Good morning, everyone, and welcome to the IONOS analyst and investor call for Q2 2025. Thank you for taking the time to join us today. My name is Stephan Gramkow, and I'm responsible for Investor Relations at IONOS. Here's what we will cover today: Achim Weiss, CEO of IONOS, will start with the business update and strategic priorities. Next, Britta Schmidt, CFO of IONOS, will walk you through H1 and Q2 2025 financials. She will also cover our outlook for the full year. Achim and Britta will then be happy to answer any open questions after the presentation. I would now like to hand it over to Achim. The floor is yours.
Thank you very much, Stephan. Good morning, ladies and gentlemen, and a warm welcome to our conference call. I'm Achim Weiss, CEO of IONOS SE , and I'm pleased to update you on our business and key topics. Our mission at IONOS is unchanged: we empower small and medium-sized businesses to succeed in the digital era. We strongly believe that every business, regardless of its size, should have access to the same technologies and expertise as large enterprises. We aim to close the digital gap by offering innovative, user-friendly, and affordable digital solutions that help SMBs grow and compete efficiently. The reliability of our mission-critical products, backed by the dedicated support of our personal consultants, has helped us to develop and maintain a very loyal customer base. In the first half of this year, we successfully onboarded 150,000 new customers, increasing our total customer base to approximately 6.47 million.
We increased our ARPU by around 4% compared to last year on the back of successful up-and-cross-selling, as well as our regular price adjustments. Customer growth, high ARPU, and strong ad tech performance, which we will elaborate on in detail later, led to revenue growth of around 19% in the first half year. Furthermore, IONOS' platform model and substantial economies of scale, combined with robust growth, have led to consistently high levels of profitability for the company, which in turn led to an Adjusted EBITDA growth of around 23% in the first six months. Our core offerings, supported by reliable and personalized customer care, continue to drive customer loyalty. We successfully added 70,000 net new customers in the second quarter, bringing our total net additions to the first six months to 150,000, as already mentioned.
This represents a significant improvement compared to the 90,000 net additions in the first six months last year, underscoring the strong alignment between our product offerings and customer needs. Our average revenue per user continued to develop positively and robustly, driven by successful upselling and cross-selling initiatives, as well as targeted price adjustments. Importantly, churn remained largely stable throughout the transition, underlining both the strengths of our customer relationship and our pricing power in the market. Artificial intelligence continues to boost internal efficiency and reduce operational costs while enhancing customer-facing features and creating new revenue opportunities through additional use cases and cross-and-up selling. AI is currently integrated in most of our product lines, with more features to come every quarter going forward. Our focus remains on hosting, boosting efficiency, driving innovation, and delivering measurable value to our customers.
For small and medium businesses in particular, our AI products provide access to advanced technology, previously available only to large enterprises. All of this with a strong emphasis on full data privacy and compliance. In 2024, we launched the AI Model Hub, a sovereign multimodal platform hosted in the IONOS Enterprise Cloud. Building on that, in April, we introduced IONOS GPT, an intuitive interface that empowers businesses to incorporate their internal documents, databases, and content securely processed and stored in our European data centers. We launched IONOS GPT in Germany. Now we are preparing for a broader rollout across additional European markets, and we will also expand the range of functions with further AI models, new features, and enhanced personalization. One of the most exciting developments in AI is the rise of intelligent virtual assistants, which we believe will fundamentally transform how SMBs operate.
Today, small and medium-sized businesses spend only 50%- 60% of their time on core business activities. The rest is tied up in operational overhead, managing inquiries, handling appointments, and running marketing tasks. With AI agents, we believe this number can rise to 80% or even 90%, freeing up valuable time for business owners to focus on what truly matters: their customer and growth. AI agents bring a number of fundamental advantages to the table. They're always available, never get sick, never forget, and can work 24 hours a day, seven days a week, across multiple channels simultaneously. This makes them ideal for automating repetitive, time-consuming, and low-value tasks across the business. We see two categories of agents playing a key role. The first are general-purpose agents that require minimal setup. By scanning a website, online shop, FAQs, or other digital content, they can start delivering value immediately.
These agents can manage first-level support, handle customer inquiries, and provide product recommendations. They also support marketing tasks, creating newsletters, helping with online ads, or generating and scheduling social media content. These use cases align closely with our existing products, like website builder, email marketing, and online shops, and offer fast time-to-value without complex integrations. The second, even more powerful category includes deeply integrated agents that connect directly to backend systems, CRMs, calendars, POS systems, databases, or reservation tools. These agents can automate entire operational workflows. A restaurant could use them for booking and shift planning, a hairdresser or doctor's office for scheduling, or a travel agency for handling requests and follow-ups. Today, SMBs typically buy only one website, but they can deploy several AI agents across their business in the future. For example, one for customer support, one for scheduling, and one for back-office tasks like invoicing or HR.
This significantly expands the addressable revenue per customer. While traditional website products generate EUR 10-EUR 20 in ARPU, virtual assistants are expected to start in the range of EUR 20-EUR 50 per agent. Over time, the use of multiple task-specific agents per customer opens up entirely new revenue layers for us. IONOS is uniquely positioned to lead both segments. We serve a large and loyal SMB customer base, have decades of experience building digital solutions, offer excellent personal support, and have the technical platform and development capacity to integrate AI deeply into all business workflows. Our first agents are already in beta, with general availability planned by the end of this year. At IONOS, we're not just adopting AI; we're shaping how AI delivers real, measurable value for SMBs.
