Atos Origin Sa

Good day and thank you for standing by. Welcome to the Attoff Group FY 2025 results conference call. At this time, all participants are in a listener-only mode. After the speaker's presentation, there will be a question and answer session. To ask a question during the session, you will need to press star 1 and 1 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press star 1 and 1 again. Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Philippe Salle, Group Chairman and CEO. Please go ahead.

Hello. Good morning, everybody. Thank you for joining us for this call of the four-year results of 2025. I'm here in the room with Jacques François, the CFO. And Florine, our CTO, because we're going to talk also a lot about technology. And of course, we're going to talk about the future of the company. So the agenda of today is four topics. The first one is 2025 Excel Business and Strategic Atlas that I will manage. Then Florine will take the floor to have a tech update. And today we are launching two things, with the launch of Atos Amplified and the launch of our Agile Peak Studios. Then I will come back on operational and financial results with Jacques Francois. We're going to have together this section. And then I will finish by the outlook, and then we'll take Q&A. So let's start with the first part, and I'm going to go on page 6. So 2025 was the year for me of the reset. Remember that I want to have three phases, I would say, in the turnaround of the company. Reset, rebound, acceleration. So 2025 is the reset and 2026 is the rebound. In 2025, first, we have a very good financial improvement with clear signs of recovery, we will see that. Second, a significant progress, of course, of our Genesys plan. And the third is that we have a positive business momentum and a commercial, I would say, traction. If we go on page seven, the key numbers of the company, first in terms of top line, the revenues at 8 billion, 8-0-0-1 to be really clear. So above the target that we have set during the Q3, the last call, in fact. Operating margin, 351 million. It's 4.4% just for information. That's the best margin we have for the last five years. And I think I'm very pleased to say that we have doubled the margin versus last year, with a decreasing top line of minus 14%. And remember that we have gathered $340 million at the beginning of 2025. Next change in cash is minus $326 million, although we have accelerated, I would say, Genesys. And we paid, in fact, more than roughly $15 million of exit costs. And it's better, of course, than our guidelines that say that it will be 350 or below. And then the liquidity, and we published this already in Jan, is 1.7 billion, which is far above, of course, I would say, the covenant that we have in our debt package. That is 650 million. So we have ample cash to finish the genetic plan. And, in fact, this year we'll be already cash flow positive and the debt, in fact, will go down. Now, if we go on page 8, you can see the inflection point in terms of the revenues. So, the fourth quarter, and that's the figure we have published, in fact, in January, was around minus 9%. So, you can see, I would say, quarter by quarter, that we have, in fact, a deceleration, or let's say, a less momentum, I would say, a better momentum in terms of revenue decrease. And then you can see also the number between and the organic growth that we have, which is around minus 9% in Q4. In terms of the OM, so the EV, the pro forma of 24, for your information, we thought we agreed that we are sold in 24, and that the constant exchange rate is 172. So you can see that we have more than double the margin, and the margin was beyond 2% in 24, and in terms of 25, it's at 4.4%. And in terms of cash flow, the net change in cash was roughly minus 700 million in 24 and we have minus 300 million in 25. If we go to page 9, you can see also that the backlog for the book-to-bill, excuse me, has also improved in the course of 25. That's 94% for H2. And the total book-to-bill, in fact, for 25 was 89 versus 82 in 24. Now if we go to Genesis on page 10, remember exactly the plan that we have sketched in May 25 with the seven pillars. uh i just decided i would say exactly what we have shown in fact in may and i will now if i go on page 11 a little bit deep dive on what we have done so on the first pillar we have reviewed the top 100 accounts and it has a i would say produced one billion plus of opportunities remember that during the cmd i have said that the number of business line per The cost was around 1.4 and we want to push it, of course, to two or above, I would say, that level. And then we have also, in terms of both the streamline, I would say, the processes, and also with, I would say, a better organization in terms of sales, with salespeople in the different GOs and EVDs. In terms of HR, so we have reviewed the bonus framework. We have launched our ATI plan, which is a share plan for the top 200. and we have started also to have a leadership culture. And this year we're going to push very hard on the AI culture. In terms of company reviews, so we have roughly 10 countries that are exited or inactive. Remember that we want to go probably above around, let's say, 40 countries. And we have also sold the seven countries in Latin America and also in Nordics, which is now in Finland. And this year we want to continue to probably close or make it active around 20 countries. In terms of portfolio review, first we have sketched, I would say, the different branding. So you have Atos Group, which is the holding, and under Atos Group you have three brands now. Atos, of course, the service LSA company. Eviden product and software, and Atos Amplify, the new name that we are launching today for the consulting arm. So we have six, in Atos, six business lines and six NGOs. And with Eviden, after the disposal of BUL, the high computing, we have three product lines. In terms of PM and GM, so in terms of project margin and gross margin, so that's the way we look at our P&L internally, your revenues, project margin, gross margin, and then EBIT margin. So we have, I would say, a plan. Remember that we are looking at 650 million of savings, and we have already achieved 80% of it. 350 will be millions. We have increased the billability ratio by three times. We are now above 80%. We have also continued, I would say, to push the offshoring. And as you can see, we have also a different, I would say, action on the reduction of the surpluses, just switching, I would say, bench people. I would prefer, of course, to utilize people on the bench. Just for information, very important, the discipline also on the new contract we have signed. Remember that we want to have a margin of 25 to 26 in the future. And in fact, if you look at the black contracts, we are already at that level. And on average, last year, we have signed roughly all the contracts on the book today, so the 8 billion plus and 7 billion plus today, is around 24%, which is roughly 2.5% above what we have done in the course of 24. And it's very important to understand that we could have probably a better book to read in the course of 2025, but the idea was really to protect the margin, and I prefer to say no, and for some tender, then I would say to have, let's say, more revenues and less margin in the future. Pillar 6 is the cost review, is the GNA, so we have done a lot of things there, and for information, we reduced the GNA by roughly 26%. Remember that the target we have in GNA is 5%. We are close to six, so we still have one point to gain in the course of this year and next year. And then in terms of cash, for the pillar number seven, the DSO is at target. We have reduced solar use by 13%. We have reduced the width, in fact, for the beam by roughly 27%. And we have managed quite very well, I would say, the capex. And without blue, in fact, we are at roughly at 100 million plus. Just on the bottom, you see also that we have a complete review of the target operative model on the top, and also the government has been simplified. Now, if we go on page 12, today we are launching, as I said, two things. I'm going to talk on Amplify, and then we're going to talk on the Adjantic Studio with Florine. So Amplify is the consulting arm of Bodhi, I would say, of Fatos. It's still the brand of Fatos because it's very linked to the services that we provide with Fatos. And the idea is that we're going to refocus Amplify on AI. The idea for us, it's a door opener, in fact, in artificial intelligence that will help us after that, of course, to push the studio that we're going to. In terms of workforce, on page 13, we are now at 63%. and also the restructuring we have done. And if you want to see without Latin America and without DUL, we are close to 57,000 people. That's the number of staff that we're going to have after the diagnosis. Last, on page 14, that's the order book. So we have the key numbers. As you can see, the renewal rate, in fact, is 92%. It's not bad. I want to have a bit more this year. And in fact, in this year, for the large accounts, what we call the large bids, over 30 million per year, we estimate that we're going to win most of them, or all of them, in fact. The number of strategic bids is 19, and it was 10, in fact, in 2024. And there's a good fraction, in fact, in cloud, cyber, and data AI, where at the business end, we are pushing more than the rest. You can see, I would say, different names on the, I would say, expansion or win. with them, we're going to do probably 250 million in the course of 26. Last, on page 15, just to also recognize that we also continue, I would say, to be recognized as a sustainability index. So we have one of you that we test on the in terms of stability. You can see this on the left. And then many are business aware from the analyst that we continue to have in the course of 25. I go quite fast, in fact, because I think it's very important that we spend some time on the technology today, because there have been a lot of buzz on AI and probably a completely crazy movement for me on the share price in the different companies, on that also. And we estimate, in fact, that for us, we are very well placed in AI. And, in fact, we don't do EPO. That is probably, I would say, the business that is going to be attacked for me by agentics. And definitely, I think that there is a big opportunity for us, in fact, with the AI going forward. So with this, I give the floor to Florin, who is in the room with us, and he's going to talk about you, about, I would say, the AI and the agenda we're going to launch today.

