simulate, and validate reality. Today, those technology have become extremely relevant and more important than ever because virtual twins represent and simulate the complex systems. Industrial world model captures the industrial context, and I think our Virtual Companions, they bring knowledge into the workflow of engineers and scientists. Together, we bring not only accelerated computing as well as AI infrastructure, ecosystem of open models, but also industrial knowledge and scientific foundation to make AI actionable in very complex environments.

Jensen Huang
CEO, NVIDIA

Accelerated computing, it's at the foundation of almost everything we're talking about, whether it's the virtual twin reasoning, perceiving, and planning, or it's the virtual twin grounded on physics and simulation and science. All of that is running on NVIDIA GPUs. The intersection between our technology has never been greater.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Why do you believe the combination between AI simulation and scientific knowledge is so important for the industry?

Jensen Huang
CEO, NVIDIA

Well, in order for us to apply artificial intelligence to the largest markets in the world, we have to understand the laws of physics. You're designing, simulating, and engineering the car. Maybe you're making the car autonomous, or maybe you're creating the factory that has robots inside that's building the car that's going to be autonomous. This entire space requires the tools that Dassault has made over the years. In life sciences, the concepts are very similar. The big revolution now is in the world of drug discovery, and so the AI would explore different proteins and chemicals and their affinity and whether it activates or deactivates a disease. You would then explore those options in the virtual world first, eliminate less likely successful combinations, and then prototype, collect data, come back and train, and refine your model.

That autonomous loop is kind of like a self-driving car loop.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

It is also already tangible in our work we do together. Consider an AI factory, a system involving computing, cooling, power, networking, physical infrastructure. What does it mean for you, and how model-based systems engineering technology can be used to scale AI?

Jensen Huang
CEO, NVIDIA

Each rack is about 2 tons, nearly 2 million parts. Inside a data center, you could have 400,000 GPUs. It operates on 1 GW of power in the latest generation AI factories. The power delivery system is incredibly complex, and you want to be as energy efficient as you can. We build it inside Dassault, and all the parts will be there, all the components are integrated. We can then put that directly into simulation, and you could prototype the entire AI factory before you build it. All the software could be running in the loop. This is the whole dream of the virtual twin. in the future, we are going to create almost everything inside a virtual twin. It is the place where you can engage AI.

Your Virtual Companions can help you explore all of the options and all the various optimal combinations, before you even break ground.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

This is what makes our partnership distinctive, because ultimately, the opportunity will be measured by what our customers can do together. The potential to accelerate innovation, to improve the engineering productivity and reduce the cost, to explore new possibilities. We are already seeing this use case emerging. We showcase Audi and Agile Robots in the factory scale, physics AI, and autonomous robotics, but also with Volkswagen, the agentic engineering workflow.

Jensen Huang
CEO, NVIDIA

Yes, it's incredible, the Audi factory. You have a robot working inside a robotic factory, building a car that itself will be robotic. Software is integrated across the whole thing. Everything was captured in the Dassault virtual twins. All these robotic systems are essentially Virtual Companions, and everything is running AI and 3D graphics, physically based. All of it is one giant harmonious system, and we're going to make these systems better, faster, and more cost-effectively, and operate them more efficiently, all at the same time. The amount of innovation that's going to come out of virtual twins and Virtual Companions, the combination of the technologies that we've been working on all these years, really the intersection between Dassault and NVIDIA has made all of this possible.

Every single factory in the world, every single logistics center, everything that moves in the future will be built in this way. It's a complete reinvention of Dassault. This is your golden ages.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

It's a good way to conclude, Jensen, because 20 years ago, we helped engineer accelerated design. Today, we are extending our ambitions to engineering, in fact, the real world. NVIDIA brings the power to scale AI. Dassault Systèmes brings the science, the engineering knowledge, and the virtual twins to connect intelligence to the reality. Together, we are expanding what industry is able to design, discover, and achieve. Thank you so much.

Jensen Huang
CEO, NVIDIA

All right. Congratulations.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

So, looking ahead, I believe the next 20 years of our partnership will be about building the foundation of industrial intelligence, connecting computing power, scientific knowledge, and virtual twin, ready to transform those industry. And because the future of industrial AI will depend on much more than the computing power, I think it will depend also on the ability to understand and simulate and manage the complexity of the real world, and to train AI. So this is why I believe what we have been doing for the last 45 years is so relevant. Now, the technology only makes the difference when people can actually use it. So let's bring this into the workplace. Imagine an engineer working on a complex industrial project. Today, that engineer navigate multiple applications, searches for informations, contact experts, coordinate different disciplines. Now imagine a complete different scenario, complete different way of working.

This engineer has a specialized Virtual Companion. He understand the objective, he access the right knowledge, he use the relevant applications, and works with other agents to coordinate activities. What's the results? The engineer focus on solving the problems, and the Virtual Companion help to organize and execute the task and the work. So, everything remain connected to the industrial platform, and I think this is much more than conversational AI. I think we are moving AI that answer questions to AI that gets job done. And there is a big difference between the two. So this is the foundation of our 3DEXPERIENCE Agentic Platform. It's really opening up new possibilities across our entire customer base, which brings me, by the way, to the business opportunity. So, the question now is, how do you translate this into growth for Dassault Systèmes?

It start with a fundamental change in how customer use the software. Historically, our application were mainly used by specialists, and AI, I think, is changing this equation. It makes our expertise accessible to many more people. It allow one person to work with several specialists and several Virtual Companions across the company. And to do this, let me introduce four of them. The first one is AURA. She is a program management. She knows how to manage complex programs, such as the one you have seen at Airbus, or the one you could also have in the pharmaceutical industry. LEO is the engineer. He's the one knowing how to solve complex problems, not only at the design level, but along the entire life cycle, which is pretty unique when you think about it.

Because usually, if you are an engineer, you are specialized in design or in manufacturing or in tests, but you will have a hard time to find one who is mastering all the disciplines. MARIE, she's a scientist. So she master material science, biology, information systems, computing. So she has all this deep breadth of knowledge across multiple scientific domain. And LOUIS is the med. He's the one knowing what clinical and patient intelligence is about. So each of them, they are bringing knowledge directly into the work. And I think the real breakthrough is not having another AI assistance, it's really having a specialized agents working together, accessing our applications and coordinating extremely complex workflow. And here is the message I really want you to remember. Our Virtual Companions, or the agent in general, they don't replace our portfolio. They just unlock it.

There is a big difference between the two. You heard what Jensen is saying. The more agents will consume more applications. Why so? Because I think our applications contain decades of scientific and industrial knowledge, and I think we are making this expertise available in a very completely new way. Today, in June, when we introduce the Virtual Companions and the agentic platform on the market, they already have the skills to work with more than 50% of the entire 3DEXPERIENCE portfolio. Somewhere next year, they will have the full capacity, the full access, and the full ability to use the entire portfolio. In the future, they will know also how to use third-party applications. Think about it, what it means. We are not building a new business from scratch. We are unlocking the economic potential of the business we have been building over the time.

This is what it means. Again, let's make this concrete. Today, only a limited number of specialists may use our most advanced engineering applications. What do we do with Virtual Companions? With Virtual Companions, many more employees can have access to specialized expertise. One person can work with several companions. Each companion can use several applications. This is creating an entirely new way of working, and the opportunity is no longer simply about selling another software license. It's really about increasing the amount of industrial work that our platform could perform. I see, we see four growth levers. The first one, more users, making more advanced things because they get access to advanced expertise. The second one is more usage, enabling Virtual Companions to use application across more activities. The third one is more value per customer, supporting a broad part of the industrial processes.

The last point is the new monetizations, because this will obviously, we will evolve how we are packaging our solution, how we are selling our solutions as users and customer value increase. Historically, the software growth was tightly close to the number of users. Now, one person can work with multiple companions, each accessing different applications and area of expertise. The value we are creating is increasingly linked to the industrial work our platform enable, not simply the number of seats. Our ambition is really to capture a fair share of this value. Of course, we need to be very disciplined in the way we turn this opportunity into revenue. Today, we have more than 100 strategic partners around the world with whom we are developing AI use cases.

This is an important starting point, but I think we need to move from adoption to deployment, from deployment to regular usage, and from usage to measurable economic value. I think our monetizations principle are very straightforward. We don't charge customer for AI simply because it's new, it's a new technology. We want to be paid for the additional value we are delivering. To explain how we are building a scalable business model, I am very pleased to hand over to Adrienne Brotons, our Head of Pricing. Adrienne, over to you.

Adrienne Brotons
Head of Pricing, Dassault Systèmes

Thank you, Pascal. Good afternoon, everyone. You've just heard how AI is becoming part of our customers' products. My role is to answer the question you're probably asking yourselves, how does this turn into revenue for Dassault Systèmes? Let me start with the key message. There is not one way for our customers to access our AI. There are three, depending on where the customer starts. Our own 3DEXPERIENCE platform and Virtual Companions, third-party agents, or an AI model. Every one of them generates revenue for us. Let's start with the first path. The first path is the 3DEXPERIENCE platform. The customer uses our roles and our 3DX Agentic platform, as they do today, under a user-based subscription. On top of that, they add our Virtual Companions, LEO, MARIE, AURA, and LOUIS, which come with an annual subscription with capped usage.

This is our more complete offer, and the one where we create and capture the most value. The second path is for customers who enter through another agent. It can be an agent that they build themselves or an agent from a third party. These agents connect to our platform. They call our own Virtual Companions and interact closely with them. The customer pays for what they use through an annual subscription with capped usage. The third path. The third path is through the AI models. When a request comes from an AI model and needs real engineering, it will call our capabilities. Here again, we are paid on usage through an annual subscription with capped usage. Our Virtual Companions, a third-party agent, or an AI model, you will have three entry points and one pricing model.

Whichever door the customer chooses, the path that must be right, the path that is grounded in physics, in rules, certifications, comes from us. It is based on our industry world models and certified applications. On every path, we bring the value that matters most, results our customer can trust and build on. On every path, we are paid for it. Let me explain why this creates a larger value source for us, and why our model is shifting towards usage. Let's start with the left side of this slide, the world before AI. A person uses our applications to complete a workflow. We are paid through a user-based subscription. In that world, growth has one main driver, the number of users. The number of users is limited by the number of people who can design, engineer, and create. AI changes this.

Look at the right side of this slide. The Virtual Companions leverage our applications to execute a workflow. This is the value of work that we could never capture before. For the first time, we can monetize part of it. Agents do something else. Every workflow an agent performs involves calling our applications. An agent does not call them a few times a day as a person would. It calls them again and again to explore, to test, to compare. More cycles, more application use. If we put these two together, what we have is that an activity on our platform that is not capped by the number of people in front of a screen. A single user can work with many agents, and each agent generates calls to our applications. The number of human users no longer tells the whole story.

