Dassault Systemes Sa Adr

Good morning, everyone. I'm Stacey Pollard. I'm here with the SOCIS stands CEO, Pascal Dalloz, and the CFO, Reuven Bergman. Unfortunately, our head of investor relations, Beatrix Martinez, could not be with us today. She's out for a couple of weeks. So I have the pleasure of being in this room again. It's been a few years since I sat in the chairs beside you guys. So it's very interesting to be a different perspective on this side of the podium. Now let me move on and formally welcome you to Dassault Systèmes' third quarter webcast presentation. At the end of the presentation, we will take questions from participants in the room and online. Later today, we'll also hold a conference call. Dassault Systèmes results are prepared in accordance with IFRS. Most of the financial figures in this conference call are presented on a non-IFRS basis with revenue growth rates in constant currencies unless otherwise noted. For an understanding of the differences between IFRS and non-IFRS, please see the reconciliation tables included in our press release. Some of the comments we will make during today's presentations will contain forward-looking statements which could differ materially from actual results. Please refer to our risk factors in our 2024 universal registration document published on the 18th of March. I will now hand over to Pascal DeLose.

Thank you, Stacey. Good morning to all of you. It's always a pleasure to be here in London and to have a chance also to interact directly with you. So we're going to review the Dassault Systèmes performance for Q3. Let me give you some, at least my reading of the numbers. I think this quarter is a solid quarter. with a healthy margin, and I think Rouven will come back on this, and with a strong EPS growth, and we continue to grow the recurring revenue part, which is, I think, the important thing, because this is reflecting the strengths and the resilience of our business model. Now, if you look at the numbers, the revenue grew 5%, thanks to a strong demand across our core industries. Our subscriptions business is up 16%, accounting for almost half of the recurrent part of the revenue. If you remember, a few years ago it was only a third, so this is growing extremely well. We hit a 30.1% operating margin, which I think is reflecting our focus on running profitable and efficient business. And finally, the earning per share came at 29 cents and growing at 10%. So behind this number, I think there are certain things I would like to highlight, which are our strengths. The first one is industrial innovations, especially transportation and mobility. We continue to expand our footprint. And despite the ongoing challenges in this sector, We have also a strong momentum behind 3D experience and SOLIDWORKS this quarter. The second thing is I think our focus on accelerating SaaS adoptions is starting to pay off this quarter, as you will see. This is driving the revenue growth and the strong market tractions. And to further support this momentum, we have established a new leadership at Centric to fast-track the adoption of the SaaS business model. Lastly, in the field of artificial intelligence, I think we are shaping the future with a powerful combination between the industry's most comprehensive data sets, the scientific rigor, the advanced modeling and simulations being combined with the real-world evidence. We call it the real-world validations. And AI for us is really not an add-on. It's embedded in the core of the 3DEXPERIENCE platform for a long time. Because you remember, the 3DEXPERIENCE platform, this is really how we are managing the knowledge and the know-how for many of our customers. This quarter, we are coming with a new category of solutions. And you remember the virtual twin as a service, the generative experience and the virtual companions. And we'll say more about this. And they are really transforming the way Our industry, our customers, they are designing, producing and operating the life cycle. Now, for the full year, we are confident enough, at least to reaffirm the earning guidance, and we expect the EPS to grow between 7% to 10%, with a total revenue rising 4% to 6% on an adjusted basis. And it's mainly new to three factors. The first one is the lower growth from Medidata, which is in line with Q3, in fact. The impact of the saturation for Centric and the volatility impacting some of the timing to close. Now, let's dig into some details behind those results. Let's zoom first on the manufacturing sectors. As I was telling you, transportation and mobility has once again proven its resiliencies. And to give you the numbers, this quarter we are growing at 18%. Why so? Because it's usually when it's a difficult time for our customers that they have to take radical decisions. And this quarter we have some. Ford took the decisions to go with us to expand outside of the engineering borders and we have signed a contract with them for the next five years to use the platform across all the different programs. I will tell you more on this probably next quarter. But there is also another very important flagship customer we signed this quarter with Stellantis. And I know some of you were expecting us to move along this way. And I will come back on this. Why those companies are basically adopting YD, the 3D experience platform, is because they are using our solution first to speed up innovation. And speed is becoming really one of the key topics. You remember a few years ago to develop the car, it was almost 48 months. Now we are talking about 16 months. So it's a little bit like fast-moving goods. And to master the complexity, you need a different approach. And this is where I think we are making a difference. Sustainability is also a topic. The electrification is driving the cycle. You know it. And more and more with the SDV, we are creating a personalized experience for the customers. And this is really the combo, if you want, of what we can provide with our solutions. We are also seeing a strong growth in defense. It's growing double-digit this quarter, where programs are becoming more complex and collaborative. And I think our 3D Expense Platform, combined with what we call the Model-Based System Engineering, MBSE, which is now a standard in the industry, is more and more widely adopted. And this is really opening a new opportunity for us Not only in Europe, but also in the rest of the world. Life sciences. The market remains unstable and challenging. I think Ruben will say more about this. We still see the new clinical trial start being contracted. Nevertheless, we landed with some big contracts this quarter. And more importantly, I think we also... being encouraged by some large win-backs. AbbVie is one of them. You remember it was one of the flagship customers of Viva a few years ago. They signed with us a contract for the next five years and I think this is the proof that what we do is