Europe stands at the threshold of a transformative era shaped by evolving global political dynamics, offering significant opportunities for the future growth of IONOS. The pursuit of digital sovereignty, specifically independence from U.S. hyperscalers and full data control, is becoming increasingly critical. This trend is set to significantly impact not only cloud solutions but also web presence and productivity services across sectors, companies, and organizations of all sizes. We are seeing a significant increase in interest around digital sovereignty practically in our cloud, sovereign cloud offerings. This growing awareness spans across companies of all sizes, including public sector institutions. However, in the enterprise and public segments, sales cycles tend to be long, typically ranging from 9- 12 months. Cloud projects in these areas uniquely require detailed planning and are often implemented in several phases.
While this means revenue materializes more gradually, the current level of interest provides a strong foundation for future growth. In June, IONOS, together with our partner O'Keeffe, submitted an expression of interest to the European Commission for participation in the Invest AI initiative. This program aims to establish up to five large-scale AI gigafactories across Europe to meet the continent's growing demand for sovereign, high-performance AI infrastructure. Our proposal outlines a facility with an initial capacity of 50,000 GPUs, scaling to over 100,000 in the long term. It is designed to support large-scale AI workloads with a fully sovereign and sustainable ecosystem, with high-efficiency cooling, European data protection standards, and deep integration with our cloud infrastructure. It's important to note that this is currently an expression of interest only, the first non-binding step in what will be a competitive multi-phase selection process.
The final rules for participation, including funding frameworks and procurement guidelines, are still being worked or developed by the European Commission. We expect a formal call for proposals to take place later this year, with decisions and legal structures expected in 2026 and initial developments from 2027 onwards. The EU has announced that up to 35% of the capital expenditure could be subsidized, depending on the quality and strategic fit of each proposal. This level of co-funding is an important signal, but we want to be clear: we will only move forward if the final framework allows for a sound business case based on commercially viable terms, long-term demand, and a stable regulatory environment. This is a significant potential investment, and we will approach it with the same discipline we apply to all our strategic initiatives.
With over three decades of experience in mission-critical hosting and cloud infrastructure and a strong track record in security, we believe IONOS is very well positioned to play a key role in this initiative. That said, we see this project as a long-term opportunity, not a short-term risk. It aligns with our broader strategy of enabling Europe's digital sovereignty and supporting the next generation of AI innovation, always under the right conditions and with clear economic value. At this point, I would like to hand over to our CFO, Britta , to talk about our financials before we start this Q&A session.
Thanks, Achim, and welcome as well from my side. As you know, with the full year 2024 results publication, we introduced a new segment reporting to enhance transparency and better reflect our operational structure. Firstly, digital solutions and cloud, which includes web presence and productivity and cloud solutions. This segment generated EUR 656 million revenue in H1 2025, accounting for approximately 73% of our total revenue. The Adjusted EBITDA margin in this segment is continuously strong, standing at 36.1% in the first half of this year. If we look at the two included business lines separately, we can see that our web presence and productivity business generated EUR 544 million in revenue in the first half, accounting for approximately 61% of our total revenue. Our cloud solutions business generated EUR 90 million in revenue, accounting for around 10% of our total revenue.
Our ad tech segment on the right side generated EUR 239 million in revenue in H1 2025, accounting for approximately 27% of our total revenue, with an Adjusted EBITDA margin of 13.3%. Overall, we are extremely happy with the first six months. Total revenues, as mentioned before, reached EUR 895 million, reflecting 19.1% year-on-year growth. Adjusted EBITDA came in at EUR 268.7 million, up 23.3% compared to the previous year, resulting in a 30% Adjusted EBITDA margin compared to 29% in the first half year of 2024, again underlining the strong operational leverage of our business model. Marketing investments were slightly higher than last year, as we are growing our business and continue to invest into brand building. Adjusting for the marketing expenses to make it comparable, Adjusted EBITDA would have been EUR 281.3 million, growing 29% year- over- year. Looking at the second quarter, total revenues increased by 18.5% to EUR 448.7 million.
Adjusted EBITDA increased by 22.7% to EUR 137.7 million, resulting in 30.7% Adjusted EBITDA margin. Adding back the EUR 8 million higher marketing expenses to make it comparable, Adjusted EBITDA margin would have been 32.5%. Let's take a closer look at the performance of our segments. The digital solutions and cloud segments achieved EUR 326.4 million in revenue in the second quarter, making marketing a 6.7% increase year- over- year, or 7.1% excluding intercompany revenues. The Adjusted EBITDA margin improved significantly to 38%, up 4.3 percentage points from Q2 in the last year. Our ad tech segment reported revenue of EUR 122.3 million, representing impressive growth of 68.3% year- over- year, based on a weak prior year and supported by the positive development of the ongoing product transition. The EBITDA margin was at 11.1% compared to 12.8% in the previous year.
As previously emphasized, we anticipate temporarily lower margins as a result of the ongoing product transition. I will provide more detail on the ad tech dynamics shortly. Let me reiterate what Achim already briefly touched. In the first six months of 2025, we added 150,000 net new customers, which is well above 90,000 net additions in the first six months last year. As outlined in previous webcasts, we expect stronger customer growth in 2025 compared to prior years. The strong performance in Q1, with 80,000 new customers, was not an outlier. In Q2, we added another 70,000 customers, continuing this positive momentum. Demand for our products remains solid. Customer growth continues to have seasonal variations, with slightly lower growth during the summer months and the strongest customer growth in the first and fourth quarter, but we expect customer growth overall at a higher level compared to last year.
