QTCOM.HEQTCOM.HEHEL
Loading
Heli Jamsa: Good afternoon, and welcome to Qt Group's Q4 2025 Results Presentation. My name is Heli Jamsa, IR Lead. And with me today are our CEO, Juha Varelius; and Interim CFO, Ann Zetterberg, to present the results. After the presentations, we will have Q&A first in the room. And if there's time left, move on to the questions from the lines. Without further ado, please, Juha, the floor is yours.
Juha Varelius: Thank you. Thank you. Good afternoon, everyone. And we have a same agenda, as always. I go a bit through what happened on Q4, and then Ann is going to go through the numbers in more detail. I'll talk a bit about the future outlook and then questions. So the Q4, we had a growth of 12.6% or 18.6% comparable currencies. And of course, IAR acquisition, which we completed -- contributed in this development. And IAR was EUR 8.1 million on Q4, and our organic growth was 6.1%. So it was a -- compared to the very difficult year we had last year, it was a decent quarter, and we were happy on that. Our EBITA margin was 35.6% and the EUR 27.5 million, and that's -- there is a decrease compared to last year, but we did have a one-off cost on the acquisition that were burdening that. I'm going to talk more about the overall market environment. But of course, the -- even the year changed, the market environment hasn't changed that much. So we had quite a bit challenges last year which affected our customers in a way that we had tariffs and whatnot uncertainties. So the selling developer licenses last year was slow, if put it on one word. So the -- on the whole year, we ended up on EUR 216.3 million, which is a increase of 6.6% on comparable currencies. So we went pretty much in the middle of our guidance. The distribution license revenue grew very well last year. There were a lot of new things coming into the market, new programs started, which ended up on the 26.4% growth. And of course, the main growth drivers, industries for distribution licenses is the automotive, medical and industrial manufacturing. The whole year EBITA was EUR 51.8 million, and there was a decrease. EBITA margin was 24%. Our personnel increased end of the year to 1,100 out of which 215 are IAR employees. So -- but we did continue our own hiring as well. The one-off costs for IAR acquisition, EUR 5.8 million. We're going to talk that also a bit later, but the -- of course, we all know that the IAR profitability has been less than the Qt. So that affects the overall profitability of the group going forward this year. We haven't disclosed the ARR before. And on the ARR we had a growth of 8.3%. And there on the small print is that the -- it is Qt and the QA developer license base and it does not include the IAR licenses and distribution licenses. So that ARR is the Qt and QA business. We plan to give that ARR number now in the future also in the half year sequence. So you can see that because one of the questions affecting our revenue has always been the shift from 1 to 3-year licenses. Of course, last year, we did see the cautiousness in our customers. So the -- it was slowness in sales, but it was also people shifting from 3 year to 1 year. So now presenting this ARR, we don't have to -- you don't have to worry about the shift from 1 -- 3 year to 1 year because we can follow the ARR. And our plan is to give that number now next time after second quarter and so on. Obviously, it's a number that doesn't change that much. We might even go on a quarterly basis if that's needed. But the -- like I said, it's much slower moving -- slower moving measurement. Well, here are some of the product-related things we did in 2025. There are always questions about AI. Is AI going to eat our lunch in a way that the -- you know that there are a lot of predictions on AI that the -- no developers are needed and AI is going to do all the code. Well, at least as of now, we don't see that development. We do see that there are a lot of AI assistants being used like we are offering them in Qt and our design studios and on Squish. So on writing test scripts, for example, you can use AI and then the Squish does the actual testing. So they help on that. But do we see that the -- specifically on embedded world that the AI would become and replace the developers, that kind of a development, we don't see as of yet. At the same time, of course, it's good to realize that I think that the U.S. companies are planning to invest EUR 500 billion, EUR 600 billion next year. So obviously, they are expecting to get something out of it. But I have -- I don't see that developers would be going away next year or even in the coming years in that sense. On the partnering side, we -- on Axivion, we do have partnerships with NVIDIA CUDA. So the -- when you're doing CUDA code, you can -- or using CUDA, you can use the Axivion. On the R&D, on the defense sector, we did have the FACE certifications and working with Infineon over there, on the AI consumer power devices. And then we are expanding our ecosystem through the Qt bridges, which will enable more languages over there basically. These are just some of the highlights that we are working on the product development. So in general, our product has -- all our products have always been very good. We get a very good feedback. So this is just to show a few examples that we do continue our R&D and we do -- we are on the forefront of product development all the time, making sure that all the Qt products are very competitive in the market, and that seems to be the case on all the customer surveys from our users. With this, Ann, please. Some numbers.
