Elisa Oyj (ELISA) Earnings Call Transcript & Summary

Hello, everyone, and welcome to Elisa's Q2 2026 analyst conference call. I'm Vesa Sahivirta, Head of Investor Relations. And here we have a very familiar team, CEO, Topi Manner; and CFO, Kristian Pullola. We also follow the normal practice, and we start with the presentation followed by the Q&A. Topi will go through the highlights of the report and Kristian will elaborate more on financials. And now we are ready to start. So I give word to Topi. Please go ahead.

Thank you, Vesa, and welcome, everybody, to this Elisa Q2 earnings call. let's go right down to the main points of our Q2. During Q2, revenue was effectively flat and comparable EBITDA was up by 1.4%, especially driven by successful execution of cost measures. We were still way down by last year's competition in terms of mobile service revenue, but that was partly offset by good progress in fixed service revenue, which grew 2.2%. The overall telecom service revenue decreased 0.7%. International Software Services, the comparable organic revenue increased by 0.6%. This was largely due to the fact that customers were cautious in proceeding in their projects, given the geopolitical uncertainties and higher energy prices. And with that, we saw some license deals being postponed further to H2. What you should also note that in IndustrIQ, we sold earlier a small business in Brazil impacting the comparable numbers. In comparable cash flow, the comparison quarter was exceptionally strong. And during this quarter, the cash flow was solid at EUR 71 million, decreasing due to higher financial expenses and less favorable working capital development. And we were happy to note that the postpaid churn normalized during the quarter, further decreasing from Q1 levels and landed at 16.7%, in line with our long-term average for Q2 churn. In mobile postpaid subscriptions, we increased with 21,000 of which 15,000 were -- 15,000 subs were related to IoT and M2M. With these numbers, our market share in terms of mobile subs remained stable during third quarter as we wanted it to be. The fixed broadband subscription base is increasing nicely in this quarter with 3,500 subscriptions, and we do experience good demand, improving demand in terms of fiber business. As mentioned, our cost measures were successful during the quarter indicating also clearly that our transformation program is proceeding according to plan. At the start of the year, we moved to quarterly dividend and following that, the Board of Directors now decided for the second installment of the dividend, namely EUR 0.60 and that will be paid to shareholders on the 29th of July. Looking into the revenue development, the revenue landed at EUR 551 million. As stated mobile services weighed down given the last year's competition also divestment of EpicTV is something to note in this one. And during the quarter, we also ramped down and discontinued our public switched telephone network impacting the revenue a bit. Fixed services, as stated and equipment sales were supporting the revenue growth during the quarter. We're happy to note that EBITDA margin improved during the quarter to 36.5% and the whole EBITDA landed at EUR 201 million. This was driven by efficiency measures, the transformation program as stated and then in part also by reduced sales and marketing cost. In terms of telecom service revenue, I already mentioned that we were weighed down by last year's completion in terms of mobile service revenue. It is important to note, though, that the mobile service revenue started to grow on quarterly basis comparing with Q1. It is also noteworthy that in Q2 last year, we started the rollout of the so-called security features, hard bundled with our mobile subs. And we started that rollout with a large cohort of back book price changes, and that was supporting the Q2 '25 comparison quarter in terms of mobile service revenue. In line with the quarterly growth in the ARPU started to increase a bit during the quarter and landed at EUR 24. When we look at our mobile KPIs now, the key message is that they returned to normalized levels. In Q2 this year, the new sales prices actually continued to increase from Q1 and exceeded the level of Q2 last year. This was a positive development on the market. However, we need to note that on the market, the share of fixed-term mobile contracts has increased significantly, and that means that there is a longer time lag than previously in terms of the new sales prices moving to book. This means that, for example, if we acquire today a customer from our competitor, that customer might be moving into our book with a delay of 3 to 4 months, depending on when the customer's fixed-term contract ends with the previous service provider. This also means that the new sales prices impact to service revenue will come in with a delay, and it will be more visible in Q4 in particular. Churn, as stated, decreased from Q1 levels and landed at 16.7%, which is a tad below our 10-year average of churn, which is 16.9%. Also the mobile sales and marketing costs, including the voucher costs