extremely critical And I think also this is a proof that what we have built as a foundation is critical for them, also for the AI-based programs. And I will come back on this. In infrastructures and cities, the demands keep growing, in fact, for autonomous and sovereign infrastructure, you remember, especially in the energy space. But we are more and more seeing new use cases or new opportunities emerging. One of them is the nuclear decommissioning. As you know, it's a big topic because you have many reactors around the world aging. And we are using our solution to do a vehicle twin as a service to manage the safety and the efficiency of this process and to manage the end of life of those nuclear reactors. So this space is really, again, a way for us to establish leadership in a domain where we are the challenger because in this space, I think we do not have the same footprint than the others. Now, let me show you some key wins. Stellantis, for you know the company, I mean, and you remember, we had significant footprints with PSA, but the rest of Stellantis was much more in the hands of our competitions. So, they took the decisions to, I mean, to standardize on 3D experience platform on the cloud, which is a thing important, for their system engineering backbone. And this is extremely important because, as you know, the system engineering is the foundation to do the SDV. And all the car players are moving along this way, and they are using our system-to-system approach as a way to standardize across all the domains to unify the bill of materials, but more importantly, again, they are building the foundation for their AI initiatives. Because one way to reduce the cycle of time to develop the car is to be much more generative, And you need an infrastructure to do this. And that's what the 3DX Plank platform is really for. So we are extremely proud to support this transformation. And it's a significant one because it's a ramp-up at the end with more than 20,000 users we need to equip with the systems. Moving to life sciences, I already said a few words. So, AbbVie, you know, it's a global biopharma. It's one of the top ten global pharma. And it's a win-back. And it's a win-back of a win-back. Let's say it this way. Because, again, a few years ago, they took the decisions to open some clinical trials with Viva. And now they are back with us. And there are a few reasons for that. One of them is the time. They were sharing with us that we are ten times faster in the way to run the processes and the clinical operations. It's also a big cost saving, which is an interesting takeaway because you remember one of the arguments which was used was this EDC technique. is becoming a commodity and it's price sensitive and the reality is the price is one thing the savings and the efficiency is another one and it's here you have the proof and the last arguments all the pharma sector a little bit like the auto sectors they are building their ai programs in order to automate in order to use in a better way the data set they have And they have seen through our platform the ability to develop their own program on top of what we do. So those are the reasons, if you want, behind these win-backs. Finally, from a customer standpoint, this is an interesting case also. Correa Hydro and Nuclear Power is the largest energy public company. enterprise in Korea, and they have launched the digital transformations to manage, as I was telling you, the decommissioning of 26 reactors. So the reactor is first generation. They are progressively replacing it with a new generation. And to do this, it's a complex process. They have to decommission this large install base. They showcased this example, this case in Korea's 3D experience forum a few weeks ago, and I was having the chance to participate in this. And frankly speaking, you should really look at it. It's amazing what they have been able to do, because it's a very complex process. Safety is at stake. Compliancy is at stake. It's a very, very sensitive process, because you have to manipulate the reactor when the reactor is still working. At the same time, you need to do it in a very precise manner. And to manage this complexity, to predict the complexity of the process, to prevent the risk, to keep track of everything, because you have to be compliant, they are using the platform and they are using the virtual twin in order to make this. So, why I pick those examples? Because behind all of them, there is a clear pattern. We are not only the partner for them. I think in many cases, we are the game changer for them. We are the one allowing them to accelerate their industrial transformation, whatever it's in the mobility, life sciences, and the energy sector. Now, let's speak about 3D universes. So, you remember, we announced it in February this year, and I was making the statement, 3D universes is not an extension of what we do. It's really a leap forward. And there are a few things I want you to keep in mind. What are our differentiations? The first one is we are building our AI engine on the large and the most structured industry corpuses. And it's the result of 40 years having almost 400,000 customers worldwide in very different sectors. Building the virtual twin of all the objects you can see on the slides. And this is a unique corpus to train our systems. And remember, AI without having high quality data is just only a noise. But if you have the right data, it's becoming a game changer. The second takeaway is the data set is not enough for what we do. You need to build AI on science. And this is extremely important because if you are only relying on patterns matching and recognitions, it's not enough for what we do. You know, the AI needs to be built on physics, biology, material sciences, engineering principles. And why so? Because when life is at stake, when you develop a drug, when you fly on objects, when you are driving an autonomous car, you cannot take risks. The system should not guess, should not hallucinate. You need to understand how the parts fit together, how the materials behave. And this is really what we have been able to build, which is an AI which is rooted in sciences. The third element is we are coming today, I mean today a few weeks ago, on the market with the new category of solutions. So you remember we presented it during the Capital Market Day. And now I'm really pleased to introduce you to our virtual companions. And in fact, it's a family of three for the time being. You have Aura, which is our business strategist, focusing on the outcome and efficiency. You have Leo, for engineering experts. And Leo is really diving deep into design and simulations. And you have Marie, who is our scientific authority, handling probably the most advanced questions on research. The interesting things, if you ask the same questions to all of them, you have different answers. So, more than a long explanation, let's look at the video.

