Good morning, everyone. I'm Stacey Pollard. I'm here with Dassault Systèmes CEO, Pascal Daloz, and the CFO, Rouven Bergmann. Unfortunately, our Head of Investor Relations, Béatrix Martinez, could not be with us today. She's out for a couple of weeks. 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. 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 Daloz.
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. 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. I think, Rouven, you will come back on this. We have a strong EPS growth. 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 recurring part of the revenue.
If you remember, a few years ago, it was only a third. This is growing extremely well. We hit a 30.1% operating margin, which I think is reflecting our focus on running a profitable and efficient business. Finally, the earnings per share came at $0.29 and growing at 10%. 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. Despite the ongoing challenges in this sector, we have also a strong momentum behind 3DEXPERIENCE and SOLIDWORKS this quarter. The second thing is, I think our focus on accelerating SaaS adoption is starting to pay off this quarter. You will see this is driving the revenue growth and the strong market traction.
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. 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. You remember the Virtual Twin as a Service, the Generative Experience, and the Virtual Companions. I will say more about this.
They are really transforming the way our industry, our customers, are designing, producing, and operating the lifecycle. Now, for the full year, we are confident enough at least to reaffirm the earnings guidance. We expect the EPS to grow between 7%- 10%, with the total revenue rising 4%- 6% on an adjusted basis. It is mainly due to three factors. The first one is the lower growth from Medidata, which is in line with Q3, in fact. The impact of the SaaS acceleration 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 resiliency. To give you the numbers, this quarter, we are growing at 18%. Why so?
It is usually when it's a difficult time for our customers that they have to take radical decisions. This quarter, we have some. Ford took the decision to go with us to expand outside of the engineering borders. 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. There is also another very important flagship customer we signed this quarter, which is Stellantis. I know some of you are expecting us to move along this way. I will come back on this. The reason those companies are basically adopting widely the 3DEXPERIENCE platform is because they are using our solution first to speed up innovation. 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. It is a little bit like fast-moving goods. To master the complexity, you need a different approach. This is where I think we are making a difference. Sustainability is also a topic. The electrification is driving the cycle. You know it. More and more with the SDV, we are creating a personalized experience for the customers. 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 is growing double digit this quarter, where programs are becoming more complex and collaborative. I think our 3DEXPERIENCE platform, combined with what we call the Model-Based Systems Engineering, MBSE, which is now a standard in the industry, is more and more widely adopted.
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 Rouven will say more about this. We still see the new clinical trial start being contracted. Nevertheless, we landed with some big contracts this quarter. More importantly, I think we are also being encouraged by some large win-backs. AbbVie is one of them. You remember, it was one of the flagship customers of Medidata a few years ago. They signed with us a contract for the next five years. I think this is the proof that what we do is extremely critical. I think also this is the proof that what we have built as a foundation is critical for them, also for the AI-based programs. 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. 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. We are using our solution to do a Virtual 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. This space is really, again, a way for us to establish a leadership in a domain where we are the challenger. In this space, I think we do not have the same footprint as the others. Now, let me show you some key wins.
Stellantis, you know the company, I mean, you remember, we had a significant footprint with PSA. The rest of Stellantis was much more in the hands of our competitors. They took the decision to standardize on the 3DEXPERIENCE platform on the cloud, which is, I think, important for their system engineering backbone. This is extremely important because, as you know, the system engineering is the foundation to do the SDV. All the car players are moving along this way. They are using our system approach, system-to-system approach, as a way to standardize across all the domains to unify the bill of materials. 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. You need an infrastructure to do this. That's what the 3DEXPERIENCE platform is ready for.
We are extremely proud to support this transformation. 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. AbbVie, you know, it's a global biopharma. It's one of the top 10 global pharma. It's a win-back. It's a win-back of a win-back, let's say it this way. A few years ago, they took the decision to open some clinical trials with Veeva. Now, they are back with us. There are a few reasons for that. One of them is the time. They were sharing with us that we are 10x faster in the way to run the processes and the clinical operations. It's also a big cost saving, which is an interesting takeaway.
You remember, one of the arguments which was used was this EDC is becoming a commodity. It's price-sensitive. The reality, the price is one thing. The savings and the efficiency is another one. Here, you have the proof. The last argument, 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. They have seen through our platform the ability to develop their own program on top of what we do. Those are the reasons, if you want, behind these win-backs. Finally, from a customer standpoint, this is an interesting case also. Korea Hydro & Nuclear Power is the largest energy public enterprise in Korea. They have launched the digital transformations to manage, as I was telling you, the decommissioning of 26 reactors.