have been normalizing during the quarter, and there was a bit of a decrease from Q1 levels. So all in all, good to see normalized levels in the mobile business in the forward-looking indicators. And in the past, with these levels, we have been able to deliver solid growth. And as stated, that growth will follow with the time lag, assuming that the market stays on these normalized levels. When we look at our business segment by segment, in consumer business, the revenue was impacted by the divestment and the mentioned phenomena in terms of mobile business, it was good to see support coming from fixed services and equipment sales. And then the cost measures were successful in consumer business and the segment EBITDA improved with 2.1% EBITDA margin hitting 43%. In corporate customers, overall, a very solid quarter. Revenue weighed down a bit by equipment sales and also the discontinuation of the network supported by fixed services and in digital services and in particular, by high-margin hybrid cloud services and data services, which was encouraging to see. The cost management was successful in this segment and EBITDA improved with a very solid 4.2% for that segment. In international software services, in turn, a more challenging quarter. Comparable revenue growth was 0.6%. And Here, we need to remember that Q2 is seasonally typically the weakest in Elisa industry business and overall in software industry. We did see some license deals being postponed to H2 during the quarter. However, it is important to note that we did not lose any deals. Nevertheless, with the new CEO, [indiscernible], we will be starting now specific measures to improve the profitability of Elisa IndustrIQ business, looking into sales, looking into boosting revenue as well as capturing synergies on the cost cost side of things. In Estonia, in Estonian market, we saw solid progress, revenue increasing 3% on the back of mobile and fixed services also some support from equipment sales. EBITDA increased by 13%, driven by the mentioned service revenue growth and then also an accounting alignment internally in Elisa. So good work, solid progress in Estonia. We continue to be focused on implementing our strategy and now especially in terms of 5G and fiber as stated in 5G and mobile services, we see normalized mobile indicators, improving the outlook. And with fiber business, we are seeing good organic demand in fiber to the home as well as fiber to the building. And then as a new category of fiber business the data center connectivity comes in offering longer-term revenue support. I will come back to that in a minute. In international software services, as stated, we will be starting specific measures to boost the revenue, take home synergies to boost profitability. In simplicity and productivity, the cost measures have been successful transformation program is proceeding according to the plan. So we will be staying focused to implementing our initiatives related to all of these focus areas. In mobile business, 5G penetration grew during this quarter, more than normally during the quarter. This was on the back of a focused sales activity. So we should not expect this 5G penetration to increase at this rate on quarters to come. As stated, in fiber business, we are seeing some good momentum. The fiber subscription base continues to grow. At the beginning of July, we acquired in Lounea area, in Finland, a fiber network provider with some 8,000 customers, a bolt-on acquisition in that space. And then related to the fiber to the building, we announced a customer win, a partner win from DNA, the largest rental housing company, [ Lumia ] plc having 4,000 apartments in Finland chose us as their fiber to the building provider. So good to see that progress in the fiber business. Related to the overall fixed services during the quarter, we ramped down successfully our PSTN network. That network has been serving as well for 140 years. And now customers have been moving to new technologies. For our fixed service business, this also means that now the drag of decreasing PSTN revenue will cease to exist. And with that, we will be having a clean sheet for fixed service revenue growth going forward. This quarter marks the start of our large-scale data center connectivity business. During the quarter, we signed first large-scale data center connectivity deals. And when we talk about large-scale data centers, we talk about newly constructed above 100-megawatt data centers. On the overall, in Finland, the data center market is growing. We see more and more projects materializing and the ultimate size of the market will depend on many variables. One of them is the overall AI super cycle development and the investments of especially the hyperscalers, electricity availability and electricity costs will be impacting the investment levels and so will the regulatory environment. But it is very clear that Finland is an attractive place for data center operators. We have an optimal climate seismically stable land. One of the most reliable and best electricity grades in the world. low electricity prices and very