Ladies and gentlemen, dear guests, welcome to the Analysts and Investors Conference of the United Internet company. My name is Dominic Großman and I'm very happy to welcome you in person here to Sofitel in Frankfurt. I would also like to welcome all participants in the webcast. I'll be happy to take you through our agenda of the day. First, Mr. Dommermuth will speak about the development of the fiscal year 2025 and give you an outlook on the current business fiscal year of 2026. After that, Carsten Theurer will give you the details of the financials. Ladies and gentlemen, after our presentation, as usual, you will have the opportunity of asking your questions in a Q&A session. So much by way of introduction, I'll give the floor to Mr. Dommermuth.
Yeah. Thank you, Mr. Gro ß man. I think, we've all met before. Those of you who haven't been there, welcome on my behalf as well. Mr. Gro ß man has taken us through the agenda. I will start with the company development. Many of you know this. We work in the internet access and applications. We offer that for consumers and business customers. In the middle, you see our assets, our team and our infrastructure. We have over 10,000 employees. Our optic fiber network is 86,000 kilometers long. We have a 5G mobile network, and we operate about 100,000 servers. We offer our products in different brands. In the consumers mainly the 1&1 brand. Other brands which we have taken over from Drillisch and in the business customers, 1&1 and Versatel.
Applications for the consumers, we offer in GMX and Web.de and mail.com. Applications for small, medium-sized companies are mainly driven by IONOS or the IONOS subsidiaries. You see them here on the chart. Companies that we have acquired over the years, Strato in Germany, for example, Fasthosts in England, others and in Spain, so on. We have some minority shares, companies that we cooperate with. Let me start with the access business. We have just heard in the 1&1 conference, 3.8 million broadband connections, 12.5 million mobile contracts. Our own Open RAN, fully virtual networks, reaching 27% of the household as per end of the year 2025, overfulfilling the requirements of the authorities. High customer satisfactions in internet access and also in the mobile range.
First place in both competitions on the Connect customer range. Customer contracts minus 17,000. In the access, minus 110,000 broadband accesses. I've just said that we get a hold on this, and the mobile contracts is a bit dampened by the migrations of our customer in the last year. Our turnover revenue, 0.8% growth. Service revenue stayed about the same. Others, especially, selling smartphones, tablets and so on, +3.7%. EBITDA in the consumer access as a total, a minus of 11.7%. I'll reason this if we look at the individual segments. We have the segments, 1&1 Mobilfunk, with, EUR 265.3 million. The costs for the same figure as in 2024.
In the segment EBITDA access, which is our end customer business, 8.1% less EBITDA with the change of the national roaming partner. This is reasoned by Telefónica activating a part of the costs. Now we don't have this activation anymore, and it's the Vodafone network has grown slower in 2025 than we had expected. Looking at the business access range, here we work with 1&1 Versatel. 68,000 for 450 km of network. 350 biggest cities in Germany and up 29,000 directly connected customers. Two handover points, two antennas or handover points, two telecom or regional carriers. The turnover increased by 2.1%. Also, the EBITDA 2.1 better, EUR 177 higher than before, despite the starting costs.
Applications business in the fiscal year starting with consumer GMX, Web.de. Well, consumer, we offer a broad consumer portfolio for information management, emails, calendars, online office, cloud storage as well for the photos on your cell phone, for example. We fund this by we differentiate by the German data protection, German servers. Last year, we won over more than 60,000 new accounts, 360,000 paid customers. Usually we start with a free account and step by step migrate the customer to a pay account. We did very well in that last year. We have 310,000 additional pay accounts. Free accounts also from the conversion is dropping due to the conversion, and we have increased the revenue by 8.1%. 8.1%, that's good.
EUR 322.6 million, or euros, driven by the better monetization and the pay accounts also due to advertising. Parallel to the revenue, the EBT grew a little bit more even, 8.7%, EUR 223.1 million. This is a business that we've been doing long years, developing well, and we are optimistic for the year to carry on with good opportunities in the future. Looking at AI, which we implement or in our services, or offer as add-on to extend the business. Business applications, the last range I'd like to present here, driven by IONOS. IONOS subsidiaries. IONOS is the leading digitization partner for small and medium-sized companies in Europe. Cloud enabler, 40 European countries and in the U.S., Canada, Mexico. A large product range, presence on the Internet.