The metric that reflects growth is now usage, how much work is done and how much knowledge is used. This is why we are moving towards usage. We follow the metric that is now driving value for our customers and therefore for us. What does this mean for the business model? The user-based subscription stays in place. It remains our foundation. On top of it, we add a usage-based layer. The result is a larger value pool, user-based plus usage-based. But this raises a question. If usage becomes a metric, what exactly do we measure? The answer is, we measure what matters most to our customers and to us, which is the value of what is delivered. This is the role of our Dassault Systèmes Blue Tokens. How does it work? When customers use AI, there are many ways to charge for it.

We could charge it by the number of words typed in a request, but it would never tell us what the customer actually received. What the customer wants is a result, a drawing, a simulation, a validated design. So we price the results. This is what the Blue Tokens does. It is our unit for the value delivered by our AI. It reflects the work done and the knowledge used to do it. The more value in the results, the more Blue Tokens. And the customers pay for that value, never for how well it prompts. Let me show you what this means in practice. We have here an example of a user asking our companion, LEO, to create a steel support frame for a hopper. On the first example, the part is simple. The results cost 45 Blue Tokens.

Now, second example, the same user on the same part writes a much longer request, which load, which steel profiles, which mounting faces, every detail. The request is much longer, but the result is exactly the same structure. So the price stays the same at 45 Blue Tokens. And finally, last example, a request just as short as the first one, but resulting in a much more complex structure. The number of words is the same, but there is more work to deliver. So the price is higher at 315 Blue Tokens in this example. So the price follows the value of the output, not the length of the prompt. Why does this matter? For our customers, it brings two things. The first thing is predictability. They purchase a prepaid annual subscription with capped usage.

The usage allowance is tailored to their use cases, industry, and business processes, drawing on 40 years of Dassault Systèmes industry expertise. Second, it brings flexibility. Blue Tokens work across the entire AI portfolio, and customers can buy additional Blue Tokens at any time. For you, as investors, it means two things. The first one is that our AI revenue is linked to the value we deliver. It's not tied to the length of the prompt. Second, it is built on an annual prepaid subscription, which fits the recurring revenue model you know. Three paths to AI, a larger source of value, a simple way to price it. That is how AI becomes revenue for Dassault Systèmes. Thank you, and back to you, Pascal.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Adrienne. I think the message is pretty simple. AI is not just creating a new product. It is increasing the economic potential of basically the entire install base. I think we are building on what we already have, our portfolio, our customers, our technology, our industrial ecosystems, and we are applying this ambition across all the businesses. This leads me to speak now and to move to life sciences and healthcare. The same transformation we have seen in manufacturing is also happening in life sciences and healthcare. The opportunity is particularly exciting. In manufacturing, physics sets the rules, but in life sciences, biology does, and human health is at stake. To think about it, you should look at the journey of a new treatment, scientific discovery, clinical developments, manufacturing, safety and regulatory intelligence, and finally, the patient.

I think our ambition is to connect those different worlds, BIOVIA for the scientific discovery, Medidata for the clinical development, DELMIA for the manufacturing, and now ArisGlobal for the drug safety and the regulatory intelligence. I think together, those capability connects the pharmaceutical journey from molecule to patients and back to science. I think AI can accelerate the discovery, improve the clinical development, make manufacturing much more intelligent, and strengthen the drug safety. The real world, the real opportunity is not simply to improve each activity separately, it is to connect the dots, to connect the entire journey and creating more value throughout the life of the treatments. Now, I also want to address a second point, which is the performance of our life sciences business. Let's be direct. Life sciences has not delivered the performance we were expecting in the recent year.

Believe me, I am not satisfied with those results. That is why we have taken action. That is why we have strengthened the leadership team. That is why we changed the operating model. That is why we brought the business together around a more integrated offer, and to make it easier for our customers to access our solutions. It is a new way of bringing our capability to the market, and I think now we have the portfolio, we have the leadership, we have the operating model in place, and I think our ambition is to return to sustainable double-digit growth, and the sooner is the better. In this context, I am very pleased to welcome Dr. Ashwini Zenooz, Executive Vice President and CEO of Life Sciences and Healthcare. She will share with you not only her perspective on the opportunity, but how we are going to capture it. Ash, over to you.

Ashwini Zenooz
CEO of Life Sciences and Healthcare, Dassault Systèmes

Thank you, Pascal. Hello, everybody. Thank you for being here. I am very excited to be in this room today. As a brief introduction, I am a physician by training, but I have spent much of my career not only delivering healthcare in large systems, but also in large, small, medium-sized technology companies. The opportunity to bring those two perspectives together is why I am so thrilled to join Dassault Systèmes, and it is a pivotal moment for our industry, and it is a great time to be here. Today I want to talk about why the time is right, why it is now, why we are in a position to win, and how are we organized for growth. Hopefully you can hear me. As I was saying, I want to talk to you about how we are positioned to win.

Once I finish, you are going to hear directly from one of our customers, who is going to hopefully bring this vision to life for you. Let me go through why I keep calling this a moment for life sciences. First, and really important, our customers are investing again. Biotech is having its strongest IPO year since 2021, and M&A and licensing are accelerating as large pharma is refilling its pipelines. We are seeing that our mid-market and our CRO businesses are recovering. It is very important for us. Second, the decision-makers believe that AI is going to transform R&D. Eight in 10 executives expect that AI is going to cut drug discovery costs, or drug discovery cycles from four and a half years to one and a half years. That is a pretty significant change.

Third, eight out of 10 organizations say that they are not ready for data access and integration. The evidence is there, but it is not connected. Simply put, the capital is there, the ambition is there, but the foundation is not there. While AI has made incredible strides for molecules to become easier to design, it has not yet made biology easier to predict. Pascal called this out. This matters because the life sciences industry has one of the highest R&D spend of any industry, but only about 10% of the clinical candidates ever reach a patient. Behind every one of those failures is a patient who was waiting. For a long time, that was simply the cost of doing science. Now with AI, we have a new path forward.

We can connect the evidence from molecule all the way to the patient, that is the opening for us here at Dassault Systèmes. Let us look at the life cycle of a therapy. Discovery can tell us how a molecule might work. Clinical development can show us how different groups of patients will respond. Manufacturing helps us scale it and make it consistent and reliable at scale. Regulatory and safety highlights benefits and risks once the patients are using it. Today, most of those teams are working independently from each other, and what one team is learning rarely reaches the next, data, process, et cetera. With the 3DEXPERIENCE Agentic Platform, the 3DEXPERIENCE platform, all of those domains are now coming together in one environment.

The team can now work collaboratively, and data can seamlessly flow across these domains so that we can simulate and predict better outcomes for the patients. Pascal talked about the virtual and the real worlds, so I want to emphasize how that works in our industry. On the left side of the loop here, this is where therapies are created. This is our virtual world. We have a lot of research, development, clinical, manufacturing. We are remodeling and simulating and predicting. On the right is where the patients are. This is where we can measure real outcomes. This is a loop in our world, because when real-world evidence flows back into discovery, scientists can actually use it to refine their models and design and decide which therapies they are going to pursue next. This is where AI comes in very handy.

Every turn of this loop gives them better evidence to learn from it, so it can predict the next therapy better. We are critical and integral, I hope, for that process, because we bring a lot of foundational science to this loop. At Dassault Systèmes, this loop is built on a very strong patient foundation. Our clinical trials have touched 15 million patients, more than 60 million patients in real-world data, and a reach of more than 6 billion people. You heard Pascal earlier today talk about, or earlier just now talk about Virtual Companions, MARIE, LEO, LOUIS, AURA. These agents can work across the 3DEXPERIENCE platform. They can access data, provide intelligence across every step of this loop. What can they really do? They can do a lot of things. They can help a development team draw on discovery and evidence.

They can help a manufacturing team see what matters most for quality. They can help a safety team now assess a new signal in the context of everything that is already known about the therapy. That's great, but what does this mean when you bring this environment to a patient? As we said earlier, AI just can't be fast or clever or superficial. It has to be grounded. Our virtual twin technology is proven. It is built on decades of modeling and simulation. I want to go through the dimensions that it spans. First, the twins are multidisciplinary. Modern therapies sit across many fields, and our science covers them all, everything from organic chemistry to cell biology to protein science to pharmacology, formulation, material sciences. Dassault Systèmes has a strong foundation in this.

Second, they are multi-scale, from molecule to cell to organ to human to cohort to population. What is learned at one scale can inform the next. Our technology leverages the best foundation models. We heard about that just now. We are also building and training our own on top of that, models that design molecules, predict how a drug will behave in the body, and whether it's safe. All of these advances are what's making our platform stronger. All of this is grounded in industrial know-how. Just to put it in context, we work with 15,000 customers across pharma, med tech, and CROs. Our technology has supported 42,000 clinical trials. We handle more than 12 million safety events per year. That is more than half of the worldwide safety events. It's not just theory.

Four out of the five drugs approved by the FDA in 2025 used Dassault Systèmes technology. It's not just drugs. More than half of the medical devices are designed on our platform, and increasing. The 3DEXPERIENCE platform brings the right knowledge, the right know-how to the right decision across the therapeutic life cycle, across the manufacturing life cycle. It is the intelligence platform that is there to improve patient outcomes. How are we going to deliver that? Many of you asked me this question yesterday at dinner. We are bringing the sector together, as Pascal mentioned earlier, under one leadership team and one operating model. Our team has deep industry expertise. We have leaders who have deployed AI in regulated workflows. We have scientists, PhDs that are doing deep science. We have R&D specialists that can bring the right solutions to our industry.

So what can you expect to see from all of us? First, we are going to sell as one sector, not as individual brands. This is important for our customers. We have evaluated where our businesses can create more value together across the 3DS platform of brands. That means bringing planning for strategic accounts, strengthening our renewals, cross-sell, upsell, expanding our enterprise platform model, and building the momentum that I mentioned earlier that we are seeing in our mid-market and CROs. There is a great opportunity for partnership here. Second, we are going to present, and we are presenting, a unified offer on the 3DX Agentic platform. We are going to prioritize the integration that helps the customers use evidence across discovery, clinical, manufacturing, safety, regulatory, quality. In fact, we have already identified integrated solutions. I probably shouldn't say this, but look for it in 2027. Third, we are going to double down on our domain expertise.