This positive development underscores the strengths and appeal of our offerings, as well as a successful execution of our strategic initiatives. At first glance, ARPU growth may seem modest. This is primarily driven by different phasing and scale of price adjustment initiatives following significant changes in our pricing structure kicked off in 2023 and rolling into 2024. As you might remember from previous webcasts, ARPU in the second quarter is typically slightly lower than in the first quarter. This is a seasonal effect, as we typically have many domain renewals in the first quarter, where revenue for 12 months has to be recognized upon renewal. Additionally, revenue growth in our cloud solutions segment is slightly lower, which is also impacting ARPU, as ARPU for cloud products is naturally higher than in web presence and productivity. In total, we are seeing ARPU and customer growth of approximately 4% each.
This puts us firmly on track to meet our full-year revenue growth target of 8%. Going forward, our pricing approach remains balanced and strategic, aiming to attract new customers, maintain competitive positioning, enhance customer satisfaction, but also implement price adjustments where appropriate. Let's have a look into the performance of the different business lines within the digital solutions and cloud segment. In the second quarter of 2025, our web presence and productivity business grew by 7.4% year- over- year to EUR 270.5 million, driven by continued customer growth and higher ARPU compared to the previous year. Excluding FX effects, which were negative in the second quarter, revenue growth is 7.9%. Our cloud solutions delivered revenue growth of 5.2% to EUR 45.1 million, or 5.9% like for like, excluding FX effects. Let's have a deeper look into the cloud business.
In Q2 2025, cloud solutions revenue reached, as mentioned before, EUR 45.1 million, representing a 5.2% year-over-year increase. The changing global political landscape presents significant opportunities for future growth at IONOS. Nevertheless, as Achim already pointed out, we are seeing increasing customer demand, but the translation of higher demand into revenue growth needs time. Looking at individual product areas, public cloud grew by 10% in the second quarter. Revenue in this quarter does not include any material revenue contribution from the ITZ Bund project. As a reminder, the majority of revenue from ITZ Bund is recognized progressively, aligned with the deployment of hardware blocks in the data centers. We have finished setup and commissioning of our cloud in the ITZ Bund-owned data centers, followed by an intensive test phase for the first hardware blocks. Technical sign-off has been granted, and we are now in real-life operations.
We therefore anticipate meaningful revenue contributions from ITZ Bund and an acceleration in growth in the second half of the year. Private cloud revenue grew by approximately 3%, while managed cloud increased by 4%. Our current focus is on driving customer adoption of our cloud solutions and to convert the high demand for data-sovereign cloud products into growth, both supporting our long-term growth strategy. In our ad tech segment, we've seen another strong quarter. After a very strong first quarter, revenue grew by 68% year- over- year in the second quarter, based on an exceptionally weak prior year quarter and supported by the positive development of the ongoing product transition, where we have seen both products performing strongly. We are closely monitoring the impact of the ongoing product migration to RSOC.
While we are confident this transition will ultimately benefit our business, we are aware that it may lead to some short-term volatility. In fact, we have seen very strong first two quarters in this year. However, it should not be forgotten that this is a new product for all market participants, and we expect that optimizations will be necessary in order to have a strong product in the long term, which in turn will result in a temporary decrease in revenue growth in the coming months. Nevertheless, we are encouraged by the better-than-expected performance in the first six months and by the RSOC onboarding process, which is well underway. We are committed to driving this transition forward and are confident that it will have a positive impact on our ad tech business.
Turning to our capital expenditures on slide 19, we can see that our total CAPEX in the first half was EUR 23 million, or 2.6% of our total revenue. This is a slight decrease from the previous year, where our total CAPEX was EUR 32 million, or 4.3% of our total revenue. Breaking down our CAPEX, we can see that our gross CAPEX was EUR 19.6 million, or 2.2% of total revenue, which is mainly from the expansion of our cloud solution capabilities. Our maintenance CAPEX was EUR 3.4 million, EUR 3.4 million, or 0.4% of our total revenue and remains low and predictable. CAPEX was notably low in the first half of the year, primarily due to the phasing of investments throughout the year. Looking ahead, we project total CAPEX for the full year at around EUR 80 million.
This is slightly below our internal initial target range of EUR 80 million-EUR 90 million, or around 5% of expected revenue. This reflects disciplined investment to support innovation, growth, and operational scalability. Let's now walk through our free cash flow for the first half of 2025. Starting with Adjusted EBITDA of EUR 269 million, we deduct EUR 10 million in adjustments, which are mainly standalone and LTI-related costs, to arise at reported EBITDA. We then include EUR 23 million in CAPEX, EUR 32 million in tax payments, and payments in relation to the settlement of the long-term incentive program obligations. The settlement was made net, which means that the obligations were settled with treasury shares after deduction of the individual taxation of the recipients, but the tax portion is then paid to the tax authorities by us. After including working capital, we reach a free cash flow before leasing of EUR 176 million.