Thank you so much, Philippe. Good morning, everybody. Thank you for joining us. Delighted to be here. So what I suggest we do for the next few minutes is for me to walk you through the way that we see the market developing in general in the space of technology. And I will double click on AI, of course. And I will share with you how we are planning on attacking and delivering this very exciting space. So I'm sure you're well familiar with the fact that global AI spending is booming. a lot of increasing AI infrastructure, AI services, AI software, AI cyber, and we genuinely and truly believe that Atos is very well placed to win in all of those areas. We do have some very strong moats. So the background and the history of Atos, as you all know, is to work and help our clients in highly regulated environment where security is a top concern, where sovereignty is increasingly becoming a priority, where the IT landscape is very complex and where a lot of the systems are truly life and death and are genuinely mission critical. And what we see happening is that in all of those environments, there is a flying wheel convergence happening between sovereignty, AI, and cyber. And the fact that we have this decades-long managed services relationships with the clients, the fact that we have really deep know-how of their environments, of their data, has truly enabled us to progress very fast into packaging those into agentic AI as a service offerings. And I'll cover this in more detail. As you know, the technology space is quite complex right now. It's evolving super fast. There are daily announcements from left, right, and center. And our clients are really hungry for a level of clarity and assurance. And I think we play a very important role as, if I'm allowed to use the word, Switzerland of governance. The ones which are able to provide secure, cross-platform, neutral, agentic AI. And we're doing this through a very exciting set of partnerships with the big players, but also through a set of unique partnerships with AI-native and sovereign startups. So let me double-click on all of this into more detail. So if we go on the next slide, what you'll see is the three big bets for Atos going forward. We believe these three pillars are going to be substantial drivers of growth for us in 2026 and beyond. The first one is mission-critical agentic AI. This type of agentic AI is fairly different to the type of AI that is most commonly mentioned in media. When you're doing agentic AI in really complex regulated environments with high level of governance, requirements around sovereignty, reliability, security, and responsibility, the type of technology and the type of services is fairly different. We're also seeing digital sovereignty be super important for our clients, and actually this is the case in North America, in Europe, and international markets as well. And what we're doing is that we have embedded digital sovereignty as a core design principle across all of our portfolios. And last but not least, cybersecurity, of course, continues to be a high area of focus. Development in AI are, of course, helping us to deliver cybersecurity services in a better, faster, more efficient way. But AI actually, of course, also opens up new attack surfaces and new potential vulnerabilities. So what we see happening is that there is this flywheel of self-reinforcement powers between AI, sovereignty, and cyber, and we believe we have the right to be winners in all of these three areas. Sovereignty, of course, requires security controls to be able to be fully enabled. Sovereignty is, by definition, more complex than non-sovereign solutions, and therefore AI can play a role to make them more affordable and more innovative. As I mentioned, AI has a huge impact on security. There is a quest to secure AI, but also to use AI to drive more security solutions. So you will hear us going forward really focusing and doubling down on these three areas. And what I would like to do is to double-click into each and every single one of those to give you a flavor of what we're doing, the success we've seen so far, what we see happening in the marketplace, and to give you a glimpse into the future. So if we go to the next slide, slide 19, We are very excited to have four Atos sovereign agentic studios come out of stealth mode in UK, in US, France and Germany. This will serve the local markets based on their needs, the focus industries, their requirements and they're all built for truly mission critical production from day one with extraordinarily high focus on the topics which make AI adoption at scale more difficult in most organizations, which is governance, sovereignty, reliability, security, and responsibility. So the reason we're launching this studio is that we see that our client spend is converging services and technology budgets into a unified value pool this value pool and the size of those budgets are increasing, of course. But there's also a very high demand for measurable value generation at scale. Everybody is sick and tired of pilots and proof of concepts and prototypes. Organizations really want to make sure that they have AI which is secure, which is reliable, which adds business value at scale. And they The main challenges in this space is governance and orchestration. And this is where we believe that Atos is an absolute key power player. And then we're actually also seeing sovereignty emerge for AI as a very high priority. So our clients are very happy to use closed black box models for governance the typical back office functions, which are important, but which are not differentiated. But they actually are increasingly becoming wary of developing their own brains, so to say. So they own their future, and they have full control of their data, the controls, and the intelligence which they're building. So we are very excited to announce a unique partnership with one of the absolute leaders in foundational models for a Gemtech enterprise, which is Poolside. And this will allow us to deploy and develop sovereign solutions in Europe, in North America, in international markets. And this will help and is helping already our clients to harness the full power of AI on their terms and without compromise. We're, of course, also working with the major market leaders here, such as Google Cloud, SAP, IBM, AWS, Microsoft. As a little anecdote, we've just received frontier partner status with Microsoft, given the fact that we were one of the leading organizations leading the path around AI and innovation. But we're also working with this really interesting set of AI-native startups, which allow us to add unique value across the entire value chain of AI. So we're using KYT, which stands for Know Your Potential, to help our clients mine and redesign processes. and to really understand the business case and the value generated with AI on a very specific and data-driven manner. We're working with the likes of Emma and M8N to create and orchestrate and manage the digital AI employees, the agents. We're working with PayEye to really be able to measure and value the highly how should I put it, movable cost of AI consumption and to have the causality and the correlation to value. And we're working with clarity around helping our clients drive this continuous change. And we're using all of this technology for our own back office and front office transformation as well. So I know I've used a lot of words here, a lot of concepts, But let me move to the next stage and try to make this very real for you with a number of client examples. So to be honest, we have more demand than we can almost handle right now. We have incredible interest in this agentic AI studios and just sharing with you a couple of examples here. One is Scottish Water, where we're working together to really transform the way they are doing operational planning, risk assessment, decision-making across the entire national and wastewater networks. And this is really mission-critical environments where AI agents are used to continuously monitor the network, to analyze proposed changes, to automatically generate contextual risk assessment. And as you can imagine, this is the type of AI which really needs to work, which really needs to be secure, which really needs to be accurate and timely. Another example is DEFRA, the UK Department for Environment, Food and Rural Affairs. Their mission is to make the air purer, the water cleaner, the land greener, and food more sustainable. So obviously a very important mission and vision. So what we're doing with