This is critical for us. It means we are going to continue to invest in our science and our technology, so the platform's understanding, the understanding of the virtual twins and the models of the therapies, the modalities, it is going to continue to deepen over time. We are going to bring the data back so it can learn. Fourth, we see an opportunity to extend into healthcare that brings us even closer to the patient. We are exploring how our platform could follow the therapy into everyday care. Pascal called this out earlier. I am going to emphasize this. For life sciences and healthcare at Dassault Systèmes, this is a turnaround story. We are already on that journey, but expect more from us. The market is improving. Our business model and our commercial execution are strengthening. We have deep bench of domain experts and new leaders on the team.

With one team, one offer, one platform, I believe that we are on a strong trajectory towards double-digit growth. With that, I would like to introduce the newest member of our leadership team here, Aman Wasan, who is the CEO of ArisGlobal. Aman, welcome to the stage.

Aman Wasan
CEO, ArisGlobal

Thank you, Ash. Great. I think it is afternoon already. Thank you for the opportunity. What a morning and afternoon, I think, hearing from Pascal Daloz and some of our customers in automobile and aviation. I think I can say that we are super excited as ArisGlobal to be part of Dassault Systèmes, which is at the forefront of advanced technology and bringing true value to our users, our customers, and to everybody who is touching the technology. I have been with ArisGlobal since 2017. We have seen amazing progress and innovation, and we have delivered that innovation to our customers.

It gives me immense pleasure, and I am super happy to welcome one of our customers and key stakeholder in the life sciences industry, Matteo, who needs no introduction, but he currently is Senior Vice President R&D, Real-World Data, and Digital Health Business within Bristol Myers Squibb, which is one of the world's top pharma companies. He is based in U.S., and he couldn't be with us here, but I am sure, the next opportunity he will be very happy to be amongst us. But he is joining us virtually. I will be asking Matteo a few questions on how he sees the application of AI in life sciences and the amazing work that we have done with BMS as ArisGlobal, bringing AI in production, and the outcomes, which one of the colleagues talked about, are super important.

I mean, we all can talk about AI, we can talk about technology. What really matters is the outcomes, right? The return on investment and what's the impact of AI in production, right? In life sciences, it's patients. Whatever we do is for the patients, so what's the impact to the patients at the end of the day? So, Matteo, I hope you can hear me. Thanks for joining us. If you can just quickly introduce yourself, that will be great.

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Hi, everybody. Matteo di Tommaso here. I lead the R&D Business Insights and Technology team for Bristol Myers Squibb. It's a pleasure to be here with you today.

Aman Wasan
CEO, ArisGlobal

Thank you, Matteo. Matteo, you have a huge remit, right, within BMS, and you are no stranger to technology, innovation, and you have worked in different organizations in the past. At BMS, with the current remit, of course, you see the application of AI across the platform, across the entire landscape and ecosystem. What really is the most important thing when you are looking at where AI should be deployed? We heard in the morning that it's more impactful when the AI is native to the platform. If you can just share your perspective, when you look to deploy AI across the entire landscape, what are some of the key tenets? What are some of the key parameters that you look at?

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Thanks, Aman. I will start by just saying that AI, you mentioned already earlier that 90% of drugs do not ever reach patients, do not ever get to the market. That is an enormous failure rate and enormous cost of doing business in drug discovery and development. The number one thing that we can do to really change the outcomes is to increase the probability of success of these clinical trials that we are running and bring more medicines to patients in a shorter period of time. So that probability of success is the top driver. Today, we are using AI not as a side project, but it really is embedded in every part of the pipeline. In drug discovery, the way you are looking at early, the target identification and how you are going to design a molecule that intervenes with the biology.

The second major phase is de-risking that trial, de-risking that asset that you have in order to bring it into clinical trials. Then you are matching with AI patients to the clinical trial so that you are putting the right patients in there and increasing the chances of success of that medicine in the clinical trial. You are using the AI to design the clinical trial, select sites, do a lot of monitoring and operational work, and de-risking. Then you are using it to make the big decisions about whether you move from phase II to the very expensive phase III clinical stage. Beyond that, you are monitoring and managing and understanding the effect and standard of care in the market.

It really is across that whole pipeline that you are looking at the able to leverage these tools to understand and use the large amounts of data that are produced across that drug discovery and development area.

Aman Wasan
CEO, ArisGlobal

Absolutely, and thanks for sharing that perspective. When it comes to outcomes, Matteo, how important are outcomes? Because as you know, and we all know, that a lot of AI trials, pilots fail because of many reasons. One, they may or may not be native. The use cases are nice, but the application of those use cases in reality, in real world, does not work. We have recently deployed NavaX AI as part of ArisGlobal within BMS, and I know it is just one of the domains in pharmacovigilance, and you have a huge remit. But I know you and I have discussed, we discussed this technology when we were building it in 2024. We worked collaboratively with not just BMS, but also your partners, the BPOs that are involved here.

Now that the technology is deployed in production and we have seen the return on investment, can you please share what are you hearing from your teams around this technology being used in real world, and what's the impact on the ground that you have heard and seen?

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Thank you. Well, I'll start by saying that one of the things you have to do to set up AI efforts for success from the very beginning is to make sure that you have the right sponsorship and you know why you're doing it, and you've designed it in stages so that you move from stage to stage to make sure that you can make good, informed choices. Agility is absolutely critical. Agility and a good picture of what you're doing and where the critical decision points need to be made. I think that this is an example where, when we're partnering, the number one thing we do is to make sure that we're working together in that sort of win-win way towards outcomes, well-defined milestones. This is the way that we've worked together, and it has been a very effective way of working together on multiple programs.

Aman Wasan
CEO, ArisGlobal

Thank you, Matteo. When AI is in production, there are changes, right? There are changes to the process, there are changes in the way the users experience the application, the front end of the application. There is change in the way the cost is handled. There's a change in the way the results are derived from the use of AI. When it comes to humans, the users, what are some of the things that you have seen in terms of impacts of AI in life sciences across the entire landscape in your experience?

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

I think there are a couple of different things that I'll emphasize there, Aman. The first is just that early on, we made a choice to roll out AI broadly, because there's a lot of value in the whole organization understanding what is this technology, how do you use it, how do you use it responsibly, how do you put it to work for you? There is a personal productivity benefit. Anybody who's used AI these days to create a PowerPoint slide or to digest a set of complex scientific literature and get some idea of where you might want to read as a human more deeply, can instantly see the productivity gain you get at a sort of personal level.

That personal productivity helps to get the organization aware and on board and understanding the technology, but it is the targeted larger programs of effort that also have the bigger enterprise return, and those are things where you are picking, where do we really need to change our business process? How do we get cost to be different, time to be different, probability of success to be different, quality to be different? Those are the major metrics that you are looking at, and step by step, making sure that you can have an impact on those metrics. That's where the real value comes from. Engaging users and co-creating across that process. Certainly, technology has always had to be very tightly coupled with business process.

There is only way to be successful today, in my view, is to work in a very agile way, co-creating, and in integrated small teams focused on the technology and the business process with one set of goals towards those business outcomes.

Aman Wasan
CEO, ArisGlobal

Well, that's well said. Now, a little bit of personal/professional question, Matteo. You have seen ArisGlobal over the years, right? We are a 40-year-old company, and we have been supporting this life sciences industry in domains that we are relevant in. What's your most recent experience with the business, and what sort of transition you have seen when we, as an organization, transition from a technology company to an innovative technology company, bringing AI into production, and scale that across the life sciences industry?

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Well, I would say there's a couple of things that are really essentials to compete. Certainly one of them is really domain expertise. We all know that you can create large amounts of code today with AI coding techniques, and that's really changing the way we can create things. It takes a good combination of business expertise and technical expertise to be able to execute effectively. I have seen ArisGlobal evolve in that way and have really appreciated both the technical depth and the business depth that you bring to the table. I think the agility is the other key here. The technologies are changing very rapidly. We look back now and we say things that we did six months ago, things that we did a year ago, we need to rethink and redo.

That ability to continually disrupt is really essential because the technology and what is possible is changing, therefore, the way we harvest value and the way we use that technology has to change as well. That evolution is essential, as well as the opportunity, I think, right now to continue to create industry platforms that do the jobs that we don't need to do, right? There's a lot of partnership that we need, and we look for partners that are complementary to us, partners that bring capabilities that all of us need in pharma so that we don't have to invent those all individually. That foresight, again, requires a lot of business expertise in order to be effective. By that business expertise, it's understanding of our business and understanding of where there's value and a good dialogue, two-way dialogue.

That two-way dialogue is essential as well.

Aman Wasan
CEO, ArisGlobal

No, absolutely. I think one of the things that was amazing for me to see personally, last month, I was at World Patient Safety Congress in Boston, and I was at the booth of Accenture, which is one of the partners, you work with as well. They had a paper that they had printed on the success that they have achieved with NavaX and delivered their success back to BMS. It was amazing for many reasons. One big reason is we have created a working model, a playbook, if you will, for the industry to follow with respect to leveraging AI in life sciences. Not just leveraging AI, but also to ensure that AI is priced well. We heard from our colleague, how important pricing is. At the same time, laser focus on outcomes, right? What value companies will derive from AI.

I think that was one of the reasons we have been able to scale it in a meaningful manner today. Most of our enterprise customers like Johnson & Johnson, Amgen, Merck, AGA, and others leverage that foundation. So I wanted to thank you personally as well for your leadership and the support of the extended team within BMS that we have been able to achieve this tremendous success. We are just getting started, and there's a lot that can be done. Just last question, Matteo. With the success that we have achieved, with the foundation we have created, with the technology evolution at such a fast pace, what are your expectations of the AI labs, the entire ecosystem for the future?

What would you do, or what would you expect from the industry such that it can create more value for companies like BMS with application of AI in future?

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Yeah, I think you actually just described some of the things that are most important to me as well, is that first of all, this laser focus on business value, business outcomes, business outcomes. We are here to make medicines for patients with serious disease, and that ultimately is where we are aiming. We want to make sure that we are using the best technologies to deliver those outcomes for patients. Working backwards from that, obviously, there are many things that we can do across the pipeline that we have already talked about somewhat today. The expectations I have is that we have really partners who really understand that value chain, who really are laser-focused on outcomes.

I would emphasize the agility and the ability to adapt very quickly, and the ability to be industry-leading in the understanding of both the technology and the science, because the science and the technology are changing very rapidly. This is a bridge where being able to have that agility, disrupt the components that you have very quickly, and be able to put in place the best technology, staying a step ahead, the pace has changed. This is the most important thing, is this agility, collaboration, shared goals, and value and outcomes focus. These are what make good partners for me, and we are looking for partners because we know that there is a lot common to the industry that we do not need to do and should be doing in partnership, and we are looking for these win-win relationships with partners always.

