Good morning, and thank you for joining our third quarter 2026 earnings conference call and event in London. I am with Pascal Daloz, Chief Executive Officer and Chairman of Dassault Systèmes, and Rouven Bergmann, Chief Financial Officer, to discuss our third quarter 2026 results. Dassault Systèmes results are prepared in accordance with IFRS. The financial figures described are non-IFRS unless otherwise noted. Some of the comments contain forward-looking statements that could differ materially from actual results. Please refer to today's press release and the risk factor section of our 2025 universal registration document. All earnings materials are available on our website. These prepared remarks will be available shortly after this call. Now over to you, Pascal.
[Non-English content] Marie. Good morning to all of you, and thank you for being here, and especially for the ones who are physically here in the room with us. I always appreciate to have this live interaction with you. I think before Rouven takes you through the numbers, I really want to start by giving you the big picture. It is about what we are building and why it matters for the next era of the industry. At the start of the year, you remember, we called 2026 a foundational year. I think it is a year where, how can I say that? It is not a year of waiting. It is a year of building. Building new category of products, building an AI-native platform, developing partnerships, all in services of the growth.
I think for more than 20 years, we have helped industry to connect both worlds, the virtual world and the real world, for them to be able to imagine, design, and test. Now, I think at the age of AI, this connection matters more than ever. Why so? Because generating an answer is not enough. You need to create, design, simulate it, test it, validate it, manufacture, often certify before it could fly, before it could be on markets. Whatever is a powerful model you have in your hand, it is not enough. You need engineering context around it. You need connected environments where AI can be tested against the behavior of the real world. This is the role of our virtual twin, our Industry World Model, to ground AI in science and in the behavior of the physical world.
This is why also 3DEXPERIENCE is evolving into an agentic platform. We are bringing AI together with industrial knowledge and human expertise, so intelligence can move from generating ideas to helping to deliver trusted outcomes. Let us start with the performance of Q3. This is important, because I think execution matters. You remember, that is what I say last quarter. I think we delivered a strong execution at the high end of our objectives. You saw it. Total revenue grew 5%. Software revenue grew 5% as well. 3DEXPERIENCE platform +25%, cloud +14%. EPS reached $0.32, up 8%. The revenue growth is now expected to be between 4%-5%, up from 3%-5%. The EPS growth is now expected to be between 6%-7%, up from 3%-6%.
ArisGlobal is obviously part of the updated outlook, and Rouven will walk through the assumption and the financial details for you. Behind this number, I think there are some priorities which are driving the progress. The first one is we are helping our existing customers to transform their businesses and to turn industry pressure into opportunity, and you can see the 3DEXPERIENCE platform and the cloud continue to perform strongly this quarter. Most important, we are also winning new customers, and this not only in one given industry or in one given geo, it is really broad-based. The second priority is we are taking a major step forward in life sciences, bringing ArisGlobal into the portfolio, building toward a unified AI-native platform, and putting a dedicated leadership in place.
I know it is a topic some of you asked many times, how we will integrate all the pieces together, and I will come back on this. The third one is we are expanding 3D UNIV+RSES to turn AI advances into practical industrial value. We see this every day in our work with customers, and we will share a few great examples, the one we saw at the Paris Auto Show or the GTC in Berlin last week. Just remember one thing. Our strategy is very clear. We deliver today, and we build for the next wave, and we connect the two. This is really the most important, and this is what we have done this quarter. The momentum is really across many, many industries, high tech, industrial equipment, automotive.
The next steps for them is not only to serve their local market, but it is really to industrialize those innovations to bring them to the global market. We see these similar shifts in many industries, robotics, pharmaceuticals, aerospace, where complexity is increasing and the ability to connect the disciplines matter more than ever. Take COMAC, the China leading commercial air manufacturers, as an example. Its mission is very simple, is to bring its first large wide-body aircraft to the global market. To make this happen, COMAC needs to industrialize the programs, need to secure the certifications to get ready to bring the aircraft to the market, and again, both in China, but more importantly, internationally. That is the key, because the global market needs to be part of the plan from day one. Not something you think after and you put it later.
This is where we enter into the game, because with 3DEXPERIENCE platform, we bring design, system engineering, simulation, manufacturing planning in one unified environment, giving everyone a shared view of the programs. With the model-based system engineering, teams can tackle the complexity very early on, and when changes are easier, less costly to make, and decisions can have almost no impact. That is how we are helping our customers to manage complex programs and turn the global mission into the reality. Now, let us look at life sciences. Good news, we start to see some positive signs. Pharma and biotech are picking up. You are seeing, like. That is an important point because last time we saw this was at the COVID time. So that is number one. Number two, we are also seeing things improving quarter over quarter at MEDIDATA.
MEDIDATA is flat. Again, Rouven, we come back on this, and we expect to return to growth in Q4. Another encouraging sign is that the customer are increasingly seeing the value of our platform approach. We saw this in Q1 with one of the large deal we signed with WCT, Worldwide Clinical Trials, and now we are seeing something similar with the rest. The idea is very simple. Instead of using different tool for every clinical trial, we give the CRO one AI platform to manage their clinical operations. The benefits are very clear. The team can work much more efficiently and get better in diversification industry, which grew at a double-digit rate this quarter. We are seeing very strong momentum, especially in high-tech and around AI infrastructure.