Atos is that we're using a new set of highly differentiated agenda solutions we have developed to rapidly modernize and transform their entire application portfolio. We call those digital transformation engineers. They're AI agents which work in collaboration with our human experts to achieve things which, frankly, wouldn't have been possible to achieve just a few months ago. So we're seeing a close to 30% time-to-market efficiency gain around how to modernize those mission-critical applications. Another example is mBank. We are really working very closely with them to develop their entire advanced digital foundation. And again, this is not AI, which is an add-on. It is mission-critical AI, which is being used to improve operational resilience, to create real efficiency in their business, to manage risk, and to really make a difference around their customer experience. And there are many, many, many more examples of this. So we're very proud about this Sovereign Agenda KI Studios. Much more to come in this space going forward. If I go to the next slide, I'd like to share with you the perspective around how we're approaching the Sovereign space. So what we see is that clients they have an increasing desire to retain control, authority, and accountability over their data, their infrastructure, their applications, and digital operations, and to have distinct compliance with all the applicable regulations to minimize dependency, exposure, and disruption risk. I think it's important to note that this is not just a European development. We see sovereign requirements being very high in North America as well, both in the United States and in Canada, and also in our international markets. And in actual fact, there are data points which point to the fact that over 80% of requirements from clients going forward are going to include a critical demand for sovereignty. And this is a massive business opportunity. It's currently estimated to be in the 40 to 50 billion euro of total addressable markets, and it is growing quite fast. And frankly, we believe that we are one of, if not the best player in this space. I'd like to draw your attention to the quote on the bottom right corner from one of the leading independent analysts, which is basically, and I'm quoting, few players can claim the unique combination offered by the Atlas Group, an umbrella of sovereignty, which provides the whole with an unprecedented career. So we are able to do this in a variety of models because sovereignty takes different shapes and forms in different countries. What UK means by sovereignty is slightly different than what France and Germany means by sovereignty, which is different than what US means by sovereignty and so forth. So we're able to offer this full spectrum of solutions ranging from enhanced native clouds, to controlled clouds, to trusted clouds, to disconnected clouds, to fully sovereign AI, as I mentioned previously. So I would like to give you, again, a little example of this. One of them is Eurocontrol. Eurocontrol, you might be familiar with. They are the organization which manage and control the European skies, they are providing a really mission-critical service to the entire continent. If their systems and operations wouldn't work and flights would not fly, that would obviously have a very, very high impact on the entire economy. So what Atos is doing is that we're one of their leading partners to ensure the strict resiliency safety, security, compliance requirements around the entire IT value chain. And we do this in a way which is coherent, is aligned with the industry regulations, and generally spans infrastructure, application, artificial intelligence, and so forth. And this is a solution where we are partnering with Microsoft as well around the Azure cloud. If we move on to cyber, that is of course a very hot topic, and it continues to be so. And what we're seeing is that AI security has really changed the game. So it's become the primary focus area for the way our clients spend. AI is truly redefining threats, defenses, and vastly expands the attack surface and cybersecurity is shifting to an always-on compliance model where our clients are required requiring very much verifiable controls and sovereignty aware architectures and again this is a space which is moving super fast and really redefining the game. So to give you a little anecdote, it is estimated that there are 80 times more machine identities in any organization today compared with human identities. And this is, of course, because of the advent of agentic AI. So that type of AI where you have agents which perhaps only need to have split-second life cycles, they need to be controlled, they need to have verifiable access controls, they need to be spanned up and potentially terminated in under a second, really redefines the rules of the game. So we believe that we are super well positioned in this space. We have very much an end-to-end best-in-class set of cybersecurity services, ranging from advisory, which Philippe just mentioned. We have very much embedded AI agents in our entire lifecycle of threat intelligence, threat detection, investigation, and response. We're also, we believe, one of the market leaders in post-quantum cryptography. And of course, with the Evident Group, we have some fantastic EU European sovereign cybersecurity products. So again, to make this real, I'd like to give you a sense of the work that we're doing with the European Commission. This is one of the most important cybersecurity services in Europe, full stop. And ATOS is on point and has won a substantial framework agreement to provide operations, incident response, digital forensics, threat intelligence, threat monitoring, Offensive security in the areas of vulnerability management, penetration testing, and red teaming. And again, I draw your attention to a number of independent analysts which continue to recognize us as a market leader in the cybersecurity space. So let's move on. to the next slide and try to give you a big picture of where we're at and how we see AI impacting our businesses. So this might be a little bit of a busy slide, so please give me a chance to walk you through it. So at the bottom of the slide, you're seeing our different historical business lines with data and AI, cyber, evidence, and so forth. And the Harvey Balls are representing the way we see AI impacting those specific business areas. So on the top row, you're seeing how AI is impacting the addressable market expansion with the full Harvey Ball, meaning it is very high expansion, more opportunities for us. or a limited partial harvible demonstrating or indicating a limited expansion. And then the AI top-line pressure row is basically a way of indicating how we see AI impacting or having the potential to impact our top-line revenue. ranging from low to high. So all in all, all in all, we see AI being a strong driver for growth in Aptos. We are very much on the offensive. We believe that AI is a game changer, and we are super well positioned in this space. But the important bit to mention here is that we have a leading position in a number of these building blocks in the colorful table below that what we are doing very successfully is to combine and recombine them into these three big bets that I've been talking to you for the last few minutes. So again, please remember the growth engines of Atos are agentic AI, digital sovereignty and cybersecurity. And we're seeing substantial opportunities and a lot of momentum in those areas. And we truly believe we have the right to win. So moving on to the last slide, as a little bit of summary, we are still going through a massive transformation. We've turned the corner. We are reimagining and we have reimagined the entire technology function in Atos. We're attracting some absolute top-notch talent. We have done a full portfolio redesign, doubling down on agentic AI and AI in general, digital sovereignty, and cyber. we have a very unique and differentiated approach to sovereignty and security we're boldly and ambitiously embracing this new world of services software where increasingly we are building very unique very specialized ai solutions powered by software to augment and enhance our services and the agentic sovereign studios which we have just launched are really a showcase of much more to come. We really look forward to sharing with you progress and a lot of success in this space. So having said that, handing over to my colleague, Jacques François, to walk you through some interesting numbers.