Aman Wasan
CEO, ArisGlobal

Oh, well said. Matteo, thanks a lot for taking the time. I know you are just starting your day. I really appreciate always the perspective you bring and the wealth of experience that you have, and thank you for sharing that experience. At the same time, sharing your view on application of AI within drug discovery and R&D IT. Really appreciate that. Thank you very much for joining us today.

Matteo di Tommaso
SVP of R&D, Real-World Data, and Digital Health Business, Bristol Myers Squibb

Thank you so much. Appreciate the partnership. Thank you, Aman Wasan.

Aman Wasan
CEO, ArisGlobal

Thank you. With that, I think we can wrap up the life sciences session. So, I think we are just getting started. There is tremendous exciting announcements to follow. We have, as Ash mentioned, a clear vision, clear strategy of where we want to go. Our customers fully understand where we want to go, and in the coming weeks to months, it will become even more clearer, the value that we will bring as Dassault Systèmes. With that, I would like to hand it over back to Pascal, and thank you very much.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Aman. Thank you, Matteo. And thank you, Ash, for these beautiful introductions. I think what I am drawing from the discussion, probably three priorities. By the way, I am giving the priority live. The first one is, I think, strengthening the business. We are clear on this. The second one is really bringing our portfolio together more effectively. And the third one is obviously improving the execution. Why? And you heard from Matteo why we are creating new opportunities through AI. That should be the game plan for 2027. Let me bring everything you have seen together before to give the floor to Rouven, because I want to summarize it. The first one, our industrial business are well-positioned for the next generation of AI-native product and processes. In life sciences, we have taken the actions needed to put the business on a strong foundation, on a strong footing.

The consumer industry, even if we did not spend time on this during this Capital Markets Day, but I think continue, and you see it quarter after quarter, to show a strong momentum. Let us turn to the financial side and address the questions I know matter the most for you guys. How does this translate into the financial results? I think we are not building our next growth cycle from scratch. And this is the title of this Capital Markets Day, "New Industry Era, New Cycle of Growth." We are entering in this cycle with a stronger portfolio, with a more predictable business model, and a new opportunity for growth. And I think our financial ambition rests on four pillars. First, larger market. AI-native virtual twins, physical AI are expanding the activity we can address.

So, more industrial needs, more complex workflows, entirely new type of product and systems. The second one is more value per customer. You have seen it, more users, more usage, more of our portfolio, and new monetization opportunities. The third one is stronger revenue, stronger recurrent revenue, I should say. You know, our subscription model and the annual recurring revenue provide a greater visibility and predictability. Rouven will come back on this, but subscription already represents 50% of the recurrent revenue. And fourth, better executions, improving business performance, increasing profitability, generating sustainable cash flow. Those are the four pillars, and I think they are reinforcing each other. AI increase the customer value, the recurrent revenue provide visibility, and a strong execution turns growth into profit and cash. But I think you need the number. That's the reason why you are here. So, where are we today?

Where do we want to go? What are the driving force between them? So, this is exactly what Rouven Bergmann, our CFO, will now explain. So, Rouven, back to you.

Rouven Bergmann
CFO, Dassault Systèmes

Okay. Thank you, Pascal, and welcome again. Now let's go through this exciting moment and section together to bring all of those elements that you have heard today into the financial model and the clarity ahead of us. So, let us start with where we are today. In 2026, and I think you will agree with me when you see this chart, we see an improvement across all the metrics, and it's driven by the solid execution that we have delivered. This was our promise when we entered the year, and this is what we did. You see now revenue is expected to land at EUR 6.4 billion. We are targeting 4%-5% growth. It's a continued acceleration quarter-over-quarter. We have introduced this year the annual run rate, and we have been tracking this quarter-over-quarter on the progress we have made.

We are expecting EUR 4.7 billion at the end of this year. And you see here, we are incorporating the contribution from ArisGlobal into the number in Q4. We're expecting 7%-8% ARR growth for the next 12 months at Q4. The operating margin has trended very positive, 80 basis points of improvement. And also here, it's a story of consistent improvement quarter-over-quarter. And with this, when you put all the elements together, we are generating very healthy operating cash flow, which is growing 8% to 10% year-over-year, at a cash conversion rate that will be above 85%. So, to summarize, the revenue, the margin, and the cash are all moving in the right direction, and this is growth and discipline together. Now, as you can see on the next slide, the momentum is building quarter-over-quarter.

This is, as I said, what we have introduced entering into this year, so this is exactly how we planned it. The total revenue growth has moved from 1% of growth in Q4 last year to 3% in Q1, to 4% in Q2, 5% in Q3. Now for the fourth quarter, we expect to continue this trajectory with growth in the range now of 5%-8%. This, of course, also includes the contribution of ArisGlobal in Q4. As I said, I want to reiterate that and emphasize it, this is the result of strong execution, and it is driven on a broad-based execution. Not only strong, it is broad-based because it is across geographies, it is across industries, and we are leveraging our growth drivers of 3DEXPERIENCE and cloud. I think this is clear. We are walking the talk. What is driving this improvement?

The business model is turning to become a growth engine. Subscription revenue doubled over the last five years. It grew 13%, 2021 to 2026 on average, and it is even up 20% when you exclude the headwinds that Medidata and our life sciences business were facing during that time. Now, you understand the impact of the strong headwinds in life sciences during this period. What is important for you to know, that this is behind us, and Medidata is back to growth in a new operating model as part of life sciences. In terms of business model, we achieved a turning point in 2026. For the first time, subscription revenue is now surpassing the maintenance revenue in absolute terms. Why is this important?

Because the lifetime value of a subscription is a multiple of the lifetime value of a license, and because it is recurring and it is not a one-of-a-time transaction. Subscription now is the primary driver of growth, and it is expected to gain share and grow even higher at a higher rate going forward. Consequently, ARR becomes now the guiding metric to project the growth and manage the business model evolution going forward. ARR has gradually accelerated, as I mentioned, from 5% growth to 6%, and it is now estimated to reach 7%-8% at the end of this year. At this point, the subscription ARR represents more than half of the total ARR. In this next phase, what is coming after, starting 2027, we expect the ARR growth to accelerate even further to reach EUR 6 billion-EUR 6.5 billion by the end of 2029.

With this, we are targeting a growth now of double-digit 10%. It is clear, the subscription annual run rate drives this acceleration, while the maintenance annual run rate remains rather flattish. How are we going to reach this path, and how do we bridge to 10% growth in ARR? This is a result of two factors. First, two points comes from the acceleration in subscription into the high teens. Today, we are growing low to mid-teens, and we are transitioning to the next phase that is going to be high teens in subscription. This acceleration is driven, of course, by the adoption of 3DEXPERIENCE, the adoption of the agentic 3DEXPERIENCE platform, of course, the AI monetization that Adrienne introduced and Pascal alluded to, and I come back to that.

The growth acceleration in life sciences, you heard about that, and the continued momentum in the consumer-centric industries, I will come back to that. The second point is, we have one point of acceleration which comes due to the mix effect, where the faster-growing subscription run rate becomes a larger share of the base. These two factors, they offset the lower growth contribution from the maintenance base, of course. I know many of you have asked, "Help me to bridge the annual run rate to revenue, and how are they going to converge over time?" Simply put, the more we transition our business model towards subscription, the more the gap between ARR and revenue will narrow. This is, at the end, improving the predictability of our business model.

Less of the revenue over time is going to depend on the one-time licenses and businesses that we still have. The ratio of ARR, it was 1.5x in 2023, and it will be around 1.4x in 2026. By 2029, the ARR is expected to reach, when it reaches about EUR 6 billion-EUR 6.5 billion, the coverage, we will cover the large majority of our revenue with the ARR, and the multiple will be 1.3x . Of course, it will continue to converge. As you can see, the upfront licenses plus the services is going to add less than EUR 2 billion to bring the total revenue to an EUR 8 billion target. Now, with this in mind, let's turn to the business growth drivers that are behind, in which we are implementing this business model. As you know, today we are at EUR 6.4 billion in revenue.

We are growing mid-single digit in 2026. By 2029, with this plan, we expect to reach more than EUR 8 billion. At that point in time, by 2029, we will grow double digit. We are expecting or we're targeting double-digit growth by 2029. The CAGR over this period from 2026 to 2029 is between 8% to 9%. Important to highlight, very important to highlight, is that every business in this plan is contributing to growth. As you see here, the industrials, which is contributing EUR 1 billion, which includes industrial innovation as well as SOLIDWORKS. For life sciences, we are expecting or estimating EUR 400 million-EUR 500 million of incremental revenue coming from the contribution of Medidata, BIOVIA and ArisGlobal. For consumer industries, which is mainly driven by Centric, we are adding EUR 300 million of growth to the company.

Now, of course, AI monetization and virtual twins, and virtual twin market expansion, they amplify these core drivers. Let's take a look at the monetization opportunity that AI creates for our business. As you heard from Pascal, industries are becoming AI native. A key element of the core strategy is to capture this potential that AI creates. Adrienne covered our AI monetization strategy, and now I would like to show you how this turns into actionable growth levers. AI agents do not replace engineering software, because the work is dependent on the physics-based certified applications. Agents need those applications, and they need more applications than people do. Two levers come to play. First, today, the monetization model is based on a human, and it's capped by the headcount.

In the AI native world, Virtual Companions, they work and they run jobs in parallel, and they do this around the clock. So the question about fewer seats turns into more work performed, and consequently, usage is the metric to price, and for this, we are using Blue Tokens. Second, Virtual Companions use many more applications to harness and train compared to humans. A human works typically in one application, in one application at a time. Just think about a CATIA engineer. In the AI native world, the agents call several applications to really master and fulfill the job end-to-end. So these two shifts, as you see at the bottom of the slide, they magnify each other. The monetization moves from seats to work performed, and the applications used per job moves from one to many. All of this is going to be a catalyst for our ARR growth.

Now, as you heard in Adrienne's presentation, the foundation remains the agentic 3DEXPERIENCE platform. This is where our clients manage their IP, and where the humans and the Virtual Companions, they collaborate, and they consume the knowledge provided by our science-based solutions. So the foundation is already established. Now, let me go through the business levers, first of all, starting with the industrials. This is our heritage, as you know, and this is also where we have the significant client base. This is within manufacturing industries, and today, it is our largest business. With 3DEXPERIENCE, we are the market leader in industrial innovation. The manufacturing industry, as we have seen in the customer testimonial and all the programs that we are running, is transforming to become AI native. This is our opportunity.