Ann Zetterberg: Yes. I am Ann Zetterberg. I am -- I have been the CFO of IAR for -- I'm on my fifth year now. And with the acquisition of IAR, I had the opportunity then to step up and become the interim CFO for Qt. And I'm going to tell you a little about the numbers then for this quarterly report. So delighted to meet you all. There will be a bit of a P&L first, maybe a little repeat on what Juha just mentioned. But we had -- in Q4, we had a growth of 12.6% and after exchange rate impact, it was 18.6% at comparable currencies and the organic growth with removal of IAR revenue, which was EUR 8.1 million, that was EUR 6.1 million. And we -- in -- for 2025, the growth was 3.5%. Exchange rate impact has been pretty bad, both for Q4 and for the full year, especially the dollar has behaved very, very badly for us. And the growth there for 2025 at comparable currencies was 6.6% and the organic growth was 2.6%. But as Juha also said, we plan to show the ARR as that shows better the yearly underlying growth for the company. It doesn't -- it's not affected from which contract length the customers chooses. As we recognize 95% of the contracts upfront, it depends -- it matters a lot if they choose a 5-year contract or a 1-year contract for revenue, but ARR illustrates the underlying growth very stably, and that is growing good for us. It was 8.3% of growth for the Qt part, excluding IAR during the year. And then looking at expenses, the personnel and year-on-year grew by 267 individuals. That's a growth of 31%. But of course, a lot of that relates to the IAR acquisitions, 215 people worked at IAR at the acquisition. And -- so that increased the headcount to 1,136, both on average for the year, but also at the year-end. And IAR contributed EUR 4.8 million in staff costs in the P&L. Under other OpEx, the IAR acquisition had some extra costs then, EUR 4.1 million in Q4 and EUR 5.8 million during the year. And also, I wanted to highlight, even though it's a very small cost, the capitalized asset as IAR has interpreted IAS 38 a bit differently than Qt has and has capitalized R&D assets in the balance sheet. Presently, there is EUR 5.4 million of capitalized unfinished assets in the balance sheet of IAR and those are expected to be finished under 2026. But this means that we will have a small positive effect on the P&L from these capitalizations, removing costs and putting it into the balance sheet. I don't expect any large amounts from this, but it is still good to understand that this is what it looks like now. Over time, there will be some harmonization within the group, so all companies look at this in the same way. And then, of course, the profitability, like Juha just mentioned, has gone down. The EBITA margins are lower both for Q4 and for the year. IAR has a lower profitability. So that contributes to that and as does the acquisition costs. But of course, when you join 2 companies, there are also opportunities for integration, efficiencies and cost reductions, which we are going to work with on starting this year. And this means that the earnings per share has gone down to EUR 0.73 for the quarter and EUR 1.25 for 2025. So then moving on to the balance sheet. A lot has happened to the balance sheet, obviously, from the acquisition of IAR. The preliminary PPA added EUR 204 million in net assets to the balance sheet. Of that, goodwill was EUR 122 million. And then there were identified other intangible assets of almost EUR 90 million. Those were customer relations, technology and trademarks. And those will be written off over 15 years. So the amortization yearly net of tax would be EUR 4.8 million. And also the PPA added, or the acquisition added other net assets of EUR 11.2 million in IAR. Some of those assets on the asset side of the balance sheet and some on the debt side sort of spread over, but the net of them all are EUR 11.2 million. Some of those assets were trade receivables then, which increased the trade receivable balance to EUR 58.7 million in the balance sheet. And there are also other receivables, which could be good to know, one booking of EUR 5.1 million as we have booked the full value, 100% of the shares to the balance sheet, as there is arbitration going on, and we are obligated to buy the rest of the shares. We are not showing any minorities under equity and so because it's only a matter of time until we own 100% of the shares. But that can also be good to know. And then the ending cash balance was EUR 40 million -- EUR 40.1 million, a little lower than compared to last year as we have made this large acquisition. And as the balance sheet has expanded, the equity ratio has gone down from 81% to 50% and also the interest-bearing debt has increased. The interest-bearing debt is EUR 143.2 million. And of those, EUR 134.4 million are debt relating to the acquisition of IAR. So we have paid off some of the debt already. It was EUR 150 million from -- to begin with. And also on the deferred tax on the debt side relating to those intangible assets that were EUR 90 million on the other side, there is also deferred tax booked on the other side which is EUR 18.5 million. So good to understand that also how the PPA affects the balance sheet. And on the short-term liabilities, there is a debt of EUR 5.1 million, which is the amount we expect to pay for the remaining shares of IAR after the arbitration is finalized. And then I can just, as a final note, say that operating cash flow then had gone down a bit, but mainly because of the profitability going down. So nothing strange about that. And with that, I suppose I'm done with the financials, and we will take questions afterwards, but I will then leave to you, Juha, to take the next of the slides.