You remember we have built expertise in the nuclear a few years ago, which was, and still one of the most complex systems to engineer. I think now we are putting that expertise to work in another equally demand area, which is this AI infrastructure. Because AI factory is much more than a room full of GPUs, you need compute, networking, power, cooling, software, and all those things need to work together. When you change one element, it could affect everything else. So you need to think about the whole systems from day one and to be ready to adapt when things are evolving. You know the cycle of obsolescence is such that things are evolving almost on a monthly basis. First for the data centers. Today, these different elements are often designed separately. So what do we do with our virtual twin?
We bring all of them together to design them together, to simulate them together, how they will work as one systems before anything is deployed. The goal is very important, is get AI infrastructure up and running faster and reduce the time to the first tokens. That's the motto, reduce the time for the first token. That brings us back to the big picture, because as AI becomes more complex, we can apply, again, the same system-level engineering expertise we have developed in some of the world's most demanding industry. This is becoming, I would say, on top of the agenda of many of the CEOs who are in this business, because for [audio distortion] , now the constraint that all the other industries, they have to basically be on time, on quality, on cost, on purpose.
We need to apply the other techniques we learn from the other industries, and this is where we are stepping in. Now, let's take a little bit of step back and look at the kind of product industry are starting to create. Think about intelligent robots, autonomous vehicles, smart wearables. They are not just traditional product with AI on top of it. They are entirely new kind of products, and that change the way we need to design them. So remember this, most of you are challenging us how we are integrating AI in what we do. But the real question is not this one, is how we are evolving to basically support this industry when they are becoming AI native. With AI native products, intelligence is built from the start. Intelligence to the physical world, and to make sure that it really works.
That is what 3DEXPERIENCE is all about. Connecting intelligence, digital intelligence to the physical world, and improve and prove that it works. To do this, we combine industrial know-how with science and AI to create Industry World Model and therefore twin at scale. In practical terms, what does it mean? It means industry can try out different ideas, see what might happen, test their decision virtually before taking action in the real world. There is an important point here. In industry, being intelligent is not enough. You also need to trust the results. That is the reason why we are different, because we bring, again, science and industrial knowledge into AI, so we can explain the result, test them, and validate them. Let me give you two very simple example. AI can describe very easily an aircraft wings, but cannot design it, and more importantly, cannot certify the wing.
It can summarize a clinical trial protocol, but that does not mean it can devalidate. So there is a big difference between the formulations and basically, if or being right are very two different notions. That is why AI needs an environment where it can learn, test against the rules of the real world. That is exactly what our virtual twins and our Industry World Models provide. To this, we will come back much more in details at our Capital Markets Day. Now, we also are seeing this shift to take place. Remember, in July, we made 3DEXPERIENCE platform advantage. We introduced all the Virtual Companions. You know them, LEO, MARIE, AURA, Louis. They have some specific skills. Skills need to work alongside with the people. Now, what do we do?
We are exploring with the early AI adopters across industries, around the world, from AI-native startups to some of the big, basically global player. So far, you can see on the slides, we have around 100 strategic customer partnerships working with us, and we are developing, industrializing the use cases. Now, the focus is really on taking this early success to the next level, and this is what we are going to do in 2027. We plan to scale up, moving from the promising use cases to solutions can be deployed more broadly, repeatedly across the customers. Let me give you a more Virtual Companions. They can take on tasks, work together, and bring the right virtual or human experts when it is needed. The most important thing is they do not just answer the questions.
They can actually get work done, drawing on specific skills, industrial knowledge, and they keep the people in control. So here you have a good example, which is what we are doing at Volkswagen. Virtual Companions are helping engineering teams to streamline their workflows and get things done much more efficiently. So we start with AURA as a coordinator. AURA is bringing action plans together, assign tasks. LEO explore the design options based on the Volkswagen's engineering heritage datasets. They share, LEO share the result with AURA. When AURA can review those results, adjust the plans, keep the tracks in what is happening next. So throughout the process, again, the people remain in control and accountable for the decisions.
The real value is not just automating the task, it is getting different skills to work together with, again, industrial knowledge and human overseeing, if you want. It is humanoid robots for Audi with Agile Robots. Here is an interesting case because this is really the physical AI in action. LEO used our Industry World Models to create the virtual twin of the factory. The robots, you know they have their own virtual twins, too. What does it mean? It means it allows the teams to explore how the two will behave together in one virtual environment, reflecting the real world physics. This is where it gets interesting. Assess the cycle times, see how production throughput will be affected. They can work through these questions virtually before anything goes on the shop floor.
The results, and we have now a lot of KPIs, fewer unknowns, better informed decisions, much more confidence, and more importantly, the time to deploy, because at the end, it is how you speed up the commissions. What is in common between the two strands, between the two examples I shared with you? It is really we can explore and learn and validate the virtual works before taking a logical AI. In fact, we are applying the same thinking process, but to a different challenge. The stakes are especially high because ultimately, it is about human health and patient outcomes. Too often, and now you start to know it, what we learn in the real world does not slide back quickly enough to inform the next scientific discovery. In fact, closing this gap is one of the biggest opportunity we see in Life Sciences.