The reactors are first generation. They are progressively replacing it with the new generation. To do this, it's a complex process. They have to decommission this large install base. They showcased this example, this case in the Korean Swedish Experience Forum a few weeks ago. I was having the chance to participate in this. Frankly speaking, you should really look at it. It's amazing what they have been able to do. It's a very complex process. Safety is at stake. Compliance 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. 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.
They are using the Virtual Twin in order to make this. Why I picked those examples? 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, whether it's in mobility, in life sciences, and the energy sector. Now, let's speak about 3D Universes. You remember, we announced it in February this year. I was making this statement. 3D Universes is not an extension of what we do. It's really a leap forward. 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.
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. This is a unique corpus to train our systems. Remember, AI without having high-quality data is just only a noise. 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. This is extremely important because if you are only relying on patterns, matching, and recognitions, it's not enough for what we do. The AI needs to be built on physics, biology, material sciences, engineering principles. Why so? When life is at stake, whether it's 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. 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. You remember, we presented it during the Capital Market Day. Now, I'm really pleased to introduce you to our Virtual Companions. 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. You have [Marie], who is our scientific authority handling probably the most advanced questions on research.
The interesting thing is if you ask the same questions to all of them, you have different answers. More than a long explanation, let's look at the video.
Which materials should I use to design this e-foil wing?
When choosing a material for an e-foil wing, you'll want something that balances strength, lightness, and resistance to water. Many e-foil wings are made from carbon fiber since it's strong, lightweight, and performs well in water sports. Fiberglass is.
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 is fiberglass reinforced polymers.
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 gm/cu cm with 10,000.
As you can see, it's not just about AI. It's about having an AI which is behaving like your team. When you do engineering activities, you need to assemble different domain expertise at the same time. If you try to converge too rapidly to the solutions, at the end, you are letting some open opportunities untapped. 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. This is also an interesting thing because you can use AI as a way to take smarter decisions and faster. Here is, again, a concrete example. It's Aura. Aura is widely used by many industries currently to deal with the tariff, with the trade policies, the supply chain issues. This is changing so much that you need almost every day to reactualize your what-if scenario.
Aura, in this case, is not only anticipating but reacting. She anticipated the turnaround, the uncertainty. She tried to manage with data-driven insights the consequences. This is important because for many industries, the margin is at stake. To keep you ahead, 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. Let's speak about SOLIDWORKS. This is an interesting, this is a very important year for us. It's a milestone because we are celebrating the 30-year anniversary of SOLIDWORKS. Why is this important? If we step back after 30 years, I think no one will debate that SOLIDWORKS is the undisputed leader in the 3D CAD. I put some numbers on the slides 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. Again, it's covering the large spectrum of all different industries we serve. It's a lot of legacy of innovations that we are keen on pushing from a product development forward. 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. The topic for us is not only to automate tasks, but more importantly, to give more time for the creativity. We have some features we are introducing and some functionalities. The first one is obviously the generative design.
The second one is what we call assistive features, which is an intelligent pattern recognition when you do, for example, an assembly. All those kinds 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, this AI approach is a way to do the docking, to bridge with the 3DEXPERIENCE platform. As you know, this topic is at stake for several years. I think now, I believe we have found the roots to connect the SOLIDWORKS large install base we have with the 3DEXPERIENCE platform. It's a way, if you want, to turn the SOLIDWORKS users into the lifelong experience partner. I think, and Rouven 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.
To conclude, I think why everything I share with you matters. There are a few things. I'm sorry. I should not anticipate your presentation, Rouven. The first one is 3D Universes is giving few and large advantages. The first one is, you remember, we are helping our customers not only to manage the full lifecycle of their products, but more and more to manage the lifecycle of the intellectual property. You should remember what I'm telling you. In this AI period, the most important asset is the intellectual property. Because everything you build is leveraging the intellectual property. If you do not have a way to manage it safely, to take it as a real asset, to manage the lifecycle the same way you manage the lifecycle of the products, you take the risk to be out of the game.
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. 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. AI is a fantastic tool to read millions of documents, to extract thousands of rules. What do we do with those rules? We do design, 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. 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. The way to do it, you remember, we are training our AI on very comprehensive data sets, which is pretty unique. Those are very high quality data sets. It's validated by the science, which is even more important. We are deploying those artificial intelligence capabilities into a secure and sovereign environment, which is what we do with 3DEXPERIENCE at scale. This combination is pretty unique on the market. It's a huge differentiation compared to many of our peers.