Which material should I use to design this eFoil wing? When choosing a material for an eFoil wing, you'll want something that balances strength, lightness, and resistance to water. Many eFoil wings are made from carbon fiber, since it's strong, lightweight, and performs well in water sports.

From an engineering perspective, you need to optimize the strength-to-weight ratio and minimize hydrodynamic drag. carbon fiber composites particularly unidirectional layups with epoxy resin are ideal due to their stiffness and fatigue resistance while remaining lightweight the alternative fiberglass reinforced from a material science standpoint the critical factors are density modulus of elasticity and resistance to water induced degradation

Carbon fiber reinforced polymers offer a density of 1.6 grams per cubic centimeter with 10,000... So as you can see, it's not just about AI.

It's about having an AI which is behaving like your team because you need, when you do engineering activities, you need to assemble different domain expertise at the same time. And if you try to converse too rapidly to the solutions, At the end, you are letting some open opportunities untapped. And this is basically what we are doing with the virtual companions, which are a way to complement and to enrich the roles we have developed. Now, this is also an interesting thing because you can use AI as a way to take smarter decisions and faster. And here is, again, a concrete example. It's Aura. And ORA is widely used by many industries currently, you know, to deal with the tariff, with the trade policies, the supply chain issues, because this is changing so much that you need almost every day to reactualize your what-if scenario. So ORA, in this case, is not only anticipating, but reacting. She anticipated the turnaround, the uncertainty. She tried to manage, you know, with data-driven insights, the consequences. And this is important because for many industries, the margin is at stake. So to keep it to your head, the system, if you want, is helping you to collaborate, is bringing you the right expertise, is telling you what are the different avenues you have in front of you in order to fix the problems at the right times. Now, let's speak about SOLIDWORKS. You know, this is an interesting, this is a very important year for us. It's a milestone because we are celebrating the 30th anniversary of SOLIDWORKS. And why this is important? Because if we step back, After 30 years, I think no one will debate that Sullyworth is the undisputed leader in the 3D CAD. And I put some numbers on a slide just to give you the proof. 8 million users. It's by far the largest design community around the world. 1.5 million commercial licenses, which is truly addressing the large company, but also the startups and all the shakers. It's almost 300,000 clients worldwide. And again, covering the large spectrums of all different industries we serve. So it's a lot of legacy of innovations that we are pushing from a product development forward. And I think now, with SOLIDWORKS, we are also introducing the artificial intelligence to build the next phase. To make it faster, smarter, easier to use, in fact. And the topic for us is not only to automate tasks, but more importantly, to give more time for the creativity. And we have some features we are introducing and some functionalities. The first one is obviously the generative design. Second one is what we call assistive features, which is an intelligent and pattern recognition when you do, for example, an assembly. And all those kind of things are really helping the users to work smarter, but not harder. Behind this, I think if there is one message I want you to keep in mind is this AI approach is a way to do the docking, to bridge with the 3D experience platform. As you know, this topic is at stake for several years. And I think now I believe we have found the roots to connect the SOLIDWORKS large inside base we have with the 3D experience platform. It's a way, if you want, to turn the SOLIDWORKS users into the lifelong experience partner. So, I think, and Ruben will come back on this, but you will see the performance of SOLIDWORKS this quarter is really extremely good. It's growing at double digits. Now, to conclude. I think why everything I share with you matters. There are few things. I'm sorry. I should not anticipate your presentation. The first one is 3D University is giving few and large advantages. The first one is, you remember, we are helping our customers not only to manage the full life cycle of their products, but more and more to manage the life cycle of the intellectual property. And you should remember what I'm telling you. In these AI periods, the most important asset is the intellectual property. Because everything you build is leveraging the intellectual property. And if you do not have a way to manage it safely, to take it as a real asset to manage your life cycle the same way you manage the life cycle of the products, you take the risk to be out of the game. And this is what we are bringing to our customers is this ability to mix the different knowledge coming from different sources but at the end still tracking who belongs to what. The second thing is in many domains we are turning compliance into a competitive advantage. If you take aerospace, if you take health care, if you take energy, those are extremely heavily regulated industries. And one of the answers to the tariff war is to put more regulations. That's the way to protect if you want certain markets. The flip side of this, if you are an industrial company, you have to manage with this complexity. And AI is a fantastic tool to read millions of documents, to extract thousands of rules, And us, what do we do with those rules? We do compliance by design, if you want. The system is checking automatically that everything you do, every design you do, every decision you do, are compliance by design. The third element, I think, generative AI is really a game changer as soon as you can trust it. And your AI, in many industries we serve, needs to be certifiable. If you cannot certify the output of what you have produced with AI, it is useless. And the way to do it, you remember, we are training our AI on very comprehensive data sets, which is pretty unique. And those are very high quality of data sets. And it's validated by the science, which is even more important. And we are deploying those artificial intelligence capabilities into a secure and sovereign environment, which is what we do with 3DS outscale. So this combination is pretty unique on the market. It's a huge differentiation compared to many of our peers. And this is, in my view, a game changer in many, many customer engagements we have right now. The last but not least, I think we are coming on the market with a new category of solutions. You have seen this morning the virtual companions. Aura, Leo, and Marie, but you will see more and more the generative experience, the virtual twin and services. We have a roadmap for this for 26th, 27th, and this will accelerate the contribution of AI in our revenue streams. So with this, I think it's time for me to hand over to you, Ruben, to give more flavor on the numbers and probably the outlook for the rest of the year. The floor is yours, Ruben.