This quarter, we took an important step in this direction with the completion of the acquisition of ArisGlobal. It is much more than adding the drug safety capabilities. It is about connecting the pharmaceutical journey, and the formula is really from the molecule to patients to the real-world evidence, and then bringing what we learn at pharma, helping the life science company, whether biotech, big pharma, to turn those real world insight into the next wave of scientific progress. To make this ambition real, we are putting a new operating model in place. In fact, we are bringing together expertise across all the disciplines, clinical development, discovery, labs, manufacturing, regulation, compliance, real world safety, medical practice also. We are unifying this capability with our 3DX agentic platform, and we streamline the go-to market. The goal is very straightforward.
It is one coordinated organization, one integrated offer, a simpler way for our customer to get access to the solution they need. To do this, I am also pleased to welcome Dr. Ashwini Zenooz to Dassault Systèmes. Ash, she is a physician and a technology entrepreneur, and she has a very strong background in digital health and AI. What makes her background particularly valuable is the fact that she brings together the two worlds, the world of the medicine and the world of the technology, and this is exactly the combination we need to bring our ambitions to life. Her role will be to lead the strategy, bring MEDIDATA, BIOVIA, ArisGlobal together in one AI native platform, connecting molecule, patients, real world outcome, and again, closing the loop back to science.
She will lead the Life Sciences & Healthcare as one business with a close focus on the science-driven innovations, the operational excellence, building ever stronger relationship with our customers and partners. So follow about her visions for the business. The missions of this integrated organization is also very simple. The entire purpose of Dassault Systèmes is not to only improve the productivity and to streamline the workflow. It is to inspire smarter treatment for healthier people. Why this is important, because at the very core of everything we do, you have the patient. Just think about the reach today. Our solutions supporting 50 million patients in the clinical trial, 60 million patients in the real world to collect those real world evidence. The medicines which are tested, and the COVID vaccine is one of the example, is used to reach more than 6 billion people on Earth.
That is the dedication of these organizations. We are also covering half of the safety event worldwide across a wide range of therapies and treatment types. Those numbers are telling you a very important story. They show the scale of the ecosystem we are building. Behind the scale, you have more than 5,000 colleagues. Many of them have a PhD, doctors. They are knowledgeable on the therapeutic domains, on the modality of the medicines. More than basically only the two nodes working closely with the regulators in Europe, in United States, in Asia. We also have built a rich foundation with experience across 42,000 clinical trial and 2 million safety event every year. Why all of this matters? Because when it comes to AI in life sciences, scale and expertise really make the difference.
They help us to build AI which is grounded in science and reflect the reality of the industry. This is an important point, distinction to keep in mind. In manufacturing, physics set the rules. In life sciences, biology is the one, and biology is incredibly complex. You can understand by looking at the document which basically we can make the process faster or more efficient. So it is really about helping the scientists to make new discovery, advanced treatments, and most importantly, improve the outcome for patients. So we look forward to share more with you at our Capital Markets Day in Paris on November 17th. Before to hand over to Rouven, let me sum up a few messages for you. We are making good progress in our priority, which give us the confidence in our updated full-year objectives.
We are also moving from our AI roadmaps to the real world adoptions, and we are taking interpersonal steps forwards in life sciences. With that, Rouven, the floor is yours.
Thank you, Pascal. Also, everyone, welcome to you here, in London, joining us in person, and also, good morning to you online. As Pascal, as you mentioned, we enter the final stretch of the year with momentum. Our focus remains squarely on the execution of our growth model and on delivering our objectives. The annual run rate is up 6% as revenue growth is at 5%, and it is at the top end of our objectives that we had set for the quarter. The operating margin expanded by 150 basis points to 31.6% for the quarter. It is well above our objective. The operating profit rose 10% year-on-year. The strong profitability is converting into cash, with cash conversion now reaching 111% year-to-date. This is driving an EPS growth of 8% and an operating cash flow growth of 20% year-to-date.
In short, we are stepping up the execution as we head into the close of the year. Let us take a deeper look into the financials of the quarter. The total revenue reached $1,552,000,000. It is up 5%, with subscription revenue up 6% in the quarter and year-to-date, and licenses growing at 14%. This license growth reflects a robust performance, a strong performance in China, and it also compares to a lower baseline versus last year. Highlights the progress in our business model transition. This quarter, we are reaching an important milestone because our subscription revenue now represents 50% of our recurring revenue. Service revenue was up 7%. This top-line leverage and continued operating discipline translated into an uplift of 10% in operating profit with a margin of 31.6%, which is up 150 basis points ex-FX. EPS was $0.32 and is growing at 8% ex-FX.
Year-to-date, this brings us to $4,617,000,000 in total revenue. It is growing 4%, with a solid expansion of the operating margin of 80 basis points. EPS is now at $0.93 year-to-date, and it is up a healthy 7% ex-FX. In summary, this quarter, revenue reached the top end of our expectations while operating margin and EPS surpassed them. Looking more closely at our recurring business growth. In the quarter, we had ARR to $4,505,000,000, or a growth of 6%.
What is important to highlight and must be understood is that the subscription run rate continues to accelerate quarter- on- quarter, and it is now moving towards mid-teens, and this is what is driving the underlying strength in ARR and the momentum. Importantly, as of Q3, the subscription run rate now accounts for half of the annual run rate, and it is up from 46% at that time of Q3 last year.