developed telecom infrastructure. And we, as Elisa, we have clear coeditive advantages in this market. We have a strongest and widest backbone network in Finland and to and from Finland. We are the market leader with excellent capabilities to build fiber connections on time. And timely delivery is really important for the data center operators. And then, of course, we also have a strong track record in operating networks efficiently and reliably. So clear strengths on this category of business. And then when we look at the data center connectivity business characteristics, the way that business is emerging, as stated, the market is emerging, but it holds significant future potential for us. At the same time, it is important to note that we are focusing on data center connectivity business, meaning the fiber connections to the data center, potentially supplemented by optics in some cases. And that means that, that connectivity bit is only a small part of the overall data center investments that the data center operators are looking at. Based on the deals that we have now signed and based on the deals that we are now in discussions of, we see attractive capital returns. And importantly, we see attractive cash flow characteristics. We think that this will represent a notable positive EPS impact to Elisa over time. But it is important to understand that this is a long-term business. When we enter into a deal, the construction periods vary, smaller deals, might involve a construction period of some months, and the larger deals include a construction period up to 24 months. So we're a long-term business in nature. The CapEx needs that will arise from these deals will be handled outside of our 12% CapEx to sales envelope. But let me reemphasize that what we see is that the cash flow characteristics of the business are attractive. Going forward, we do not plan to disclose individual deals in this category of business. But I stated in emerging business opportunity that has now reached an important milestone with first large-scale data center connectivity deals being signed during the quarter. Looking into other aspects of our business. In Home Services, during the quarter, we reached a new agreement with MTV, a Finnish commercial TV company after lengthy negotiations, which also included a blackout period for our customers. Nevertheless, it's now good to see that this agreement is in place. We will be deepening the collaboration with MTV over time, also developing our offering to customers together with MTV. When we go into the corporate space, we clearly see an increased need for security solutions. During the quarter, we entered into a pilot agreement with Finnish Potergard and a drone company called Sensor Fusion. So testing a drone monitoring and drone charming solution, which is an important and intriguing entry to mission-critical defense business related to drones. As stated, in international software business in Elisa IndustrIQ, the revenue and profitability was below our expectations due to the delays that we saw on the market. It is important to note that the order intake developed positively. And the order backlog as such, was strengthened. In this part of the business going forward, during the -- this calendar year, we now expect to see revenue growth in the range of 5% to 10%. And as mentioned, we will be doing these specific measures to improve profitability and secure improving profitability from last year's levels. We also had a good customer win during the quarter in Spain, [indiscernible], big telco chose our software solutions and then that demonstrates the quality and the competitiveness of the solutions that we are having in the telco space. In terms of sustainability, Time Magazine and Statista once again selected us to the list of 100 most sustainable companies in the world. This time around, we ended up on the 61st place, and this was indeed third time in a row demonstrating our long-term commitment to sustainability work. And then when we go into outlook and guidance, our guidance for this year remains unchanged. So revenue, we expect to be at the same level or slightly higher than in '25. And comparable EBITDA, we expect to be within the range of EUR 815 million to EUR 845 million, the midpoint there being EUR 830 million. In terms of assumptions related to the guidance, we now expect the telecom service revenue to grow within the range of 0% to 2%. And please note that this is related to the outlook for calendar year of EUR 26 million. We do see the mobile indicators normalizing. And with those indicators in the past, we have been delivering solid mobile service revenue growth but it comes in with a delay being especially visible in Q4. And the same postponement phenomenon is visible in Elisa IndustrIQ. At the same time, the totality of this means that our cost measures are progressing well. And then what is noteworthy also is that the data center connectivity opportunity will offer long-term support for the telecom service revenue development. So with that, I will hand over to Kristian to cover the rest of Q2. Thank you.