Juha Varelius: Thank you. So 2026. Well, I think the first big thing is that the -- during the next 3 years, as you know, the IAR has been on a perpetual model. And our -- during the next 3 years, we are -- our target is to shift that into subscription model. That's roughly the -- by the way, the same plan we did have the -- early on with Qt when we did this a few years back. And if this goes as planned, the IAR revenue will be going down this year. So it's going to be decreasing this year. And then depending on how aggressively that goes down this year, then the swing back will be bigger next year. So -- but it's the early phases. So we've started the journey. We have now a couple of months behind us. So it's to make -- exact predictions at this point is there is a bit of a room for that estimate still. The -- well, the -- I think it's -- the market has been uncertain so long, that the uncertainty will definitely continue. As we know, there are a lot of global tensions going on as we speak, and that's what we're looking this year. Some of our customers are in a challenging environment. The -- like in automotive, the Chinese automotive manufacturers are putting a pressure on the European manufacturers. And at the same time, there are tariffs that's obviously going to continue all this year. And so on and so forth. So I think that on the industries, the automotive will be in challenge, Medical will not so, and the industry automation seems to be doing pretty well. Defense is doing really well. So in -- if I now look at the 2 of our biggest industries, they are actually medical and defense at this point of time. So they've grown quite substantially over there where they've been. The long-term growth prospects, well, like I said on the AI, this software really defines the value of the products. Each product will have software going forward and the new versions of it we don't see on embedded, that the AI would be eating all that market away far out from that. But we do see AI improving our own products on many respects, and that's what we are implementing. So before we've given our estimates that -- we've given you a range, but we gave up on that range. So now we're saying that the -- our full year net sales will increase at least 10%. So we're saying that, that's the floor, but we are not giving a range. So we're not giving the upper part guidance. So that's a bit different. And we're saying that our operating profit margin will be at least 15%. So again, we're saying that, that's the floor. We're not giving the upper range. So we've -- we're not giving those ranges anymore. Going forward, we're going to start after Q1 or after Q1, we're going to start giving you more info on the -- on how the -- well, we'll start sharing this ARR, which will give you a better understanding. You don't have to worry about the shift on the 3 to 1-year licenses. And then we're looking at the -- we're going to give you more on the revenue per product, so you get a better understanding on the -- how the licenses -- distribution licenses are coming. So we're looking to open up that a bit. I don't know if it's going to make your life any easier because there is a lot of fluctuations. But at least you can then see that fluctuation. So the -- we've been listening -- what you've been asking and -- so that's the -- but more to come on that later. I think the ARR actually will help you more than seeing the license -- distribution license sales and whatnot, but the -- more than that. So do you have any questions? Okay.
Matti Riikonen: It's Matti Riikonen, DNB Carnegie. A couple of questions. They are very simple. Do you expect the legacy Qt business to decline in 2026?
Juha Varelius: Simple answer, no.
Matti Riikonen: Okay. Do you have a rough estimate of how much IAR's revenue would decline in '26 versus '25, if you give a broad range? You say it's going to decline and you say that you don't know yet, but roughly where is -- where are your thoughts at the moment?
Juha Varelius: Well, double-digit.