We have been adding one point of share per quarter since Q3 of last year. This is why it is an important milestone, because it is shifting our business model further towards growth acceleration. Equally, we saw continued growth in cloud bookings and the contribution from multi-year subscription deals with high and cloud together in our core industrial business with CENTRIC and MEDIDATA, all generating sequential growth contribution. The top 15 deals accounted for about 80% of the net increase in subscription annual run rate. Let us turn to the growth drivers. 3DEXPERIENCE was 25%, cloud up 14% in the third quarter. 3DEXPERIENCE cloud, now our core business, growing 46%, and it is pretty consistent for the quarter and year-to-date.
The growth in 3DEXPERIENCE and cloud, as you heard by Pascal, reflects strong customer demand and increasing focus on transformation, on transforming the operation, and for the preparation of industrial AI at scale. 3DEXPERIENCE connects our customers' data, the applications they use, and the knowledge they have. With AI, we are virtualizing the real world, leveraging virtual twins to provide intelligence to engineers, scientists, and business users. Our agentic platform. This quarter, clients such as EDF, Mahindra & Mahindra, Eaton, Embraer, RER, Vertiv, among others, they highlight this momentum and progress. Let's take a look at the geo performance. Looking across the geos, Asia was excellent with a growth of 12%. It was led by strong growth in China. Europe performance was resilient. It is growing at 9% in the quarter. In Americas, the execution was well.
However, the growth was down by -3% due to a tough year-over-year comparison. Outstanding performance in China. We saw good momentum in aerospace, including COMAC, as you mentioned, Pascal, as well as strong wins in the high-tech sector. In Europe, it is a healthy growth, and it was driven by renewals, mainly in the U.K. and France. However, we also offset some softer performance in Southern Europe and Central Europe. From an industry perspective, the performance was broad-based. It was supported by good momentum with strong growth in 3DEXPERIENCE as well as subscription. As mentioned, the Americas was negative growth, -3%. It is a result of tough comps when we look at high-tech and industrial equipment are performing and growing well. Our growth drivers performed well with strong growth in cloud in the Americas region.
Important to highlight is that we are advancing fast with strategic clients, creating concrete plans to adopt our Gen7 portfolio. Vertiv is an excellent example, highlighting the potential that Gen7 creates, and it helps us to expand our reach in new industries, building sustainable AI factory infrastructure. Overall, the Americas is more back-end loaded than other geos this year. I think that is important to keep in mind. Let's move to our performance by product line. Industrial Innovation grew a strong 5%, and it had this strong comp base that was more visible in the Americas. Nevertheless, Industrial Innovation up 5%, very solid, driven by 3DEXPERIENCE and cloud, with ENOVIA leading the momentum. We saw several competitive wins across both core and diversification industries, and it really highlights the strength of our integrated platform with good momentum with CENTRIC and SOLIDWORKS.
SOLIDWORKS performance was broad-based across geos, with a unit growth of up double digits. Our traction on AI is very solid, with over 40 use cases now available in the SOLIDWORKS portfolio. This is driving the move to subscription and from CAD into simulation and manufacturing. CENTRIC delivered another strong performance in Q3 with several significant wins, including a global leader in sports merchandising and licensing, as well as a significant U.S. fashion and outdoor brand. This reinforces our confidence in the CENTRIC business to remain at the forefront of consumer-driven transformation across food and beverage, retail, and sports apparel. It is all augmented by CENTRIC end-to-end platform and AI. Moving five points, and we are reaching two important milestones. First, MEDIDATA improved sequentially, and it was flat year on year. It is on track to deliver growth from Q4 onwards.
The shift to enterprise platform business model is well underway. We are replicating the deal we signed with WCT in Q1. In Q3, we signed a significant expansion with Everest. Overall, we are expanding our pipeline of transformational platform agreements across CRO partners, mid-market, as well as enterprise clients. This is resulting in a positive impact on renewals and bookings growth over the last quarters. From a solution perspective, it is worth highlighting that Patient Cloud and our Study Experience. We completed the ArisGlobal acquisition in the second half of September. You heard this already from Pascal. This is, ArisGlobal revenue will be reported as part of the Life Sciences product line going forward. The revenue contribution is not material for the third quarter 2026. One more point.
With Pascal, I share the excitement to have Ashwini Zenooz join and lead our teams across Life Sciences, and I look forward to working with her to reveal and deliver the full potential of our business in this new chapter that we are starting. Now turning to our cash performance. We have generated strong operating cash flow year to date, $1,567,000,000 billion . This is up 20% ex-FX and up 18% supported. This was driven by strong operating performance with higher billing activity and good collections. The third quarter cash flow from operations was up over 77% on a strong operating $1,388,000,000 billion , driven by strong operating cash flow. All of the above resulted in good cash conversion at 111% year to date, which is well ahead of last year at this time. Now to complete the picture.
Cash and cash equivalents reached $3,814,000,000 billion at the end of September 2026. Year to date, operating cash flow has already funded the $1.5 billion ArisGlobal acquisition, leaving cash and cash equivalents broadly unchanged. This also includes the repayment of the $0.9 billion bond, which was due in September. Our net cash position of $1,049,000,000 billion provides the financial flexibility to continue to invest selectively in our growth and value creation levers as we move forward. Now, let us turn to our full-year objectives. First of all, let us take a step back. Since we started the year, we have progressively increased our top-line growth rates from 3% in Q1 to 4% in Q2 to date, as we enter into the fourth quarter. This gives us the room to narrow our full-year guidance range from previously $120 million to now $60 million while we are integrating ArisGlobal.