Subtracting EUR 8 million in lease payments gives us EUR 168 million in free cash flow after leasing, which is well above the EUR 151 million in the first half of 2024. Further, we made EUR 27 million in interest payments and executed EUR 37 million in share buybacks. After factoring in these items, comparable free cash flow for the quarter stands at approximately EUR 105 million. In the first six months of 2025, net debt was reduced to EUR 786 million. This includes both external bank debt and the shareholder loan from United Internet. In the second quarter, we repaid EUR 70 million, reducing the shareholder loan to just EUR 100 million, which we plan to further reduce over the coming quarters. Our net leverage ratio improved to 1.6x net debt through Adjusted EBITDA compared to 2.4 x a year ago.
The weighted average annual interest rate stands at 4.9%, which will go down further with the upcoming repayments of the shareholder loan. This improved debt profile supports financial stability and provides flexibility in the future. Looking into the outlook for 2025, for our core business, digital solutions and cloud, this is unchanged, and revenues are expected to grow by around 8%, with web presence and productivity growing at around 7%- 8%. We now expect cloud solutions to grow by around 10% compared to around 15%- 17% before. We remain confident about the midterm growth opportunities in cloud solutions and expect our public cloud to get around 20% growth over the next quarters. Adjusted EBITDA margin is still expected to be around 35%, up from 32.9% in 2024.
Based on the better-than-expected business development in the ad tech segment in the first quarter, as well as our ongoing cost control, we already specified our outlook for 2025 in May and increased the expectation for revenue growth in the ad tech segment to around EUR 400 million compared to above the previous year's level before. In May, we already increased the outlook for the total Adjusted EBITDA from EUR 510 million-EUR 520 million. Due to the overall positive development and, as mentioned, the continued cost discipline, we are now increasing our guidance again and expect Adjusted EBITDA to grow by approximately 17% to around EUR 530 million. That concludes our presentation for today. Hopefully, we provided you with a comprehensive overview. We will continue to work hard for our customers, improve our products and services, and strengthen our market position.
With this, I would like to hand back to the operator to open the webcast for any open questions.
We will now begin the question and answer session. Anyone who wishes to ask a question from the webinar may click the Q&A button on the left side of the screen and then click the Raise Your Hand button. If you connected via phone, please press star followed by one, in which case you will hear a tone that confirms that you have entered the queue. If you wish to remove yourself from the question queue, you may press the Lower Your Hand button from the webinar or star two on your telephone. Anyone who has a question may queue up now. Our first question comes from George Webb, Morgan Stanley. Please go ahead.
Hi, and morning, Achim and Britta. I've got a few questions, please. Firstly, just as we think about ad tech, could you give us an update on how far through the transition you are at the end of the third quarter? Maybe you won't be willing to give this, but if you could add any color on what you're seeing in terms of growth rates on the AFD side versus the RSOC side, that would be interesting. Secondly, on cloud, I mean, I guess what's driving the downtick there on your expectation for 2025, and we wouldn't have expected, I guess, ITZ Bund to have deliveries in Q2. That doesn't look like the surprise, but I guess that public cloud business, even in the NICS, going at 10%, you know, is not strong compared to some of the regional players and certainly not the hyperscalers.
What's driving that downtick and how do you think about your overall cloud growth rate? Lastly, on the revenue target for digital solutions and cloud, you've kept that at 8% for the full year, with that cloud guidance reduced from 16% at the midpoint to 10%. I think that's about a 70 basis point headwind. Are you implicitly expecting a slightly stronger performance from the core business compared to previously, or is that just rounded in that 8% number? Thank you.
Let me maybe start with the ad tech business, and Achim, you might take the cloud question. The transition from AFD to RSOC is well underway. We saw RSOC overtaking AFD in May. The picture changed a little bit after that with more quality initiatives coming in from Google, which the market is currently tackling. Overall, as mentioned, we are still optimistic to see the EUR 400 million growth, largely driven by RSOC compared to AFD, which was definitely stronger in the first half, but it will not be very strong in the second half.
Yep. The same question was about the cloud growth. First, again, the cloud segment or cloud business segment comprises a lot of different services. There's the public cloud, there's private cloud, MSP services, there's a VPS service, different things in there. A few of them are not strategic. You know, we acquired them, for example, and we worked hard to acquire some MSP business with that. This is not strategic, so it's not really growing much. We have now a much larger partner network for the cloud than trying to do it ourselves. The other reason, so that the mix effect shows lower numbers than what is the most important part of the public sector, the public cloud growth.
The other thing is that we are, of course, we're pretty busy with really installing and getting the ITZ Bund project operational, which just happened basically a week ago, with all the testing and final approval of the ITZ to get this live. Now the first customers are on that platform. It's kept us pretty busy. That's why, like Britta said, we expect now a faster growth again in the cloud segment for the rest of the year.
Answering then your second question, we would see that presence and productivity being in line with expectations, but with the cloud solutions uptake, which we would see over the next quarters, that adds up to the rounded 8% revenue growth for the year.
That's helpful. Maybe one last question, Achim. When we think about the AI gigafactory initiative, and let's see how it pans out, from the investor side and certainly from our analyst side, how should we be thinking about the steps from here in the timeline and kind of headlines that we should be looking for?
I'm sorry. From the timeline, I think now is a lot of commenting phase. We are talking to a lot of politicians now in Germany, for example, and also in the E.U., because they are trying, they are making up their mind how this program is going to be set up. Is it the PPP, like public-private partnership? Is it going to be some other sort? What are the requirements for funding and so on and so forth? Basically, every other week, we send some paper and some statements of what we think the direction should be. I think that's continuing until the end of the year.