Matti Riikonen: All right.
Juha Varelius: Low double-digit.
Matti Riikonen: And third question before I give the mic to somebody else. How will IAR's fixed cost base develop in 2026?
Juha Varelius: Simple question, longer answer. The -- well, I mean, we're not looking to increase the IAR cost base. So what you're going to see now is that the IAR -- the revenue decline really depends on the -- how well can we go on a subscription, and we try to go as aggressive as possible. So the -- if I say low double-digit revenue decline, somewhere there, right? I don't know yet, but somewhere there. And then 2027, I do expect to see a double-digit -- high double-digit growth on the -- maybe close to 20-something, to give you an idea how it's roughly good work, right? On a cost level, when we see costs, obviously, we're not going to be increasing costs because the prices are increasing, right? But the -- we do have some R&D-related initiatives over there where we think that we're going to be increasing cost, and they are related into the fact that the IAR is very much on a functional safety critical environment in automotives and whatnot. We are looking for a product development that we can broaden that segment roughly, to put on a broad perspective. And then we have few places where we're going to -- mainly on sales, we're going to increase sales costs, but we're talking very modest cost increases on the IAR side. So if you look at the old IAR, I know you have the numbers from there, we're looking very -- we're looking some cost increases, but fairly modest over there. But still, if you model that -- the revenue development on IAR numbers with that revenue dip, you're going to be seeing that the EBITA contribution for the whole group this year is going to be pretty much breakeven or even slightly negative. So we're not looking for -- first of all, on the guidance, we are -- those are the bottom lines. They are the floors. They are not the -- we see that, that's the bottom, bottom, right? So we do expect a bigger numbers. And then the IAR negative contribution will be on this year, but when it swings next year, we don't -- there is no need to increase costs for that because it's basically a price increase. So it will swing the IAR EBITA. Well, it's a license sales. So everything that the revenue will be increasing will go directly to bottom line. So that's the implication. On Qt Group, we are -- well, you call it legacy group. So the time changes. But -- so we'll figure out the better name than legacy. Anyway, the old Qt, we're not expecting organic decline, and we're not expecting that the -- what we saw last year, the bottom line, we're not expecting there to see a declining EBITA that we had last year. So the -- and that's the bottom performance, right? So we expect that the bottom performance be last year level and higher from there. So that's kind of the overall picture. So it's maybe not that gloomy than you were first thinking. I don't know how gloomy you were, but that's my educated guess. But thank you for the simple questions.
Matti Riikonen: That's all from me so far.
Jaakko Tyrväinen: Jaakko Tyrvainen from SEB. On distribution licenses, what happened in the sales in Q4 since I recall that the commentary after the first 9 months performance was rather moderate also in this revenue stream? So I'm curious whether there were some customers filling up their inventories in terms of distribution licenses? And how should we look at the revenue stream for '26?
Juha Varelius: Yes. Thanks. So well, maybe later on the first Q when we open up a more broad distribution, you're going to see -- But the distribution licenses is really hard to predict because the -- it's not like this -- I mean, quarter-on-quarter like last year, it went like -- well, first quarter, second quarter boom and then up again, and that changes every year. So the quarters are not alike. So you can't expect that what was last year and second quarter is going to be the same. And that makes it difficult. And as you know, the distribution licenses go that -- some customers buy them afterwards, telling that how much they [ chipped ] and some people buy a chunk on prebuy. And that's why it's hard to predict. On a general level, we can always see that we know that the -- some big new products, productions are coming into the market, then we know on a yearly level, what's going to happen. So last year was on that sense, very good. So if you look last year numbers and distribution licenses for this year, I would take them slightly down. That's my expectation for this year given the market volatility, given the -- what's the customer demand in Europe and whatnot. So the -- I mean, at the end of the day, our distribution license revenue comes from product, what the consumers are buying, right? So the -- that's in a general terms, it follows. And we do have -- we are in 70 industries. We are both on commercial devices like industrial automation, robots and whatnot, stuff that goes into hospitals, stuff that goes into factories, but also on consumer goods like auto, cars and whatnot. So that's where the fluctuation really comes. So I would not put on my model same growth this year than we had last year. This is going to be substantially lower. So same number or a bit below. That would be my best guess. And -- that's a guess.