We need domains for that, website for that, e-shops, online marketing tools and so on. Also products concerning business office organization, business email with opportunities to administer different accounts, office suites, online storage, for example, to store your emails legally, correctly or big data management, AI assistance, cloud infrastructure, a broad range, all at higher security standards developed in Germany and the EU, all conform with the data protection laws. Customer contracts have increased here, driven mainly from growth abroad, with 5.25 million. The domestic contracts are 4.8 million. Germany good and still ahead, but the international business growing stronger. We could compensate in the business applications the drop of the dollar. The growth here good.
EBITDA 19.8%, scaled to 464.1, 35.2% EBITDA margin. An overview to what I said, 700,000 new customer contracts, up to 29.72 contracts. Revenues grown by 1.2, 1.9. EBITDA 2.4. EBITDA due to the investment EBITDA is a little less, 11.5% less, due to the EUR 42 million of depreciation. EPS EUR 1.23, a plus of 43% due to the difference in the taxation. What are we gonna see in the future? This year, we want to increase our turnover to EUR 6.25 billion. The EBITDA should end up at EUR 1.45 billion, EUR 170 million more compared to the year before.
Cash CapEx, we see at EUR 600-EUR 650 million. We've just said this at 1&1 before. We've invested a lot into decentral and regional data centers. That's gonna slow down a bit. We have 350 already up and running, so the cash CapEx is a little lower here. What's gonna change in 2026? Besides the figures, we are gonna report by segments. We are going to use the access business after selling the 1&1 Versatel anew, the segment consumer access since January 2026. When we meet next time on the quarter figures, it will be set up anew, and it is then called 1&1 Mobilfunk. 1&1 mobile network will be taken out, and consumer access will then be called consumer and small business, showing that we address small businesses as well.
There's gonna be a new segment, which is then 1&1 Mobilfunk network together with business access, and it's gonna be called Enterprise and Networks. With Versatel, we address large customers, and besides the Versatel network, we have the mobile network in this segment as well because we think that is a clearer structure for you having the pure, let me call it resale business, consumer access. Consumers buy at Enterprise and Networks. Enterprise and Networks is the wholesale business due to the networks that we have. This is the idea of the new segmentation. Okay, and now I'd like to ask Mr. Theurer to explain the financial figures to us.
Yeah. Welcome again. Basically, everything has been said again. I can tell and explain it to you in a different perspective, from a cash flow point of view. Let's start with an overview of the key figures for the group. Mr. Dommermuth already said that we were able to increase our customer base. We have added 0.7 million users, mostly from consumer business, i.e., IONOS. IONOS with Business Applications with 460,000, and consumer applications. You can see that the pay accounts on the next line with 310,000 customers who we could convert from ad-financed free accounts to pay accounts.
That's why, the ad-financed free accounts are slightly decreasing because we converted them to pay accounts, which works very well and gives us a nice recurring revenue in this business model. Overall revenue increased by 1.9%. We had an EBITDA growth of 19.8%. That's a main driver on the group level. Sorry, EBITDA was 19.8%. That was a mistake. That refers to the next line, EBITDA growth. We see a growth here of 19.8% of Business Applications where we have an increase to 2.4%. What's different here, as we heard with IONOS earlier, due to the IFRS 5, we have a discontinued business unit, Sedo, within the P&L that was disregarded.
In the balance sheet it's included and in the cash flow as well. We see it in the cash bridge later on, that those are the rules of the IFRS 5. The EBITDA, all figures without the segment EBIT. We can see the effect of the higher depreciations here. We have a slight decrease of 1.9%, compared to 2024. Cash flow, we have a positive trend and significantly increased cash flow by EUR 70 million. This comes from cash flow from changes in business activities. We have a positive cash flow. This is due to our payment of contingency payments.
That is a volume that we paid EUR 200 million in advance and will be used up over time over the next few years. That means that it's no longer a cash payment. Was made over the last five years already. The net cash inflows from operating activities, we see a net outflow through investments. We have a dividend, the catch-up dividend that we paid last year. We have the acquisition of the 1&1 shares with over EUR 200 million that we managed in 2025. We have buyback program for our shares that started in the last business year and extended into this year with IONOS.