When you just think about the market that we have here, it is EUR 1.6 trillion in R&D spend across the different end markets, and most of the operations still rely on labor and physical testing. Now, with virtual twins, every point of share that moves from the physical world to the virtual world, for us, is a point of market opportunity. This market expansion, together with Virtual Companions, with this, we are capturing the value of work performed, and we sustain our growth in the industrial sector. It is driven by the momentum in 3DEXPERIENCE and cloud. So to conclude on the industrial sector, this takes our business from around EUR 4.8 billion in revenue today, which is growing mid-single digits, to about EUR 5.8 billion in 2029, and adding EUR 1 billion in revenue, and it is expected to accelerate by 2 to 3 points over this period in terms of growth.

Now, next to life sciences. You just heard from Ash, and now let me walk you through some of the financial expectations. For us, the opportunity is to build the path to reach double-digit growth over time. 2026 has been a transformative year, as you heard. We achieved two milestones. In Q3, we stabilized the overall performance in life sciences from a revenue perspective and growth perspective. Secondly, we announced not only in the summer, we announced the acquisition, and in September, we completed the acquisition of ArisGlobal. So this path forward really is supported and driven by the new operating model that we are going to put in place and that Ash outlined to you.

It also includes that there is a return to growth for Medidata, and it is expanding our opportunity overall by leveraging this end-to-end AI platform, which we are bringing to our clients. This together will drive our life sciences growth towards 8%-10%, adding EUR 400 million-EUR 500 million in revenue by 2029, and we will be reaching EUR 1.7 billion. Of course, on ArisGlobal, we continue to expect double-digit growth, and it will be fully organic from Q4 2027 onwards. Let us take a look at our business for consumer industries, where, as you know, Centric is the main driver of growth. We are operating in a very large market. I want to remind you on that. It is a very large market, but the momentum continues to build.

The business, also important, has been fully transitioned to subscription, and there is a healthy and continuous cycle of large customer renewals in front of us. The growth is broad-based across geographies, and we continue to expand our reach by serving more and more and expanding to more and more verticals. This business started in apparels and fashion industry. It is transitioning fast to beauty, to food, as well as retail. Second, we are addressing more domains. We are expanding to e-commerce. We are reaching the customers of our customers. We are attracting more users from retailers and brands, expanding to suppliers and manufacturers. Of course, the usage of AI for personalization and virtual twins to simulate product market fit is a major need for this industry.

Overall, the consumer segment revenue is set to grow around 20% CAGR, and on the current baseline, adding about EUR 300 million of incremental revenue over this period. To wrap up this section, coming back to what I said, I repeat, every business is contributing to growth. It is powered by AI and virtual twins. Our business model is the growth catalyst providing us with the visibility and the confidence to reach EUR 8 billion in revenue by 2029. Now that we have discussed our plan to accelerate the top line, the revenue growth, let us focus on the productivity of the company. Our objective remains to expand the operating margins while continue to invest to build the future. We have a strong operating model with leverage. You know that. This drives roughly 400 basis points of operating improvement.

However, we plan to reinvest about 100 basis points, or approximately EUR 250 million over these three years to further support our growth over and above the operating leverage that we already have. The net margin improvement is 300 basis points, and this brings the operating margin to 35% in 2029. Of course, productivity levers span across all our functions, R&D, sales and marketing, G&A, and AI is accelerating this transformation. For example, in R&D, we prioritize, of course, our investments to advance the adoption of our Generation 7 solutions, and we transform to implementing AI-native R&D operations. We optimize, of course, overall the resource allocation to have the best and highest impact. In sales and marketing, we take a factory approach to streamline our go-to-market while we are leveraging AI assistant operations for planning.

In G&A, we apply AI and automation across our business and our administration systems, and we will see these benefits on a recurring basis starting to come over this three-year period. All of this, the top-line expansion, the consistency around ARR, the profit margin expansion, is of course translating directly into a strong cash flow model. Cash conversion is already improving to more than 85% in 2026. As you see on this slide, we are targeting to reach 88%-91% in 2029, which then will result in an operating cash flow of approximately EUR 2.5 billion. What are the levers that give us this confidence to drive us to this number? I mentioned the higher share of ARR, which consequently results in a much better alignment of our revenue and cash.

The operating leverage and margin expansion to 35%, and here, of course, AI-driven productivity is helping us, and the continued discipline on capital intensity and capital spend, which helps us to free up cash from working capital. Putting all of this together, as I mentioned, the operating cash flow is expected to grow from today nearly around EUR 1.8 billion in 2026 to reach EUR 2.5 billion in 2029, and this entails a CAGR of 10+%. The next step is, let's take a look at our EPS, which is the bottom line, and we are targeting to grow EPS in this updated financial plan with a CAGR of 10%, reaching EUR 1.80.

This reflects all of the elements I worked through, the 8%-9% of revenue CAGR, the operating margin expansion, including the reinvestment to the business, which is important, the disciplined capital allocation at the end with no dilution, and the normalized tax rate. This plan stands on a solid base. As you see, 2026 is already delivering proof points. At our recent earnings, we raised our EPS growth guidance to 6%-7%, up from the initial 3%-6% when we started the year. This gives us great confidence as we are moving into this next cycle. Now, let's take a look at our capital allocation. The framework is consistent and simple, and it matters even more in the landscape of AI-native players. We have three key priorities.

First, invest in growth, with a focus on R&D driving organic growth and innovation, AI technologies, of course, and disciplined M&A. We have demonstrated that as a company for many, many years. Second, it's important to maintain a strong balance sheet and an attractive investment-grade credit rating. If you combine the first two priorities, ArisGlobal, the acquisition we completed in September, it ties very well to those priorities. It's disciplined M&A, it's very strategic because it closes the loop from molecule to patient, and we were able to fund it from this year's operating cash flow. Third, return capital to shareholders through dividends as well as buybacks to offset dilutions. The use of cash follows also a very consistent approach. The rule is approximately 60% of cash is invested to fund our growth, and 40% goes to shareholders via dividends and buybacks to avoid dilution.

Now let me come to the final slide to summarize this financial outlook and our priorities. As you heard, our priority is to accelerate growth. We have reached a turning point in our business model transition in 2026, where subscription now accounts for 50% of the recurring revenue. It is growing at double digits. We are now entering the stage where subscription growth accelerates ARR growth, and this is giving us great visibility and good confidence to reach EUR 6 billion-EUR 6.5 billion in ARR by 2029. The growth in ARR translates to growth in revenue. Consequently, we project 8%-9% revenue growth over the period. Every business is contributing, industrials, life sciences, and consumers, all powered and supported by the increasing AI-native solutions that we bring to our clients.

Our strong operating model provides us the flexibility to improve the margin while also invest into future growth. We will add roughly one point of margin expansion each year to reach 35% operating margin. The strong growth in ARR, together with the margin expansion, they drive improvement in the cash conversion. As a result, we expect to reach EUR 2.5 billion in operating cash flow. Finally, EPS is set to reach EUR 1.80 by 2029, growing at 10% CAGR, and it is led by this strong operating performance powered by our business model that serves us as a growth engine. Now, we are achieving all of this while we invest into our business to fully capture this once-in-a-lifetime opportunity that AI presents. Simply put, the time is now. Back to you, Pascal.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you, Rouven. Before to open for the Q&A, let me summarize what you have seen during those two hours. We have explained the opportunity, and we spent the time to explain, because I struggled a little bit discussing with you the last 12 months to convince you about the fact that what we do is relevant in the AI age. It is not a disruption, it is a lever for us. I think we have shown also how well we are positioned to capture this market. You have seen the testimonial is not only us telling this, the customers, the partners from different geographies. Rouven has presented the financial framework to make it happen, and he gave you all the building blocks for you to basically assess if we walk the talk, and I think it is very comprehensive and it is very well articulated.

Now, I want to focus on what matter most to you, I think, as an investors, our financial commitments. Our objective for 2029 are very clear, more than EUR 8 billion revenue, 35% operating margin, EUR 2.5 billion in cash flow, and EUR 1.8 EPS. If there is something I want you to remember, it is only one thing. We are targeting a double-digit growth, whatever is at the revenue level, at the cash flow level, at the EPS level for 2029. Those are, I think, ambition objectives. They are all supported with the same ambition, structural growth, more predictability, strong profitability, sustainable cash flow generations. I want to make one thing very clear. This is not growth at any cost.

I think you heard Rouven telling that we are reinvesting, we are expanding, we are building the net growth relay because the plan will not end in 2029. We need to continue to invest. I think we are building a sustainable and profitable growth model, a model which is based on disciplined investment, stronger recurring revenue, and better executions. We have made the investment. We brought our AI-native Generation 7 solution to the market last year. We have introduced our Virtual Companions and our agentic platform this year. We have strengthened our recurring revenue model over the last three years. Also, we have taken the action to improve the performance of our businesses where we had some headwinds. As a CEO, I take all those commitments very seriously.

For me, the opportunity is clear, the technology is clear, our financial ambition is well-defined, and now it's about execution. It's about delivering what we committed to you. On these statements, I think it's time to open the Q&A sessions to the floor. For this, I will welcome all the speakers on stage to join me. Thank you.

Rouven Bergmann
CFO, Dassault Systèmes

Okay. I guess I'm going to try to manage this. There's so many hands up. Okay, thank you for the light. Let me start with Michael.

Michael Briest
Analyst, UBS

Thank you, Rouven. Michael Briest at UBS. I guess a couple from me. Previously, you've obviously had 3DEXPERIENCE cloud targets very clearly enunciated. Could you talk a little bit, even if you're not going to have targets, but there's a cloud transition ongoing as well as subscription that normally depresses the growth rates as you flag, but you're adding EUR 1 billion to the industrial software part, and that seems ambitious in light of that transition. Can you maybe walk through that? Then in terms of M&A, you said 40% of cash in the last three years, or three years including this, have gone on acquisitions. Is that what we should expect, and what will that mean in terms of the growth contribution? Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Rouven, maybe you give the number. I want to put this question into the context of the strategy. What you have seen with our AI-native applications will not happen if we are not moving to the cloud. The question is not anymore, are we walking the talk? Are we delivering the percentage you are expecting? It is, we are building the foundations for the next growth relay for the company. Why I make these comments? Because when we engineered the plan a few years ago, I think the market was in a different situation. The PLM market was probably one of the few moving slowly to the cloud, moving slowly to become software as a service, for one specific reason.