Jaakko Tyrväinen: Understand very well. And just to confirm, Q4 was strong in distribution license?
Juha Varelius: Yes, yes, yes, I was a good. So last year, on distribution licenses, Q2 was very good. Q3 was very weak. Q4 was good. Q2 was, if I remember correctly, the best on distribution licenses last year and does not mean that, that's going to replicate. It really goes like this.
Jaakko Tyrväinen: Good. Then on the ARR, thanks for sharing that to us and the growth of 8% there. Could you give some color on how much of this was pricing and how much was coming from the effect that customers changed from 3 year to 1 year, which obviously should have kind of positive pricing impact on the ARR number -- annualized ARR number?
Juha Varelius: That -- Very good and detailed question that I -- those numbers I don't have. We can come back later, but those I don't have out of my head. But the -- on general level, I can say that there was some shift from 3 year to 1 year, if I look on a whole year number, but it's slowing down. That shift is slowing down. But definitely, what we saw through all the year was the fact that on renewals, the -- what people used to do is that they had something and then they renewed older licenses. Nowadays, customers are counting that how many developers we really have, how many licenses we really need. And in general, money has been very tight. I mean our customers are very -- they're very tight on money. So they are looking all the costs. And on many R&D budgets are such now that the R&D budgets are not growing, but the -- so if they do something additional, they need to stop doing something old.
Jaakko Tyrväinen: Good. And then my final one on the possible AI disruption also in the embedded side, I heard what you said. But could you give us for -- kind of for a dummy explanation why the embedded world, what are the barrier entries for AI native solutions to break in?
Juha Varelius: Well, as of today, what we see, first of all, that you have lots of safety critical, you have lots of functional safety type of things like car brakes and whatnot, you need certificates and there are -- there is a very tight regulation what you need in order to have software. So you can't just ask AI that do me a car brake system, thank you and implement it, right? The second is that the -- on embedded, the software goes into products, right? And in products if you need to do a product recall, that is really, really expensive. So you have to be fairly certain that what you're doing. The third is that the embedded is fairly slow moving. There are huge companies building these cars and all these devices, medical devices and whatnot. So the time of the change and how secure they need to be that if I'm building this medical device, that nothing really goes wrong. So they change relatively slowly, right? Whereas if you think that on a website that I want to do a mobile application, I do a mobile application, if it works, great. If it doesn't work, it doesn't matter so much. So the -- it's kind of a different environment. And then if you think about coding, just building the software is -- it's one part of the process. You need to define what you want. You need to discuss with people that what are we building, what this product is doing and on and on and on. And AI is definitely not ready for that yet, right? So the -- where it's really going to end, we'll see, but that's what we see as of today. So there are -- we see AI as assistance and the -- like if you're designing something, you can use AI to give you creative ideas because as people, we tend to start looking one-way street. AI can open up your creativity and whatnot, but yet you're still using tools. So my prediction is that the next phase you're going to see on SaaS environment and the products like ours is yet another pricing change. We're doing this just to mess you up, right? So -- but yet another pricing change. And the pricing change is going to be that -- the pricing, I think, is going to go more towards from that the -- what has been built, how much the tool has been used rather than a deficit, right? So that's where I see the AI is going. And I had a -- one breakfast discussion and the person pointed out that the -- remember a couple of years back -- this person said to me that remember a couple of years back, everybody in Finland were talking that the -- even grandmas need to learn coding because software is in every device and everybody needs to learn how to code so that we can use these products, and they were all kind of coding school starts and whatnot. That was 2 years ago. Now everybody is talking that developers are -- nobody needs developers anymore. So there is a bit of a hype on the speed of the change. I mean, over time, of course, AI is going to -- 10 years from now, AI has changed a lot how we work, but -- and live our lives. But in the near future, I don't see much of effect. Then on the -- and this is the Qt development, right? On the IAR compiler business where you need all the certificates and whatnot, there is no way you can use the AI for a long time. And then on our testing tools, well, whatever you do with AI, you need to test. So I see that there's going to be more and more software that needs to be tested because you can't rely on AI. So the testing business is going to grow substantially as a market.