Thank you, Pascal, and also welcome from my side to our call today. Thank you for joining us online and here in the room in London. Let me start with three key messages. First, top line growth and margin expansion are our top priority. Second message, the 3DEXPERIENCE platform is driving our business model shift to subscription and recurring revenue growth. This engine is working well with 16% growth of subscription this quarter. The third message is we are mission critical, as you saw in the examples, to our clients. In fact, in 2025, we are winning significant contracts with many of the top industrial companies across the world. And this is laying the foundation to long-term value creation with cloud and AI. It is these powerful long-term partnerships that give us confidence in our long-term targets. Now before I dive into the specifics of the quarter, a few more things to summarize briefly for you. Our financial results for the quarter were solid with 5% revenue growth and an expanding operating margin, which is up 100 basis points, and 10% growth in EPS. Industrial innovation is driving the growth of 9% in the quarter and 8% due to date, while metadata and Centric were softer than expected. As discussed previously, the repositioning of metadata is ongoing. The change of the model to reduce the dependency on clinical trial activity will take time, as we are doubling down on the enterprise and the PLM opportunity in life sciences. And for Centric, we are accelerating the SaaS transition, and to this effect, we have promoted a new leadership team, as you heard from Pascal. Now, looking at the full year, we adjust our revenue outlook to 4-6% XFX in line with our current trajectory of 5% top line growth year to date. At the same time, we maintain our EPS growth target of 7-10% XFX, and this is thanks to the strengthening of the operating margin, driven by additional efficiencies we are generating in the business. With this in mind, let me take you through the details. In Q3 and year-to-date, total revenue software were both up 5%, XFX. Recurring revenue was strong, up 9% in the quarter, and it highlights a very solid acceleration when compared with 7% year-to-date. Subscription revenue growth was 16%, and it was driven by new deals signed in the quarter and the increasing visibility from large contracts that are ramping. As a result, subscription revenue now represents almost half of the recurring revenue base. It's up three points from last year, and starting in 2026, subscription revenue will surpass maintenance revenue in absolute terms. 3DEXPERIENCE was the growth engine behind that, up 16% in Q3, and the signings of Ford and Apple contributed to the strength in subscription growth. Upfront license revenue declined 13% as our clients continue to adopt the subscription model at an increasing rate. The best proof of this is that recurring revenue now accounts for 84% of the total software revenue year-to-date. The operating margin improved 100 basis points for the quarter and is driving strong EPS growth of 10% thanks to the productivity gains and cost discipline. In fact, OPEX was up 3.1% in the quarter, and we continue to rebalance resources to support our growth strategy. Now turning to the growth drivers. In Q3, we saw very good 3D experience revenue, and it's now representing 40% of software revenue year-to-date. The growth was broad-based, up 16%. Cloud revenue was 8% in Q3, 7% year-to-date. 3D experience cloud revenue grew 36% in the quarter and 29% year-to-date. The key wins for 3D experience cloud, such as Ford, Jumco, Daralla Automobili, and Stellantis, demonstrate the value of the platform for our clients, where transformation is critical, as is the need to leverage AI. Now let me review the Q3 actuals versus our objectives briefly. Total revenue came in at 1,461,000,000 in the quarter, mainly affected by currency headwinds. Excluding currency, growth was 5% at the low end. Operating margin was 30.1% and above the objective, due to 60 basis points from performance and a negative currency effect of 20 basis points. EPS was 29 cents, driven by better operating performance against a small currency headwind. Now looking at the geographies and product lines. The Americas were 7% in Q3, with good performance in transportation mobility, high-tech and aerospace and defense during the quarter. Europe was a bit softer at 4% in Q3, with double digital growth in southern Europe, solid performance in France and also Germany. This was supported by subscription momentum, especially in aerospace and defense. Asia was up 4%. India has an outstanding quarter. Korea was up double-digit. Here again, strong performance of transportation mobility as well as aerospace and defense. China experienced softness in Q3, but also on a tough comparison base when looking at last year's number. Now let me review the performance of our product lines. As mentioned previously, Industrial innovation delivered excellent results in 2025 across key domains led by Katia, Inovia, and Damia, as well as Simulia, highlighting the value of the 3DEXPERIENCE platform is delivering to our clients. So it's broad-based across domains. We are mission critical to the transformation of our clients with superior capabilities to generate virtual twins. Life sciences growth was lower than expected. It was down minus 3% in the third quarter, with metadata impacted by continued study start declines, but importantly, continuing to gain market share. Overall, from an industry standpoint, the volume of business continues to face pressure. When we entered 2025, we had assumed that volumes would stabilize, helping to support our forecasted growth in the second half. Conversely, we observe a decline in high single digits in phase three studies, and mid-single digits decline across phase 1 and phase 2 since the beginning of this year. While we are expanding our market share, the impact of the decline in study stats is not yet compensated by the growth from the expansion with our enterprise and mid-market clients, who proved resilient. As you heard from Pascal, we had a major metadata platform win back, the top 25 pharma, Appli, After a brief period with a competitor, FB decided to return to metadata for all clinical trials, leveraging AI everywhere. This validates the trust clients place in us and the value of the metadata platform. Additionally, in Q3, we expanded partnerships with Sanofi, you see the press release this morning, and also expanding our business with IQVIA, including PatientCloud. Looking at life scientists outside of metadata, the opportunity to win with PLM is our clear priority. For the first nine months, growth is up double-digit, highlighting the strong potential of our portfolio to address the challenges of this industry. Now moving to mainstream innovation. Growth in this segment was mainly driven by SOLIDWORKS, as you heard. The shifted subscription is well underway at SOLIDWORKS. Centric growth was slower than expected in the quarter. due to some shifted renewals and we saw an acceleration in the share of clients adopting the SAS model. Now turning to cash and the balance sheet IFRS items. Cash and cash equivalents totaled 3,910,000,000 as of Q3 compared to 3,953,000,000 at the end of 2024. This decrease of 43 million on a Euro basis was driven by a negative currency impact of 269 million. At the end of the quarter, our net cash position totaled 1,321,000,000, a decrease of 138 million versus a net cash position of 1,459,000,000 at the end of last year. Now let's take a look at what drove our cash position at the end of the third quarter year to date. We generated 1,334,000,000 in operating cash flow for the first nine months versus 1,348,000,000 last year. The cash conversion from non-authorized operating income was 97% for the first nine months. Cash conversion is a top priority and we expect the conversion to improve going forward. And starting Q1 2026, we expect working capital to support cash conversion reaching the 2024 levels with the potential to improve further. As discussed previously, 2025 operating cash flow is impacted by significant contracts that we signed in the quarter, as well as higher payments related to tax and social charges, as well as negative effects. For the full year 2025, we now expect operating cash conversion to be in the range of 78% to 80%. To sum up, operating cash flow year-to-date was mainly used for investments, $581 million, of which $240 million was for acquisitions, $216 million for the purchase of the centric non-controlling interest, with the remainder of capex of $123 million to support our cloud growth. We paid $343 million in dividends and made a net repurchase of treasury shares of $186 million. For any additional information, you will find the operating cash flow reconciliation, our presentation that we published this morning. Now let's transition to our financial objectives for 2025. Net-net. Our year-to-date revenue is up 5%. For the full year, we now adjust our revenue outlook to reflect this trajectory and expect growth of 4% to 6% XFX for both the total revenue and software revenue was 6% to 8% previously. In absolute terms, we are adjusting the full-year revenue outlook by approximately 140 million to the midpoint. This reflects an impact of 30 million from Q3 and an FX impact of about 20 million. The remaining data can be explained by three factors. A, the lower growth from metadata in line with Q3 performance. B, the impact of the SaaS acceleration at Centric. And last but not least, we also factor in an increasing macro volatility with the potential to impact the timing to close large transactions. Please also remember that we had a high comparison base in Q4 of 2024. Now looking forward, the change of model for centric is ongoing. Sorry, the change of model for metadata is ongoing. And we are confident as well into the accelerated SaaS transition of Centric, given its strong positioning in a very large market and clients are endorsing it. For industrial innovation, we have built a very strong foundation in 2025, where we signed significant contracts and we expect in 2026 to expand on these partnerships, transforming with virtual twins and generative experiences. And, last but not least, the SOLIDWORKS momentum is strong. Recurring revenue outlook remains stable. It's at 7% to 8% growth. And underscoring what I said at the beginning, we are implementing a sustainable recurring growth model with increasing visibility. Above all, I mentioned the strength of our operating model highlighted by the margin improvement As such, we are maintaining our EPS growth expectation of 7% to 10% growth, or €1.31 to €1.35. To achieve this, we expect Q4 OPEX to continue to trend in the same range of Q3, delivering margin expansion of about 100 basis points, which is driven by ongoing productivity initiatives, having the right people at the right place to make it simple. So this is all based on FX assumptions for an average rate for the year of euro to dollar at 1.13 and euro to yen at 166.7. Now briefly on Q4. As you can see, the revenue range of 1 to 8% is fairly large. This is predicated on potential uncertainties in the timing of deal closing, mainly for the upfront licensed business. while subscription growth of 8 to 12% is solid on a high comparison base. Operating margin is expected in the range of 37.2 to 38% and EPS growth of 7 to 17% XFX to hit 41 to 45 cents EPS for the quarter, reflecting the ongoing operating leverage. Now, as I reflect on the performance so far this year, I want to highlight that our operating model is resilient, and we apply strict financial discipline to support our long-term growth. We occupy a unique leading market position that makes us mission critical today and tomorrow for our clients. Profitable growth and improving cash conversion, as mentioned, is a top priority with clear objectives to show results starting 2026. AI and cloud are two main growth drivers. We are confident we will deliver on their ambitious growth targets. We are committed to continue to invest right for innovation, for clients, and for shareholder value. Now Pascal and I look forward to take your questions.

Thank you. Online participants can ask a question by pressing pound key five on the telephone keypad. We pause for a brief moment and take questions from participants in the room first.

