Dassault Systèmes SE (DSY) Q4 FY2025 Earnings Call Transcript & Summary

Good day, and thank you for standing by. Welcome to Dassault Systemes 2025 Fourth Quarter and Full Year Earnings Presentation Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to our first speaker today, Marie Dumas. Please go ahead.

Hello, and thank you for joining our fourth quarter and full year 2025 earnings conference call. I'm Marie Dumas from Dassault Systemes, and I'm with Pascal Daloz, Chief Executive Officer; and Rouven Bergmann, Chief Financial Officer. Dassault Systemes' results are prepared in accordance with IFRS. The financial figures discussed on this conference call are on a non-IFRS basis with revenue growth rates on a constant currency basis unless otherwise noted. Some of the comments on this call contain forward-looking statements that could differ materially from actual results. Please refer to today's press release and the Risk Factors section on our 2024 universal registration document. All earnings material are available on our website, and these prepared remarks will be available shortly after this call. I would like now to hand over to Pascal Daloz.

Thank you, Marie. Good morning or good afternoon, everyone. Thank you again for joining us today to review Dassault Systemes performance for the fourth quarter and the full year of 2025. Let me start with some opening comments. At Dassault Systemes, I think we don't only manage for the quarter. We build platform for at least decades. I think if I have to qualify 2025, 2025 was a year of transition. And 2026 will be the year of execution. So they are foundational years, years when we prepare the next cycle of growth, scale and the long-term value creation. Now let's start with the facts. 2025 was a disappointing year for you, but also for us. We finished at the low end of our objective with 4% growth ex FX. And I think this performance does not meet the standard we set for ourselves as part of our long-term plan. We own that. Now having said that, I think we are moving the company forward. Well, what moves forward? First, 3DEXPERIENCE and the Cloud. We delivered significant wins this year and we did also a lot of competitive displacements. And in 2026, we will build on this momentum. Second, MEDIDATA and PLM CENTRIC, they faced challenges in 2025, and they weighted our results. I think we are seeing early signs of recovery at CENTRIC. And for MEDIDATA, we are investing for the long term. Because Life Sciences in a way, undergoing a fundamental transformation, from an efficient document-based processes to an AI-powered Virtual Twin. This is a shift which is structural and these structural changes are taking time. The third element is, in 2025, we have introduced 3D UNIV+RSES. It's a new environment where Virtual Twins and AI converge, connecting the virtual and the real world together in a seamless, dynamic look. In 2026, we want to turn this vision into concrete value. And finally, we remain disciplined on our costs. Rouven will give some detail about this. Why? Because we want to continue to invest in our future growth. So execution matters, but return as well. As we are entering in 2026, we are scaling our transformation plan along 3 strategic pillars. The first one is the product offerings. We are really reshaping our portfolio, accelerating towards the 3D UNIV+RSES. And really, you should keep in mind that AI is not an extension of what we do, it's a redefinition of the entire portfolio the same way we did almost 15 years ago with 3DEXPERIENCE platform. Second, the go-to-market. We are strengthening the go-to-market executions with a targeted end-to-end engagement, especially in Life Sciences for the top largest 50 accounts, and consumer industry, more precisely to the formulated products. We are also transforming our partner ecosystem to generate demand, not only to distribute license. Third, as our customers accelerate their adoption to subscription and cloud, we are introducing the annual run rate, ARR, reporting in 2026. Why we do this? Because I think subscription is now representing half of the recurring revenue, and this is giving a better visibility into health and the momentum of our recurring revenue base. In parallel, we are also evolving beyond seat-based pricing towards value-based model for the AI powered solutions. Now let's step back and look at the market realities. In every industry, we serve manufacturing, life sciences, infrastructure and cities, most of them are under intense pressure. Supply chain volatility, rising regulation, aging infrastructure. And there is an urgent need for breakthrough innovations. These are not only constrained, they are catalysts. And this is where Dassault Systemes step in. So in manufacturing, we see 2 realities. Traditional sector face margin pressure and demand uncertainty. At the same time, defense and high-tech are making bold investments where complexity and collaboration are the new normal. And here is really where the 3DEXPERIENCE platform is becoming the de facto standard, reducing program time lines to under 18 months in transportation and mobilities, delivering 25% to 40% efficiency gains in aerospace and defense, and cutting error by more than 1/3 to half in high-tech through pre-build simulations. In Life Sciences, the pressure is also intense. Tighter regulation, rising R&D costs and the shift toward precision medicine. So the customer, they need a new operating model. With our lab-to-manufacturing solution, we help them reduce operating costs by over 30%, while we are turning compliance into a competitive advantage. In Infrastructure & Cities, demand for autonomous and resilience system is also accelerating. So the data center demand will double by 2030. The nuclear infrastructure requires safety commissioning in many countries, and city needs resilience by design. Now across every sector, our customers prove one thing. We don't just talk about AI, I think we deliver it. And value is a great example in transportation and mobility. You know all of them. They are global leaders in automotive technology, from ADAS to electrification systems. Together, we are pushing the boundary with generative experience. Here, AI does not assist, it cocreates by training virtual twins on synthetic data. We generate thousands of design alternatives optimized for performance, cost and compliance. And we do this before one single prototype is built. In Life Sciences, our partnership with Catalyst shows how the industry can be reinvented. By moving from static document to data-driven virtual wins, Catalyst is really redefining the CRO models, with clinical trial becoming more agile, patient centricity and continuously optimized. In infrastructure, with TECHNICATOME, we are redefining how next-generation nuclear system are designed and operated, using virtual twins to connect the entire ecosystems, ensuring the trustability and the compliance by design. Now let's move to 3D UNIV+RSES. Remember, last year, we introduced 3D UNIV+RSES. But what does it mean in practice? 3D UNIV+RSES are not application the way you know what Dassault Systemes is doing. They are knowledge factory. This is really how you could see them, where knowledge could be enriched, when the know-how is scaled and where results can be trusted. And AI is the engine. But it's not an LLM story. It's not about large language models. Because the LLMs, they don't know how to build drones. They don't know how to design humanoids. They don't know how to discover cell therapies. And this is precisely what our customers do, and we help them to certify it. So what we are doing, we are building industry world models, models that understand how the real world works, and also how you built it. And why? Because it's built on physics, it's trained on decades of industrial knowledge, continuously validated by our virtual twins. And the result is explainable, certifiable and trusted. So this is why our partnership with NVIDIA matters. Together, we are combining the virtual twin with the AI factory and an accelerated computing. Launch the video, please. [Presentation]