When you deploy a PLM system, you deploy it for 20 years, and when you manage the life cycle of the products, you have some life cycles sometimes exceeding 30 years. This is true for a plane, this is true for a nuclear plant, this is true for a submarine, just to name a few. So, that was the complication in a way. Now, AI suddenly is changing the mindset. There is no single conversation we have, including with the most reluctant people to move to the cloud, which are not now considering that they need at least to do a supplemental approach with the cloud. So, that is the reason why, Michael, interesting questions, but in a way relatively irrelevant to understand the trajectory of the company, because anyway, this is happening. So maybe, Rouven, if you want to add a few comments on this.

Rouven Bergmann
CFO, Dassault Systèmes

Yeah. Look, I think exactly that is why we transitioned to ARR. This is why we are converting our business to be better, more predictable, and increase the share of subscription overall. The move to cloud is a necessity in the world of AI. So, it is going to happen, and the share is increasing. We show it every quarter, the progress that we are making. You see it every quarter, and we are reporting on this, and we will not stop that because it is walking the talk. Setting up another objective and trying to reconcile this number with that number is sometimes not helpful, and I think it just creates too much discussion on the wrong topic. But you can be assured that the share of cloud in this model is constantly increasing.

Also, the share of 3DEXPERIENCE and the 3DEXPERIENCE agentic platform, which is today more than 40%, but in the next three years, it will reach more than two-thirds. Also, the share of cloud is going to add at least one to two points per share, over one to two quarters every quarter. So, we are on this trajectory. Keep in mind also, one of our largest cloud businesses in life sciences is also back to growth and will continue to contribute to that. So, I think we have all of those elements now pointing in the right direction, and this is really making the change as we are entering into this new era of growth that is really prepared by our business model that has now reached this critical point where subscription is the higher share in the overall annual run rate.

The last point you mentioned on M&A. I think we have been very consistent. We have been doing M&A for strategic purposes or tuck-in M&A from time to time. In this plan, we will follow this capital allocation framework, the 60% and 40%. There is no change to that. This was also the same at the time when we acquired Medidata. So, we have that strength on our balance sheet with an investment-grade rating, strong support overall, and that allows us to make tuck-in acquisitions and play this balance sheet to our strength as we are building, not only for 2029, but also beyond. So now, I think let us turn to the right side. Mo, over to you.

Speaker 11

Great. Thank you very much for the presentation today. I had a couple. First one for Pascal. As we think of the shape of this growth, it is sort of a bit more back-end loaded, so I wanted to kind of better understand the thought process behind that. Is it just being a bit more conservative in the more near term? I know in the past, you have also alluded to certain things on go-to-market and sales that kind of needed to be improved, so just curious to understand, is there some prudence built in for that? Then also on the sort of the AI monetization, it sounds like maybe you are depending on that to come much more towards the back end of the horizon. So what is perhaps the reasoning for that more back-end loadedness? One for the life science team.

There is obviously the market improvement that you have alluded to, but you keep talking about the kind of platform approach. So, what in your view is going to be the big differentiation in the marketplace, but also there were some product launches. So what needs to be further done to kind of get to that sort of ambition of sort of double-digit growth? Lastly, for Rouven, can you help us kind of understand some of the mechanics? Because licenses, obviously, you will still have, but how do we think of the license and the maintenance contribution over the midterm? Is there going to be a bit more of a role? How do you then get to this sort of margin expansion? Where are the additional savings? Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Well, we start more with the first questions. Is the growth backloaded or not, or to which extent? There are two components. One is, Rouven made it clear, the business model is a growth engine. The more you move to subscription, especially with the large customers we have, they are committing on multiple years. I mean, they are not committing only for one or two years. Automatically, you have this translation of the bookings into the revenue, and that is the reason why Rouven gave you the kind of book-to-bill ratio in order for you to adapt and fine-tune your model. So this is obviously the major explanation behind this. The second one is, are all our engines on fire? No. You remember last year at the same times, I told you I am not pleased with the indirect channel.

Still a topic, but we made some progress since then. We have two indirect channels, the one which are historically selling SOLIDWORKS, we call it CRE, and the other one, which are historically selling CATIA V5, we call them CPE. I think we made a significant progress with CRE. CRE now is delivering consistently. The growth is broad-based. We can see in term of number of units or in values, and we have a good performance of most of our partners around the world. This is not yet the case for CPE. If you look at the track record the last, I do not know, eight quarters, I think we move between low single digit to mid-single digit, and we are in this between. There is a fundamental reason, and it took us some time to understand this.

In fact, we believe it is only one channel, but the reality, we need to double-click on this. You have a fraction of the engagement which are really volume-based. What they are selling is CATIA, in a way, the same way the other channel are selling SOLIDWORKS. You have a fraction of this engagement which are serving the ecosystem of our most advanced customers. When those most advanced customers, they move to the 3DEXPERIENCE platform, they move to the cloud, they move to AI, you cannot only serve them with a pure reseller model. You have to be much more sophisticated. You need to be a system integrator in a way to connect with their information systems. You almost have to be an hyperscaler to advise them on the cybersecurity, on the cloud architecture, and so on.

You need to almost be an engineering services company to design with them the workflows, the new engineering workflows they have to do. You need also to be a consulting company to manage the changes. So, it is becoming a little bit tough for only one partner to do everything. So it is an ecosystem we have to architect, and that is the reason why, by the way, we hire Pierre Barnabé. Pierre is coming with a deep background in this industry. He is now in China this week. He is doing exactly what I am saying. This is what we are planning to change in order to have all the engine on fire. Last but not least, AI. This is true that we did not factor too much AI contributions in the coming years or quarters. Let me tell you why. We move from adoption to usages. I think this is extremely important.

Why so? Because if you do not have a clear picture on how the solution is used, we will not price the right way. So today, we started the adoption program, it is going well, and we will continue to foster the adoptions in 2027 with the idea that at the end of 2027, we will have a much better understanding on the usages. Based on the usages, you remember what Adrienne explained to you, we want to price on value. So based on what the people they really do, we will understand the real value we are creating for them. So this is how we will price, and this is how we will basically fine-tune the model in order to get most of the benefit of the contribution of AI. If you factor this, you have this curve Rouven showed.

Rouven Bergmann
CFO, Dassault Systèmes

Life Sciences.

Ashwini Zenooz
CEO of Life Sciences and Healthcare, Dassault Systèmes

Okay. Guess I am up. I will hand it to Aman. I appreciate your question. I think the biggest call-out I will make is that we have made a very strategic acquisition within the last two months. In the past, we have been competing with the rest of the industry on document-based, on Medidata, and we have been focused on the clinical components. Adding safety, and we will bring new products to market that will add to safety, and you will see that going into next year, we are now able to connect the dots from trial data into safety. That is really critical for our customers because that data can now go back into our agentic platform that we have been building on discovery. We have been working on discovery and lab in BIOVIA with our customers, but now with these synergies, they are really multiplied.

We are now able to take this data back to our customers and support them as they think about drug discovery. In the past, platform plays would have been rebuilding the platform, taking years to rebuild the applications. That is not the case today with the agentic infrastructure. We are able to have agents talk to each other, like Adrienne talked about earlier. We are able to have sub-agents that are super specialized that we can train and give them special skills, like MARIE and LOUIS, that has the intelligence that can build on. We are able to bring the data back without having to rebuild applications. These are technology synergies, but they are really significant for our customers because they do not have to take discovery data and re-enter them into clinical data. They do not have to take clinical data and then have to retransfer them into safety data. These transfers are critical for them.

We are a better partner to our customers, and we are going with the enterprise play. Of course, there are going to be customers that prefer some modules that are separate from ours, and that is why we are going to have the platform approach, but the modular approach that we can work with them. I guarantee you, we have best-of-breed solutions in what we have together today. Many of our customers are using that. Like I called out our metrics earlier, we are still leaders in the market. We have room to grow. Yes, we do. But the room to grow is we can innovate on top of what we have. We do not have to rebuild. I think that is important now with the agentic infrastructure underneath, and that will show. We are already cloud-based. We are ahead. We can predict, and we have reproducibility with ARR.

So, our conversations with our customers is, how do you make it frictionless for them for the purchasing of all of our products? It is not just safety, quality, regulatory, clinical, discovery, lab. We can now bring the power of manufacturing, which is so core to what Dassault Systèmes has been doing for 40 years. Our customers want that in med tech. Expect that to grow in our business. Expect us to bring medical device and technology with safety together. There are a lot of synergies that we can bring together going forward. Is it going to be immediate? No. We have to do a lot of work now to continue the progress that we have made on our product teams, on our sales teams, but also the ecosystem of builders.

Just like the other partners that we have, CROs are incredible partners with us for our business, and we are going to continue to invest and work with them. The strategics, the SI partners are also critical for us. We have a lot of those partnerships coming. Technology partners are really critical for us, like NVIDIA and others. You will see more of that coming. So expect to see more synergies over the next one to two years here.

Rouven Bergmann
CFO, Dassault Systèmes

Cool. Maybe I close this, your multiple questions more with the last part. How dependent are we going to be on license and maintenance in the future? I think giving you this framework for 2029, you see how we are converging to a more recurring model, and less and less of the growth is dependent on license. The support revenue essentially is going to remain flattish over time as more and more of our existing clients are transitioning to subscription. This is where the agentic platform will be deployed. One additional point which is important to understand in all of this is that all the new Generation 7 portfolio is online, it is cloud-based. There is no subscription and there is no maintenance business attached to that. It will all be fueling the subscription growth, and that is also not available as license.

It is an ongoing online service that you have to subscribe to, so there is no such thing than license business as we are transitioning. By design, the license part will further and further The contribution of the license part to our overall growth will further and further narrow, and be further and further less as subscription is the growth engine. I go back to the left side, and I see Laurent was asking a question.

Laurent Daure
Analyst, Kepler Cheuvreux

Thank you, Rouven. I have two. The first one is, I would be interested if you could share with us feedback from the 100 plus customers that are already using your AI use cases, and in particular, on a quarter-to-quarter, the ramp up of their consumption. That is the first one. The second one is for you, Rouven. I am struggling a little bit with the acceleration in ARR on the subscription, but you still have the ARR on the maintenance flat and not coming down, while I guess you are expecting some transition from your customer into the new model. How can you help us to understand why it should stay relatively stable and not declining? Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I will start maybe with the first one. Quarter-after-quarter, the consumption of the Blue Tokens, remember, we just introduced in June. I have only one quarter. Do not expect me to give you a quarter after quarter, over quarter, because right now, we do not have enough to answer to these questions. Nevertheless, when I look at the volume of token being consumed in a quarter, it is significant, in fact. The also interesting thing is, it is not equally split across all of them. This is also an interesting lesson learned. We could have expected, in fact, because we do almost the same kind of adoption program with all of them, but some of them are basically deploying faster than the others. This is also a second topic we should master.