Hi, it's Adam Wood from Morgan Stanley. Thanks for taking the question. Maybe just to start off, you finished off, Ruben, identifying that it is a reasonably large range for the fourth quarter in terms of revenue growth. Could you maybe just talk a little bit about what is in there at the bottom and top end of those ranges in terms of pipeline conversion, assumptions on big deal closings? I mean, at the bottom end, are we assuming that none of the big deals close just to give us a little bit of a feeling for what's in there and how conservative that bottom end is? And then maybe just secondly, Pascal, you talked about the huge breadth of customer data that you have that you can train models on and use for AI. First of all, could you just talk about how challenging that is where customers are still on-premise? And then how much does that force them and accelerate the shift to cloud with the impact that has on the revenue transition? Thank you.

Thank you, Adam. I take the first question. Can you hear me well? Is this working, this microphone? Yes, there is a wide range on license. The recurring subscription part is fairly consistent compared to our performance year to date. I think that's important to note. On the range, the low end of the range is de-risked with large transactions. We have a long list of large deals that we have all validated extremely detailed to see where they can fall and the size of those transactions in different scenarios. What I said, given that increasing... macro volatility and the timing that the impact on timing of closing this could create We were prudent to reflect at the law and a more conservative and prudent perspective of large deal contribution So the the mid-range the midpoint requires some of those large deals But we have the potential to do better because our pipeline is strong but it's depending on on the timing of closing of those large deals and the size of those large deals and

Coming back to your question about the transition from the on-prem to the cloud and how it is linked with AI, definitely AI is accelerating the trend. And there are a few reasons for that. One is because no need to wait to have transition everything before to start AI. And the way we do it, we do what we call supplemental. So when you have a large install base, a large deployment of the 3D Express platform on-prem, we come with an instance on the cloud in order to basically enable all this AI new category of services we are developing. And this is really accelerating the trend. And you have seen in the number, it's 36% growth this quarter. The cloud related to 3D Express platform, it's 30% since the beginning of the year. And it's extremely correlated also with subscriptions accelerations with 16% this quarter. So in a way, this is helping the transition. And if you remember a few years ago, we were convinced the collaboration will be the catalyst for the people to move to the cloud. I think AI is the way to go.

Great. Hey, Ruben. Hey, Pascal. More from GS. Firstly, just it's encouraging to see on the industrial business there is pretty good momentum, particularly with Stellantis, Ford. As you look kind of into next year as we think of some of the headwinds and the tailwinds. How should we think about the growth across the different segments of the business? Because obviously in mainstream centric has a transition still to navigate. On the life science side, it sounds like kind of visibility is still reasonably low. but the industrial business is sort of ramping. So how should we think about sort of the sort of puts and takes for growth next year? And then secondly, as we think about the life science business, have you sort of, clearly it's a sort of behind plan, How do you think about the kind of strategic sort of view of this business over the medium term? You're willing to kind of ride it out or is it something that perhaps, you know, maybe you need to kind of change the scope of to try to extract more the growth areas that are probably better positioned? Thank you.

Do you take the first one, Hovind? Yeah. I take the second one. Okay.

In terms of the building blocks more, you know, when we look at the trend of 2025 industrial innovation up 8% year-to-date, 9% for the quarter, very much supported by the strong growth in 3D experience adoption, that's a very healthy and sustainable trend. That was always our objective to convert that growth into recurring revenue and subscription growth. As I mentioned, that in year-to-date, there is, and in Q3, there is always good contribution from new deals that we are signing, but also contribution from deals that we have signed in previous quarters that are ramping and are contributing to growth in the current quarter. With many of the significant deals we signed in 2025, this will be the case in 2026. So from an industrial standpoint, manufacturing industries, including solids, for solid works, because in solid works the momentum is also favoring, I think we are fairly confident in our ability to continue to transform these huge industries. And in a way, many of the deals that we signed are a starting point for what's expected in 2026 and beyond. So without giving you guidance for 2026, but I think from that perspective, the 2025 trends are healthy and stable and sustainable. Related to MediData, the growth profile, yes, it's very much affected by the volume business as we are changing the model to become more enterprise and more sticky by really looking at an enterprise solution to transform life sciences with the objective to generate evidence and outcomes faster for patients. And that's not just in the clinical trial, it's in research, in biology, but also in manufacturing and quality management and the whole life cycle of real world evidence and trials and patients. So the opportunity is large and we are making the changes to be in a better position in 2026 now. For 2026, I think we want to be cautious on the growth contribution from that part. We're not expecting decline in 2026 from life sciences. I think we are in a better position in 2026 than 2025. That's our starting point. And for Centric, the situation is difficult in 2025, but it will improve in 2026. The SAS acceleration is... is imminent. It's already happening as we are speaking because customers are transitioning faster to the SaaS and cloud solutions than to the on-premise and we expect around mid-teens growth for this business next year. And if you add all of that, I think we are We should enter 2026 with confidence. Of course, the macro standpoint is going to weigh, and we have to assess that, but the building blocks are in place and are shaping. That's the message to you.