What does it mean for the guidance range? For the full year, we update our revenue outlook from 3%-5% to now 4%-5% revenue growth. The new range of EUR 6,399,000,000 million -EUR 6,459,000,000 million reflects an increase of EUR 103 million to the low end and an increase of EUR 43 million to the high end of the previous range. With this, our midpoint is increasing by EUR 73 million, which documents our confidence in our 2026 full-year outlook. Reflecting this positive momentum, we are also raising the operating margin outlook, which now reflects a range of 32.5%-32.7%, whereas a 32.2%-32.6% prior. Finally, EUR 5 to EUR 1.36. Now let us zoom into the fourth quarter objectives. For the fourth quarter, we expect revenue growth of 5%-8%, and this includes a two-point contribution to growth from ArisGlobal.
We expect the operating margin in the range of 37.2%-37.8%, with EPS at EUR 0.42-EUR 0.43, which is up 3%-7% ex-FX. For the exchange rates, Q4 guidance now assumes $1.14 per EUR and ¥170 per EUR, and a tax rate of 17.4% for Q4. To summarize, combining the anticipated growth in Q4 combining the anticipated growth and acceleration throughout the year. To finish, let me summarize the key points. We continue to execute on our growth strategy. As you can see, we are delivering sequential improvements quarter- after- quarter, and our growth drivers show the strategy is working. The shift to subscription is well underway, and our 3DEXPERIENCE and cloud growth is accelerating. As we enter Q4 2026, and we look ahead, our focus remains on execution and delivering on our numbers.
We continue to advance our growth drivers, accelerating AI, and shape our life sciences business to deliver sustainable growth. We will continue to expand margins and improve cash conversion to generate consistent long-term value for clients, our customers, the company, and shareholders. This will be at the center of our upcoming Capital Markets Day. With this, Pascal and I look forward to take your questions. Thank you.
We will first take questions from the room and then online.
Thanks for hosting us in person. It's George from Morgan Stanley. A couple questions. Firstly, on the Virtual Companions, could you share what you've been seeing over the last quarter in terms of the ramp-up or usage of those different Virtual Companions? Then the second one on the agreements you're doing on the MEDIDATA side with WCT and now with Everest. As I understood it from what I think you said, they make it a default platform or a default system of choice for their trials. How does that work if a sponsor or drug company comes in and mandates usage of a different platform or a different solution from a different-
George, for the question. I will take the first one, and maybe, Rouven, you can take the second one. The first one is relatively simple, is all the customer who move to the latest release. Obviously, it is the case for 100% of the one being cloud, on cloud. For the one who are on-prem, obviously, they need to do the upgrade, 12% and 14% of the total install base, whether it is SOLIDWORKS or CATIA and 3DEXPERIENCE platform. They are enabled with the Virtual Companions. The second thing is, for all of them, what we have done is a discovery offer. Obviously, because we have to promote the Companions, specifically LEO and AURA, for them to play a little bit with it, without having to charge some senior champions in order to start to build the momentum for 2027.
Right.
This is where we are currently. It is well received. That is probably the most important thing for me. Again, the 100 strategic customers I was referring to are really the one where not only they are using extensively the Companions, but they push us to go a step forward of what we can do with it. Here to use our applications. LEO could also use applications from the third party. We have to train LEO to do it. That was not the number one priority for us when we came with the Virtual Companions. Very rapidly, it became that this is also extremely important for us to basically have this ability if we want to orchestrate the full development processes or the full production processes, having all those things being connected. That is where we are.
Very lot of startups also in the robotics domain, which are probably the one adopting our solutions the fastest because they need to develop the solutions, but also they need to test it, and the virtual environment is really useful for them to make their AI, in a way, capable to be embarked into the products. The second part? Thank you. For MEDIDATA and what does it mean for the CROs?
Yeah. No, thank you, George. Those type of deals we are signing are the enterprise platform deals. What is really important for CROs, but also for our enterprise clients, is for them to standardize and the time to first patient in and first patient out of a clinical trial, or last patient out, first patient in. What we are able to bring to them is efficiency. We have an integrated platform. AI is accelerating their workflows, it is accelerating evidence generation, and they see the benefit of standardizing on our platform. CDS, Clinical Data Studio, is agnostic. It works with any EDC. That is a huge benefit for CROs, and it creates like a platform of platform for us to aggregate and manage all the operation. I have seen that evidence in customer client discussions with WCT, where I was directly involved in Q1 and now with Everest.
It's a main driver for those CROs that are growing fast. They're taking market share. Interestingly, some of them are also private equity-owned, so there is clear metric behind productivity and outcomes. The large investments they are taking with a clear commitment to revise, and we are part of this.
Rouven, the question is also the relationship between the sponsor and the CROs.
Yeah. So when a sponsor has a preference, right? Again, first of all, CDS is agnostic, right? We're not shutting the door. We cannot do that. But the benefit of an integrated platform and the reducing the time to first patient in, is what's driving the standardization and where we have a strong position to gain market share. Of course, we want to increase our share of wallet, our penetration with this approach. But it's an open approach. But we see a benefit, I think, to our adoption.
Next question.