This is, in my view, a game changer in many, many customer engagements we have right now. 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. You will see more and more the Generative Experience, the Virtual Twin as a Service. We have a roadmap for this for 2026, 2027. This will accelerate the contribution of AI in our revenue streams. With this, I think it's time for me to hand over to you, Rouven, to give more flavor on the numbers and probably the outlook for the rest of the year. The floor is yours, Rouven.
Thank you, Pascal. 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. 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% year to date, while Medidata and Centric were softer than expected. As discussed previously, the repositioning of Medidata 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. For Centric, we're accelerating the SaaS transition. 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.
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. 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. Starting in 2026, subscription revenue will surpass maintenance revenue in absolute terms. 3DEXPERIENCE was the growth engine behind that, up 16% in Q3. The signings of Ford 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 it's 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 3DEXPERIENCE revenue. 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. 3DEXPERIENCE cloud revenue grew 36% in the quarter and 29% year to date.
The key wins for 3DEXPERIENCE cloud, such as Ford, Jamco, [Darrala Automobile], 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 billion 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 $0.29, 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-digit 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 had 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 2023 across key domains led by CATIA, ENOVIA, and DELMIA, as well as SIMULIA, highlighting the value the 3DEXPERIENCE platform is delivering to our clients. 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 -3% in the third quarter, with Medidata impacted by continued study start declines, but importantly, continuing to gain market share. Overall, from an industry standpoint, the volume business continues to face pressure. When we entered 2023, we had assumed that volumes would stabilize, helping to support our forecasted growth in the second half. Conversely, we observed a decline in high single digits in phase three studies and mid-single digits decline across phase one and phase two since the beginning of this year. While we are expanding our market share, the impact of the decline in study starts 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 Medidata platform win back, the top 25 pharma, AbbVie.
After a brief period with a competitor, AbbVie decided to return to Medidata for all clinical trials, leveraging AI everywhere. This validates the trust clients place in us and the value of the Medidata platform. Additionally, in Q3, we expanded partnerships with Sanofi. You see the press release this morning. We are also expanding our business with IQVIA, including PatientCloud. Looking at life sciences outside of Medidata, 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 shift to subscription is well underway at SOLIDWORKS. Centric growth was slower than expected in the quarter due to some shifted renewals.
We saw an acceleration in the share of clients adopting the SaaS model. Now, turning to cash and the balance sheet IFRS items. Cash and cash equivalents totaled $3.91 billion as of Q3, compared to $3.953 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 billion a decrease of $138 million versus a net cash position of $1.459 billion 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 billion in operating cash flow for the first nine months, versus $1.348 billion last year. The cash conversion from non-IFRS operating income was 97% for the first nine months. Cash conversion is a top priority.
We expect the conversion to improve going forward. 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, we now expect operating cash conversion for the full year 2025. We now expect operating cash conversion to be in the range of 78%- 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 in 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%- 6% XFX for both the total revenue and software revenue, versus 6%- 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 delta can be explained by three factors: A, the lower growth from Medidata in line with the 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 Medidata is ongoing. We are confident as well into the accelerated SaaS transition of Centric, given its strong positioning in a very large market. Clients are endorsing it. For industrial innovation, we have built a very strong foundation in 2025, where we signed significant contracts. We expect in 2026 to expand on these partnerships, transforming with virtual twins and Generative Experiences. Last but not least, the SOLIDWORKS momentum is strong.
Recurring revenue outlook remains stable. It's at 7%- 8% growth. Underscoring what I said at the beginning, we are implementing a sustainable recurring growth model with increasing visibility. Above all, I mentioned the strengths of our operating model, highlighted by the margin improvement. As such, we are maintaining our EPS growth expectation of 7%- 10% growth, or $1.31- $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. 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%- 8% is fairly large.
This is predicated on potential uncertainties in the timing of deal closing, mainly for the upfront license business, while subscription growth of 8%- 12% is solid on a high comparison base. Operating margin is expected in the range of 37.2%- 38% and EPS growth of 7% to 17% XFX, to hit $0.41- $0.45 EPS for the quarter, reflecting the ongoing operating leverage. As I reflect on the performance so far this year, I want to highlight that our operating model is resilient. 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. Pascal and I look forward to taking your questions.
Thank you. Online participants can ask a question by pressing the pound key five on the telephone keyboard. 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, Rouven, 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? 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. 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?