Thank you very much.

So thank you, Florine. I will take the lead. And in fact, Florine, you're right. We're going to have a special press release on the Atlantic next week on the 11th. We're going to come much more in detail on what exactly we're going to do in the coming weeks and months, of course. So now going back on the topic number three on the presentation. So we go to page 26. So you can see the revenues of 25 versus last year. We produce also the pro forma. We have the foreign exchange and scope. Scope is world grade, of course, in 24. And as you see, minus 14% in terms of sales. If we go to page 27, you have the 8 billion between AEDM and also ATOS in blue. And then in the different countries, Germany is number one. And if you look on the right, this is the 7.2 billion, that's the pro forma of 25 without Latin America and without bull. And you can see that the base, we're going to rebound for this year. And you see now, I would say, what is the split of revenues between S&P 500 and S&P 500 of course much smaller on the 12 million plus and I would say at those with the different geodes. Now if we go to page 28, I'm very proud to say that we have doubled the margin in terms of edicts and in terms of percentage more than that. So pro forma in 24 we have 172 million of edicts and last year we touched 351, so it's more than doubling. margin at 4.4 percent and as i said that's the biggest margin we have since 2021. now if you look at the operating margin by geography on page 29 i will not go into details but you can see on the left column that's the the results of 25 and on the right that's the pro forma without bull and without latin america so that's the rebound for the 7.2 billion and 314 That's the base, in fact, of the rebound for 26 years. Now, I will go very quickly on the different business units. But you can see that in Atos, for the six NGOs, we have done quite a very good job. Germany, we start first, minus 10% on the top line. So, first year, we know that some of the clients, for example, Deutsche Bank, have decided to exit. For example, for Deutsche Bank, it was a re-internization of their platform, so it was nothing to do with Atos. Germany is for the first time probably of many years on a positive territory. And as I said to you, this year we're probably close to $100 million. I think the budget is $90 million. So we have, I would say, with Genesys, more to come, of course, in the course of 20 years. I go to North America on page 31. That's the area that has been touched more, I would say, in terms of top line. A lot of clients. have been frightened in the course of 24, and they stopped, of course, some of the contracts. But as you can see, of course, the EBIT in terms of quantum is less than 24, but in terms of margin, we are double-digit, and I think it has been a very good job done by the U.S. team. Now, if we go to France, the decrease is around minus 10%. And also, I would say, however, we have a decrease in terms of supply, So it's a bit more in fact. And of course with Genedys there is more to come in the course of 26. Now UK and Ireland also is an area on page 33 where we have had also a big hit. It's like in the US in fact it has been a tough year because of a lot of clients stopping to work with us and stopping contracts. But as you can see, we have been flat in terms of EV, in terms of 83, that is 82, but in terms of margin, we have increased the margin by roughly 1.6%. International market is gone also by minus 15%, but we have more than doubled the profitability. We have done a very good job, in fact, in the Genesis transformation in different countries in the Middle East, in also South Europe, and also in Asia. And last, Benelux, where I would say probably we have been the more resilient in terms of And so we are on page 35, minus 4% in terms of organic, inorganic growth, so a decrease in terms of organic, let's say, and a very, very good job from the team on the bottom line. As you can see, we are multiplied by 10, which is easy, with a margin around 7%. So as you can see inside, despite, of course, the top line, I would say, pressure, we have been able, I would say, to manage very well at the bottom line. Last slide on 36 is from Evidem. Of course, this is the part that is growing and many of the advanced computing activity. This activity was using money in fact in 24 and we have done quite a good job to restore some profitability. It's still too low for me, but definitely there is more to come in this business year. With this, I hand over to Jacques-Francois to go more on the P&M and Valentin.