I can proceed, yes. Two from me. The first would be around just in mainstream innovations, the sort of split of performance by SOLIDWORKS or even reverse. Then, I guess, Pascal Daloz, competitive dynamics. There have been some changes announced this quarter, and I assume you see more opportunity than threat in that. But can you talk about your game plan, how to capitalize on that? Thank you.
Thank you, Michael. I will take the first one. So for mainstream, SOLIDWORKS performance is very consistent throughout the year. SOLIDWORKS is growing above group level, 6%+. It varies from quarter to quarter, but continuously performing at 6%+, with also expansion in double-digit growth in terms of users that we are winning. In terms of business model transition for SOLIDWORKS to finish the new clients is about one sort, which is cloud. That is well underway, and it is progressing and increasing. When you look at the mainstream overall between SOLIDWORKS and CENTRIC growth was very strong in the first six months. We said that it will start to slow down a little bit in the H2. But Q3 was again a very good quarter. We indicated around 15% growth, it was better than that. And we have also good visibility and pipeline for the fourth quarter.
Overall, CENTRIC is growing at a higher rate than SOLIDWORKS in the mainstream segment. And mainstream overall consolidation, which is happening in the industrial software, is a very interesting one, by the way. And there are few takeaway. The proposal to acquire PTC by Schneider Electric had to be put in front of for Siemens, because at the end, that what they try to do is to link the IT space with the IT space. Right. It has been the strategy followed by Siemens at the time when they acquired UGS. It was also the— Why it is a different game we are following.
We do not want to create a strong connections between the world of the controllers of the automatisms with the world of the PLM. We want to create a tighter integrations between the world of creations, the world of the life cycle management with AI and intelligence.
When you think about it, we are only a two pure play now on the market. Right. Remember a few years ago, we were a significant number of competitors having this agility and flexibility to bring these innovations. Now if you look at many of them are part of a big industrial conglomerates or others are part now and deeply linked with the EDA space for the specific for the simulation. So what does it means? It means what I started to share with you is how we are building the 3DEXPERIENCE platform to become agentic. I think having PTC acquired by [audio distortion] have been much more a constraint or a problem for us. So that is the reason why, again, at a strategic level, my view, this is cleaning the landscape.
This is giving us the agility like a pure play has, and it is giving us the unique positioning we are looking for. There are also direct consequences on this. As you may know, Schneider Electric is strong on the process industry, more than on discrete manufacturing, where PTC has a footprint. So this is obviously opening the competitive position, and you can count on us to put something in front, the medical device, for example, or the fashion industry. Remember, PTC is one of the competitor with Windchill, against CENTRIC, and this is opening this market. Last but not least, PTC was the last U.S. player. There are certain sensitive domain in the U.S. where being U.S. native
or European native, you do not have the same access. My view, this is also reopening some doors for us. More to say, Michael.
Thanks. A question from Fred Boulan, Bank of America. Commercial and pricing strategy. I know it is early days, but it would be great to understand any kind of KPIs or insights you can share on traction and to what degree we can start to see contribution to revenue next year. Second, from a cost perspective, how should we think about AI investments, in terms of R&D, et cetera, OPEX, CapEx versus AI efficiencies that you can actually leverage? Thank you.
Thank you. Thank you, Fred. I will start then.
Sure.
On the AI pricing and monetization, how do we track that? Right now we are more in a project mode, an enter mode. It is important to gain footprint and momentum to show the progress that can be achieved. We already said at our last Capital Markets Day that we have a financial objective for revenue that we will be generating with our Gen7 portfolios, 3D UNIV+RSES. 2027 would be the first year, because at the end, what we are achieving is client commitments. Those client commitments will be around our 3DEXPERIENCE agentic platform. The 3DEXPERIENCE agentic platform continues to be user-based. It can be humans, it is the agent, it is a user-based platform that offers skills and domains and access to knowledge. Agents, we know, are much more efficient in consuming and materializing knowledge and know-how.
The usage on our platform will be much, much larger than what a usage typically is for humans. That part will be measured based on the units of work, the tokens consumption with Virtual Companions and generated experiences. Those two dimensions continue to expand on the user-based agentic platform and the agentic capability to consume all the knowledge and know-how and the skills and domains. For that, we already released the Virtual Companion competencies that represent about 50% of our portfolio, and we are advancing very fast to complete that to north of 80%. That is all full in motion, and as you see evidence with the 100 projects we have. I think from that perspective, we are ready to execute. Should I cover the cost part?
Yeah.
From a cost standpoint, I think the best way to. You can look at this in terms of CapEx spend, of course, agentics. When you look at our year-to-date performance, we spent roughly EUR 100 million, a little bit more on CapEx, so not materially different than last year. There is no exposure in 2026. However, I expect that in our R&D organization, we have a model today in place that allows us to trade off token consumption AI across our expense categories. Our current OPEX is growing less than 3% on a normalized basis. There could be some volatility quarter to quarter, but overall, I think we are very well set around 3%. That gives us room for margin expansion.
At the same point in time, room to continue to invest, because that is very important to ensure that this growth wave that we now have in front of us, we can execute, and we do not have to catch up. That is the mindset we have. I think we are good positioned.