That takes us to a free cash flow after lease, which has significantly increased by EUR 18 million from EUR 47 million in 2024 to EUR 320.6 million in 2025. Cash flow bridge from EBITDA to free cash flow. You can see we have a slightly increased EBITDA compared to 2024, and I'll focus on 2025 now. We had AdTech added because it's necessary for cash flow considerations. We had a similar CapEx, a little bit less than last year. You can see EUR 730 million right now. Investments into the extension of our network, mostly, and you can see the biggest difference here in taxes. What is that due to?
Well, you can see last year, 2024, we had a normal tax as it were by the end of or at the beginning of 2025. Backdating to beginning of 2024, we had Versatel integrated into the tech group. We had an advance payment for these taxes that we didn't pay back in 2024, but in 2025. The advanced payments have a positive effect on the taxes in 2025, so we have a positive effect, and therefore we have this big discrepancy because these two figures. We have the working capital, pretty constant in 2025 compared to 2024 of EUR 508 million, and leasing minus leasing.
That takes us to free cash flow of EUR 320 million of just EUR 270 million more than 2024. Now let's take a look at the balance sheets. I won't read out every detail here. Maybe the most important topics you can see that the fixed assets grow fast. We have EUR 350 million added here, particularly through our investments in our fiber optic network. We had the contingent payments from telecom. We're using those up over the years, so we're activating those now. Nearly identical balance sheet same despite increased intangible assets.
Let's take a look at the liabilities and equity, and we can see that we've increased those by EUR 350 million. Trade accounts are decreased. Liabilities due to banks have increased, particularly in order to finance the CapEx, but also due to the share buyback and our dividend had a negative impact on our cash flow. We have a net debt of EUR 3.2 billion at the end of 2025, with a leverage of 2.48 at the end of 2025. The other financial liabilities have slightly increased, the leasing additions that we've seen, then lower refinancing liabilities on the other end.
Equity has slightly decreased, but the ratio is still at 43.6%, so that's still a pretty solid equity rate, equity ratio. That takes us to the end of the financials here, and we're happy to take your questions if you wish to ask any.
Okay, so much concerning the presentation. Let's start with the questions and answers. Please use the microphones so that everybody can hear you, and also indicate your names and companies. First of all, on the left side, Karsten Oblinger, DZ Bank. I have two questions concerning IONOS. First of all, concerning the data center that was planned since the beginning of last year or the summer of last year. Is there any news? We never heard anything about it.
The second question is concerning the revenue development for cloud solutions. That's a bit behind schedule. Could you give us some reasons for this and maybe give an outlook for the next couple of years? What are the current plans?
Well, we're not the IONOS board. I'm on the supervisory board, but I'll try to give you the best answer I can. If you speak of the planned data center, you're probably referring to the potential EU gigafactory. We're still waiting for the tender, which is being delayed again and again. The last thing I heard was end of March. Once we get the tender documents, then we or the IONOS management can determine whether it suits or not. Concerning cloud solutions, I can tell you very little about this.
As far as I can tell, the business is going very well so far, but I really can't say too much. Well, I might be able to give you some information there. We had some headwinds in the last couple of years, but this leads to tailwinds right now. We've had a quite good start in the first quarter, and we're quite optimistic that we can grow by double digits, particularly in the public space, and so we're quite optimistic looking forward there.
Next question, Polo Tang, up front here.
It's Polo Tang from UBS. Just a clarification question in terms of tax, because your tax stepped down quite notably to less than EUR 100 million in 2025. I think it was EUR 250 million in terms of 2024. How should we think about the tax charge going forward?
Yes, we have to make a distinction here between cash flow. I showed that in the cash flow bridge and what is declared in the P&L. We have two different effects. With the cash flow, we had the impact of the United Internet's, and in the second step, the sale to 1&1 of Versatel. We have the tax group Versatel, including 1&1. In 2024, it was for UI because the tax effects were similar for UI, and now we have the same thing for 1&1 Versatel. Due to the losses of Versatel included into 1&1. In 2027, we'll go back into a normal level.
In 2026, it will have an impact on the P&L concerning the taxes. The main special effect was cash back from the advanced payments from the year 2024.
Next question on the right-hand side. Yes, Mr. Dommermuth. My question also refers to IONOS, basically from investor to investor again. Since August, there have been opportunities for AI. We discussed them and everybody understands them. But since August, the opposite has occurred, because everybody believes that it is a threat to the IONOS business model, and I would like to know what your current view of this is. It won't surprise you that I'm quite relaxed in this context. I hope it's not a misjudgment.