Okay, thank you, Philippe. Good morning, everybody. Now that Philippe has gone through the drivers of our business operational performance, let me walk you through the P&L items below operating, as well as the cash flow statement and the balance sheet. So, as Philippe indicated, our operating margin amounted to €351 million in fiscal year 2025. We incurred reorganization and rationalization charges. for €642 million in total, of which €540 million reorganization costs as we made significant progress in the execution of our restructuring program and €102 million provision related to leases and real estate asset impairment. We impaired €166 million of goodwill this year as a result of the upcoming disposal of the advanced computing business. Other items reached negative €331 million. They included losses related to some onerous contracts, €423 million, and litigation provisions, €145 million. The net cost of our debt reached €333 million, up from €178 million last year, reflecting our new debt structure, post-24 refinancing, and including fixed interest as well as the amortization of 2024 fair value adjustment. Other financial expenses were 102 million euros in fiscal year 25 due to debt, lease, pensions, and provisions on non-consolidated investments. As a result, our net income group share amounted to minus 1.4 billion euros. On the next page, we see the cash flow generation. which improved significantly year on year from minus €735 million in 24 to minus €326 million in fiscal year 25. We generated €883 million OMDR in fiscal year 25, and we expensed €170 million in CAPEX and €278 million in LEASES. Our change in working capital requirements once we neutralize for the working capital action, you recall that the unsolicited cash received in advance from some customers, this amounted to a positive 33 million euros. It essentially reflected a lower activity level in 2025. Going forward, we expect further sustainable working capital improvement. Our cash restructuring expense was 445 million euros. As expected, cash out accelerated in the second half of the year. Tax paid was 31 million euros and cash cost of debt 160 million euros. Other contracts and litigation amounted to 157 million euros. As a result, our net changing cash was limited to 326 million euros, better than anticipated, despite higher restructuring costs cash at 445 billion euros. Now the net debt as at December 31st, 2025. The net debt was 1.8 billion euros compared to 1.2 billion euros as at December 31st, 2024. beyond free cash flow. It reflected the impact of the change in working capital actions for 43 million euro, negative forex impact for 104 million euro and other elements such as the peak component of the debt. Net debt consisted firstly of cash and cash equivalents for 1.265 billion euro and secondly borrowings for a nominal value of 3.64 billion euro. As of December 31st, 2025, the group financial leverage ratio was very similar to the end of 24 levels at 3.17 times. I remind you that our target is to reduce leverage below 1.5 times at the end of the year 2028. Thank you, and I now hand over back to Philippe.

Thank you, Jacques-Francois. So, let's go to the section which is the outlook. So, on page 42 first, we want to come back on what is at host now, without bull, that we will give the keys at the end of the month, in fact, end of March, without also Latin America, that we have sold, and the closing is expected, in fact, in April, and also the small DHS share that we have done in the Nordic. So, on the left side, you can see that the revenue is 7.2 million. Operating margin is 314. And that's the pro forma, as I said, the new perimeter. Roughly 57, 58,000 people without bills in Latin America and 54 countries of operation. And as I said, we want to be below the 40 threshold, so we continue to reduce the perimeter in this topic. On the right, you can see the different business lines. The different geographies, number one market is North Germany, North America number two, France number three, and the UK and Ireland number four. And as you can see, these four countries are worth more than 70% of our total revenues. And then you can see also the industry. Now the financial ambition is on page 43. And I know that a lot of people are waiting this moment. Guidance for the three elements, which is top line, bottom line, and cache. On the top line, we are looking for a positive organic growth. That's the project that we have internally. But we want to say that there is also a downtime scenario possible that is limited to minus 5%. So it's very important that we are cautious. We don't want to over I would say confidence. It's very important that we deliver the numbers that we announce. And that's why we say that, of course, the budget is, and our target internally is to grow. It could, I would say, there is less good news in terms of those lines. The maximum we can see this year is minus 5. And remember, it's only at minus 14. So, of course, the first half year will be negative. We estimate that it will be probably around minus 9, minus 10% in Q1. And then it will, of course, stabilize in Q3 and rebound in Q3 or in Q4. Operating margin around 7%. So it means that it's indeed, let's say, around 500 million. So it's an increase by 50% versus 25%, which is very important. And we are on route, I would say, to the journey to touch the 10% margin by 28%. And the positive net change in cash, so it's without bull, without, I would say, the divestiture, of course, of bull. So it means that with the cash that we're going to produce this year, plus, of course, the cash we're going to have from the M&A, the debt will be reduced. The EBIT will increase, so the leverage, for sure, is going to decrease strongly, in fact, in the course of 26. And we are very, I would say, very confident that we're going to produce cash this year. accelerated in the course of 25. Now for 28, we continue to say that the three phases, as I said, reset in 25, rebound in 26, accelerate now in 27 and 28. We continue to see an acceleration of the 9 between 5 and 7. We tried probably to do better than that. Still looking at another acting machine around 10%. And, of course, to be below the 1.5 percent by the way we calculate this in the course of 28. So to have a profile of BBB and BBB probably in the course of 29, that's the goal we have. Now, if I have to sum up what we have said today with Florine and Jacques-Francois, so it's here on page 44. First, we have restored the foundations of PATOS. We are very pleased to say that we have met the financial guidelines that we have set. We have done a lot of job in fact in the commercial strategy and I definitely think it's going to yield a lot of results. In fact, I would say the Genesis cost, it's a one to two year effect. We have done most of the time in 25, we will finish in 26. And the rebound, it's a two, three year effort. We have done a lot of job in 25. We're going to see some of the Results in the course of 2016 are definitely things that we're going to accelerate in the course of 2017. As I said, the Genesys plan, we have done roughly 88% in terms of scaling. It's a pro forma, so we have, of course, partner rates in the P&L of 2025, and there is more to come, of course, in the P&L of 2016. Second, I think we are very well-placed for the AI journey, and I think Atos has a unique position. We're going to reinforce, as I said, the three tech pillars that Florine has set. So Agents with AI, the launch of the studios, Mortal Connectrix, Sovereignty, and Silt and Cyber. And remember also that we have launched also the consulting with Rebranded, I would say, the Atos Amplified. So today we are announcing the launch of Amplify and also the launch of the . And we have, in fact, a new website that you can see . And then we have quite a promising outlook on the right part of this page. So stabilization in 26 with a rebound in H2 and then acceleration of supply and, of course, production of a lot of cash in the core When we can probably review M&A, we'll see there are targets that are interesting, but it's also possible that we do probably less because we estimate that with Atlantis, we have a lot of opportunities we're going to have, and we probably will try also to invest in the company more in our studios. With this, I turn to the Q&A session that is open, and then I will give the floor to Florine or Jacques-Francois, depending on your questions.

Thank you. If you would like to ask a question, you will need to press star 1 and 1 on your telephone and wait for your name to be announced. And to withdraw your question, please press star 1 and 1 again. Thank you. The first question today is from Frederic Boulin from Bank of America. Please go ahead.

Hey, good morning. Two questions for me. Interesting discussion on your AI offering. We'd be keen to understand how you define your competitive edge versus your key global competitors and players. I know broadly looking at your targets, the kind of growth ambition. what kind of upside do you anticipate and have you penciled in on that kind of segment versus potential pressure on traditional, I mean, digital transformation, as you mentioned on that slide? And maybe as a second question, is there any, you know, any kind of areas of your business where you do see kind of, you know, margins going down on new projects. I mean, you mentioned some of competitive bid where you walked away. But where you do see already today, you know, Gen AI driving some price deflation. Thank you.