That's at least the idea of the European government to say, okay, we'll figure this out until the end of the year, and then they will start with the real tender, and then we need to see if we want to apply or not. There's plenty of time for us to get prepared. I mean, we're doing our business as usual. There's a lot, I mentioned, a lot of AI things coming along, so there's huge demand for AI capabilities. It's in our plan in some ways anyhow. If that project works out well and makes a lot of sense for us, we would take the 35% subsidized. Why not? That's basically how we look at it.
That's great. Thanks, and good luck for the second half.
Thank you.
Our next question comes from Toby Ogg, JP Morgan. Please go ahead.
Yeah, hi, morning, and thanks for the questions. Perhaps just the first one just on the AI agents. You mentioned there are EUR 10-EUR 20 sort of traditional ARPU, and then I think it was EUR 20-EUR 50 ARPU that you mentioned with virtual assistants. Could you perhaps just walk us through the mechanics of what the approximate buildup of that could look like for a customer if they were to, say, move from EUR 10-EUR 20 ARPU today to EUR 20 -EUR 50 over time with agents? Just how you're thinking about the revenue model for AI agents. Is this going to be a monthly subscription fee on top of, say, the baseline website builder, or do you envisage any consumption-based revenue streams? Just second question, just again on the AI gigafactories in your submission here. Appreciate it's still early. You talked about up to 35% of CAPEX being subsidized.
Any sense around how the rest will be split, what the total spend could look like, and any early thoughts on how you plan to fund that? Thank you.
Okay, so with the AI first, it's a very early phase. The products will get released, though, not just with our companies, but honestly, there's a lot of people working on it, I'm pretty sure. The pricing in the market is not really set yet. That's our expectations. That's how we would start. Depending on the AI agent, if it's a small, kind of small agent doing just a very specific task, it might be cheaper than having one very powerful AI agent taking care of half of your business, of course. It will be very likely a monthly subscription, like everything else we do. It will be on top. It's just different separate products.
Now, we are enhancing our website builder product, for example, with AI capabilities all the time, and we have separate models, separate modules as well, which are not a monthly subscription, but AI agents are really a separate category, new product line coming along. It will be just much higher in ARPU, and typically, companies can use more than one. In websites, you usually need one. Emails, you need a lot. Shop system, you need one. AI agents, the more we build, I think the more we'll find use cases, or customers will find use cases to use the AIs. Compare a, let's say, EUR 50 a month AI agent taking a job, it's like a mini job for now, which is in Germany, for example, like EUR 500 or EUR 800. The comparison is, it's not completely out of pocket.
It's a huge saving for companies with all the advantages of 24/7, never sick, and so on and so forth, multi-language, multi-everything. I think a price of EUR 50 for a virtual AI employee is really on the cheap side, rather. We will see how the market is developing. It's really everything is at the beginning. We'll see how the dynamics go, how we start offering, what the adoption of the market is, what the competitors are doing. For sure, everyone is now on the brink of having new product lines with additional revenue coming in. It's not subsidized, and it's less packaged within other products. For the AI factory question, how you finance it, for such a huge AI factory in the full scale, we'll run roughly EUR 4 billion-EUR 5 billion. That's, of course, a very, very substantial investment. You have to deduct the 35% from the government.
We have O'Keeffe as a partner, and O'Keeffe is very specialized in building data centers and running data centers on behalf of others. They will probably take care of all the shell, which is EUR 1.5 billion bytes old. The rest we'll split in the consortium. We're not the only one. We have partners like O'Keeffe and others. We'll see how we finance it. There are infrastructure funds we could use. In the end, we're not planning on having a huge amount of CAPEX for this project. To be clear, the financing strategy is not made up yet because we don't even know how the structure with the government should look like or could look like. This is all work in progress, but I think it's pretty clear that we don't want to invest a huge amount of money. We built this up in a very sensible way.
Yeah, let me maybe add on this. As Achim already mentioned throughout the webcast, obviously, we are looking into the business case now, trying to get our heads around the economic model in discussions with the European Commission, et cetera, trying to understand how it will work out. Overall, I think we are, in terms of debt, if we just look at our deleveraging profile, we do have a sensible amount of financing available if needed. As Achim mentioned as well before, it depends as well how many partners are in the consortium, et cetera, but we think we are well positioned, and we see a strong support from the banks we are working with.
Great, thank you.
Next question comes from Dhruva Shah, UBS. Please go ahead.
Morning. Thank you for taking the questions. I have three, if that's okay. First, just starting with cloud. Cloud growth was muted again this quarter. We know ITZ Bund revenues should boost that in Q2, but excluding that, in terms of the underlying growth, do you expect that to accelerate mid-term? You've talked a bit about digital sovereignty as a theme, but perhaps could you be a bit more granular and share some details on how you'll transform that thematic tailwind into more tangible financial growth, both for the top line and for profitability? That's on cloud. Second is on WPMP, the core business. Customer acquisition was strong for another quarter, but ARPU growth slowed. In terms of the mid-single-digit price rises in that segment, how have they landed so far? Are you seeing any impact on NPS or volumes?
Finally, the third question is a bit more broad in terms of financials, but the federal cabinet in Germany has recently approved 30% accelerated depreciation now through to the end of 2027. I was just wondering if there are any early indications of whether the maintenance and cloud investments by IONOS are likely to qualify for this and what the potential uplift to earnings could be. Thanks very much.