Thank you, Topi, and good day also from my behalf. In Q2, group revenue was essentially flat year-on-year at EUR 551 million. Within that, we saw expected mix effects. Mobile services, the EpicTV divestment, consumer digital services as well as the traditional fixed PSTN continue to weigh on the top line, while fixed services, equipment sales and energy spare services supported revenue. . Despite the slightly grind in revenue, comparable EBIT increased by EUR 3 million to EUR 201 million and the EBITDA margin improved to 36.5% from the 35.8% in last year. The main driver here was really the operating cost savings, which came both from the ongoing transformation program as well as from disciplined cost controls more broadly. Comparable EPS increased from EUR 0.59 -- to EUR 0.59 from EUR 0.57. And all in all, we delivered solid profitability and better EBITDA margins even with the slight revenue decline. When it comes to the second half development, we did see positive development in the global -- in the mobile indicators in Q2, as Topi discussed. However, as said, the financial impact will be coming in with a lag and will be especially visible only in Q4. This will have an impact on the normal seasonality that we have seen over the years, where Q3 EBITDA has been stronger than Q4. This is not going to be expected this year. Q4 is expected to be stronger this year. CapEx for the quarter was EUR 72 million, down from EUR 76 million last year. the allocation was very consistent with our strategy, main investment areas, remaining 5G coverage expansion, fiber build-out and IT systems that support simplification, customer experience and productivity. Some part of the fiber CapEx was implemented through the JV that we established last year. You might recall from Q1 that we emphasized strict CapEx discipline and focus on technology leadership. That approach continues. We are investing where we see a clear long-term value and attractive returns while avoiding discretionary and lower spend, especially lowering CapEx for older technologies and old systems. All in all, we are maintaining disciplined investment levels while funding infrastructure and IT that underpin the future growth. Comparable cash flow in Q2 was solid at EUR 71 million, down from an exceptionally strong EUR 130 million in the prior year quarter, which represents a 37% decline. There are 3 main drivers behind this. First, lower CapEx compared to last year was a positive for cash flow. Second, higher paid interest costs weighed on cash flow reflecting both higher interest rate environment as well as the fact that we, this year paid a larger portion of the whole year interest costs in Q2. And then thirdly, and most importantly, net working capital development was less favorable than in Q2 '25. And that quarter benefited from very strong positive net working capital movements, especially on payables. In Q1, we highlighted that net working capital had developed positively for 5 consecutive quarters. In Q2, we saw flat development. We continue to focus on working capital efficiency and managing interest costs to support cash flow in the second half and beyond. All in all, solid cash conversion in a tougher comparison quarter with room to improve in working capital and a normalized cash interest cash interest impact going forward. Our capital structure or our capital structure remains efficient and clearly within our target ranges. Net debt at the end of the quarter was slightly down from the year-end. Net debt to comparable EBITDA was at 1.8x, which is well inside our target range of 1.5 to 2x. Equity ratio was 39.1% above our minimum target of 35%. We have a balanced maturity profile with bonds, loans and undrawn revolving credit facilities. As indicated earlier this year, we are focusing on proactively refinancing our 27 maturities to maintain cost-efficient and diversified funding. Both S&P and Moody's reconfirmed our strong investment grade ratings during Q2. Return ratios remain at good levels. Return on equity and return on invested capital are both running in the high teens, consistent with our ambition to deliver industry-leading returns through strict CapEx discipline and strong focus on cash flow. All in all, Elisa's balance sheet is strong. Our leverage is comfortably within targets, and we are well positioned to continue investing in growth, paying dividends and maintaining solid returns to shareholders. With that, I hand back the call to Vesa for Q&A.