So in fact, Frédéric, you have to understand, I think the slide of Florine, which I think is not the most important, but I would say the page number 23, the way we look at it is very simple. In fact, AI is going to touch the company in two types of impacts. There is an impact on the coding, so the digital application, where we definitely think we're going to go faster. and cheaper. And that's why we say there is an upward to negative impact. But here, in fact, what we see is that it's not going to impact the top line that much, but we're going to produce much more for the term price. And what we see from CIOs and the budget right now is that they are accelerating, in fact, their plan because there is a lot to do, in fact, in digitization in many companies. And, in fact, we can probably do probably twice as much that we were able to We definitely think that in fact with AI coding and testing is very simplified and we can produce much more than we have done in the past, with probably less people. But for us, I think if there is no impact on the top line, it's just the fact that we're going to accelerate the project and we're going to provide more. And there we definitely think that it's a big opportunity because we definitely think that with AI we're going to provide more services or accelerate, for example, some work that we ask and we say by the client. So we don't see for the moment, for example, for a big tender, we're going to announce one probably in the course of March, a very big one, and it's a very long contract on CMNI. In fact, the margin is cut because also we apply also Adjantic on our own delivery. We pass, of course, some of, I would say, the savings to the client, but we protect the margin of that loss, in fact, in the future. So that's why we say we are quite positive. Yeah, sure, thanks.

So if we go to slide 17, I'll give you a summary of it. So I think one of the key differentiators is the fact that we have this very long relationships and know-how with a number of really important clients. And what we've been able to do is to bottle up this decades-long insights and data from running hundreds, if not thousands, of managed services and long-running engagements into a series of agents which are sitting on unique Atos foundational models. So if you remember previously in the presentation, I mentioned our collaboration with Bullseye. So we are creating a frontier level model, which is Atos native, which packages up this know-how developed of the processes and the data built over decades. which we're providing on a genetic as a service model. I think the, you know, the other differentiation we would have is this experience of working in highly regulated, secure, mission-critical environments. So you'll need to remember that, you know, most of the time when people talk about AI today and agents, it's around, you know, things like, customer service or B2C or call centers. And AI is frankly fairly easy to implement in those environments. The accuracy just needs to be good enough and to be very direct if the customer who calls a call center does not get the right answer. the sky does not fall down. On the other hand, the type of agentic AI that we specialize in, like the super mission-critical one, it's a completely different ballgame in terms of robustness and industrialization. So if our AI agents would not work properly when there is a flooding in Scotland, then we have a serious problem. If the AI solutions that we're creating together with your control would not work properly, well, then you have massive flight delays in Europe and the entire economy loses a billion dollars a day. So I think this know-how we have based on our heritage of working in areas which some people consider non-sexy, if I'm allowed to use that word, it's turning into competitive advantage for us. We really know how to work, how to make AI work in those environments. And, you know, you see some of the recognition we have in this space. So ISG has recognized us as an absolute leader in advanced analytics and services. We've just made a leader in all four market segments with Nelson Holt. around transforming business operations with Gen AI and so forth. So to summarize, we're neutral. We're the Switzerland of governance. We know how to make AI work in this super difficult environment. And we have bottled and packaged this know-how into unique models and unique agents, which nobody else would be able to replicate.

Thank you, Fred.

Next question. Thank you. We'll now take the next question. This is from Nicholas David from OdoBHS. Please go ahead.

Yes. Good morning. I have three questions on my side. The first one is regarding the cash guidance. Can you help us reconcile how this net change in cash to expect for 2026 is comparable to what could be a free cash flow to equity definitions. What could be the difference between the two in terms of cash collection or cash outflow? The second question is regarding the provision you have passed in 2025. There were 23 million on the contract, notably. Can you present if it's just a cost of a run on last year and it was linked to cash out last year, or is it a provision for multiyear upcoming losses on the contract you identified? And do you expect more in 2026 if you review more contracts? And also regarding the litigation, when do you expect the potential cash out? And the last question I have is what would be your strategy regarding the debt refinancing given that the debt market for tech companies is getting more tight right now. Thank you.

Okay, I will just answer the last question, then I give the floor to Jacques-Francois for the first. As we say, the door is open for us to renegotiate the debt after one year. In fact, it was on December last year, 2025. And as you... What we are... What we have done is that we are prepared, I would say, to take any opportunity to refinance the debt. And as you said, right now, the door is closed just because the markets are not in good shape. So we will wait until there is an opportunity, so we will see. It could be March, it could be, I would say, in different other periods. I think the message is that we are ready. to do part of the refinancing as soon as the door is, I would say, the window is opening again, we will probably decide an opportunity on this one. Okay. So we'll see what happens in the course of 26. I don't have a crystal ball. It's difficult also because, of course, you said for the tech, it has been shaky, I would say, in Fed. Now with Iran, I'm not sure it's going to be less shaky in the course of March. So let's wait and be patient. But if there is an opportunity, we're going to take it. Now, for the two first questions, I'll let Jacques François to answer them.

Yes, Nicolas. So the net change in cash is the way we call internally this free cash flow, which you're referring to. There are no reasons for differences. Just in our guidance, we are excluding the repayment of debt. We keep in there the interest to serve the debt. that the repayment of debt is excluded, so is ethics impact, so is M&A. So that's the first question. Second question is regarding the provisions for onerous contracts and other items basically. So in terms of onerous contracts, Philippe has mentioned quite regularly in the course that we have still a couple of significant black accounts on which we are losing some money. We have the duty to assess these contracts regularly. Of course, management is going to mitigate with action plans to reduce the losses, and to be clear, we're also trying to exit, but so far we are bound. So in our reviews at the end of fiscal year 25, we have decided to provide more for future losses. So at this stage, you should not expect additional provisions to be added in 26, because the review we have done is quite prudent and should be comprehensive to cover all the future. And in terms of litigation, well, by definition, it's a bit uncertain and it doesn't depend on us. So I'm afraid I cannot give you really a timing for the cash out of this provision. But you will recall that the bulk of the litigation provisions has already been booked in H1-25. So there is not so much which has been added in the second half of 25.