Maybe let me start with the cloud. Yeah, we do expect a growth in cloud. First of all, you know, the ITZ Bund now is live, so they start using the cloud, and cloud is consumption-based in the end. That will drive the revenue. The other thing is, like I already said, the pipeline is filling because there's a lot of interest now. People, you know, a lot of larger companies started talking to us about, you know, what could you offer? What's sovereignty in your case? We can be, I think we are more sovereign than everybody else in Europe because the full stack is developed by us. It's either open source or our own software. Engineers are sitting here in Berlin. This is more European, I think, than it gets, or nothing that gets more European than what we do. Let's phrase it that way.
Now it's kind of the discovery phase. Large companies who are usually with the hyperscalers now look around and see, okay, these are alternatives. What can you offer? What are the workloads which would transition? Now we're starting the light, you know, sales cycle, in some cases means like testing something, you know, looking at if the system works well in our environment and so on and so forth. We expect these leads which are coming in right now to materialize over the next 9- 12 months, which is typically a cycle where, you know, from thinking, doing the tests, looking at the transition project, moving some workloads, starting to generate noteworthy revenue. From a profitability, I think we're still keeping it roughly break-even. That's what we did in the past.
I think right now is not the best time to, or would be not the greatest time for trying to pull out a lot of margin because now it's developing. That market is now developing. We should not underinvest. I think the investment is, you know, what we earn, not more and not less. You don't cross-finance it heavily or anything. I think that's the way to go forward.
Exactly. Let me comment on the web presence and productivity with the ARPU growth in Q2. I think there are several things to keep in mind. First of all, as mentioned during the webcast, Q1 is usually extremely strong, whilst Q2 therefore looks a bit odd. Let's keep that in mind. Obviously, just given how ARPU is calculated, when a lot of new customers are coming in, which are usually on a lower ARPU, as we have 6- 12 months starting discounts, and the ARPU of the customer is building up over time, that is partially diluting ARPU growth. I would more look into how ARPU is developing overall. As I said, it's growing roughly 4% year- over- year, which we believe is strong. Price adjustment initiatives are still rolling in, so we would see them as well stronger going forward.
In terms of NPS, yes, we do see a slight downtick in NPS, not unexpected. We are confident to get it back to levels which we had before. Keep in mind, it's only very slow, but obviously, you will not avoid that some customers might be unhappy with NPS. This is reflected in the NPS.
Surprisingly positive level.
It is not to an unexpected amount, and it is not like we do not get it back up. We are still confident and around the churn levels as well, which are in line with expectations. All fine, I would say. Could you please repeat your third question, please?
Yeah, of course. Thanks for the first two. That's very, very helpful in terms of the color. The third is just on the accelerated depreciation in Germany, the 30% accelerated depreciation through to the end of 2027. I'm just wondering if you had any early indications of whether the CAPEX that IONOS could qualify for that program and whether there's any rough indications of potential uplift to earnings as a result of this.
No, the question is no.
Thanks very much.
You're welcome. The answer is no, sorry.
The next question comes from Sarah Roberts, Barclays .
Hi, good morning. Just three from me, if that's okay. Firstly, just as a follow-up, your cloud guidance of 10% year-on-year growth implies a 2H growth rate of about 13%. I just want to understand how much of that is coming from the ITZ Bund contract and how much of that is coming from underlying cloud drivers. As a follow-on, can you talk through what you're seeing in terms of cloud demand in your underlying core SMB cohort? It seems as if enterprise growth drivers are pretty strong in the mid-term from digital sovereignty, but I think in the past you've spoken about SMBs being a little bit more hesitant, particularly in light of the fact that macro is a bit uncertain. Is that still the case? Any color you can add here would be great.
Secondly, my question is on the Adjusted EBITDA upgrade, about EUR 530 million versus EUR 520 million previously. I just want to understand in a little bit more detail what are actually the main drivers of this coming through. Is it simply a factor of cloud expectations are lower and that comes with lower margins? Is this a mix effect or is there underlying cost improvements that we should be aware of? Thirdly, there's been a bit of noise with the rise of low/no-code AI web builders that have been rapidly gaining user base over recent months. I just want to understand whether you're seeing any changes in the competitive environment and how you are thinking about the potential threat from these new AI startups that are coming through over time. Thank you.
Let me start with the cloud ITZ Bund and the uptick over the next quarters. A large portion will be driven by ITZ Bund. As we mentioned before, we do see a strong demand, especially as well from the SMBs. Let me phrase it that way. We are more cautious now in turning it into revenue as we see from the past that those projects can take longer. Even if those customers are locked in into cloud, it takes longer to develop the revenue over time, given their capabilities from smaller customer size in order to build up their own cloud. That takes a little bit longer. Especially on the smaller side of customers, we really see strong demand coming in and see a lot of new customers joining us in cloud solutions.
In terms of the Adjusted EBITDA upgrade, this is mainly a driver of cost discipline, where we do see a strong development driven by several initiatives which we internally started. Obviously, as well, the very strong development of ad tech in the first quarter adds its bits, still being very profitable. Overall, it's more cost improvements rather than any and the operational leverage in the business, I would say. We are growing well. Nothing unexpected, I would say, if you look into the business model and how strong it is from an operational leverage perspective. Last question, I think.