Felix Henriksson: Felix Henriksson, Nordea. Three questions. Firstly, on Q4. The revenue growth organically accelerated a little bit, and we discussed about the distribution licenses being strong, but was there anything else that improved? For example, the lack of large deals that we saw in earlier quarters, did that -- did those sort of come back at all?
Juha Varelius: No. no. If I look on the regions, the -- I would say that the -- we're doing well in APAC. We're doing okay in Europe. We have room for improvement in Europe in some markets. And then in general, we have lots of room to improve in the U.S. So the majority of our softness has been in U.S. And then we come to the point that the -- if we talk about the AI or if we talk about that the -- is there a competing product or is there a price change? What I see in the market is that we're doing fine in APAC, we're doing fine in Europe and the main softness we have is in U.S. and even in U.S., we have some teams that are doing okay, but then some teams are really suffering in that respect. So that's why I'm fairly confident that it's not about AI eating our market because if it would be, it would be eating our market everywhere globally, right? And this is more a local softness we are having. Same thing for prices and competition because we have same type of -- in APAC, we have the same industries and same type of customers we have elsewhere. So our softness basically has come last year that we've been a bit soft, been a U.S. related, right? And I'm very confident that we can fix that and get the efficiencies over there on a better shape.
Felix Henriksson: Right. And then on the guidance, you mentioned that you're no longer giving those ranges, upper end. Can you expand on that a little bit? What's changed with your guidance philosophy in a way that triggered that change?
Juha Varelius: Yes. I wasn't very good at that last year.
Felix Henriksson: Okay. So maybe more conservatism in that way?
Juha Varelius: Yes, yes, absolutely. Yes. Well, hey, we gave 2 profit warnings was not on my plan.
Felix Henriksson: Fair enough. And lastly, on distribution licenses, I mean, we've started to see memory prices going up and there are some supply constraints emerging that potentially are impacting your customers, I presume. Do you think that's a sort of potential headwind when you look ahead and what are your customers saying when it comes to this?
Juha Varelius: No, that's a downwind because the -- that's where Qt really signs. The fact that if you use Qt, you can do more with less memory. So that's the -- I mean, that's been the basic promise since the beginning. So the -- with Qt, you can have the same performance with the lower-end hardware, and that's the main selling point we are having as of today. And so higher price is better for us because at the end of the day, our customers will have to build those products anyway. So then it's a question of that how -- what kind of performance they want, what kind of end user experience they want. And that's where we sign. And that's where like Android doesn't sign, right? You use Android and you need a lot more hardware than using Qt and so on and so forth. Same goes with Unity. So most of our competitors, they may be in some use cases like Unity, Unreal, they might be able to do a better 3D visualization or it looks better, but it consumes so much hardware that if we go on a lower-end hardware, we can beat them. And you can get good enough, you can get a fairly good performance and a lot lower hardware using Qt. So that works for our benefit.
Antti Luiro: It's Antti Luiro from Inderes. One question. We know that the last year's growth was quite sluggish and there is still uncertainty around this year. So is that affecting your own investment plans? Or are you just keep on going with all the growth investments that you have planned before?
Juha Varelius: Yes. I mean, yes, we will continue our investments, yes, for sure. That's the -- no doubt about that. And we do have these few areas where we see the -- well, first of all, I think that we wouldn't be here in the first place if our products wouldn't be so competitive. So we need to keep them in that way. And then, of course, we are exploring the opportunities that the AI is opening up, and we need to do product development to have AI agents in our own products and so on and so forth. And you're going to hear product releases as we go forward this year. So yes, definitely, we're going to do that. Then at the same token, like the -- Ann was saying over here that we just merged 2 companies. And of course, we are going through all the processes, we're going through the -- that where can we be more efficient. So we've grown very fast. We are 1,200 people. And the -- so we do have also the efficiency programs, if you like. But it's -- so it's not all more, more, more. It's also efficiencies at the same time, and that's very much on and stable as well.
Waltteri Rossi: Waltteri Rossi from Danske Bank. A few questions about AI. Did I hear correctly that you said that you might change your pricing model in the future due to AI?