You've seen that last year we increased our EBITDA at IONOS by 20%. The management gave guidance of increasing it by another 10% this year. I'd say business is going well there, but it doesn't necessarily mean that it will continue to do so. You mentioned quite rightly that the question remains whether we are on the right track and the big issue is what about AI. If you look at IONOS figures, we had a share price of up to EUR 43, which rose very fast, but then dropped very fast as well. We have to see that our peers had more severe losses. GoDaddy have a third of their peak value.
Another competitor is down to a sixth of their peak value. The entire segment has come under pressure. Question is, will AI be a benefit or detrimental to us? I think at the end of the day, it'll be beneficial. Why? Because we see the opportunity for a new business field, AI for small and medium-sized enterprises. I think we're excellently positioned there with a number of our customers, the type of our customers, our sales capacity, and we can see that it's going very well. It started very well. We're at the very beginning still. We are increasing the number of AI agents. We try to increase our sales efforts, and I see a lot of opportunity for additional business. At the same time, we're including ever more AI into our products.
Making our products, in programming, we gain a lot of efficiency. Now, that's my personal view of things. It doesn't have to be right. Now, where does this skepticism come from? It comes from new possibilities of building a website with so-called vibe coding. There are small new companies are excellent in doing this. Lovable keeps popping up here. I don't see why that should affect us much because we ourselves. Well, the value creation involved in making a website, building a website is not our main value creator. We make the infrastructure available. We will also offer vibe coding in a few months from now. I don't see any reason why this business should bypass us.
I think those are in the field of creating websites as a core business, they are more affected. Even if you build a website via Lovable, you still need a domain for that. If you build a website with Lovable, you still register the website with us because through a cooperation with an American supplier called Entri, you wind up with Entri again. We can not only sell the domain, we have a customer relation then, and we can sell more products. The website has to be hosted someplace, and with Lovable, there's a premium model. You can use a lot of products free of charge. The next level is $25 and the final premium level is $50. That's much more expensive than with IONOS.
I told you about the share price development of Wix, which crashed by five sixths, and they have this premium pricing model as well, and that's not our model or IONOS, the IONOS model. I think we still will continue to have this business model. We have a good customer growth. We expect even better growth this year, but maybe we just don't see the problem or the risk. We see opportunities more than risks. Would it be an option, IONOS has its own share buyback programs that you increase your share with United Internet, or don't you see that as an option? Well, if you ask me what I would like to do, I'd go ahead with this, yes. If you ask me what I could do with that, I would delist IONOS.
If you ask Carsten Theurer, he'll tell me immediately about the high debt rate and all the things that we need money for the auction, et cetera. We have to see. I hope that I have the right view of things, but if I have the wrong view, it might be a mistake that might cost us billions. 50% of the IONOS shares are with us due to the buyback, and we're happy with that. If management were to desire more buybacks tomorrow, I'd always agree with it.
We could, if we are to increase our stake, I'm on board. To say, "Okay, let's buy back everything," my heart says, "Yes, I wanna do that," but I think it's a good thing that I can see that we don't have these opportunities and that we have other investment opportunities as well, and that is why we have to keep our powder dry. That's not on the agenda right now.
Okay, thank you very much.
Next question, just beside.
Kali Voda Research. Hello. Mr. Dommermuth and Mr. Theurer, you are active in 14 European countries, but US and Mexico, Canada as well. Now the last three, US, Canada and so on, there are the hyperscaler Amazons, Microsofts and Google Clouds. I think they take care of small businesses as well, don't they? How are you moving forward in that segment? Can you say anything about the margins or the investments that you intend to place there? Possibly you have a critical size already. Where are you at or in the US?
US are our second most important country at IONOS, but it's not a country where we would offer our cloud infrastructure. Our cloud infrastructure is mainly done in Germany, France, Spain, England, European range. The focus really is Germany, because here we have a sovereign cloud addressing a part of the market very well. We can't compete with Google, Microsoft or Amazon in the full range. That's our target. We offer our web presence products, the productivity tools, IONOS. The business is growing well. We've had a nice growth in America last year. This year started well also. It's a good country, good money, we grow good. I think we have more opportunities, but I don't see them in the cloud infrastructure range.
Our USP, the European sovereignty, is not worth anything in the U.S. We just cooperate by size. We just compete by size, and we wouldn't have an opportunity against the big giants. If one of the hyperscalers uses it, isn't it difficult to win over the products that you want to sell in the US? No, I think our products mainly address small and medium-sized companies, and these are not the Amazon customers or Microsoft either. We sell everything concerning the website, the domain.