The most interesting question for me is not this one, it is, are we creating value or not with it? Because at the end, that is the reason why we do it. Here is the findings. On all the use cases we are targeting, it is either a factor two or a factor 10 improvement. I repeat, it is either a factor two or a factor 10 improvement. That is the reason why you understand why we want to take the time to price it adequately. Because otherwise, if we rush, we can basically miss the real opportunity, which is to really price the value of the outcome those guys are producing. When I was telling you we are building a virtual twin factory, what does it mean?

I have the operating model, I have the factory, if you want, with the AI factory in conjunction with the virtual twin factory, and with the Virtual Companions, I have the workers, the virtual workers. It is pretty new. This is the combination of the two we want to assess. The last comment I can make, the adoption is going across the board, so it is not only one specific geo, it is really everywhere. We are basically working either with large customers, but also startups. The last comment maybe I can make is, it is extremely easy for a CATIA user or a SOLIDWORKS user to adopt it, which is also very important because for many years, my dream was to democratize what we do, right? You remember, I was speaking about mainstream of the mainstream.

That was also maybe the initial plan with 3DVIA, is there a lowest, let's say, positioning of the products to have a mass adoptions of things we do. I think with AI, including with the Frontier model being connected directly to our applications, we have a unique way to democratize what we do. This is all of this we are observing for the time being, and I will report to you in Q4.

Rouven Bergmann
CFO, Dassault Systèmes

Okay, and thank you for the second part of your question on ARR and contribution of subscription and the support run rate. You are right to call this out. Over the next two to three years to support run rate, we expect it to remain flattish. And of course, if you go beyond that, as a subscription and more and more contracts will convert, it is expected to come down year-over-year over year. But keep also in mind that this part of our business is very sticky, and it takes quite some time for clients to convert. It is not happening overnight. So, to assume that this is now falling down step by step pretty rapidly is the wrong assumption. It is very sticky. The majority of this revenue is related to the V5 portfolio across our applications.

And those V5 applications, they can also be connected, and as you saw in the business model that Adrienne presented, they can very well be connected to Frontier models. So their lifetime is probably longer than you expect. Nevertheless, they will be connected to the platform, they will be able to do many things. That is why we have been taking this approach to consider it flattish, maybe marginally decline on a quarter-to-quarter basis, but overall to remain flattish. And really, the acceleration comes from the adoption of subscription, the expansion on the cloud, and the net effect is at the end what matters. That is the three point of acceleration.

And that can either be delivered through faster acceleration of the subscription side or slower slowdown on the maintenance run rate, or a faster slowdown on the maintenance run rate in combination with a higher growth on the subscription run rate. So you need to give us this flexibility. What we provided is a framework, and of course, we will adjust this as we learn more, but we feel confident with the assumption and the outlook. Jay, you were very fast raising your hand. Up to you, over to you.

Jay Vleeschhouwer
Analyst, Griffin Securities

Thanks. Thank you, Rouven. Please, Jay, Griffin Securities. Couple questions. Number one, let's think about your growth equation somewhat differently as P times Q times V or plus V. With regard to the P, for the last number of years under your current model, you've been raising your prices pretty broadly 2% to 5% a year every summer with the annual price increase, more or less, right? Under the new world, how are you thinking about the headroom you may have for higher price increases than that on average over time? Do you think that the new model gives you more capacity to raise prices than you've been doing to date? On the Q part, in your annual reports over the last number of years, you've cited new customer or new user acquisition as among your highest priorities. Maybe you could talk about that.

I assume it's still a high priority, but maybe talk about how you go about that. Part number two, you've got one of the largest R&D armies in the industry. Not necessarily the highest R&D budget, but the largest armies spread across all your segments and brands. As the portfolio is evolving, which it so obviously is in multiple ways, how are you thinking about better managing R&D in terms of time to market, effectivity, and so on and so forth? Lastly, you cited the relationship with NVIDIA, but everyone has a relationship with NVIDIA, right? More or less. So in what other ways do you think you can have partnerships that are more specific to you or more differentiated for you, rather than with one company that everybody works with?

Pascal Daloz
Chairman and CEO, Dassault Systèmes

You asked so much question, I don't remember the first one.

Rouven Bergmann
CFO, Dassault Systèmes

The price.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

The pricing. It is an interesting point. Let me step back a little bit, and you have the story, so you will understand where I want to go. Let's say before we face the huge inflations periods, after COVID times, the real price increase was attached to when we were migrating to a new version, from version 4 to version 5, version 5 to version 6, and version 6 to 3DEXPERIENCE platform. If you remember, we, for example, from the move from V5 to 3DEXPERIENCE platform, we came with this named user model, which was a way to attach the usage of the license to a users, designated users. This is basically the way we were improving our prices over the time.

When the hyperinflation came, what we did, we have introduced in many of our contracts basically an index, and we are following this index. This is true for the large contract we signed, but it is also true for the mainstream market. This is how we were managing and we are managing the pricing today. Right, Rouven?

Rouven Bergmann
CFO, Dassault Systèmes

Yeah.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Now, the question is, the good one, is the agentic approach, the AI approach, a way to change it? I think this is, again, what Adrienne explained. She said, "We do not want to price the capability anymore. We want to be able to price the output. We do not want to price the amount of work. We want to price the complexity of the output." This is the chance we have, and I do not want to miss this one. You remember we pushed this idea for many years. The only way we had in the past was to have success fees. Frankly speaking, that was difficult because the way people are using our software was not dependent on us. The way they were structuring their processes was not in our control.

The only way to commit ourself on the end result was, okay, let's define some KPIs, reduction of the cycle times, productivity gains, and we take a piece of it. Now, everything is in our control because the Virtual Companion, we know what he's doing. The Generative Experience is basically the workflow they're going to follow. It's on the cloud, so basically we have all the metrics and all the KPIs to price in a different way. This is the reason why, again, we are so sensitive to this topic. The second question was?

Rouven Bergmann
CFO, Dassault Systèmes

Yes, the new customer acquisition.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

The new customer acquisitions. Let me give you some statistics. We have 400,000 customers. I repeat, we have 400,000 customers. Do you know how many customers SAP they have? 200,000. Do you know how many Salesforce they have? 120,000. Why I am saying this? Because, of course, if we have such a number of customer, the acquisition of the new customer is still the engine. The way it works, it usually start with the SOLIDWORKS install base, because this is how we are establishing the footprint. In the past, we had almost 60% of the growth which was coming from new customer acquisitions, and 40% was coming from the upgrade of the install base. Today, we are basically revert.

I think we have 60% of the growth coming from the upskill, upside, or upgrade of the install base, but we still have 40% of the growth coming from the new customer acquisitions. It is part of the engine, and this is the first driving force. The second one is the industry approach. We are well-known in the auto sector, but believe me or not, you have newcomers still in this industry. You have seen many Chinese emerging. All of them, or most of them, are our customers. We are diversifying in the consumer industry. You have seen it with Centric, and we report to you almost every quarter how many brand we are winnings. Today, if I remember well, it is around 50,000 brands around the world who are using Centric PLM, which is amazing.

It is a large market, so it is really an untapped market, and we could continue to expand. By the way, now PTC being part of Schneider, I think it is even better for us, because they will be probably the focus on this sub-segment of the market. Life sciences, we have 50,000 customers only on the med device. Only on the med devices. It is again a market where and most of it is coming from SOLIDWORKS. Many of them, they need to do a clinical trial for the category 3 of their med device. Many of them, they have the same obligation than the pharma. They need to follow the adverse effect, the safety cases. So, only few of them are really doing a pure lifecycle management. This is, again, a market where we have untapped reservoir to continue to expand.

So, to answer your questions, yes, it is still part of the growth engine.

Rouven Bergmann
CFO, Dassault Systèmes

Yeah, and then another question was related to the R&D organization effectiveness.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

This one is a very interesting question. You have two different ways to look at it. First of all, we are not hiring coders since day one. I repeat, since day one, we set up the company, the goal was we want to have a high level of engineers. And there is a difference between an engineer and a coder. An engineer is the one who understands the problem and is able to solve the problem. The coder is the one translating how you are going to solve the problem into code. So, why I am making this? The vibe coding is eliminating the second category, not the first one. And if you could do more with vibe coding in order to accelerate your delivery systems, I think more than ever, we are committed to do this.

Florence Hu-Aubigny, our Executive Vice President for Research & Development, she has set up some objectives for the R&Ds to speed up basically the delivering process, not only to do tactical things, but also to go faster with more capabilities, more rapidly, have an earlier feedback loop with the market. Now, in terms of productivity, how do we factor this? I do not plan to reduce my R&D capacity. I repeat, I do not plan to reduce my R&D capacity. And why so? Because there are so many industries, so many challenges of the industry which are not well served today. It will be a complete mistake to do this. Hiring someone, training someone, building the loyalty with the company, it takes so much time that I think if you have this capacity, you should keep it. But you should give them the ability to do more with less.

Less resources, less time, and to do more with what they have. How you will see this? If you look at the track record, basically when we are growing at 9%, 8%, 9% organically, we are increasing our headcount at around 4% to 5%. That's basically, it's easy. You look at the quarterly result or maybe the annual report, and you have all the information. Now, you have seen when we are growing at 7% to 8%, the growth of the headcount is much more around 2%. And that's the way we will basically plan, knowing that we have attritions anyway, every year. To give you some statistics, we used to hire 5,000 people worldwide, gross, net 2,500 people. Now, we are basically hiring 2,000 gross and much more around 1,500 or 1,200 net.

Not taking into account the acquisitions, because ArisGlobal by itself is bringing 2,000 people, right? Okay. NVIDIA.

Rouven Bergmann
CFO, Dassault Systèmes

NVIDIA.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

There is a NVIDIA question.

Rouven Bergmann
CFO, Dassault Systèmes

Absolutely, it's important.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I will debate this. NVIDIA, they do not have a direct sales force, right? That is true. They are selling through the ecosystems. Their model is by design an ecosystem one. That is the reason why they are partnering with all the different players around the world. But they are not inventing with all of them. We are one of the few inventing with them, or co-inventing with them. That is how we are making the difference with the others. The net of this is what we have not only your people supporting how to integrate their technology and so on, but also we have access to their lab, we have access to the high executive within the company in order to fine-tune what they are doing. There are many technologies they are putting on the market which are not yet ready, believe me.