The second part of your question, Mo, is if we consider life-sensing still being strategic for the associated enterprise. Ultimately, this is the question you ask. And the answer is yes. And there are a few reasons for this. One, if you look at who are the industry spending the most in research and development, life sciences and high tech are the two. In the previous century, it was the auto and aerospace. In this century, they are the two spending the most. And if you remember, the core market we serve is really the innovation space. And we are obviously serving the one spending the most in innovations. So from market attractiveness, there is no doubt. The second reason is because we did not diversify in the life sciences only for the purpose to expand or to diversify the markets. It was also a way for us to learn new scientific, at least to develop new scientific foundation. Let me tell you why. If we want to address the sustainability challenge, we need to understand how life is generating life. Right? This is, it seems maybe a little bit far from the day-to-day numbers, but from a scientific standpoint, this is extremely important for us to crack science. how life is designing things. And my bet is the next generation of generative design will, from a scientific standpoint, will come from this space. So this is the second reason why this is so important to continue to invest and to crack this sector in a good way. Now the question, what are we doing to change the game? Ruben already answered partially to this question. The first one is we need to minimize dependency on the volume of clinical trial. That's obvious because right now the model is extremely sensitive to this. So when the market is booming, we are getting the full benefit of it. You have seen it during the COVID time and just after the acquisition of Medidata. But when the market is shrinking, basically you are penalized. And I know every quarter I'm repeating this. The worst of the worst is, in fact, we are gaining market share. But you have a hard time to figure out because you see the number decreasing. In a way, we are reinforcing our position into this space. So the way to do it, there are two different axes. One is to be more sticky and less dependent on the number of clinical trials, which is the enterprise approach, which is nothing more than the PLM approach we are applying to the sector. And we see a lot of traction downstream. You know, all the topics which are related to the manufacturing, to the supply chain management, how to accelerate the transfer from the lab to the production system are extremely critical. The reason why we signed with Sanofi the extension was they have 12 molecules in their pipelines. They need to basically put on the market in the next three years. They want to speed up the ramp-up for the production and to gain almost a year compared to what they used to. And the way to do that is very simple. You do most of the ramp-up production when the molecule is still at the lab level in terms of development. So you do what we do in other industries, except we do it specifically for the life sciences. So this is one axis, to be enterprise-wide and to focus on the downstream and basically climb up if you want the value chain. The second one is, Medidata is a medical platform. The 3DEXPROMPT platform is an enterprise platform, but Medidata is a medical platform. And if you look at what kind of information we have into the systems, those are the medical insights for, right now, only the clinical trial, and the intent is to expand the usage of this medical platform when the patients are under treatment. So this is what we call the patient-centricity, because there is no reason we cannot follow the patient when the patient is taking the drugs. And we can follow this, we can follow the adverse effect, we can make prescriptions, how to take the drugs, when is the appropriate time. There are many, many services we could imagine around this way. And this is the strategy we are building around MyMidiData. those are the two axis we are using as a way to if you want to be more sticky and less dependent on the volume of things now this is requesting to change the offer right and this is what we are doing in the way we report the number there is a little bit something which is hidden In fact, you do not see the traction of the rest of what we do in life sciences, because in the life sciences and healthcare line, we are only reporting basically metadata and biovia, but we are seeing more and more delmia, enovia, simulia, and also kaki and soliguas in the med device, and this is reported into the industrial innovations. So if you combine all those things, the picture is, in fact, better.

Hi. Good morning. I'll repeat my question. So the first question is on MediData. How are you seeing the dynamics in the U.S. evolving? It would appear that some of the overhangs, regulatory overhangs, have been reducing of late. And separately, we have also seen some of the CROs in IQVIA, ICON reporting decent booking numbers of late. So where are you seeing incremental growth headwinds for MediData coming from? And as we go in 2026, are you seeing a better visibility or some improvement in the decline in stats that we have seen in 2025?

Yeah, Balaji, thank you. I think the IQVIA and ICON outlooks were mixed, to be fair. So I don't think that anyone is saying we are yet through the decline in clinical trial activity. For sure, when we just look at clinical trial starts, Balaji, and the public available number that where all pharma companies are reporting their clinical trials that are starting, the numbers are down. And as I mentioned before, for phase three, they're down significantly, and this is where the lion's share of value is concentrated. For software vendors as well, also for CROs, because this is where there's the largest operation, most of the people are involved, and it is where the lion's share of the value is created. This is what's affecting us, and we have yet, to Pascal's point, to show that we can rebalance that headwind that we are facing from the volume decline with growth by creating a more sticky offering, connecting the dots across the life sciences enterprise to have that growth outweighing the decline just from the volume in terms of number of trials started. This is the challenge that we are facing. This is what we're seeing right now in our numbers reflected of minus 3%. At the same time, when I look just at the segment level from enterprise and mid-market, both parts are growing. But also for those two parts, they have less clinical trials in their portfolio than what they were doing in 2019, even before COVID. So we are already rebalancing with our offering and increasing our footprint with more value that we are creating for clients that we are able to monetize. But when you then include the pure volume part, still it overweighs and it reduces the growth in this quarter to minus 3%. I think regarding the U.S. regulation, Right now, biotech funding is still not great. And that's also a reason why there's less trials started in the US and in Europe. But we see an increasing trial activity, for example, in Asia, specifically in China. And that's a market opportunity that we are also addressing. But it's a different market with different economics. And now looking into 2026, It's difficult to predict what trial starts will be in 2026. I think what we should assume at this point is that our mix in 2026 is improving versus 2025 to be in a better position to offset that volatility. And offerings like Clinical Data Studio are an enterprise offering. They are not clinical trial related. And this offering is going very well and we are leading with this offer. in the industry. One important part of the App-V announcement is the AI everywhere part. So when you are making decisions today as a company to think five years out, AI is at the center. And our AI strategy is resonating very well. And this is a catalyst for 2026. So you hear I'm a bit more optimistic than what we're seeing in 2025, but it's too early to declare victory.

Thanks a lot for a very comprehensive answer. If I can have a follow-up question. So following up on the AI debate that we have right now, and I appreciate it is a bit different for vertical software companies. But are you seeing your clients taking a pause in decision making also on account of trying to understand where the debate moves of software versus foundation model and also as your own 3D universe is often mature. So are customers also weighing decision and taking a pause in decision making?

So it's a very good question. In fact, for many of our large customers, they started the AI initiative two years ago, in fact, by doing a lot of by themselves or sometimes partnering with startups. After two years, they are coming to the conclusion it's promising, but you need to integrate this foundational model in the way you operate the company. And we are at this point. Let me give you an anecdote. I was with Ford a few weeks ago. And the CIO was telling me he stopped almost all the AI initiatives because now he wants to rationalize. That he wants to rationalize in the productive way. He said, obviously, it's a lever for us. We have investigated many use cases. Now we need to focus on the one being more promising. And usually the way to do this is very simple. You look at the moonshot, the one really changing the game. If you focus AI on the things which are making some improvement in what you do, you will never have your payback because AI is costly. However, if you focus on things you cannot do or you can do with a very different level of efficiency, the payback is there. And we are really at this stage. So for us, I will say it's driving against the adoption of the platform as the data lake. They are building more and more on the foundation because there is no need to redo the job. We have already done it. And more importantly, it's already integrating in everything they do. Because at the end, if you want to design the car, you still need to get here. The fact that CATIA is driven by an AI engine is one thing, but you still need CATIA to produce the model, to produce the geometry, to produce basically the instruction for the shop floor. This is really where we are game changer. And this is extremely difficult to do without having our foundation, in fact. So to come back to your questions, I think, yes, there is a pause in a way people are doing less experiments. But now they are taking the decision to focus on the core use cases which are really productive and making the moonshot, I call it the moonshot, I mean having a significant level on the efficiency or opening new avenue. For example, this is what we are seeing in the material science. There are certain things you cannot do if you do not have an AI engine to do that. We are at this crossroad, but we are much more benefiting from this than something else.