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? Just walking with the 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 impact on timing of closing this could create, we were prudent to reflect at the low end a more conservative and prudent perspective of large deal contribution. The mid-range, the midpoint requires some of those large deals. We have the potential to do better because our pipeline is strong.
It's depending on the timing of closing of those large deals and the size of those large deals.
Coming back to your question about the transition from the on-prem to the cloud and how it is linked with AI, definitively, AI is accelerating the trend. There are a few reasons for that. One is because no need to wait to have transitioned everything before to start AI. The way we do it, we do what we call supplemental. When you have a large install base or large deployments of the 3DEXPERIENCE platform on-prem, we come with an instance on the cloud in order to basically enable all these AI new categories of services we are developing. This is really accelerating the trend. You have seen in the numbers, it's 36% growth this quarter, the cloud related to 3DEXPERIENCE platform. It's 30% since the beginning of the year. It's extremely correlated also with the subscriptions accelerations, with 16% this quarter. In a way, this is helping the transition.
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, Rouven. Hey, Pascal. Mo from GS. Firstly, it's encouraging to see on the industrial business, there is pretty good momentum, particularly with Stellantis, Ford. As you look into next year, as we think of some of the headwinds and the tailwinds, how should we think about the growth across the different sort of segments of the business? Obviously, in mainstream, Centric has a transition still to navigate. On the life science side, it sounds like visibility is still reasonably low, but the industrial business is ramping. How should we think about the puts and takes for growth next year? I mean, and then secondly, as we think about the life science business, have you, clearly, it's behind plan. How do you think about the kind of strategic view of this business over the medium term? You're willing to ride it out?
Or is it something that perhaps maybe you need to change the scope of to try to extract more the growth areas that are probably better positioned? Thank you.
You take the first one, Rouven. I take the second one.
OK. In terms of the building blocks, Mo and we look at the trend of 2025 industrial innovation up 8% year to date at 9% for the quarter, very much supported by the strong growth in 3DEXPERIENCE 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.
From an industrial standpoint, manufacturing industries, including SOLIDWORKS, because in SOLIDWORKS, the momentum is also favoring, I think we are fairly confident in our ability to continue to transform these huge industries. In a way, many of the deals that we signed are a starting point for what's expected in 2026 and beyond. Without giving you guidance for 2026, 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. 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. The opportunity is large. 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 a decline in 2026 from life sciences. I think we are in a better position in 2026 than 2025. That's our starting point. For Centric, the situation is difficult in 2025. It will improve in 2026. The SaaS acceleration 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. We expect around mid-teens growth for this business next year. If you add all of that, I think we should enter 2026 with confidence.
Of course, the macro standpoint is going to weigh. We have to assess that. 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 sciences still being strategic for Dassault Systèmes. Ultimately, this is a question you ask. The answer is yes. There are a few reasons for this. One, if you look at who are the industries 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. If you remember, the core market we serve is really the innovation space. We are obviously serving the ones spending the most in innovations. 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 foundations. Let me tell you why. If we want to address the sustainability challenge, we need to understand how life is generating life. This 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 how life is designing things. My bet is the next generation of generative design will, from a scientific standpoint, come from this space. 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? Rouven already answered partially to these questions. The first one is we need to minimize the dependency on the volume of clinical trials.
That's obvious, because right now, the model is extremely sensitive to this. 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. When the market is shrinking, basically, you are penalized. I know every quarter I'm repeating this. The worst of the worst is, in fact, we are getting 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. 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. We see a lot of traction downstream.
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. 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. You do what we do in other industries, except we do it specifically for the life sciences. 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 3DEXPERIENCE platform is an enterprise platform, but Medidata is a medical platform. 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. The intent is to expand the usage of this medical platform when the patients, they are under treatment. This is what we call the patient centricity. There is no reason we cannot follow the patient when the patient is taking the drugs. 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. This is the strategy we are building around my Medidata.
Those are the two axes we are using as a way to, if you want, be more sticky and less dependent on the volume of things. This is requesting to change the offer. 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. In the life sciences and health care line, we are only reporting basically Medidata and BIOVIA. We are selling more and more DELMIA, ENOVIA, SIMULIA, and also CATIA, and SOLIDWORKS in the med device. This is reported into the industrial innovations. If you combine all those things, the picture is, in fact, better, Mo.