Juha Varelius: Yes. I said that, that's probably going to be the first change that we see on AI that the SaaS models pricings will start changing more from based on the consume of the tool rather than the deficit. I did not say that we're going to do that change, and I did not say that we're going to do that change this year, but I said that, that's what I see that the -- how AI is going to be affecting SaaS companies in general that the pricing will change going forward. I don't see that AI will be taking over the tools business per se.
Waltteri Rossi: Yes. I understand, but no time line for that?
Juha Varelius: No, no, no, no, no.
Waltteri Rossi: Okay. But that would imply in a way that there is at least a big threat on your developer license sales?
Juha Varelius: No. No, I don't see it that way. I see it that the -- that's going to be the effect that the SaaS business will go more towards that, that the people are charged at how much you use the tool rather than the per deficit. I see that development coming. But no time line, definitely not this year, next year or so. No.
Waltteri Rossi: Okay. Well, next one is still on AI. What would you say is basically Qt's value proposition for the customers because there's the argument that AI will make developers' work more efficient. So that's kind of eating up your -- one of your value propositions. So what else do you basically offer for the customers?
Juha Varelius: Well, we offer a tool that they can build their graphical user interface or applications. And as we are here today, AI is not capable of that. So you need the human and you need the tool. And then it's debatable that when will AI be able to do that, if ever. And then we see that if you need certifications, you need -- like on defense, like in automotive, on many industries, on medical, there's a long list certifications you need to meet. So who's going to train an AI that will meet those certifications and make sure that AI does the things every time in that particular manner and everything is met. I mean, that's years away, if ever.
Waltteri Rossi: Yes, yes. I agree. But still on that, actually, a follow-up. We know that programmers are already today using AI assistance. But are you saying that you don't see your customers yet using them?
Juha Varelius: No. And you see a lot of developers using AI on the web technologies. So if you want to do a simple mobile application, you can do that or you want to do web pages, you want to do your own homepage, you can use AI for that. But of course, they are so simple that you can -- if you want to do your own web pages, you can have -- they are on a web already. So what AI does is that it takes a web page and then it produces a new web page, right, that you can do. But on embedded building on products, no.
Waltteri Rossi: One last on AI. So can you please elaborate how Qt is currently using AI in the framework? Or do you have add-on or something?
Juha Varelius: Yes, you can have add-ons. You can have assistance over there that helps you getting started. For example, on the -- on testing, you can use AI, that it helps you doing the testing script and these type of things. So it helps you kind of where you can think that it helps you building a bit of a story or text, but yet still you have to read it and modify it. That's what we see as of today.
Waltteri Rossi: All right. Then one last question on the usual 1 to 3-year licenses. So do you have a number on how much that shift from 3-year licenses to 1-year licenses impacted last years?
Juha Varelius: No, but we're going to give you the ARR, so you can start following that.
Matti Riikonen: Matti Riikonen, Carnegie, a couple of questions more. They are even more simple. First of all, you discussed the capital -- capitalized cost policy so that you would basically go towards Qt's policy, which I read that you would not any longer capitalize some costs that IAR has. Should we expect that there would be none whatsoever on the capitalized costs in '26?
Ann Zetterberg: Because we have unfinished assets in the IAR balance sheet, and we need to continue capitalizing on those according to IAS 38. So there will be some capitalization of R&D assets during 2026. But we expect that those will be finished under 2026 assets that are not finished. And after that, we will harmonize between the companies so we can find a common application of IAS 38...
Juha Varelius: Because -- I want to come over here, so we have you in the camera as well.
Ann Zetterberg: Yes, sorry. I apologize. Yes, about the capitalization since Qt and IAR has handled the IAS 38 application very differently. So IAR has been capitalizing R&D assets into the balance sheet, which increases the profitability and then you write off the assets over time. And after the acquisitions, we kind of cleaned out the balance sheet. But the assets that are not finished, they are still there, and we will have to follow IAS 38. We will have to continue to capitalize on those until they are finished. Otherwise, we don't follow the bookkeeping rules correctly, and we don't want to break them. So that will happen. And in that time, we will evaluate and harmonize between the companies so we can have a common approach to this. And then I expect that we will not capitalize anymore, but I cannot 100% tell you that, that will happen. But we will have to have a common approach anyway within the group on how we handle this adjoiner [indiscernible].
Matti Riikonen: What kind of magnitude of capitalizations do you think there would be in '26?