Yeah, the no-code paper also for the SMB in the cloud. I think we doubled the inflow per quarter from last year for small, medium businesses. Of course, ARPU is lower, and you know they need a little more help, but that's exactly what we are very good in, you know, compared to the hyperscalers. We have very much closer ties to small, medium businesses by those support levels for them and so on and so forth. It is a valuable customer group now picking up substantially. For the AI, low-code, no-code, you know, lovables and these kinds of companies, actually, they are our big friends because lots of the agents we're building, lots of the AI features we're building, we're using or we're partnering with small, smaller companies who have great ideas, who have already started with a product. We don't have to develop everything ourselves.
I don't see them as competition. Of course, they are in some ways, but they tend to, you know, line up with us. Why? Because we have a huge customer base. We have support. We have billing. We can do mass market. We can, you know, support millions of customers, and they have just a product idea. From having a product idea and maybe a prototype on the technical side to come up to scale and have six and a half million base customers to talk directly to in such a marketing machine and brand machine that we have, that's very appealing to them. I am very happy for all the small customers, our small, small, you know, startups and smaller companies with good ideas coming to us.
I think we're working with like roughly 20 different companies together right now to see what other products they offer and how we can integrate them in the next 6- 12 months. Basically, every week there's one more or two more. We don't see them as a threat. We see them as an enabler for us to be even faster with more agents and more AI features to come to the market sooner.
Great. Thank you.
Next question comes from Stephane Beyazian from ODDO BHF. Please go ahead.
Yes. Thank you. I've got a couple if that's possible. Let me start with the AI gigafactories. First, if I understand well, you target the whole package across Europe, but would it be fair to assume that you could be only selected for one factory, for instance, at a cost of EUR 1 billion? You also mentioned that you need guarantees on long-term demand. How could you be guaranteed that? What makes you confident that you can be successful in AI cloud? I guess my question there is, you know, how close that business is to your current cloud solutions business? I've also got another question on ad tech revenues.
I was just wondering if you already see the signs of lower activity or if the visibility is so low in that division that actually you may end up surprised positively again as much as what you did in the first half. Thank you.
Okay. For the AI gigafactory, the European government wants to support five. They're talking about four to five, but I think it's going to be five different AI gigafactories, and each of these gigafactories is around EUR 4 billion to EUR 5 billion. This is a huge project across Europe, and I'm pretty sure there will be one, maybe even two of these gigafactories in Germany. There will be for sure one in France. There probably will be one in Spain. That's what the idea of the European Commission is, to really boost or foster Europe and give some, you know, help for companies like us to establish a huge European AI cloud market. We are applying only for one. We're not doing it across Europe. We only apply for one in Germany because we think that's the, you know, it's our strongest position here on our home turf.
I know lots of politicians. I think we are most suited to win the German pitch instead of trying to do it in Spain, where we also have, you know, companies and stuff. I think, you know, Germany is the most obvious one for us. You're talking about the sales guarantees. Of course, we don't get sales guarantees from the market, but we can get sales guarantees from the public sector. That's part of the initial idea of the E.U. to say, okay, you know, we want to also, we want to subsidize 35%, but we also want to guarantee a certain amount of workloads. This is all part of the package being worked out right now to see, okay, how much will they take off the resources?
For example, if you have 50,000 GPU cards, will there be a base customer or anchor customer for like 20,000 of these GPUs for five years? That would be nice because it's much easier to calculate on that level. It really fits very well in what we're doing because we are already selling AI cards, GPUs, and services on top of these cards, especially on all the new AI agents and all the products coming along. All of them need our cloud services. All of them need our AI services. It's basically the world we would have anyways. We would not build these huge data centers at once if we wouldn't have the load, but we would gradually walk into it anyways.
Now with the subsidized or the possible subsidized of the E.U., it would make much more sense, or it could make much more sense to now engage in this project. That's why we're looking into it and find out what the requirements in the end are and if it's a suit or is it suitable for us.
Let me comment on the ad tech. As I mentioned before, AFD is definitely on a lower trajectory by now, with RSOC still being optimized from both sides, from the Google side, but as well from our side, from the market side, so the platform providers. There are a lot of optimizations in the products going on. Overall, as we have already seen on the slides, we expect revenues to be slower in the second half of the year than it was in the first half. That's the current outlook, what we are currently seeing in the market and how the AFD business is currently developing with a lot of, as mentioned, optimization ongoing in the RSOC product.
Thanks for the clarifications.
You're welcome.
Our next question comes from Gustav Froberg, Berenberg. Please go ahead.
Good morning, everyone. Thank you for taking my questions as well. Three short ones from me. Firstly, on the cloud as well, I mean, growth is slowing. You mentioned it a little bit, but you are not thinking about cutting back CAPEX, incremental OPEX, etc. Would you consider reinvesting it at higher ROI levels elsewhere or conversely doubling down to reaccelerate any growth? Be curious to hear about CAPEX and OPEX in cloud. On the data center strategy for cloud solutions more broadly, I'm just curious about an update on your footprint there and what the pipeline looks like in terms of your actual data centers for the next 12- 24 months and whether or not we'll see any step changes in CAPEX. Last one on AI.
Could you tell us a little bit more about the uptake of AI solutions that you have already rolled out to some of your products and how you see this trending? What level of customer penetration should we expect for AI going forward? Thank you.
Let me start, and Achim, please chip in. In terms of cloud solutions, we think with investing what we earn, we are well positioned to drive the future growth, as well covering the demand, which is now stronger than it was before. We believe in terms of data center capacity, we are pretty well set up.