Okay. So I hope you had a chance to see the video. I'm not sure because I'm blind here in my conference room. But in case you have not, the video was basically presenting the purpose of the partnership, with, by the way, the voice of Jensen. Jensen was on stage with me last week at the 3DEXPERIENCE World we had in Houston, a very, very successful event. And why this is important? Because we are building the foundation of industrial AI, to enable, in fact, 3 things. In research and innovations, we need to create foundational model who understands the causality, not only just a correlation with a statistical approach. It doesn't work. Second, for the factory of the future, we need to invent software-defined autonomous factories, continuously optimized through the simulations. And last but not least, for the designer, for the engineers, we need to develop skilled virtual companion, not chat bots, but industry trained experts. So that's what we do. And in 2026, we are turning this into a reality with our new category of solutions. The first one, the virtual companions, they are not assistants. They are experts. They scale the knowledge. They democratize the expertise. They turn complexity into productivity. The second one, the generative experiences, is really where AI and codes, if you want the best practices. Whatever the best practices, it's coming from the science, the manufacturing, the engineering. And more importantly, we are compliance by default. The third one is the virtual twin as a service. Here, we are not selling any more of the software, the tools. We are selling the end results, which is the Virtual Twins, and it becomes an outcome. So in parallel of this introduction of the new categories, we are also evolving our business model, from seat-based licensing to value-based monetizations. Why? Because we want to use this new generative solution as a way to unlock 3 powerful levers. The first one is expanding the adoption with the virtual companions, mainly using -- mainly building a usage-based business model. We want to monetize the know-how with the generative experiences. And as I was saying, we want to sell the outcome with the virtual twin as a service. So now I would like to leave you with one message before to hand over to Rouven. Dassault Systemes is really undergoing a profound transformation of its business model. Transitioning digitally towards a subscription-driven future. The shift is powered by not only the acceleration of the cloud platform strategy, but also, and probably above all, by the evolution of our offering with 3D UNIV+RSES, the new category of solution I just mentioned: the virtual companion, the generative experiences and the virtual twin as a service. This transformation is not incremental. It's a fundamental one. But it's also highlighting one thing: the resiliency of our model. Because we continue to grow at a moderate pace, yes, but with a sustained momentum. And you should remember that in our industry, many of our peers, they struggle in such a transformation. So we are executing, we are innovating, and we are advancing. And I think our guidance for 2026 is reflecting not only the conviction we have, but also the confidence we have. Now it's time for me to hand over to you, Rouven.