And in terms of Black Accounts, there are no new Black Accounts. So don't worry. And we are, as I said, we are trying quite a very easy project. And we are still managing the last two accounts in the UK. Again, one account should finish mid-2027, so that's the goal, that is to stop one. And the second one, we are in negotiation also to stop it, but the end of the contract is 2034.

All right, that's very clear. Thank you.

Thank you. We'll now take our next question. This is from Sam Morton from Invesco. Please go ahead.

Hi, good morning. So in the release, I think you talk about considering to repurchase bond debt. Can you talk a little bit about what that would look like? Is there a particular tranche that you're looking at, or is that just sort of repurchasing across the board? And then I'd like to dig into the refinancing. Obviously, the window is challenging at the moment, but... When you think about the refinancing, is this a piecemeal approach or are you looking to do all of the refinancing of the first lane and the one and a half lane at the same time?

So I would say on the refinancing, the goal is first to refinance the 1S because it's 13% and we definitely think that we can be much cheaper. Right now it's a D-minus. And also with a positive outlook. And then after that, if we can do 1L and 1.5L, of course we will do both. I would say it will depend on the depth of the market. But I would say 1L is more important for us, just because it's too expensive. The 1.5L, in fact, is cheaper. It's around 8% plus in terms of yield. So 1L is the priority. But if we can do 1L and 1.5L so that we can stop also the, I would say, the procedure that was in place since 2024 for Atos, we will try to do both. But I would say the priority is 1L.

Jacques-François probably wants to... Yes, I'm on the reverse of Bond. So forgive me, I'm not going to give you a straight answer. However, I can tell you that what is guiding our action is, you know, we're making a standard calculation of value. And we are targeting the instrument where there is the better value.

Okay, sorry. Can I just dive into that? So would you look at the lowest cash price or would you... I mean, what's the philosophy? You're looking at the lowest cash price or you're trying to facilitate the refinancing? I'm just trying to understand how you think about it.

Well, in the ULD, which is going to be published next week, you will see that in 2025, we have already bought a little bit of second lean bonds, very tiny amounts because it's not very liquid. But we have put a little bit of 2L already in 25. Now we are looking at the NPV IRR. The first reason, the first objective is to look at what's generating more money. Because today we consider we're a little bit in excess cash, we have some big proceeds coming on, namely with the proceeds for the closing of the advanced computing division in a few weeks. So we are trying to make the best use of our money.

Great. Thanks a lot.

Thank you. As a reminder, if you do have a question, you can press star 1 and 1 on your keypad and wait for your name to be announced. The next question is from the line of Derek Marcon from Bernstein. Please go ahead.

Good morning, all. Thanks for taking my questions. I've got four questions, if you'll authorize me. The first one is on the range given for the guidance, you gave for the guidance, so minus five, zero plus, or positive. Can you try to, could you try to help us understand the difference between the low end of the range and the upper end of the range? At the bottom of the range, does it take into account significant revenue reduction with Siemens? And can you also explain us where you land with Siemens in 2025 versus 2024, and what do you expect in 2026, just to understand if it's an important moving part in the construction of this range? My second question is on your commercial momentum. If we look to the full qualified pipeline number at the end of 2025, it does not improve much compared to previous quarters, despite effects. So I'm trying to understand here what KPI do you have to, let's say, assess a much better, as you said, not Q1, but maybe Q2 or Q3 or Q4, And do you see really this momentum improving quarter after quarter? Because unfortunately, on our side, we can't see that through that number. My first question is on CapEx. So as you said, really good performance in 2025 on that side. Do you expect CapEx to remain at the same level in 2026? Or will you be impacted by the massive price increase on memories? And what percentage of the CapEx? of this 150 plus is linked to server plus memory, hardware, let's say. And that's it for me.

Okay. So on your first question on Xemans, we're seeing revenues of 25 or 300 million, and this year we anticipate 250 plus. So it's only 50 million, so it's less than 1% in terms of impact on the top line. Remember that with Demand, we work with three different entities, in the healthcare segment, Energy and Siemens Agile. And in fact, we do a free 150 plus and 50-50 with the two others. And in fact, I would say there are also different dynamics with the different accounts. But as I said, this year will be a 250 million plus because some of the contracts will stop also in the course of 2025. But there is no, I would say, big impact on Siemens, as you can see. Now, between minus 5 and 0 plus, as you say, we want to be cautious this year. I don't want to say we're going to go and sign it today. The goal, of course, for us is to do it. But we want to be a little bit cautious and give you a range between minus 5 and 0 plus and 5. Let's say between minus 5 and plus 1. And then you will pick the number you want. But I think it's a cautious stance in the beginning of the year, and we will have probably more to give in the course of this year. For the qualified pipeline, you're right, it's stable, but I think it's much more quality, I would say, for wind than it was one year ago. And in fact, what makes me, let's say, more optimistic is that the wind ratio is increasing right now. So I would say that the qualified is a pipeline where we are quite confident we can make a lot of wind in this pipeline. And then your last point was what? Capacity. After that, I would say for Eviden, the chips are not a big problem for us. And in fact, for some of our data centers, most of our contracts, we pass, I would say, the increase that we see from our providers directly, I would say, to the client. So there is not much risk, in fact, in terms of capex. The capex we are looking for this year is at 100 million plus. So that's the target that we have for this year.

Can I add just a small follow-up? Because on your explanation on AI, very helpful and interesting, and I'm on the same line than you about compensating price deflation with volume on most activities you are doing, but I was wondering if this reasoning can apply or could apply digital workplace and cloud and infrastructure and modern infrastructure because here I struggle to understand you will get this price deflation for sure but I don't see where the increased volume will come from probably you can explain that so it's a great question so actually if we go into the modern infrastructure so we see