Maybe I could add a few things. We have deployed the agentic approach across the development teams. We have deployed Cursor, and all the 12,000 engineers we have within the company, they are using Cursor extensively. How do we see the benefit of this? Of course, you have all the productivity KPIs you can imagine, the time to write the codes, the time to do the tests. But if you want to modelize it, I think the best way is to look at the increase of the head count. 500 people. This year, we divided by two. So growth is around 2,000 people and net slightly below 1,000 people. That is really how we are basically taking the benefit of it. We do more with a little bit less people we need to hire in order to fulfill the capacity.
Hi, it is Charlie here from Jefferies. Just two questions from me as well. Firstly, can I touch on the guidance? In Q3, you outperformed the midpoint of expectations by about a percentage point, and that is roughly EUR 15 million of revenue. You will get an incremental EUR 40 million of revenue from ArisGlobal in Q4. Taking those together, that is about EUR 55 million of incremental revenue. About a EUR 25 million downgrade to Q4.
Does that mean that some deals moved from Q4 into Q3, or is that just an abundance of caution to make sure you continue to deliver on Q4? Secondly, a slightly technical question, actually. You are reporting ARR growth of 6%, but recurring revenue growth of 3%. What are the technical levers that account for that disconnect? I would have thought recurring revenue growth would be closer to ARR growth. How should we think about that 3 percentage point gap as we move into 2027? Thank you.
Okay. First of all, let me give you a clearer picture on the guidance. Yes, two things I can confirm, a few things are not correct what you said. We overperformed Q3 by roughly EUR 15 million. So I can confirm that. The number that you have in mind for ArisGlobal is not correct. It is more like 2 points of growth in Q4. It is not EUR 40 million, it is less than that. We talk about EUR 30 million to EUR 35 million of contribution to the fourth quarter. If you put that together, we are raising our midpoint by about EUR 73 million. The remaining data is pretty much currency effect. That is the first part. I want to be very clear. The Q1, Q2, Q3 performance puts us clearly on track to hit our midpoint.
Plus the ArisGlobal contribution in the fourth quarter, which is in the range of EUR 30 million-EUR 35 million, brings us the visibility that we now need in order to reduce the range from EUR 120 million to EUR 60 million, raise the low and increase the high, and raise the midpoint by EUR 70 million. It is really a mix of ArisGlobal and FX adjustment. But it is not a downgrade of our organic perspective. That is not the track record we have set for the first nine months of the year and the visibility we have for the fourth quarter. I think you hear our confidence and our conviction on this. Now let us move to ARR.
Can I just. I thought you said that ArisGlobal was going to do EUR 160 million.
I said 2 points of growth to Q4. Okay?
But on the last call, I thought you said it was going to do about EUR 160 million of revenue for the year.
I said $175 million. That was the number.
EUR 160 million.
It's not EUR 160 million. It's based on the currency that we assume in our guidance. You are putting all the numbers, I need to correct them. Let's go to ARR. For ARR, you're right. The current revenue growth has 3%, versus 6% rate. When you look at the definition of ARR, what's important to understand it is less so in maintenance, but really in subscription revenue. We have always some seasonality because of timing of renewal, as much as subscriptions have an upfront part in our revenue recognition. That can skew the subscription number, and that creates some volatility in year-over-year comparison. That is not in our ARR reflected because the ARR always takes a 12-month run rate based on the order entry. All the orders that we have received on a 12-month value going forward. It's a forward-looking metric.
That has been really consistent now for a few quarters now with 6%. What's really happening inside the ARR is that the share of subscription is increasing steadily at 1 point per quarter since Q3 2025. Whilst the maintenance portion is declining, and also because the maintenance growth is coming down from around 3% to 0%. This is accelerating, of course, the share gain of subscription overall as this mix is shifting. Again, I think just to summarize, the revenue number is a backward-looking number, where the maintenance is pretty evenly recognized as the ARR looks forward. While subscription has more volatility, and now with the ARR, we really straight line that. We have a clearly 12 months run rate of mid-teens growth on subscription. That is the difference.
There are differences in the way it is calculated, and we think we are very explicit about that. Of course, we expect as a share of subscription is increasing, that the recurring revenue recognized will approximate more towards the ARR while the ARR is accelerating.
Thank you.
Balajee Tirupati from Citi. Two questions from my side, if I may.
Yes.
Firstly, on competitive dynamics, how do you see the evolution of AI labs in terms of their model capabilities, including in design as well as acquisitions to extend their capabilities? For example, the Orchestra by acquisition by OpenAI. It would appear the vision around orchestration of being becoming orchestrator of different workflow within life sciences would somewhat overlap with your own vision of becoming operating layer for life sciences. The second question, version getting ahead of the previous target of it being in line with 2024 levels.
Looking ahead, do you see the strength as sustainable, or there were some one-off tailwinds in the quarter? Thank you.
Thank you for the question. The best way to answer to it is orchestration is not science. You need both. Why this is important? Because orchestrating the usage and the different apps in a coordinated workflow is necessary since when you want to improve the productivity of the existing things. If you want to create new things, you need to combine the two. In a way, this acquisition, my view, is much more a topic for Veeva than for us, because they are the workflows. We are the science. That's the biggest difference for me. What does it mean, nevertheless, is if you could combine the two together, you are creating a unique position, and that's what we do. That's really what we do. Do not underestimate that when you do an agentic approach, you have several things. Obviously, you have the harnessing capability.