Thank you, Pascal, and also welcome from my side to all of you on this call. Welcome. Have a good afternoon and good morning, depending on where you are. Before reviewing the numbers in more detail, I would also like to share 3 key themes that defined 2025. The first one is our core industrial business was resilient in '25 with strategic client wins. However, we faced the backdrop of tough comps and a complex macro environment in the fourth quarter. We are focused on strengthening our growth model and operational excellence, and we have identified the challenges and we will now execute to deliver. The second message is centered around the business model evolution. The 3DEXPERIENCE platform continues to drive the transition towards cloud and subscription, with an increasing share in recurring revenue. As AI adoption accelerates, our business model is evolving beyond the traditional seat-based pricing towards usage and value-based models. And to better reflect this shift, we will begin reporting an annual run rate, or ARR, and I will talk about this in more detail later. And third, as Pascal said, 2026 will be a year of execution where we will strengthen our foundation. Our full year guidance for total revenue growth of 3% to 5% provides us the room to navigate current challenges as well as to prepare the organization for a new era of growth. With this in mind, let's review the financials for the quarter and the full year in a bit more detail. In Q4, total revenue rose 1% ex FX to EUR 1.682 billion, with software revenue up slightly, plus 0.3%. We navigated a complex macro environment with weakness in France and Germany, mainly in the auto sector, plus headwinds at MEDIDATA and CENTRIC. We have taken the actions to address these issues, which also I will cover shortly. Recurring revenue rose 3% in Q4, with 4% subscription growth, and services was up 11%. Operating profit for the quarter was EUR 622 million, with a healthy operating margin of 37%, up 90 basis points ex FX, thanks to productivity gains across the group, which we had initiated entering in the year. EPS was EUR 0.40, up 9% ex FX. For the full year '25, we saw total revenue at EUR 6.240 billion, along with software revenue growing at 4%. Recurring revenue grew 6%, and it's making up now 82% of our software revenue. And subscription was up 11%. We delivered good profitability in 2025 with an operating profit of nearly EUR 2 billion and an operating margin of 32%, achieving 40 basis points of improvement versus last year. The EPS was EUR 1.31, and it was up 7%. Now turning to the growth drivers. The 3DEXPERIENCE platform is at the core of our growth strategy and the foundation to review the power of AI for industry. 3DEXPERIENCE revenue grew 10% for the full year and it's making up almost 41% of eligible software revenue. As expected, the fourth quarter was impacted by a strong year-on-year comparison. On top, we faced a weak auto sector in Europe. However, important to highlight, we signed several strategic 3DEXPERIENCE deals that have the potential to further expand over the course of 2026 and 2027. This will generate future revenue and it helps to build the momentum in ARR, which I will come back to shortly. Cloud revenue at the group level grew 9% in Q4 and 8% for the full year. The 3DEXPERIENCE, Cloud grew a strong 38% and 32%, respectively. This strong growth highlights the value of our platform for clients where transformation is critical, as is the need to leverage AI. Now looking at the geos and product lines. The Americas rose 3% in Q4, with a good performance in high-tech and transportation and mobility. For the full year, the Americas was up 5%, and it was impacted by the weakness in Life Sciences and Home & Lifestyle. Now for the core industries in the Americas, the growth was 10%. Europe declined minus 5% in Q4, but it was up 2% for the full year. The weakness in the quarter was against a strong comparison base and with softness in France and Germany, which, as mentioned before, was mainly driven by the challenges in the automotive sector that we encountered in Q4. Meanwhile, Southern Europe was resilient and Northern Europe gained momentum with strong performance in High-Tech. The performance in Asia was robust. It was up 6% in the quarter and 5% for the full year. The growth was driven by Transportation and Mobility and High-Tech, with strong momentum in Korea and India, while Japan delivered solid growth. China had a softer quarter on a backdrop of tough comps. Now to the product lines. Industrial Innovation was up 1% in Q4 and 6% for the full year. As noted, the quarter was impacted by the lower growth in 3DEXPERIENCE and the particular challenges in Europe. But overall, the full year saw good momentum led by solid traction with our brands SIMULIA and ENOVIA and continued solid growth with CATIA. We are confident on the resilience of our core business, which is led by the cycle of 3DEXPERIENCE adoption while, at the same point, we are preparing for the next wave of growth with AI-based virtual twins and companions. On Mainstream Innovation, Q4 was up 1% and 2% for the full year. Growth again was driven by the strong momentum of SOLIDWORKS, which was up high single digit in Q4. And as expected, CENTRIC was down double digits in Q4 on a high comparison base. 2 effects played a role here for CENTRIC. First, we had some shifted renewals; and second, we saw an acceleration on move to cloud. We expect a marked recovery this year in 2026 with new management in place and a robust pipeline that's building going forward. For Life Sciences, the growth was lower than expected. Revenue was down minus 4% in Q4 and minus 2% for the full year as we faced continued headwinds for MEDIDATA, which I will cover in more detail shortly. Outside of this, MEDIDATA signed several strategic account win-backs over the course of the year. This included the likes of Novartis, Merck KGaA, AbbVie and Gilead. It highlights our competitive advantage as we build a strong foundation and expand our footprint within the large pharma enterprise. Now as we look ahead, we believe, and we're convinced, that the time has come to transform the biopharma industry from a document-based approach to a virtual twin based operations. It is our Life Sciences vision for the long term. Therefore, let's take a holistic view on our Life Sciences industry software revenue as it is today. So this first includes MEDIDATA as well as the 3DEXPERIENCE portfolio adopted by pharma and medtech. And in order to better highlight the growth dynamics, we are differentiating the direct model and the indirect go-to-market model. The direct enterprise business accounts for about 70% of our total Life Sciences revenue. And this business grew 3% in 2025. Within that direct enterprise business, MEDIDATA enterprise grew 1%. However, this growth was impacted by Moderna, one specific client, which was adjusting its run rate to reflect lower study volumes. If we exclude this adjustment on the Moderna renewal, the enterprise business for MEDIDATA was up 6%. Meanwhile, the 3DEXPERIENCE business grew 7% in the enterprise segment. Now let's take a look at the -- at what we call the indirect business or volume business where we are selling through CROs. This accounts for the remaining 30% of our Life Sciences business as of today. And this one is declining by 5% year-over-year. Our market saw a lower study start volume of minus 7% year-over-year in 2025. Now the revenue was down minus 5% in this segment on a reduction of study starts of minus 7% across the industry. Now importantly, we continue to expand our market share in this segment of Phase II and Phase III trials. So what actions are we taking to reinvigorate the growth? To address the enterprise, we are setting in motion