a quite substantial uptake around the work that we're doing based on the sovereign movement. So it is a quite complicated area where clients need a lot of help, everything from advisory to try to understand which workloads they do sovereign and which version of sovereign. and to move and redesign both the application and the infrastructure space from those areas. I would also say that we have substantially improved our partnership with a number of the hyperscalers, so we're driving a lot of additional new joint go-to-market campaigns and solutions in this space, which is acting as a as a net positive. And I would also say that on how the modern infrastructure, actually AI is opening up new opportunities which historically wouldn't have been possible to do for our clients. So as AI is making the modernization and the digitalization of legacy applications possible in a way which, you know, frankly, again, wouldn't have been realistic or cost-efficient in the past. That drives substantial requirements for infrastructure and cloud modernization. So AI is actually a tailwind for us in cloud modern infrastructure. And when it comes to digital workplace, we are expanding the type of services we provide in digital workplace. So again, AI is to some extent a headwind because some of the services which we historically would have done with people are now done by agents, but we're able to improve our margins in that case. But we're also seeing AI acts as a multiplier. So one of the key demands we see from clients is how to have their people truly be able to use AI constructively, usefully, and in a meaningful way. So we're actually adding AI enablement and AI capabilities as part of our digital workplace services. We're also using AI to make the digital workplace experience a lot more enhanced to help with self-healing. So we're basically adding additional services, additional value-adding services in our digital workplace portfolio, which again are quite nicely balancing those tailwinds or nicely balancing the the headwinds we would have had traditionally with, you know, digital labour replacing human labour. I hope that answers your question.

Yeah, very clear.

Understood. Thank you. We'll now take the last question today. Please stand by. And the question is from Laurent Dor from Kepler Shufra. Please go ahead.

Yes, thank you. Good morning, gentlemen. As for Derek, I have also four questions. First, if you could come back on the way you have built your revenue plan for 2026. I mean, if you start the year with the first quarter close to minus 10, and you're not going to have much easier comps the following quarter, Does it mean that you're expecting to win sizable deals that will start during the year? Or what makes you so confident that you're going to end the year with strong growth in order to offset the first quarter? Then my second question is, first, thanks for the clarification on Siemens. But if you could share with us exactly your relationship with your clients as of today, and in particular, I understand that you have two more years of business, but do you already have a visibility on what's going to happen for that client as of 2028? And the last two questions, one is on the one-offs at work. Which timing do you expect the P&L to start to be quite clean with limited restructuring and provisions? Is it 2027? And the final question is on the nice improvement you were expecting on EBIT. Could you share a bit the building blocks to go from 4% plus to 7%? The main savings, that would be helpful as well. Thank you.

So on your first one on demand, so we have what we call the card that was signed in 2020. It was a five year plus two year contract. We just answer tender and one project. So in fact in 28 we continue to answer the tender and win some of the projects. In fact and when you look at the back end in fact we have all the revenues for 28 and 29 for some of the projects that we have won in fact in the course of 25. So we say it's a normal client. There is no need I would say to resign a trial whether or not it doesn't make sense. Also because, in fact, in the crowd that we had signed in 2020, you should know that there was a signing bonus that makes, in fact, the margin of the contract not that good. And, in fact, now the margin has been restored in the course of 26, so we are quite happy on it. And the idea for me is to continue with Siemens like all other clients. There is no specific agreement that we need, I would say, with Siemens. And remember also that, as I said, Siemens is three different entities. with three different IOT clients. So in fact, we have also client partners evaluating the different entities of finance. Now, for your second question, of course, if we start at minus 10 and we want to be positive, there is no magic. We need to be at a strong growth in Q4. That's the anticipation that we have. I cannot go into details on which contracts we want to win or not. It's too difficult to do that. And I'm not sure it's very useful. But, of course, I would say the goal that we have in our budget is to be roughly at zero plus in Q3 and then have an acceleration of the growth in the last quarter. I would say that if we had been able to get zero growth, it's a very good result for us because it means that we have now, let's say, growth going forward in the course of 2017. The bar is high, Laurent. Don't say it's an easy one. Please be careful on that. Don't estimate that everything is easy. But, of course, we have ambition, and we definitely think that we have the pipeline and the projects to rebound, I would say, in the course of Q3 and Q4. For your other questions, I don't remember. Yeah, go on.

Well, I think, Laurent, you were asking when do we stop the one-offs and when do we have a P&L which is clean. Well, I think already 26. I mean, for me, the numbers we are publishing now are taking everything we know into account. So, of course, in 26, we still have the continuation of the Genesis Restructuring Plan, because we said that we booked a large chunk in 25. If you remember, the full envelope was 700, so we're still a bit below, so there is still somehow a portion of that to come in 26. But beyond that, I would say that 26 already should be expected to be clean. That's the question on the one of your last question. I don't know if you want to take it, which is the further, you know, the building blocks of the path to the 9% to 10% margin.

I think, yeah, well, first we were at 6% in H2. Remember that we have roughly 200 million of savings after January in the P&L coming this year. So if you restart with 300 million plus, plus the 200 million, We're going to have an increase of salary, now it's an impact of 17 million plus. So your reading growth is 314 plus 200 minus 70, and then plus the other actions that we are going to take in the course of this year. But that's why we are quite confident on the 7% margin.

Philippe, if I could add on a new scope, a question on the seasonality of the margins, because you improve nicely from first half to second half. But do you have part of that as coming from seasonality or going forward you expect when you will have stabilized the operations to have a similar margin level between the two halves?

In fact, it's going to be always more margin in H2 than H1, but with less evident bullies out. And that's most of the explanation why H1 and H2 are very different. It's not going to be the case in the course of 2026. You will see a more stable revenue and I would say EV strength between H1 and H2, but usually, and all the companies, it's the case of there is more margin in H2 than H1. But not, I would say, like it was in fact in the course of 2026. To a smaller extent, yeah. And remember that I said already in the CMD last year that the There will be close to zero, I would say, non-recurring expense in terms of cash in 2028. We cash out, I would say, Genevieve. So this year we estimate that it's going to be between 150 and 200 million. We have done 450 last year. And then the rest in the course of 2027. No more, I would say, cash out in 2028. Same thing for the litigations. We estimate that... Most of the litigations will be done. And then for the Blatter Comp, as I said, there will be probably only one in 28. So it will be, I would say, a small impact in terms of cash. So the emit will be clean this year, but I would say in terms of cash households, it will be clean in the course of 27 and 28.

Great. Thank you.

Thank you. There are no further questions at this time, so I will hand the conference back to the speakers for any closing comments.

Okay, so thank you everyone for this long call. We are very happy, as you have seen, I think the focus was on technology today because there were a lot of questions on our industry and also on Atos. Have a good day, and of course, we will talk to you probably for Q1 and in the coming months, and we, of course, focus to the rebound of the company. Have a good day. Bye-bye.

Thank you. This concludes today's conference call. Thank you for participating and you may now disconnect.