Thank you, Kristian. And now we move on to Q&A part, and we ask for first question from the conference call lines, please. .

I had 2 questions. Apologies. The first one is a little worthy, but just wanted to dig in on your telecom revenue growth visibility. So just check, it looks like it sounds like the downgrade on your guidance for telecom revenue growth this year. Is [indiscernible] seeing a greater degree of these 12-month fixed term contracts that were signed during the second half of 2025. That mean that first half '26 improvement in the market you're seeing just isn't translating to better growth until those fixed term contracts and towards the end of 2026. That's our understanding at least. So the question is, why weren't you aware of the degree of importance of those 12-month contracts that have been sold during the second half of '25? And how confident are you now that you have a good enough grip on the market dynamics and tariff structures? And if you are now confident, can you tell us if that mobile service revenue growth could come in positively in the third quarter of '26? Or do we have to wait until the fourth quarter? So it's just a question around the visibility that you have on what's going on in the market and your confidence that we now guide to that improvement in the back end of the year. And then the second question is shorter. It's just on the IFS growth. Are you totally sure that the lower revenue growth you're seeing is not at all structural in terms of headwinds? And is it all macro related? Just wondering if there's a risk here that you're just missing out on some AI-related revenues as your customers reallocate spend towards that, and that's just not being spent with you.

Thank you, Andrew. So if Kristian, can you start with the first one, then I follow with the second.

Yes. So on the TSR. So there has been a bigger lag impact from the competitive environment than what we anticipated and modeled that at the beginning of the year. And that is why we updated this year's outlook for revenue growth there. This is more going to be -- so the positive impact from the market stabilization and the price increases that we have seen will be more visible in Q4 and not as much visible in Q3.

Yes. And to add to that, I think that the market phenomenon in this one has been that on the whole market, all players include it, it seems that the share of fixed term contracts has been increasing quite a bit. And that share with competitors is something that they do not disclose and have not disclosed. So this has been sort of an unknown factor to all players on the market during the year. And now the empirical evidence points to the share of the fixed-term contracts being clearly increased on the overall market increasing the time lag that we are seeing in this part of business. But eventually, with the normalized levels of mobile indicators, the revenue will follow. Add to the industry part of the business, whether the revenue slowing down is structural. We are, of course, observing this very, very closely, and we are listening to our customers very, very closely in this one. And we do not see AI-related impacts in this one. We are dealing with mission-critical software for our customers. be that telecom software or be that industrial automation software. And with that, we have a clear moat in our software business. The delays that we have now experienced with the license revenue clearly related to more practical matters like production customers, projects related to investing in new production facilities being postponed. We do not see deals being lost and that is important to note related to your question.

Can I just a quick follow-up on the kind of surprise on the amount of fixed term contracts, I understand that you don't see the contracts signed by your competitors. But surely, you saw the amount of fixed term contracts that you guys were signing in your commercial offices, et cetera, know what's going on in the market and where the competition is. So I understand there was a change in market dynamics. But given that there's been some volatility in market dynamics, how confident are you that you have a grip on the tariff structures, et cetera, that are being signed by your competitors now given that so much of the price engagement in Finland is below the counter. Just trying to just gauge that degree of confidence that investors can have in that quarter improvement.

I think that if we go back and look at what has happened in the last 12 months since last summer in the Finnish mobile market, clearly the reman of the share of fixed-term contracts increasing significantly starting from summer last year, has been impacting the way revenue comes in for all players. So that is clear as such. What we do see now is that the important indicators in mobile business are normalizing, as you saw from our presentation. And therefore, we do have a line of sight to clearly improving MSR going forward. Q4.

The next question comes from Paul Sidney from Berenberg.

I also have 2 questions, please. First one, really following up from Andrew's question on Finnish mobile. You made it clear in the past few quarters that it's not acceptable for you to lose market share. And I just wondered, have you seen competition the competition back off because of this stance over the past few quarters. Is that why the market environment has improved because you've taken this stance and push back on promotional activity? And do you intend this stance from Elisa to continue going forward? Or would you consider giving the market a bit more room to breeze is the first question. And then just secondly, I was intrigued by the fiber acquisition that you've in that land, it's obviously very small 8,000 customers. But are you seeing the regional fiber players starting to really struggle given that they really like scale. And if there are opportunities going forward to make more of these bolt-on acquisitions that are obviously very value creating.

Yes. If I start from the fiber part, I think that the fiber regional fiber players are open to discussions related to consolidation, and we see fiber assets at play on the market. We have strict conditions for value creation and for geographical location. But if we see assets on the market that are meeting our strict criteria, then we are willing to do similar bolt-on acquisitions that we did in Lapland in this case.