Hi. Hi! Good morning. I'll repeat my question. The first question is on Medidata. How are you seeing the dynamics in the U.S. evolving? It would appear that some of the regulatory overhangs have been reducing of late. Separately, we have also seen some of the CROs in IQVIA, ICON reporting decent booking numbers of late. Where are you seeing incremental growth headwinds for Medidata coming from? As we go into 2026, are you seeing better visibility or some improvement in the decline in stats that we have seen in 2025?
Thank you. I think the IQVIA and ICON outlooks were mixed, to be fair. 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, and the public available number that pharma companies are reporting, their clinical trials that are starting, the numbers are down. As I mentioned before, for phase three, they're down significantly. This is where the lion's share of value is concentrated, for a software vendor as well, also for CROs. This is where there is 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.
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 - 3%. At the same time, when I look just at the segment level from enterprise and mid-market, both parts are growing. For those two parts, they have less clinical trials in their portfolio than what they were doing in 2019, even before COVID. 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.
When you then include the pure volume part, still it overweighs, and it reduces the growth in this quarter to - 3%. I think regarding the U.S. regulation, right now, biotech funding is still not great. That's also a reason why there's less trials started in the U.S. and in Europe. We see an increasing trial activity, for example, in Asia, specifically in China. That's a market opportunity that we are also addressing. It's a different market with different economics. 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. Offerings like Clinical Data Studio are an enterprise offering. They are not clinical trial related, and this offering is going very, very well.
We are leading with this offer in the industry. One important part of the AbbVie announcement is the AI everywhere part. When you are making decisions today as a company to think five years out, AI is at the center. Our AI strategy is resonating very well. This is a catalyst for 2026. You hear I'm a bit more optimistic than what we're seeing in 2025, but it's too early to declare victory.
Thanks for a very comprehensive answer. If I can have a follow-up question. Following up on the AI debate that we have right now, and I appreciate it is a bit different for vertical software companies. 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? Also, as your own 3DEXPERIENCE universes is offering mature, are customers also weighing decisions and taking a pause in decision-making?
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 it 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. 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. He wants to rationalize in a productive way. He said, obviously, it's a lever for us. We have investigated many use cases. Now, we need to focus on the ones being more promising. 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. We are really at this stage. 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. More importantly, it's already integrating in everything they do. At the end, if you want to design the car, you still need CATIA.
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 a game changer. This is extremely difficult to do without having our foundation, in fact. 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 lever on the efficiency or opening new avenues. 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. We are much more benefiting from this than something else.
Hi, it's Charlie Brennan here from Jefferies. Apologies, three questions from me. Firstly, I'm struggling to match the narrative onto the actual numbers. You're attributing the weakness in the quarter to Medidata and Centric. Medidata 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? 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? Traditionally, it's tough to accelerate growth when you're moving to subscription. Do we need to think about a phase in 2026 and 2027 with accelerating license declines?
Do we have to think about that in the shape of growth going forward? 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 2026 to disrupt that 84% cash conversion?
Okay, thanks Charlie for the questions. Let's start going 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, Medidata 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's volatility to that.
They'll still get through the recurring line?
Of course, but you know when studies are ending, they stop recognizing revenue. There's a lot of studies ending. There's new studies that are starting. If you're down -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. They're in a market, as Pascal said, where the volume is stable, right? 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 may say, represents about 30% of the overall business. That business is down high single digits, and that's impacting the quarter, high single digit to low double digit for that volume business.
It's offset by the increase in subscription contracts from deals that we are expanding and winning in the market. That's to the Medidata part. There is no magic to that other than this.
Subscription acceleration.
On the 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 licensed subscription and a portion in the cloud subscription. That has a higher impact on the in-quarter revenue in the subscription line. It's still recurring and it's building over time.
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. Of course, when you look at our subscription on a quarter-to-quarter basis, 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. Rounding up the point, there's a contribution from deals signing in the quarter that are hybrid, where we have on-premise and cloud portion, where the cloud is over time and on-premise has more of a point-in-time impact.
We are seeing ramping deals from deals that we have signed before. 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 licensed subscription where you upfront revenue and it impacts the license part. That's not driving the subscription business, Charlie. Going forward, as we are transitioning this business more to an ARR model, to a SaaS model, it will support the subscription growth. That's the whole point of what we are doing. From a cash flow perspective, 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.