The harnessing is really how you connect the different things together. You also need the reasoning. The reason why we have different companions is because all of them have a different reasoning. An engineer does not have the same reasoning as a scientist, does not have the same reasoning as a doctor. That's where we are making a difference. The second thing, you need the knowledge base, and the knowledge base is the virtual twin. If you want to give AI the context of everything you do, you need the virtual twin. Remember my story with Tesla driving the virtual car in the virtual city, but basically using the real algorithms, which is now embarked into the car. That's the virtual twins, the knowledge base. Then after, you need the application. I think we have the application.
We have a pretty extensive set of applications to cover the entire life cycle. Last but not least, you need the models, whatever the frontier model or the open weight model. Here is an interesting thought. I do not know if you have seen, by the way, Jensen Huang last week. He made a very comprehensive explanation about this. You could use both. By the way, the world is going to consumptions and the quality of the results. That's point number one. Point number two, if you have an open weight model and you have the datasets, you could exceed the performance of the frontier model. I repeat, if you have the open weight model, like Nemotron or others, in conjunction with a proper dataset to train them, you could exceed what a frontier model can do.
This is our strategy, and this is how we are differentiating ourselves. On the quality of the harnessing, the quality of the reasonings, the knowledge base with the twins, the deep set of applications, and how to fine-tune the open weight model to be basically extremely precise on certain use cases. I think with this, we have enough to differentiate ourselves.
Good. To your second part on the cash. First of all, we are very pleased with the progress of our performance on operating cash flow and cash conversion. Our target was in 2026 to be back at the level of 2024 after some headwinds we were facing last year. Now we are ahead of this target. Take a look into the fourth quarter. There are a few puts and takes for the fourth quarter cash flow. But right now I am looking at a performance that is similar to 2025. If you do that, we are well ahead of the target that we have set with 84%. That gives us a new starting point entering into 2027.
Clearly here we will be. Right now I see us more in that range of 85%-90% of cash conversion, and we will move further upwards. That is the simple answer.
We will take one more question online and then we will be good.
Thank you. To ask a question, you will need to press star one and one on your telephone and wait for your name to be announced. To withdraw your question, please press star one and one again. Our last question for today, just one moment, comes from Laurent Daure from Kepler Cheuvreux. Please go ahead.
Yeah. Good morning, gentlemen. Thank you for taking my questions. I have two, in fact. The first is on MEDIDATA and the return to growth we are expecting for, especially at renewals. So I was wondering, on the side of the massive bookings you had recently, all those headwinds are going to move in the coming quarters and whether for 2027 we will have less headwind offsetting the new winds. My second question is back to your guidance for the rest of the year. You have been improving the growth rate by one point every quarter, but when I take the midpoint of your fourth quarter guidance on slightly easier comps, it is a bit softer than Q3. Is there anything specific we should be aware of or any challenges you are going to face in the first quarter, or just a bit of cautiousness? Thank you.
Maybe I start with the second question. Thank you, Laurent. Look, the guidance is always a range. You can take the midpoint. Of course, you are right, and that gives us confidence. The performance year to date, we have delivered midpoint or better than midpoint. On an organic basis, the 3%-6% puts us squarely at the midpoint of our guidance that we have set at the beginning of the year. That has been our objective. There is a range, right? There is an opportunity to do better than that, but I think the midpoint is very secure, and we have clear visibility towards that. Do you take the first question, or should I continue?
Please take it.
Too many data. That also still impacts our 2026 performance. As we are entering into 2027, we have a much cleaner slate of growth where those headwinds of prior renewals are more in our base than they were in 2026. More importantly, it is to highlight that the growth bookings performance has been very consistent year to date to build our order entry and starting point for 2027. Very different than when we were in 2026 with growth. As I mentioned before, Q2 was -4% for life sciences. We are now +1%, so we are improving by 5 points. For MEDIDATA, it is about 3 points of improvement, plus. Next quarter, we are expecting an improvement sequentially of about 2 points. Then we enter into 2027, where we should expect further improvement in terms of growth.
That is the trajectory, and that will also be the guidance philosophy for 2027, will be very much based on the order entry and our backlog that we already have in place to make sure it is a baseline forecast and we can build growth on top.
Okay. Very clear. Thank you.
Thank you, Laurent.
Thank you. That is all the time we have for questions today. I will hand the call back to the speakers for closing remarks.
So thank you again for all the questions. To wrap up, let me leave you with three key messages. The first one, I think we are delivering, and we consider we had a strong Q3. We are almost to continue to expand and use our 3DEXPERIENCE and the cloud. The numbers speak by themselves. The second thing, I think, bringing AI into the real world with 3D UNIV+RSES, our agentic platform, the Industry World Model, the virtual twins. I think we are helping the industry to put AI in work, in the real work environments. This is, I think, the most important thing, because all of our customers are becoming AI native. The third point, I think we are making important steps and progress in Life Sciences. ArisGlobal joining us.
Dr. Zenooz joining us as well to take the leadership and to foster the ambitions to create. These priorities are closely connected. They are definitively helping us to build a stronger customer relationship, focus on our investment, and create value on the long term. That is why we are building together with our customers and employees. I think this is for the benefit also of you as shareholders. So we look forward to sharing more again with you at our Capital Markets Day in Paris, November 17. Remember the date. Hopefully, we will have the pleasure to see all of you. Thank you, and hope to see you on the road and no later than the Capital Markets Day. Goodbye. Thank you.