Hi, it's Charlie Brennan here from Jefferies. Apologies, three questions for me. Firstly, I'm struggling to match the narrative onto the actual numbers. You're attributing the weakness in the quarter to metadata and Centric, but metadata is a recurring revenue business and recurring revenues actually beat expectations. Centric is partially a licensed business, but it doesn't feel big enough to account for the size of the license decline. Is there anything else going on there, maybe a change in revenue allocation between the two lines or what else went wrong in the quarter to justify the shape of the numbers? Secondly, I'm hearing accelerating subscription as one of the themes coming across. Is that just centric or is it more broad than that? And traditionally it's tough to accelerate growth when you're moving to subscription. Do we need to think about a phase in 26 and 27 with accelerating license declines? And do we have to think about that in the shape of growth going forwards? And then thirdly, I should probably sneak one in on cash flow, 84% conversion in 2026 is a surprisingly precise guide given the recent track record on cash flow. Are you confident that that takes account of all of the working capital terms on deals that you're going to sign in 2026? or is there scope for payment terms on deals in 26 to disrupt that 84% cash conversion?

Okay, thanks, Charlie, for the questions. Let's start, go through this one by one. On the quarter... There's nothing else than what I outlined. I don't know where the disconnect is, but maybe I just reiterate it in simple form. Yes, metadata is a recurring business, but it also has a volume aspect of clinical trials that are starting and ending in a way they're not recurring, right? I think we were always clear about that. There's a subscription part where we contract over a period of time, And then there are studies that are starting and ending. And there is volatility to that. Of course. But, you know, when studies are ending, they stop recognizing revenue. So there's a lot of studies ending. There's new studies are starting. If you're down minus 3% in a quarter on 250 million, you can do the math in terms of how many this is. But there's a number of trials. We are running thousands of trials, Charlie. So they're in a market, as Pascal said, where the volume is stable. Where we can gain growth through market share expansions, it's a solid generator of growth. And this has been the model of MediData for a long time. Now today, the volume part or the consumption business, you might say, represents about 30% of the overall business. And that business is down high single digits. And that's impacting the quarter, high single digit to low double digit for that volume business. Now it's offset by the increase in subscription contracts from deals that we are expanding and winning in the market. So that's to the metadata part. There is no magic to that other than this. Under subscription acceleration, as I said in my remarks, it's twofold. It's deals that we are signing in the quarter and RAMs that are contributing to the growth from deals that we have signed in previous quarters. You know, we are transitioning our installed base to cloud, but in many cases it's not a 100% transition. In the case that Pascal mentioned in the presentation, the company Stellantis, that's 100% for that part, right? It's not contributing to subscription this quarter. It will contribute over the period of time. But we have other deals where we have on-prem and cloud hybrid deals where there is a portion in the license subscription and a portion in the cloud subscription. And that has a higher impact on the in-quarter revenue in the subscription line. But it's still recurring and it's building over time. And this hybrid structure of deals helps us to get away from the subscription license where the upfront portion is most significant and helps us to spread revenue more equally over time to create a more recurring base. And of course, when you look at our subscription on a quarter-to-quarter basis, it's and I know where you're coming from, it's not sequentially up every quarter because of the on-premise part of subscription, which we well understand that depending on the start time of renewal or renewal dates, there is a fluctuation from quarter to quarter on our subscription business. You can go back years and you can see that. So, rounding up the point, there's a contribution from deals signing in the quarter that are where we have on-premise and cloud portion, where the cloud is over time and on-premise has more of point-in-time impact. And then we are seeing ramping deals from deals that we have signed before. So also here, there's no impact from the Centric part. For the multi-year deals of Centric that we have recognized over the last quarters and last year specifically and before, That revenue is part of upfront license because it's a license subscription where you upfront revenue and it impacts the license part. So that's not driving the subscription business, Charlie. But going forward, as we are transitioning this business more to an AR model, to a SaaS model, it will support the subscription growth. And that's the whole point of what we are doing. From a cash flow perspective, well, I think 2024 is an outlier in terms of several effects that we are facing related to tax impacts, social charges that are higher compared to 2024. That is all going to be in our base in 2025 compared to 2026. So we don't have those one-time effects any longer. At least they are not foreseeable at this point in time. And as it relates to the ramping deals that I talked about on the subscription line, they will generate significant higher cash in 2026 than in 2025. I have that level of visibility. Now, that's the baseline for the assumption to be back at the 2024 levels. Now, is there possible variability? Yes. But the baseline assumption is the 2024 performance.

Maybe one additional comment I should make. Charlie, there is no trick. I think my commitment is very simple. I want to continue to gain market share in all the industry we serve. And I think quarter after quarter, I hope I'm proving to you that this is what we do, including in sector where it's extremely competitive. And the second thing is we are accelerating the transitions to subscription and to the cloud. That's what we do. So my view, we are doing the right things. It's the appropriate time. Again, we were pushing this for a few years ago, but the market was not ready in our space for the cloud. Now it is. They are. And if you remember, the subscription used to be a third of of the recurring revenue three, four years ago. Now it's 50%. And we are on a path in the next three years to be almost two-thirds of the recurring part of the revenue. So I think we are walking the talk. That's what the commitment to do. This is what we are doing. We are redirecting the deal. And I think at the end, the numbers are reflecting this extremely, I mean, transparently.

Thank you. We have unaligned questions coming from the line of Laurent Dorre at Kepler Chevron. Your line is open. Please go ahead.

Yes, good morning, gentlemen. Three quick questions. The first, if you could elaborate a little bit, giving us an update on your pipeline of large deals by maybe verticals and your discussions with those clients, the long sales cycle, is it just a macro or is there anything else on the discussion you have with them? My second question is, if you could give us a bit more color on the change in management at Centric and also on the 15% course you're expecting for next year, the visibility you have on that given that you will continue to move to subscription. And my final comment, my final question is when you refer to a couple of years to rebalance the life science business, Do you see a risk that maybe for two or three years that this business ends up being kind of flattish? Thank you.