dedicated account teams to focus on pharma transformation with platform and AI. These teams are formed and in action across all the geos. And for the indirect volume business, the goal is to reduce our exposure to volatility in the volume business. To this end, we are evolving our pricing model and terms and conditions to monetize continued data access. Now in order to leverage AI models, this is necessary. When you want to optimize the design of clinical trials, you need to be able to access the data and the platform. Now turning to the cash flow and balance sheet items. Let's start with the operating cash flow. We generated EUR 1.630 billion in operating cash flow year-to-date, up 1% compared to last year on a constant currency basis. Indeed, despite a challenging environment marked by FX headwinds and new tax regulations, we saw resilience in cash generation. As previously discussed, we absorbed approximately EUR 40 million in 2025 driven by equally the hike in employer contributions on share-based compensation and the new exceptional tax contributions for large companies in France. Excluding this, operating cash flow grew 3% ex FX. In the first half of 2026, we expect the working capital to be positively impacted by the collection from large subscription deals that we signed last year. Now to the free cash flow. It was up 2% ex FX. CapEx investments were lower by approximately EUR 30 million due to lower investment and leasehold improvements versus 2024, while investments in cloud and IT infrastructure were stable. Cash conversion remains a top priority. We reached 82% for the full year '25 versus 84% last year. This -- the year before, 2024, this is ahead of our previous estimates, mainly due to higher collections. In 2026, we expect cash conversion rate to improve driven by cash collections and better alignment of billing to revenue. Now to complete the picture. Cash and cash equivalents totaled EUR 4.125 billion at the end of '25, which compares to EUR 3.953 billion at the end of '24. This increase of EUR 173 million includes a negative full year currency impact of EUR 263 million, which is mainly due to the weakening of the U.S. dollar to euro over the period. The net cash position reached EUR 1.530 billion at the end of Q4 '25. Now any additional information, you will find in the operation -- and the operating cash flow reconciliation in our presentation, which we published earlier today. Now let's transition to the ARR disclosures. As we previewed with you at our Capital Market Day in June 2025, we are now introducing an annual run rate, as a key -- we are now introducing the annual run rate, ARR, as a key metric to reflect our continued transition towards a subscription and cloud-based business model. We believe that ARR provides a consistent view of the underlying run rate and the health of our recurring revenue base, while eliminating the volatility from revenue recognized. As such, ARR is a snapshot reflecting the 12-month recurring value derived from all the active contracts at period-end. This includes software subscriptions, cloud, SaaS hosting as well as support. And it excludes future commitments. In the appendix, you will find a detailed definition of the ARR and how the methodology is applied in 3 illustrative examples. Growing at an average of 6% over the last 2 years, the ARR highlights the consistent execution and growing subscription and cloud, driving the growth of our recurring business. It is also more closely tied to invoicing and cash flows from those deals. Now in Q4 2025, ARR reached almost EUR 4.5 billion, with EUR 104 million of net new ARR in the quarter. This highlights the consistent performance in signing new cloud and subscription contracts, while, as mentioned before, the revenue is highly dependent on the timing of revenue recognition. In '26, we are establishing this new metric in our reporting and plan to guide starting in 2027. And during the Capital Market Day, which will be scheduled for November this year, we will outline the steps in the context of our 2029 financial plan. And as we look ahead, it's clear that the trajectory to accelerate growth is linked to the shift in business model. Now let me discuss very briefly the levels of ARR first -- of ARR growth with you. First, the mix effect, which is driven by the faster growth of subscription versus the maintenance ARR. The second growth lever is the growth in 3DEXPERIENCE and Cloud as AI-powered Virtual Twins and Virtual Companions boost our 3D UNIV+RSES portfolio. Third, within Life Sciences, we expand the footprint, and we create the next-generation clinical trial platform powered by AI. And finally, with CENTRIC, the ARR growth has a long runway ahead. So now with this, let me turn to the financial objectives for '26. We expect the total revenue and software revenue growth of 3% to 5% ex FX for 2026. Importantly, this guidance marks a tipping point. In 2026, the share of subscription revenue will surpass the maintenance revenue. Hence, we are providing also the ARR to better reflect the growth dynamics, not yet as a guidance, but to show the momentum. The operating margin is expected to achieve a 40 to 80 basis points improvement ex FX, which takes us to the range of 32.2% to 32.6%, as we continue to balance investments and margin expansion, leveraging our operating productivity gains. We see EPS growing at 3% to 6% ex FX or EUR 1.30 to EUR 1.34. This is all based on FX assumptions at an average rate for the full year of dollar to euro at $1.18 and yen to euro of JPY 170. Now turning to Q1. We expect 1% to 5% growth for both total and software revenue, and operating margin is expected in the range of 29.2% to 30.7%, and EPS in the range of EUR 0.28 to EUR 0.31. Now finally, I would like to share some key assumptions, which is -- which are underlying our guiding framework for 2026. We expect 3DEXPERIENCE and Cloud momentum to remain broadly in line with last year, driven by continued expansion within our installed base and ongoing market share gains. We're also focused on entering into new markets and accelerating the monetization of virtual companions and virtual twins. From a geographic and industry standpoint, the demand in Americas remains healthy, while Asia continues to show resilience. In Europe, we see a solid pipeline development in Southern and Northern regions, which is partially offset by continued weakness in the automotive sector, which we mainly observed in Germany and France. And this could potentially impact the timing of decision-making within quarters. The defense sector represents potential upside. Now within our mainstream business, SOLIDWORKS continues to deliver mid to high single-digit growth in both revenues and users. For CENTRIC, we expect a return to low teens growth, supported by execution against a strong pipeline and a higher mix of Cloud revenues. Life Sciences is facing a transition year, with actions underway to position the business to return to growth starting 2027. On margins, we expect continued improvements driven by productivity gains from AI initiatives and operational excellence. These initiatives are focused on increasing the flexibility and reallocating investments towards top line growth. In conclusion, in '25 and '26, we're laying the foundation for our next phase of growth. I want you to remember 3 things. First, 3DEXPERIENCE platform is at the core of industry transformation, creating a long runway of growth. Second, on AI, we are introducing new categories of solutions, which goes beyond productivity gains. It's about creating new possibilities. And third, we are taking actions to scale our operations with one single goal in mind: to generate sustainable growth. And now with this, Pascal and I look forward to take your questions.