This is not the focus of the hyperscalers. Well you never know. Of course, maybe, Google says, "Well, I have to sell to small and medium-sized companies." Okay, but they are. Google had domains. They sold their domain business. Google has tools to make homepages. There's nothing really new, but it's not a part of Google Cloud. Google Cloud has different contexts.
Thank you.
The next question may be on the left side. Mr. Dommermuth, I'd have a question concerning the AI question we just raised. I do share your relaxation as far as IONOS is concerned, but I'm not quite sure what the consumer applications are concerned. What we do see is more and more AI-based inquiries, less search engine-based inquiries. Maybe you could illustrate a bit how consumer applications could be concerned if this trend continues. Well, I'm even more relaxed concerning consumer applications. What do I see?
Yeah, well, you let me break it down. If you say there's less research, it's traffic that you have on a website that offers content, where Google, where you participate from Google search. Now media tell us that many pages are losing traffic because AI has fewer content websites presented in the search result. However, that has no effect on us because we just look at the incoming traffic. If we sell advertising, the customer logged on.
It's not the Google traffic that we live on. We do have a little, of course, but it's you know, just behind the decimal. It's an opportunity for us actually, because if you do an advertising campaign in Germany to get reach, the more the others lose, the less you can get around us. We see this. We see that media planners start to look at us, and the loss of the traffic is rather a benefit for us without me being happy about it. We have the lock-in traffic. People want to check their mails or access the webpage, but they don't take Google, give me a hint on what I do if I have an infection. That's when you visited a website before, which is called your health.
Now Google tells you what pill to take, and the website has no traffic. We don't have this problem. Search marketing, no problem. Concerning AI itself, I see big opportunities, and we do this step by step, putting this in the applications that you can have your mails translated in different languages. You have a smart search, giving you context, allowing you to give context so that they can look at your emails or with a couple of keywords, you get a proposal for your mail text. We have this as beta, partly even rolled out. We have that, and we are doing more of that this year. I see the opportunity to take AI agents offering them to private people. As we've talked about IONOS before, saying they have a high reach, they can use it to distribute their AI agents.
I'm not talking about ones that you have to build yourself. In AI, we have an assistant for phone assistants. We've got a partner in America, one partner in Germany, and we integrate these products. I think that's good because that makes us fast, and we can react better to the different market requirements, apart from. We don't have to build up the know-how ourselves for every little detail. That applies to the consumer range as well. I've talked with the colleagues on the board lately, and as an example, to illustrate this, we talked about an AI agent for travel planning, saying, "I want to go somewhere three weeks, Mallorca, Ibiza, whatever, somewhere warm, two children, need a rental car, hotel should be this and that, along seaside.
Create a proposal for a travel. This is available already today. There are people offering. I looked ChatGPT, they showed me a company, 24 people in Berlin focusing on exactly this. How do they get reach? How do they get users to book their journey with their tool? This is where we get into the game. We could integrate them into our service and generate AI reach for this AI application. You learn about the customer. That means we can extend our data reach. We know they want to go to a summer holiday in Mallorca. Okay, that could help us in marketing, advertising, and you sell it as a premium service, but I don't know that today. If somebody has a platform to distribute the products to consumers in the German reach, it's us.
We can do this better than BT, you name it. We are better. We are the best in that section. This is why I do see opportunities. Last but not least, if you say Mr. Dommermuth is a dreamer, he's always optimistic, so he just puts the AI and sells AI agents on top. If it doesn't work, if he doesn't have range tomorrow because people, ChatGPT, write emails, whatever, he has no traffic anymore on his paid pages. He won't need a tool to write English emails or Spanish emails, and he won't need AI agents for travel planning because he's got no traffic. I got an idea for that. I'll just close the service and do it as a free service. I close it as a free service and charge for it.
There are services, a lot of examples for email service who used to be free and then, and then, cost something. The email address sticks because it's used everywhere. There's about 20% of the customers who are ready to pay if we do this. This is where we have the participation in Open-Xchange. They do that. They buy emails where Telco says, "Look, I did this. It's not a business model for me anymore." They buy the emails, switch off the free service and say, "Customer, you can keep your email. It'll be much better in the future. Won't be any advertising, a great hotline. You have more storage, but you'll have to pay." Then the revenue would rocket, would skyrocket.