Okay, so Laurent, I will take it, and Ruben, feel free to add whatever you want at the end. So the pipeline coverage is two times, which is good. You know, for Q4, usually this is where we are. And it's relatively balanced between the large deals and the mid-sized deals, which is also important because when you have too much on the large deals, this is sometimes difficult to manage. In terms of industry contribution, it's relatively consistent with Q3. So you still have a fraction which is transportation and mobility centric. We have a large part also coming from aerospace and defense. And we also have good visibility on industrial equipment. So that's for the core industry. And again, the pipeline coverage is definitely not the topic. What we observe... And we have been explicit about this. Sometimes one or two big transactions can shift from one quarter to another one, independently of us. And that's what Ruben is mentioning when he says the geopolitics is basically putting some volatility on the time to close. But it's only a question of time to close. It's not a question related to the pipelines. Coming back to the Centric management change, in fact, it's very simple. You know, Chris is turning 70. Chris, the founder of Centric, is turning 70. For a few years, he was preparing Fabrice Canonge to take the positions. So we say it's the right time. You know, we completed the acquisition of the remaining piece of Centric. So from basically... A timing standpoint, it was appropriate to make the changes right now. And as part of the new setup, the new leadership, we have put these transitions to the cloud as one of the objectives for the team. And the billion threshold, which is the size of this business we want to achieve in the coming years, also one of the objectives for this new team. The last thing is related to centric performance for next year.

Life sciences. The next two to three years. What is our expectation? The rebalancing of life sciences.

The rebalancing is already happening again. So except, this is what I was telling you, we are reporting in a line which is not making visible for you. So probably something we need to change for you to have a visibility, to understand how the momentum is going in order to basically balance between the volume-based reports business versus the enterprise-based business in life sciences. Now, if we step back a little bit, I think the booking growth is good for midi data. So the public is not the booking. It has to accelerate, obviously. But it's against this termination of studies and not having a new one starting again, which is the topic. I do expect we are reaching the bottom. I told you this last year. I remember. And again, it was based on facts because we were tracking all the pipelines. The way we do this, we look at how many phase one, how many phase twos. We make some assumption about the move from one to another one. And this is how we are computing, if you want, the potential new studies starting every year. Now, there is a big change, and you highlight it. We see Asia contributing to the trend, especially China, which was not the case in the past. And we were relatively dependent, as you say, on the U.S. dynamic for the creation, or at least the most promising molecules coming on the market for the phase three. Now we see basically this being much more balanced between the different continents. And this is also giving hope for me because we see, and if you track it, we are seeing a lot of investment in the biotech in China, but also in Korea as well, Japan as well. So I do believe if we combine the two together, we will be in much better situations. And Centric, because that was also the question. Again, we had this massive renewal last year. That's the reason why we have the base effect this year. And if you combine this with the fact that we want to accelerate, I want to accelerate the transition to the cloud and the SaaS business model, this is creating the gap. But this basically in 2026 will be in a much better situation because we will not have the base coming from the big renewal. Anyway, the trend and the acceleration of the cloud is already happening. So that's the confidence I can share with you.

Yeah. Thank you. Our next question comes from Frederick Bolon at Bank of America. Your line is open. Please go ahead.

Hey, Pascal. Hey, Ruben. Thanks for taking the question. I've got two and a short clarification. Firstly, around AI, if you can spend a minute on your commercial model of the offering you've presented, any kind of attach rate you foresee on the midterm view. Second, coming back on the free cash flow side and your 84% conversion from next year, any specific moving parts or action plans you want to call out to underpin your confidence in free cash flow acceleration? And then short clarification on May data, can you confirm the comment you made on similar growth or similar revenue decline? Is this a comment about Q4 versus Q3 level of minus three? Thank you very much.

So when I take the first one. So the way it works for the AI new category of solutions is very simple. You remember the portfolio is structured around role, processes and solutions. So in front of the role, we have the virtual companions and the virtual companions are there to augment the role and to extend the roles. The generative experiences are there to basically automate the processes. And the solutions, ultimately, is what we want to do with the virtual twin as a service. So keep this in mind for the purpose of the clarity. Now, how do we price each of them? The virtual companions is priced on a fraction of the cost of the people we are either augmenting are basically substituting sometimes. That's how we price. For the generative processes, generative experiences, it's a usage-based model. So it's a token-based, like many companies do. And why so? Because I really want to ease the adoption and to accelerate the adoptions with this consumption model. And it's something we master relatively well because it's almost the same approach we have for simulation for a long time. And for the virtual twin as a service, it's an outcome-based model. Because at the end, you are not selling anymore the tools, you are selling basically the end result of what the tool is producing. So it will be an outcome-based model. Now, from an attached rate standpoint, it's still a little bit early because we came on the market with this. But we could expect that for many roles you have on the market, being used right now, you will have an extension with the virtual companions, for sure. You could expect that for processes which are the most complex one, The generative experiences will be a way to accelerate significantly the time to market and the efficiency. This is true for the design. This is true for the manufacturing. This is also true for the compliance, as I was highlighting it. And virtual twin as a service, it's something we do specifically in the new industry because they are not equipped. Usually, they do not have all the skills. and we are gaining a lot of time by doing so. One example of what I'm saying, in the life sciences, when we are speaking about the manufacturing systems and the production systems, More and more, we go straight with the virtual twin as a service, which is an easy way for us to deploy our solutions and to reduce the time for the adoption. That's how we are. Basically, pricing and how we are planning. And you remember what Ruben said. He said the contribution of those new category of solutions, we are expecting half a billion in the coming plan, which is ending in 2029.

Okay. Thank you, Frederick. I'll go through the cash flow question. Regarding the 84%, what are the kind of puts and takes and level of visibility and action items that we have underway? I think first, important to mention is we have a certain level of visibility from large contracts that we have signed. where there are clear payment terms that are going to drive cash in early 2026. So that gives us a clear perspective on the puts and takes between 2025 and 2026, which I call the timing effect that we had. So that's one part. We also have some other non-recurring payments in 2025 that will not recur in 2026. We also have visibility to this. Now, above that, When I look at our DSO and the impact of the DSO in context what we just discussed on Centric, Centric has been a big driver of the increase in DSO, has contributed to the increase in DSO, I should better say. And now as we are moving to a recurring model, we will see the benefit of that also in terms of better aligning revenue and cash. So the conversion from that perspective should also, will benefit from this change that we have decided. And then the last point is we are applying strict discipline on cash management, and that will have an impact also in 2026, and I already see that happening in 2025. The last point regarding the metadata comment, yes, this was related to Q4. So the trend of Q3 to be expected similar in Q4 2025.

So this is concluding this morning session. So thank you very much for the one being there with us in London and for the people being connected. Look forward to seeing you on the road, either Ruben or myself. We will do some roadshow in the coming weeks. And see you no later than early next year. Thank you very much.