And then could you please repeat the mobile service question that I'm completely sure that what was your main point related to that.

Yes, sure. You've made it very clear that you're not prepared to lose market share and have back in the second half of last year on the competing that you faced. But I was just wondering, do you think that the improvement we're seeing is because you pushed back? And is that a stance you expect to continue going forward?

Yes. I think that -- I mean, of course, the overall market dynamic is an equation of all the actions that all players are taking on the market. I think that what is worthwhile to note in this regard is that we do not see the mobile virtual network operators having a big impact on the market. So the competitive dynamic has been especially a dynamic with -- between the 3 established players on the market. When it comes to market shares, when you look at the postpaid mobile subs during this quarter, consumer corporate included our market shares were stable. So we are keeping our market shares. Clearly, the market has returned to normalized levels. At least in our case, we have not seen competitors disclosing their numbers of Q2 yet. But the bottom line is that when we look at our mobile indicators, no matter whether we look at new sales price, whether we look at churn or whether we look at sales and marketing cost, whether we look at the net adds we see normalized levels. And of course, that is very encouraging.

That's great. But can I just have a quick follow-up, please. You mentioned returns when you're talking about acquisitions and clearly, the data center projects are going to make an acceptor return. Have you disclosed or could you give us an idea about what the hurdle rate is for these projects and acquisitions?

So we will make sure that the capital returns are attractive, and cash flow profile is attractive. Nothing more to add there.

SP-4 The next question comes from Ondrej Cabejšek from UBS.

I have also got 2 questions, please, one also on the -- especially mobile service revenues were coming from, I guess, a bit of the opposite angle. So what we know is that last year, the competitive environment started deteriorating in the third quarter then was the worst in fourth quarter. What we also know is that a lot of the people who got on promotions at that time have these promotions for 12 months. And after 12 months, they should be kind of rolling off of these discounts on to regular pricing. I think previously you expected something like a 3-month lag in terms of the kind of full pricing to ARPU translation due to these contractual issues that you pointed out. Now you seem to be more in line with the peers saying that's going to be a bit longer than that. And I think all of that is clear. But at the same time, you're flagging that there is an expected improvement in 4Q '26, and this is why I struggle to understand the situation a bit because if the worst promotions were given in 4Q '25. And there is, as you say, something like a 4-month delay in terms of when you sign the contract up until when the pricing is actually effective then shouldn't it mean that 3Q '26 is when things deteriorate further, given the slag, then the impact of those 12-month promotions actually last for a year, and only mid-2027 is when ARPU start to really pick up again as people roll off of these promotions. So I guess I'm just confused with kind of trying to put all of those pieces of information together, if you can help me, please?

Yes. If we decompose that a bit. First of all, if we look at the price levels of new sales as of now, we are significantly -- on significantly higher levels than we were during the most fierce campaigning in Q3 and Q last year. So there's a big difference. And that, of course, if that moves even partly to the price level of those fixed-term contracts being renewed during the fall of this year, then that will be very supportive of Q4 mobile service revenue. So that is something that we need to keep in mind. I think that the time lag, especially comes from this onetime effect of our share of fixed term contracts as well as competitor share of fixed-term contracts rapidly increasing during the fall of last year. And that also means that the overall volume of number transfers between competitors on the Finnish market during the first half has been a little lower. And therefore, the weight is smaller in terms of impacting the MSR when it comes to new sales.

I guess what you're saying is that the step down from the promotional activity will be in the base kind of 3Q, but then the underlying trends, which are still the healthy kind of 4G to 5G upsell and so forth will eventually kind of be the higher impact and overall, a positive one, combining those 2 things together starting 4Q. And then as we progress in 2027, potentially a best-case scenario, you continue to have the positive underlying trends. And as people roll off these kind of 12 months discounts with that for whatever months delayed, there should be a double positive starting kind of say it or like 2Q '27 or something mid-'27. So is that roughly the correct understanding?