Yeah, I can take this. I mean, we operate like 32 different data centers right now, and 11 of them are our own. We just built one in the U.K. two years ago, went operational, which still has lots of capacity. You know, we built this in different steps. For the cloud business and stuff, I think we are pretty fine for the next two to three years from the capacity management. There's nothing to expect like building a new huge data center or anything else. The AI project, the AI gigafactory project is different. From the natural way going forward, there's nothing to expect, really. Uptake in AI solutions, I mean, the whole AI agent thing is coming to market soon. There's no value numbers right now.
For the core products where we also have AI integrated, we see, for example, in the website builder product, there's, I don't know if the number from my heart, but I think 60% of the customers are using the AI version already, or you can have both ways. You can do it manually, like in the old days, which are not so old, but then you can also do the AI. You can choose, you can just switch between. We see like 60% already using the AI approach to create the website. This is not an additional product. It's just an enhancement of the website building software we have. It's a bit difficult. There are a lot of other projects where we have AI integrated. In many ways, it's part of the base feature and then something on top. It's a bit difficult to come up.
I think we don't really report these numbers, how much of the revenue is already AI, and try to split everything in all the product lines. I think we are not.
It's pretty hard because it's integrated in most of our product lines, to be honest. It's more or less seamlessly integrated.
Yeah. From the usage, I can tell you, of course, I mean, I guess every one of yourselves makes this experience using AI more and more every day. You know, even if it's only ChatGPT for, you know, writing a text or something. There's no way back. The usage is increasing every month, in all the different product lines, all the AI features. First, because they're getting better, more feature-rich, and the content is getting better, what the AI generates. The quality is improving every quarter. People are getting more and more used to using the AI support. There's only a way forward. It's only going to be more. At some time, we might release separate numbers for AI usage or revenues, but we don't do this today.
Okay, great. Thank you.
As a reminder, for questions from the webinar, please use the Q&A button on the left side of the screen and then click the Raise Your Hand button. If you connected via phone, please press star followed by one. Our next question comes from Nizla Naizer , Deutsche Bank. Please go ahead.
Thanks. I have two questions from my end. The first is maybe a bit of a combined one. Related to the AI topic earlier, Achim, could you remind us what's the investment from IONOS's end to deploy these agents? You mentioned that you're working with several partners. Is the potential revenue that you get from an incremental step-up of this revenue high margin or low margin compared to your typical WPMP margin? Some color there would be great. Linked to that, when you look at your leverage, where it is, would you even consider inorganic opportunities where you acquire certain partners? How are you thinking about the future M&A strategy with this sort of opportunity in mind? That would be question one. Question two, 70,000 net ads in Q2 was a lot more than we expected and quite impressive in our view. Where are these customers coming from?
Are there certain geographies that are doing better than others? Any color on maybe how the U.S. is also performing linked to that? How should we think of growth in the next two quarters on the customer base, which I guess is important for the long-term story? Thank you.
Yeah. AI costs and margins, I think the margins are going to be huge. Typically, the AI costs are low because it's only inference and services. Training a large language model or training a big model is very compute-intensive. It's very costly. Using the ready-made model, with the weights already calculated, just the inference, running a query through that model is much cheaper and does not take a lot of CPU resources for a simple question or for a daily task. We don't see that there is going to be a huge cost involved in these services. If you talk about EUR 20-EUR 50 per agent, there will be a huge margin, much higher probably than everything we've seen so far in different product lines because that's hardly any cost associated with it. With the partners, you said inorganic opportunities, of course, there are.
The first step is not like going out now and trying to buy a lot of small startups. The idea is, partner with them, see what works, and then we can still do the next step. We could either program the whole thing because we would have the capabilities. We really want to find out what the market is expecting or accepting. What are the use cases now? What are the adoption rates? What is the gate product? It's just product development right now still. Once we are on a more clear roadmap, we can think of, is there one of the partners which wants to sell hidden server, we can acquire a stake. There are lots of opportunities. Especially if you're one of the high or the main partners of these startups, which is our intention, being one of the largest customers of them, we have tons of opportunities.
That's what we did in the past as well in some cases. Yes, there are organic opportunities. The 70,000 customers, do you want to answer us with that?
As we already mentioned throughout the webcast, the Q2 numbers haven't been an outlier. We see continued strong customers joining us. Keep in mind, Q3 might be falling a little bit behind due to the seasonality, which we are usually seeing, and then a stronger Q4. Overall, we aim to continue the growth which we are currently seeing in customers. I think you asked around geographic split or something, if I'm not mistaken.
Yeah, that's just fine.
We do see the U.S. being relatively strong, as well as Germany and the U.K. Currently, Southern Europe is a little bit behind. However, this is nothing which we haven't seen before. The countries fluctuate a little bit. We are investing our marketing spend based on an ROI, so CLTV over CAC. Countries might fluctuate where we do see the best ROI coming in. Overall, I would say broadly in line, most of the countries with U.S. particularly strong, and Germany and the U.K. as well for the first half.
Very helpful. Thank you.
More than welcome.
Ladies and gentlemen, that was the last question, and this concludes our Q&A session. I would now like to turn the conference back over to Stephan Gramkow for the closing remarks.
Yes, thank you, operator, and thank you all for joining today's call. Please feel free to reach out with any follow-up questions you might have. Have a great day. Stay safe, and goodbye.