Just take our customers, and let's say they pay 5.60 EUR per year, minus VAT, 50 EUR by 20% of our customers. I can't even calculate it. EUR 300 million. With advertising, I think we do EUR 120 million. EUR 180 million plus. We could reduce complexity. All the advertising, targeting, bidding, data, and so on. We got 1,000 people working for that. Simply a cost, an email that costs the customer much simpler. It's not the plan. We wanna go forward. We're not planning to do that, so no misunderstanding here, but we want to get into the products, do the agents and all of that what I said. If it all didn't work.
I would say whatever how the customer writes their emails, they will need the email address. That's the thing with IONOS. Whatever they do with their website, they, it has to be hosted. They need a domain, they need marketing that has to be compatible upwards and so on. I'd say, and that was when I started my monologue here, saying I'm even less concerned here than with IONOS. I think we'll do it well, and I have a plan B in case it doesn't work.
Next question on the right-hand side.
Dr. Essos. Gareth of Stutzen, Core Game. The last aspect was very interesting, basically added value, which is your plan B. My question is on 1&1. I saw your passion for IONOS, just now, and your willingness to buy back shares if Mr. Theurer wasn't sitting on the money bag. The question is whether the competition isn't just as enthusiastic about the possibility of purchasing United Internet or 1&1. I know that the money back, of course, is sealed up right now for the moment.
Well, of course, I spend maybe one day a month with the supervisory board of IONOS. In the past, we were one company. IONOS. I used to be the CEO of IONOS, and therefore, I spent many years with the hosting business. It's a beautiful business with many opportunities. Yes, there are moments where I say, "Oh, this telecommunications business with this slow market growth and all the problems that we had over the last few years," maybe it wouldn't have been the worst thing if I'd never seen it, if I'd always stuck with hosting.
You never know. For you as a shareholder, you compile your own portfolio. A medium-sized company doesn't want to stand on one leg alone. If I focus on hosting alone and tomorrow Google or Amazon say, "Well, we can do much better, and I will give it all for free," then I'd focus on telecommunications. I'm very happy with the portfolio. It's very balanced. There's always one wheel that turns more slowly than the others, but we do have four wheels that we drive on, and not only one.
I think as a somewhat more elderly entrepreneur, I'm very happy about it. Our passion is for hosting, for the access business, for the mail business. I always have sparkling eyes. I see opportunities everywhere. I give you the last anecdote now. Years ago, I was standing at a blackboard when we were only one company back in the days of Deutschmark. Said, "If we do it right, we can earn DEM 100 million for access, DEM 100 million marks, Deutschmarks for hosting and DEM 100 million for email business." The others were just glaring at me. You see, that was a while ago, but the ambitions were ones that we would earn DEM 100 million in each business field.
We went way beyond that by now, and I think we're still not at the end of the line. There's still opportunities. We just have to accompany every transformation. We have to participate in the AI transformation now, and we have to become much more efficient within internally. We have a lot of AI projects. You can see that the number of staff is ever decreasing because we're becoming ever more efficient, and this will continue. Programmers are becoming more efficient. Customer care will be much more efficient. We move into new fields where in marketing, in accounting, where we can consider evermore how can we do it faster, better and smarter. We're still at the beginning here. We can see that we're increasing on the top line, and at the same time, we're decreasing the costs.
If we do it right, AI will be a positive thing for us. Maybe explicitly on purchasing, 1&1 as well. That would work in terms of your desirability. Well, if you have the opportunity, would you increase your share? Well, last year we did increase our share, when packages became available, and you said, "Okay, we will make an offer." That took us to 86.5%, and we feel happy about this. If somebody were to offer us another package tomorrow, I'd take it as well. I'm still optimistic. I wouldn't buy every single share individually, but if somebody comes along and says, "Okay, I have another package. Are you interested?" We'd go for it. Okay. I thought that was a strong statement to make.
Thank you.
Well, there's another question up front here.
Straight off, Hans de Klein. I heard your enthusiasm vis-a-vis AI with great pleasure, and it might be a little bit as important as the internet was when it was first introduced. Now, your company is called United Internet. Would it be possible that you make something much bigger out of AI than AI agents and with IONOS as they did with the web business?
Well, I don't think we'll find or develop any large language models. That's way too much effort, and it's not really possible in terms of the power costs in Germany, and the competition is way ahead. We will focus on the agents, and we'll have to think about where do we use them and how can we use AI to improve our products by installing or introducing them into individual products and tools. But I don't think that we'll become an AI company overnight now.