Maybe rather than confirming your kind of thinking, maybe the way to think about this is that the lag on transfers is actually longer than the lag on renewals. And that's, in a way, maybe the dynamic to reflect here. And so when it comes to renewing the 1-year contracts that were entered into last year, there will be a similar lag there as there was when we had transfers to us some of which were fit transfers. So they didn't kick in at sales. They only kicked in when those contracts ended. And that's what is creating this dynamic where we are feeling the headwind now, and there will be a tailwind then going into the second half, especially visible in Q4.

Okay. If I may, a second question. Just on the flow-through of the cost savings, I think on a net -- I think we're seeing that on a growth basis if I kind of break things out, obviously, like in the past 2 quarters, you had high single-digit million savings year-over-year in employee costs. And I guess some underlying cost savings in other areas, including the commercial cost. But the net translation to positive EBITDA growth on a stable supply and is still pretty limited. So I was just trying to understand if that is primarily the responsibility of the top line in terms of the dilution of these efficiency gains. Or is there something else going on maybe under the hood in terms of reinvestment, et cetera, that we cannot really kind of appreciate from the outside?

Maybe a couple of dynamics. First of all, when we talked about EUR 40 million cost savings, that was across the board, kind of CapEx and OpEx. And clearly, the OpEx is more visible in the in the P&L. Yes, there is an element of reinvestment here also. And the flow-through to EBITDA from those cost savings is impacted by the headwinds that we are seeing from the from the revenue line. So it is, in a way, a mixture of all of the above that you listed. The program is on track. Most of the savings are in and in that sense, that has been a great help when it comes to being able to grow EBITDA both in Q1 and Q2, even in this revenue environment.

The next question comes from Fredrik Lithell from Handelsbanken.

I have a few small detailed question. The PTSN that you're closing down, do you foresee any further costs relating to that in coming quarters or you have everything behind you now in terms of cost? Or do you have any write-offs or something to do on old equipment would be interesting to hear. The second question on data centers and your investments. I appreciate you are careful on sort of the return metrics and all that stuff. But can you give us some time line on when you do your CapEx work and when you foresee your positive cash flow to contribute to the group. What's the time lag between those 2.

So on the PSTN, this is now kind of material behind us. So it will not be a drag to our revenue compare anymore in a similar way as it has during the ramp-up period and the material parts of the costs have been booked. And we are now against those kind of provisions, dismantle some of the air cables and so on. So I do think that don't expect us to refer a lot to that anymore as we go forward. When it comes to the data centers, I will not give you much more detail. I will only repeat what we said earlier that the overall return -- capital returns and the cash flow profile from those deals is attractive. And again, when we say that, we look at both inflows and outflows. Then we made a separate statement that the outflows when it relates to CapEx will be done outside of the 12% CapEx envelope that we have. But the cash flow attractive comment refers to both inflows and outflows.

The next question comes from Artem Beletski from SEB.

I will actually ask one by one source? And maybe the first one, I just wanted to double check comment that was made actually by Kristian earlier at the call. So do you really expect that the EBITDA in Q4 will be higher compared to Q3 this year because I think looking at historical seasonality. So the difference has been opposite roughly by close to EUR 10 million. Q3 always been better. So is this year really so much Q4 loaded in terms of gross outlook.

Good remarks, that is what I said. So that was the intention of the communication. Q4 is, as we see it, given the dynamics this year going to be better than Q3. And in the past, it's been the other way around.

Okay. Very good. And maybe then the second question was relating to some nice first deals what you have done around data centers. And maybe in terms of business opportunity, could you provide could you somehow frame it? So you mentioned about dealers covering at least 100-megawatt capacity data center investments. So how much of this type of projects could contribute to your revenues? Or what is the business opportunity maybe there are a general phenomenon, what we see on Finnish data center market is that many of these mega projects are actually built over a long period of time and in many phases. So how it works in terms of connectivity CapEx being done? Is everything done basically upfront or those investments also gradual over a longer time period in this big project.