Ann Zetterberg: It will not be a lot. Those assets are almost finished. They're EUR 5.4 million there now. So I don't expect there to be any huge capitalizations. As you saw during Q4 on those assets, we capitalized EUR 200,000. So it wasn't a lot. So you can -- then -- it can go up and it kind of go down a little depending on how much work the R&D department puts into various projects. But I don't expect it to be -- I mean, anything that affects the profitability much, but it can be good for an analyst to understand that this is a difference from how Qt has handled it before.
Matti Riikonen: Right. That's helpful. Then second question is about the annual recurring revenue disclosure that you plan, which is an excellent idea. How long into the history will you bring that? So is it possible that you would bring maybe a couple of years' history so that we could start to track it already from there and not just from here on because, of course, in the ARR pattern, the history is what counts and current day is less interesting if you don't know the history?
Juha Varelius: We already gave the last year number, right?
Matti Riikonen: That's not a very long history.
Juha Varelius: Well, it's the last year. Well, we'll look into that. Yes, great question. We didn't think it that way, but we'll look into it. And on capitalization, it's like Ann said, that there are a few projects we need to continue. But of course, in general, going forward, on a longer term, we're not looking to capitalize. So we rather implement the Qt policy going forward and not capitalize the development. Yes. It's a better way.
Jaakko Tyrväinen: Jaakko Tyrvainen, SEB. A brief follow-up on the profitability dynamics and IAR part of that. Let's say, the revenue is down double-digit something, like you said, Juha, would this imply that IAR as a stand-alone would be at breakeven or even red numbers in '26?
Juha Varelius: Red.
Ann Zetterberg: This is...
Waltteri Rossi: Sorry, one last one from me. You said you are going to continue...
Juha Varelius: You were?
Waltteri Rossi: Waltteri Rossi, Danske Bank. You said that you are going to continue recruiting this year. So could you elaborate a bit on where exactly are you going to recruit? Or you said invest, but...
Juha Varelius: Regionally or by function.
Waltteri Rossi: Say again, I didn't...
Juha Varelius: I mean regionally. Well, I mean, I think that the -- we do have a few markets where we're going to be increasing personnel, probably the U.K. is one, and these are small numbers, but then they add up for our Italian business. More or less in Japan, we're going to be increasing the personnel, the China probably. And the -- so in particular markets, I think in the U.S., we're pretty much on a headcount we like to be at this point of time. On R&D, there are these new technologies like the one that interests you a lot, which is the AI. So of course, on these new technology areas, we -- instead of trying to learn them ourselves, we are hiring people. So we do have some of these new technology areas. If I look in general on the R&D, the Qt is very well staffed. The QA business function itself, it's still on the investment mode. So over there, you're going to see pretty much on each and every function. So a bit of marketing, a bit of sales, a bit of product management, a bit of the R&D. On IAR, we are strengthening some of the R&D functions over there. So IAR, I would say that the most personnel additions will be on the product R&D side and then some on sales. But when you have so many different locations, you add up and then you get the personnel increase, that's basically what we're looking for.
Heli Jamsa: Thank you so much. I believe that concludes all the questions from the room. And as we are running out of time, I give it back to Juha for closing remarks.
Juha Varelius: Okay. That came quick. So thank you very much for being here today. And the -- as we go into this year, like I said, the -- one -- the very big item for us this year is going to be the subscription change on IAR. So going from perpetual to subscription, that's the one of the core things we're doing. And of course, integrating IAR into the Qt family. So we're going to be a bigger, happy family. We are also looking forward this year that, yes, it's going to be a challenging year. I'd like to emphasize that the guidance we gave was not a range. So we just gave a bottom line that what is the floor level where we expect to be this year. Of course, we are expecting to be better on those numbers. And the -- on the profitability side, we're not looking on Q2 decrease on profitability nor on sales, but the IAR subscription change will affect our profitability this year. And so that's why the lower guidance. Where it's going to end up then that how aggressive can we be, remains to be seen. In any case, the 2027 for IAR will be a revenue growth year and a profitability year, then the question is that how steep is that curve over there. It's still very early phases to see that how rapidly we can drive this subscription change. I think with these words, thank you very much.