We don't have those one-time effects any longer, at least they're not foreseeable at this point in time. As it relates to the ramping deals that I talked about on the subscription, they will generate significantly higher cash in 2026 than in 2025. I have that level of visibility. That's the baseline for the assumption to be back at the 2024 levels. Is there a possible variability? Yes. 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 industries we serve. I think quarter after quarter, I hope I'm proving to you that this is what we do, including in a sector where it's extremely competitive. The second thing is we are accelerating the transitions to subscription and to the cloud. That's what we do. My view, we are doing the right things. It's the appropriate time. We were pushing this a few years ago, but the market was not ready in our space for the cloud. Now it is, they are. If you remember, the subscription used to be a third of the recurring revenue three, four years ago. Now it's 50%.
We are on a path in the next three years to be almost two-thirds of the recurring part of the revenue. I think we are walking the talk. That's what the commitment is to do. This is what we are doing. We are redirecting the deal. I think at the end, the numbers are reflecting this extremely, I mean, transparently. Charlie.
Thank you. We have online questions coming from the line of Laurent Daure at Kepler Cheuvreux. Your line is open. Please go ahead.
Yes, good morning, gentlemen. A few quick questions. The first is, 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 the 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% growth you're expecting for next year, the visibility you have on that, given that you will continue to move to subscription. 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 Rouven, feel free to add whatever you want at the end. 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, let's say, the mid-size 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. 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 a good visibility on industrial equipment. That's for the core industry. Again, the pipeline coverage is definitively 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. That's what Rouven is mentioning when he says the geopolitics is basically putting some volatility on the time to close. 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. We say it's the right time. We completed the acquisition of the remaining piece of Centric. From basically a timing standpoint, it was appropriate to make the changes right now.
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. The billion threshold, which is the size of this business we want to achieve in the coming years, is 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 are our expectations? The rebalancing of life sciences.
The rebalancing is already happening again. Except, and this is what I was telling you, we are reporting in a line which is not making it visible for you. 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 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 Medidata. The topic is not the booking. It has to accelerate, obviously. It is 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, frankly speaking. I told you this last year, I remember. It was not based on facts because we were tracking all the pipelines.
You know, the way we do this, we look at how many phase one, how many phase twos. We make some assumptions about the move from one to another one. This is how we are computing, if you want, the potential new studies starting every year. 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. 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 this being much more balanced between the different continents. 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.
I do believe if we combine the two together, we will be in a much better situation. 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. 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. 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. That's the confidence I can share with you.
Can I jump one more?
Yeah.
Thank you. Our next question comes from Frederic Boulan at Bank of America. Your line is open. Please go ahead.
Hey Pascal, hey Rouven, 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 a mid-term 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? A short clarification on Medidata, can you confirm the comment you made on expect similar growth or similar revenue decline? Is this a comment about Q4 versus Q3 level of - 3%? Thank you very much.
Rouven, I take the first one.
Yeah, I'm going to close out, yes.
For the cash flow. 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. 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. The solutions, ultimately, is what we want to do with the Virtual Twin as a Service. Keep this in mind for the purpose of the clarity. Now, how do we price each of them? The Virtual Companions are priced on a fraction of the cost of the people we are either augmenting or basically substituting sometimes. That's how we price. For the generative processes, the Generative Experiences, it's a usage-based model. It's a token-based, like many companies do. Why so?
I really want to ease the adoption and to accelerate the adoptions with this consumption model. It's something we master relatively well because it's almost the same approach we have for simulation for a long time. For the Virtual Twin as a Service, it's an outcome-based model because at the end, you are not selling any more of the tools. You are selling basically the end result of what the tool is producing. It will be an outcome-based model. Now, from an attach rate standpoint, it's still a little bit early because we came on the market with this. 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 ones, 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. The 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. We are gaining a lot of time by doing so. One example of what I'm seeing 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. You remember what Rouven said. He said the contribution of those new categories of solutions, we are expecting $0.5 billion in the coming plan, in the coming plan which is ending in 2029.
Okay, thank you, Frederic. 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 2026, early 2026. That gives us a clear perspective on the puts and takes between 2025 and 2026, which I call the timing effect that we had. 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 of what we just discussed on Centric, Centric has been a big driver of the increase in DSO, or has contributed to the increase in DSO, I should better say. 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. The conversion from that perspective should also benefit from this change that we have decided. The last point is we are applying strict discipline on cash management, and that will have an impact also in 2026. I already see that happening in 2025. The last point regarding the Medidata comment, yes, this was related to Q4. The trend of Q3 to be expected similar in Q4, 2025. We are done.
This is concluding this morning's session. 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 Rouven or myself. We will do some roadshow in the coming weeks. See you no later than early next year. Thank you very much.