[Operator Instructions] And now we'll go to the first question, and it comes from the line of Frederic Boulan from Bank of America.

Pascal and Rouven, 2 questions, please. First of all, on M&A. Key message this morning around appetite. Can you share a bit more around your thinking here? Are you thinking bolt-on or is the appetite for larger deals? Any specific gaps in capability that you need to address or want to address? And then second, more broadly, when we look at the execution problems we've seen in the last 2 years, can you discuss some of the main priorities and changes you're implementing from management or KPI perspective? Anything around compensation criteria for management that you can share with us? I mean, in particular, do you think the -- you kind of focus on non-adjusted EPS is the right criteria to align your long-term strategy with value creation for shareholders?

So related to the M&A, again, we have different scenarios. But I will say between the last transaction, the way we did at the time of MEDIDATA and the bolt-on we used to do, acquiring company doing EUR 10 million to EUR 20 million, there is something in between. And this is the -- in-between we are targeting. That's to answer to the first part of your question. The second one is the M&A for us is again not a way to bridge the gap for the growth. It's to build the foundations to sustain the growth over time. So if I step back a little bit and I look at the different phases we had from an M&A standpoint, 15 years ago, we were using the external growth as a way to build the product lines. This is how we have created the brands, combining the external growth with our organic development. We will continue to do this, and we have opportunity to complement the domain expertise with some extension. Then after we did some major move, to be much more industry-focused. This is how we lead the decisions to acquire MEDIDATA and CENTRIC, to build a presence in the diversified industry to rebalance in a way, the portfolio of customers, not to have too much dependency on aerospace, autos and industrial equipments, but to be more diversified in the consumer-related industry as well in the Life Sciences. By the way, both sectors are spending a lot in innovation right now. Third element is what's next. What's next is 3D UNIV+RSES, right? And 3D UNIV+RSES is about how we are leveraging artificial intelligence in conjunction with industrial data sets. This is what has driven -- what is driving our thinking process. And again, as a capital allocation, I want to use the capital as a way to not only accelerate the organic growth, but also to build and to expand and to create buyer to entry in everything we do. So that's the answer. The second topic related to the management and KPI. There are certain decisions we have taken. So if I look at what didn't work in 2025, let's start with CENTRIC. You know the story. We basically had to rebuild the management, starting with appointing a new CEO. And now if I look at the management team, almost half of them are new. So clearly, I think it has been done last year. Now if I look at Life Sciences and MEDIDATA, the point on which where we want to accelerate the changes is on the go-to-market. I think Rouven's explanation about the performance we have in the large enterprises, in conjunction with the situation we need to fix with the volume business linked to the CROs is requesting anyway the different go-to-market. So that's the reason why for the large enterprises, we are combining the 2 go-to market, the traditional one of Dassault Systemes in conjunction with the MEDIDATA one, in order to have one single team, one client executive to take care of the top 50 largest accounts we have. And the reason is because in everything we have, we have the ability to position our solution at the enterprise-wide, not only at the domain or the department level. That's one thing. The second one, we are also changing the leadership for the sales. And we have a newcomer who's going to take this responsibility starting Q2. He's already with us. He's coming from AWS, and he was in charge of the Life Science sector for the entire AWS worldwide. Now speaking about the next topic, which is the indirect channels. Because you notice that the growth where we are growing mid-single digits versus the high-single digits, it's basically the indirect part. So I have appointed 2 senior executives in order to help me to accelerate the transformations. One is an operational officer in order to accelerate basically the efficiency, to develop the -- to redesign the ecosystem, because with the new solutions coming, anyway, we need a different kind of ecosystem to continue to expand. And the second one is we need also to invent the frameworks, the new terms and conditions, the new contractual terms, the new commissioning systems. And this is the reason why I also appoint a transformation officer, a guy coming from McKinsey, used to run basically the practice in Europe for the High-Tech and Media sectors. So he is very aware about all the benchmarks and he has this knowledge to put in place. That's a concrete example of what we do from a management standpoint. Now from a KPI standpoint. I mean, my #1 objective is to, again, accelerate the transition to subscription. That's the reason why Rouven is basically coming with this new KPI, the [ ARR ], because without it, you will have hard time to understand the performance of Dassault Systemes. And the proof of what I'm saying is look at Q4. Q4, you will conclude that we grew only 1%, flattish, on the software side. And if you look at the ARR, it's EUR 105 million plus, amounts of EUR 1.5 billion software revenue, which is 7% growth. So you need to have this KPI as a way to understand the dynamics and the growth trajectory we are building. So that's point number one. Point #2, I'm sensitive to the cash conversion as well, right? And again, the subscription model is changing the equations. When we used to be mainly licensed, we were upfronting most of the revenue. Now we have to take into account the fact that with the Cloud, we spread the revenue. And then after you have the ramp-up from a user adoption standpoint. So that's also another indicator, which is becoming something important. And we are also incentivizing -- not only the finance, Rouven and his team, on this, but also some of the geo leaders who are running the business. I hope with this, you have a clear understanding of what we do.