I was told that I should rename the company from United Internet into United AI, then our stock price would double, but I didn't go for that.
Did I understand correctly that large language models with Aleph Alpha you can see how you can fail with the huge investment there. But my question was that the AI agents and what you do could be so big that it might be an important business field.
Yes. Of course, if we do it right, it could be very big. Yes.
Are we too slow? If you ask me, we're always too slow. Here we are behind our opportunities. You can see it over the next few months, we're rolling out a lot of things there, and I see a lot of opportunities there. The tickets are larger here than what we have in hosting. In hosting, I get a good website for EUR 10, but for one, two, three AI agents that run my office, who handle my work, I'm willing to pay more for that. I see huge opportunities there.
Our meetings are always about this question, "Okay, people, let's focus more on this, push the other stuff aside, fully focus on, showing what we can do there." We can see over the first few months, the figures are fast expanding but at a very low level. We need to boost this and to learn as we go along of how to do this best. We're well established for that. We can host the AI agents, we can integrate them into our processes, we can market and service them. I see huge opportunities. We're excellently placed there.
May I follow up on that question, Mr. Dommermuth? Now that we're talking about AI, I have another question on this topic.
We've covered all various business fields. Now, concerning 1&1, because I think it was Telefónica who said that they need to take more computing power to the network. My understanding is that they need to install more data centers for AI, and they have. They have the data crunching ability. But I was thinking, where does the network, the mobile network require AI, or what could AI contribute there? Well, my understanding, but I'm not a technician, of course, I see two levels here. One being you will have efficiency gains in areas still handled by people today. Let me give the example of network operating centers. That's where we have people 24/7 watching where something breaking, and they alert repair crews, and that is something that you can automate. You don't need people anymore for that going forward. That will lead to efficiency gains.
You won't see that happening in the network, but in the core areas, the core places operating the network. If you read the press releases, then Telefónica says that we have such a great AI, they can help us identify capacity utilization levels in the network. You don't need AI for that. I don't know what they could do better with a huge data center with AI. I can't really see that. What we would do is we say we have space for applications, and these applications could, of course, be AI applications, no question. The benefit is we're faster nearer the antenna because we're only 10 km away. That's a benefit.
Otherwise, I don't see any AI specific aspect that needs to be far-fetched because it can't run on the core. I don't see that. Yeah. Well, I've had the same feeling about this. Yeah, but it sounds great if you say, "Oh, let's do AI everywhere." As was said, that'll drive the stock price.
I can't see any more questions. Oh, there is one. One more question. Another question on consumer applications. Last year, EUR 123 million EBITDA. I'd like to know what your expectation is for this year. Where are we headed? The idea was whether you want to sell this. Is that still on the table or not? Because your sales expectations aren't realistic. Or what is your view there, Mr. Theurer? What's the bandwidth for the free cash flow for this year?
Maybe concerning the consumer applications, they are growing very well because we have natural growth, organic growth, and because we're winding down a shared service. IONOS was handling that for us, and they want to focus on their own business and don't want to do that anymore. 1&1 Mail & Media has bought these servers from IONOS, so they have to handle it, and that increases the EBITDA. In the past, they paid on a monthly basis. Now we have the upfront investment. That's why the EBITDA is much higher, but it only comes from growth. Partially it comes from the fact that we operate our own servers. I think we're over EUR 140 million by now. Is that realistic?
Yeah, 100, more than EUR 140 million . That's what we expect there. Sale. I think it was in 2022, when we were asked, we want to sell a business, but we didn't get far, and it's not on the agenda. I said before, never say never. Maybe tomorrow somebody comes along and says, "Oh, this is so great. I gold-plate it for you." Then, of course, we have to listen in the interest of all investors. But it's not a plan. The business is going well. The last year was good. The year before was good. This year will be good again. We have a lot of good ideas. Free cash flow around about EUR 600 million. Very rough estimate. We're coming from just under EUR 500, as we saw.
You have the EBITDA improvement with 1&1, the forecast. Around about EUR 600 million ballpark figure. Any other questions?
I can't see anyone wishing to ask a question. I hope I didn't overlook anyone. Thank you very much for your interest, your questions, and I would like to close this conference, and I invite you to have a coffee out in the foyer and see you.