They count on us to industrialize what they do in conjunction with us. This co-development and co-engineering activities is something we know how to do with them. We have done it for the last 25 years, and we will continue to do it.

Rouven Bergmann
CFO, Dassault Systèmes

Okay, I think we have maybe time for one or two more questions. Charlie, why don't you go next?

Charlie Brennan
Analyst, Jefferies

Thanks. It is Charlie from Jefferies here. I will do two quick ones if I can. Firstly, there has been a lot of questioning on the linearity of the revenue growth. Can you just touch on the linearity of the margin improvement? It feels like you have got some AI investments at the moment going on, but the AI revenue won't flow for a few years. Should we assume that the margins are backend loaded like the revenue growth? Then just when we think about the investments, can you just touch on go-to-market and marketing? I have spoken to a number of your partners who are frustrated with some of the marketing messages from Dassault. Even today, I think we have heard about Gen 7, we have heard 3D UNIV+RSES, we have heard 3DEXPERIENCE agentic platform. None of those is a product that customers can actually buy.

Do you think that you need to invest in simplifying the marketing message?

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Let me react. First of all, this is completely wrong what you just say. You can buy the 3DEXPERIENCE agentic platform. You can buy the Virtual Companions. You need to change, Charlie. I think I'm starting to be fed up with the filth you spread on the market. You are not professional when you do this. Do your job, please. There are price lists, they are available on our website. You can go. Maybe all the partners are not yet ready to sell it, which is a different topic, and it refers to my previous comments. Now, I think if you heard very well what I just say, I pronunciate only once 3D UNIV+RSES. I spend time to speak about the virtual twins, the 3DEXPERIENCE platform, even if the platform is becoming agentic, and Virtual Companions.

How we are differentiating ourself compared to the Frontier model, the Industry World Model. I think it's not too much to ask you to master four concepts, guys. That's my view. Now, coming back to the marketing, yes, we could improve, of course. Especially why? Because for a long time, the human was generating the content. That's the reason why we have, many of us, we have marketing organizations to generate content. Now, AI is relatively changing the things because it's easy to produce a video, it's easy to produce PowerPoint presentations, a website, and you could be much more tailor-made. You could basically being much more personalized in the way we want to address our customers. That's what we are doing, and this is how we are moving forward.

Rouven Bergmann
CFO, Dassault Systèmes

Okay, now to the-

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Let me give another comment. Please do not make the confusion between speed and consistency. Think what we do, we are consistent. You have seen +1, +3, +4, +5, between +5 to +8. This is consistency. The consistency is also on our operating profits. I think we are well-equipped. The plan has been well-designed to not make the confusion between speed and consistency. Now, Rouven, you can give number if you want.

Rouven Bergmann
CFO, Dassault Systèmes

That is, I think, exactly the consistency you heard. This message is really what you should take away. When we give a three-year outlook, we cannot expect that we give for every year a number. That is why it is a three-year plan. We are, I think, very clear on where we want to be by three years. I also do not, I would not put the headline over it that it is backend loaded. It is not. It is consistent, and there is a progressive acceleration over time, and that is not backend loaded. It is not a hockey stick plan. It is consistent. This year, you see where we will land, 4%-5%. Next year, on an organic basis, I say, right now, we are at the mark of 5%.

Next year, we will move towards that mark, and then the year after, we are going to move beyond that mark to be closer to 8%-10% by 2029, and reaching the 10% in 2029. That is the trajectory on what we are on, and I think it is very consistent, and it is very much aligned with our investment strategy. I think one final word with regards to the need to invest into AI and how it impacts our profit margin. I think this argument, we heard this also at the time when we started to invest into OUTSCALE and cloud. It was not a real argument because the margin during that time has consistently evolved. We price the cloud for value. There is a lot of value that comes with it. There is a lot of value that comes with AI and AI factories, and it will be properly priced.

While there is some upfront investment to this, there is also a return on investment, and we will be able to ensure that we sequence that in the framework that we have provided to you. The 80 basis points of margin improvement, we have the operating leverage as we accelerate our top-line growth, and we will price for value. That is what we have done in the past, and we will continue to do. Fréd.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I think this will be the last one.

Rouven Bergmann
CFO, Dassault Systèmes

Yeah, the last question.

Fréd Boulan
Analyst, Bank of America

Thank you. Fréd Boulan at Bank of America. First question, you've rebased 2029 ambitions today versus your plans. Can you share with us your level of conviction and visibility on that versus where we were last year? I feel much more conviction, but if you can share a bit more granularity on what drives that. Maybe a question for Rouven. Today, you basically doubled your ambition on annual margin improvement, so it would be versus the previous plan. So it would be great to have a bit more granularity on the drivers. Maybe a question to Ash. Any specific investments, either in products or distribution, you think are necessary to achieve your ambitions? Thank you.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

I will start with the first one. Again, we are one of the few companies giving a five-year guidance. I repeat, we are one of the few companies giving a five-year guidance. Most of our peers, at best, they give three years, and sometimes they give only one year. Why I'm making these comments? Because I think as a CEO, I don't care if I achieve the number year five or year six or year four. I want to achieve it, and I want to be sure that we have built a trajectory to continue to grow. That's my first way to answer. Now, guys, you have already integrated this. The consensus is at EUR 1.6, and we are telling you the EPS will be EUR 1.8. So, is it a rebase or is it an upside compared to what is currently in the market? Think about it.

The third, I think where we spent time with Rouven the last few years, when we transitioned the model, we did not have the instrumentations to guide you with high level of precisions. I think the work Rouven and the finance team have done to develop the annual run rate concepts and how they are basically producing it and applying it to guide our businesses is much stronger. I think now we have a much better visibility in everything we are giving to you. That is my comments. You will draw maybe the conclusion that the confidence is higher.

Rouven Bergmann
CFO, Dassault Systèmes

I think for the margin part, Fréd, that you see how we have been able to expand our margin in 2026, even at a lower revenue growth. You can see the leverage we have in the model. We have certain parts of our business, where the operating margin is still fairly below the company's operating margin, though there is room to improve the margin as these businesses are growing. I put it in parentheses, looking one part is into consumer-centric industries, where there is enormous amount of scale in front of us that can be leveraged also on the operating side. I give maybe with this second comment, I hand over to Ash, is when I think about life sciences, the real focus is now on top-line growth acceleration and less on margin.

Ashwini Zenooz
CEO of Life Sciences and Healthcare, Dassault Systèmes

To that point, I would say I have spent the last two months and will continue to through the end of this quarter, really focusing on getting our distribution teams aligned, hyper-focused on how we are bringing the products to market to our customers, not from a product perspective, but really on how do they see the contracts. We do not want them to see eight contracts. I think you and I talked about that yesterday. We want them to see one logical contract that comes from Dassault Systèmes or from our brand that represents all of the incredible products that we are bringing to them. But going into 2027, I am going to be spending a lot of time.

I know we talk about AI, and that is the sexy stuff, and we could do a lot of demos, but the real work that needs to get done is the integrations on the back end and creating the synergies. Again, we have done this incredible strategic acquisition. It can add a lot of value to us, to our top-line number, but also to our product portfolio, and we really need to capitalize on that. So much of the first quarter or two of next year is going to be hyper-focused on how do you bring the integrations of all of the products that we have today to make it more usable and useful to our customers. It is very customer-focused and customer-centric going into next year.

Pascal Daloz
Chairman and CEO, Dassault Systèmes

Thank you. I think this is concluding the Q&A sessions. Maybe I will just take a few minutes to do the closing remarks. I really want you to leave with three messages. The first one is, I think we are building the industrial infrastructure for AI. This is opening up new possibilities that were difficult to imagine just a few years ago. Making AI work in the real world require more than algorithms and computing power. I guess you got it. It requires science, engineering, industrial knowledge, trusted virtual twin, and this is what we have done for the last 45 years. This is what we have built for the last 45 years. We are not simply adding AI to our portfolio. I think we are connecting the intelligence to the reality. That's what we do.

The second one is AI is unlocking the economic potential of our entire businesses. You have seen new users, new applications, new industrial workflows, new opportunity to create and capture the values. I think our Virtual Companions are really making our expertise, if you want, accessible to a much broader group of users. Our 3DEXPERIENCE platform is now reaching further into the customer businesses. You heard from the customers themselves. I think we are creating new opportunity for relationship we have built over many, many, many years. At the same time, I think we are strengthening our existing businesses, improving the execution. I think Rouven spoke about it. We are not relying on one single source of growth. You have seen all the businesses are contributing, and our mission is really to make the whole company stronger at the end.

The third, I think we are entering in a new financial cycle from a stronger positions. We have made the investments, we have built the technology, we have strengthened our business model, we have taken actions to improve the executions, and now we have a clear financial commitment for 2029. Our commitment is to turn this opportunity into profitable growth, strong margin, sustainable cash flow generations. They are not just ambition, I think they are commitment, and we really intend to deliver them. Now, let me finish with probably a much more personal note. When I look back our 45 years journey, and as far as I'm concerned, at least the last 25 years, because this is when I joined Dassault Systèmes, but I think there is one thing which has never changed since day one, since Charles Edelstenne, Bernard Charlès created, co-founded this company.

Our belief is that the technology should help people to do things they couldn't do before. Think about engineer designing the next generation of aircraft, the scientists working on the discovery of the new treatments, the people developing autonomous systems, the industrial teams building the factory of the future. Behind all of those innovations, there are people trying to solve incredibly complex problems. This is the reason why we are here. We are here to help them to succeed. We did 3D design at the time to do complex products. We helped engineer to imagine and design products in the virtual environment. Then we connected the design and the manufacturing through the PLM to provide one integrated industrial operation systems. Then we brought people, processes, and knowledge together through 3DEXPERIENCE platform because we knew at that time the knowledge will become the new currency.

Today, what do we do? We are taking another step with our AI-native virtual twin and our agentic platform. The technology is changing, the possibilities are expanding, but the missions remain exactly the same, to help people to turn ideas into the reality. I believe AI gives us an extraordinary opportunity to take this mission to a complete new level. For 45 years, we have helped industry to imagine the future. I think today we are connecting intelligence to the reality, and tomorrow we will help industry achieve things that were previously impossible. Of course, we will deliver the financial commitment we made to you. This is the next chapter of Dassault Systèmes. This is our commitment. Thank you, and hope to see you in the coming weeks and days.

For the one who will attend the Barcelona conference tomorrow or the day after, we will be there. Thank you.