The question comes from the line of Nicolas David from ODDO BHF.

I have 2 as well. The first one is, could you help us reconcile the fact that you plan to accelerate the move to subscription going forward? And the fact that in your guidance for this year, in the end, you expect a very robust license sales, actually better than in '25. And oppositely, subscription growth slower than '25 in '26. I understand that you don't expect MEDIDATA to weigh more in '26 than '25. So it would be interesting to understand why you don't expect a better subscription growth and weaker license. And by the way, about license, I'm also quite impressed by how narrow is the guidance here in the context of this transition. For instance, in Q4, you had a very wide guidance range. So let's -- could you help us understand why this time you were able to guide very narrowly? And the second question is regarding AI, could you help us understand if you see AI as a factor which is lengthening the time it takes from clients to sign contracts because it includes another variability, another variable or another angle of discussion beyond Cloud, beyond subscription and so on? And is it something which is a good opportunity for you, but in the short term, lengthening the sales cycle?

Okay. So thank you, Nicolas, for your question. I will start, Rouven, and feel free to add whatever you want. So the first part is -- so let me tell you the things. You noticed, by the way, we put the guidance in the appendix. And the reason is the following. Yes, we are accelerating on subscriptions. And then after, we need to guide the market with concrete KPIs we have and basically some early indicators we have in our systems. So what do we have? We have a pipeline. And in the pipeline, many of the opportunities we have are still declared as a license opportunity. And we know that when we are transactioning and when we are contractualizing, this license is usually turn into subscriptions. So that's something you have to keep in mind, right? So now, Rouven and I, we have to build the system on something which is credible. It's not something coming from our hat, if you want. So there is a twist here. The second one, there is mix effect. Why there is a mix effect? Because if you look at the different geos, we are massively overall -- I mean, the subscription is bigger in Americas than the license right now. This is also true in Europe. But this is not true in Asia. And just because we are anticipating a slowdown in Europe right now given basically the uncertainty we have on the auto sector, automatically, Asia is overweight. So there is also a second twist here. And last but not least, you have the rev rec topic. Because again, when we have some customers contractualizing subscription, multiyear, on-prem versus cloud or versus hybrid, this could have also an impact in which line we recognize the revenue. So that's basically the reason why I think we were communicating on the convictions on the software growth. The split between the 2, you still have some moving parts. Nevertheless, the [ ARR ] is calling for an acceleration of the subscriptions. So I hope I give you reasons for you to reconcilize if you want the guidance with what is happening. Now on the AI. In fact, it's the opposite, Nicolas. I think AI is helping us to shorten the cycle. Why so? Because it's automatically on the cloud. It's relatively hot in many engagements we have right now. And AI is calling for very short return of our investment. So I think this is, my view, a way we can use as a way to accelerate certain transactions more than to slow down, if you want, the decision process. So that's the reality we are seeing, and this is my experience for the last, let's say, 9 months with the newer solutions we are bringing on the market.

And the next question comes from the line of Jay Vleeschhouwer from Griffin Securities.

Pascal, I'd like to start with some corporate or strategic questions. You noted some changes in management and go-to-market for CENTRIC and Life Sciences specifically. But for the remainder of the business, the bulk of the business that we would call engineering software, are there any other changes that you've considered? What I had in mind, for example, is that you have an unusual structure with regard to the brand CEOs. It's unlike what we see elsewhere in the industry. Perhaps there are more disadvantages than advantages to having that structure that something ought to be considered. And then similarly, with regard to go-to market, is there anything that you're thinking of doing with regard to CPE and/or CRE with regard to the engineering software markets? Then a couple of other strategic questions, please.

Okay. So let's start -- first of all, good morning, Jay. Let's start with the first one. Why do we have brand CEO? It's because if you look at domain by domain, we do not have the same competition. And if I want to keep the focus to stay at basically the best of breed in everything we do, I need to have a leadership with having the clear focus on the domain expertise. And you will basically accept my comments, if you look at the diversity of the domain we are addressing, you do not have one single competitor. None of whatever, it's Siemens, PTC or now maybe the EDA players in the simulation space, none of them they have exactly the same scope of what we are covering, between life sciences, between manufacturing, between engineering, between simulations, between life cycle management. So this is the primary reason why we are keeping this. It's not creating complexity, as you imply. The reality, it's the best, let's say, coordination mechanism I have with the research and development, right? Because -- and you remember last week, Jensen, when he was on stage, he was joking about the brand, right, about the diversity and the funny name we have. But at the same time, he was recognizing the extremely specialization of them, which has a lot of value in this new AI game. So this is my answer to the first part. The second one is, yes, we have 2 indirect channels. And I heard many times, why you are not converging them? Let me explain to you why it's [ valid idea ]. There are 2 fundamental reasons. One is because we are not asking them to do the same thing. CRE is about volume. It's about footprint. It's about short cycle of time. It's about cloud. It's about packaged products. So this is what CRE is about, right? And today, we have 400,000 customers. We are more than 45 million users worldwide. The vast majority of them are coming from this machine, which is a volume-based. And why this is important? Because as soon as you have the footprint, you can build on top of this. You can diversify in terms of products. You could accelerate in terms of transformation. You could bring the services on top of this. And this is the reason why we have a second indirect model, which is so-called CPE, because they are the ones who are building the road map to value up this large installed base we have, especially in the ecosystem of the large OEMs. Because, as you may know, with the direct sales, we are addressing almost 4,000 accounts worldwide. And this is important to capitalize on the ecosystem of Boeing, Ford, JLR, all the big OEMs we have. And the ecosystem is not only the Tier 1 and the Tier 2, it's really all the ecosystems. And this is where CPE is extremely instrumental to do that. So I have no intent to merge this. The second reason, the business model is different by nature. Because if you are volume-based, it's a high velocity, low-touch, and the cost of sales should be relatively limited. CPE, it's almost the opposite. It's high touch, low volume, but it has to be extremely skilled. So the cost of sales is obviously not the same, and this is the reason why you cannot merge the 2 together. This is the second reason. And last but not least, depending the industry, we have different ways to tackle. If you take the [ car ] industry, clearly, auto, aerospace, CPE is almost an extension of the direct sales. And if you look at a very fragmented industry like industrial equipment, net devices, construction, clearly, CRE is really the engine to tackle this market. Long answer, but I think with this, you have the strategic thinking, Jay.

Okay. Secondly, the NVIDIA partnership is perhaps more comprehensive than any partnership we've seen DS announce previously. It covers a variety of technologies and ambitions. What are the executables as you see it for you, and particularly as it relates to R&D in terms of making sure that this partnership works? And then as well, 4 weeks earlier, NVIDIA was on stage with your counterpart at Siemens. How do you distinguish your relationship from that relationship? And then finally, on the strategic front, there were some questions earlier about M&A., perhaps we could be a little bit more specific about that. You have, I think, 2 strategic needs or dilemmas perhaps in terms of the completeness of the portfolio. So when we think about, on the one hand, your needs in simulation where the competitive environment has materially changed, and then the other, Infrastructure and Cities, which of those 2 do you think might be more important for you to solve? And then not quite the same question, which do you think might be easier to solve?

Okay, tough questions. But I will start with NVIDIA. So I think there is some differences, by the way, there are some differences compared to what we do with NVIDIA and some of our peers. And by the way, if you look at the keynotes I did last week, I asked this question to Jensen. And he gave some answers. The first one is from a research and development standpoint, we are building the world model together. And you know the next generation after the large language model will be the model who understands the physics, who understands the world we live in, but also how to create this world. So you need to understand the physics on one hand and the engineering and the manufacturing on the other hand. I think we are, by far, the best partner combining both, by far. So one element we are doing together, we will codevelop the world model to -- for us to be able to build the industrial part on top of it. And for that, we are reusing some of the precompute model NVIDIA has already built. This is one thing. The second one, there is a difference between the agent and the virtual companions. The agent is really a way to automize the workflow. The companion are the ones to explore the different alternatives and all the different options. So to do that, we need a specific computing infrastructure. And we are again specifying what we are -- what we need to NVIDIA for them to deliver it in order for us to scale. Last but not least, we also have a common go-to-market. You know that NVIDIA, they are not selling direct. At least, they are selling, especially in the industry, they are selling indirect. So nevertheless, they are widely used already in our installed base. So we are also aligning our forces in order to tackle some specific accounts in order to scale. And my last comment, you notice that NVIDIA is not only a partner, but it's also a customer, because they bought our Virtual Twins for their AI factories. And this is extremely important because, as you may know, it's a gigantic investment that the entire planet is ready to do with a factory which is at the level of complexity which is far beyond now what the nuclear plant is about, and with less maturity on the technology. So you need this virtual world in order to connect all the different pieces together, the computing power with the energy supply, with the supply chain, with basically the buildings and so on. So it's a very complex ecosystem you have to aggregate. And without virtual twin, in fact, most of those projects will be at risk. And that's one of the reasons why NVIDIA is investing on our solutions, in order to build the virtual twin of their AI factory, including the blueprint. And this blueprint will be used by the entire NVIDIA ecosystem to deploy their technology in all the AI factory around the world. Now M&A., the choice between simulations and cities and construction. At the end, I will consider both, but probably with a different approach. Let me explain to you why. I think AI is also changing the simulation world, because with what we call the virtual twin of the behavior, which are nothing more than an extension of the surrogate model. We have a way to democratize the simulations. It's something, obviously, we started to develop organically, but there are start-ups around the world we could acquire in order to accelerate and to speed up basically the completeness of the portfolio for all the different disciplines. And this is extremely important, Jay, because we -- as you may know, for a few years, we are combining modeling and simulations together, which is pretty unique because if we use a surrogate, we have basically a generative approach in order to use the simulation and to use the simulation as a way to create the 3D, not the opposite, to create the 3D, and then after, to run the simulations, which is simulations proof, if you want, by design. So this is something we are doing. Again, keep in mind, we will continue to invest and there are probably start-ups, teams around the world who could nicely complement us on that. Infrastructure and cities, the topic is different. We have the solution, but we do not have the go-to-market. At least we have a certain go-to-market. Now, the way, and you know the way I'm approaching it, I do not want to be the general purpose channel. I want to be specialized by domain. I want to subsegment this market. In a way, I think what is important, I want to have a dedicated team for the complex capital-intensive projects, wherever, it's a nuclear plant, and energy systems, a civil engineering infrastructure, right? But I also want to -- data center or AI factories, right? And you need to verticalize and you need to bring basically not only the software, but the expertise which goes with it. And for the volume, remember that most of what you have in the buildings, whatever, is the windows, the walls, the roof, has been designed with our software. And most of it, by the way, is coming from SOLIDWORKS. So we also need to have an approach which is how could we capitalize on all the parts, all the contents, all the equipment, which has been designed with our systems and could be repurposed, if you want, in the context of the buildings. And I think with AI, we have a way to change the game. But again, I still need to probably have more people, more expertise on the market in order to accelerate the penetration of this sector, which is extremely fragmented, by the way.

And now we're going to take our final question for today. And it comes from the line of Balajee Tirupati from Citi.

A couple of questions from my side as well. You have mentioned the business evolution from seat-based pricing to value-based pricing model, which obviously is quite important in the age of AI and agents. Could you tell us how compressed this transition could be? And what changes in offering and go-to-market motion you are required to make given the ongoing subscription transition is already underway? And lastly, where are your customers in readiness to accept changes in pricing model without seeking certainty in pricing?

Okay. So I spent time to explain this morning, but I will try to summarize it. First of all, this new category of solution is not an extension of what we do, it's a new category of solutions. Meaning, if you have a virtual companion, the virtual companion will continue to use CATIA, we will continue to use SIMULIA, we will continue to use DELMIA. So why I'm saying this, it's because we have to segregate in a way the value independently of the tools for those companions. And this is the reason why we came with new, what we call, new units of value. Many of my customers, the biggest problem they have is what? Either they struggle to hire people, right? And they will be extremely, extremely happy to have virtual companions, having the highest degree in mechanical engineering, the highest degree in electronic design and the highest degree in basically advanced computing and simulations. So that's what the virtual companion is about first. The second one, we have many customers, they have a capacity in engineering who needs to be a little bit flexible. Because when they have a lot of projects, they need extra. And when basically they have less projects, they need to adjust. It's a way to give for them the flexibility from a human standpoint. Why I'm saying this, because at the end, we are not forcing the customer to buy the virtual companions, right? They are free to decide if they want to buy it in conjunction with CATIA or if they want to basically stay with what they have. And the fact that the 2 values are completely separated is much more easier to articulate in a way. So the last 9 months is, in all the engagements we have been in, I think we have been able to prove that the value is real and we have been able to prove that we can price it independently of what we do. Now it's probably a little bit early to conclude that everything is set because we need to have this approach in all the industries we serve, in all the domains we serve, but the framework and the model is in place, in order for our go-to-market people, I mean, the people on the field to make proposal with prices, with terms and conditions.

So would it be fair to understand that your existing tools in CATIA, ENOVIA, et cetera, they would continue to be having the same pricing model? They will not be moving towards more consumption-based pricing?

No. No, because -- let me tell you why, because I think what I want the usage base is basically the companions. And why? Because again, it's a way for me to accelerate the adoption. If you look at my biggest constraint right now, if you take CATIA, for example, there are certain domains where the number of people who have the expertise to use the software is quite limited. And believe me or not, this is one of the biggest obstacles we need to remove. The virtual companion is an answer to this. And I think it's very sad not to mix the 2, which will give me the 2 legs, the one which is the user-based centric approach with a traditional overall process and solutions, and the user-based approach which is much more AI-based. And I think the value is a combination of the 2.

There are no further questions for today. And I would now hand the conference over to Pascal Daloz for any closing remarks.

So again, thank you very much for your participation and for your engagement today. Again, if we step back a little bit, the real question right now in our industry is AI is fundamental and this is changing a lot of things. And the question will be who are winners, who are the losers? I think with what we started to show to you today, we are building the foundations. We are putting the execution in place, and we have a determination to win, not just to compete but really to lead the industrial AI. So that's the reason why in November this year, we want to have a Capital Market Day. Usually, we do it every 2 years. But I took the decision to have one this year, and not in June, but more in November time frame, to link, in fact, these visions we just crafted last year and how to connect it with the long-term plan we have. I think this is an essential and we will provide a lot of details and information on this. In the meantime, I think Rouven and Investor Relations are looking forward to see you on the road when they will do their roadshow. And as far as I'm concerned, I will see you no later than the next quarter. Thank you so much.

This concludes our conference call. Thank you for participating. You may now all disconnect. Have a nice day.