QTGPF (QTGPF) Q3 2026 Earnings Report, Transcript and Summary

Hello, everybody, and welcome to QT Group's Q3 2026 Results Presentation. My name is Hertta Närvänen, and I'm the Communications Lead at QT Group. I'm here today with Juha Varelius, our CEO; and our CFO, Ann Zetterberg, who will be presenting the results. After the presentation, we have time for questions, first, starting from the room and then if time permits from the line. That's enough of me. Now I'll give the stage to Juha. Please go ahead.

Good afternoon, everyone, and welcome to QT Group's Q3 results. We're going to go through the very traditional business highlights and financials by Ann, and I'm going to talk about the outlook and guidance for 2026. So if we go on this current quarter, we had a pretty good, okay quarter. As you know, the third quarter is usually our slowest one. And this year, we ended up on EUR 51.7 million with a growth of 27%. So top line growth was satisfactory. We were very happy about that. EBITA margin was 11% and the EUR 5.7 million on that, we -- that does include a one-off cost of EUR 5.3 million, which was for the change management negotiations. We still have some way to go on those, but the -- kind of are minor thing to close. So our expectation that includes all the one-off costs. So there won't be any more on those on fourth quarter or later on this particular process. So I do expect that we're going to close that process during this quarter pretty much. And on ARR, the ARR growth was 32%, but that's, of course, IAR effects pretty much on that one. So we did have a pretty good growth on Qt license sales as well. And if we look on the license sales, we do have a bit of a fluctuation on different regions. So this time, the best performing region was really on the Europe. APAC was a good second. And like we've had now some time, the U.S.A. has been the slowest moving on our regions. And if we look on the segments, there is not a whole lot of change on that. As you know, we are in 70 different industries. And medical and defense, well, defense is particularly hot segment globally overall now. If we look in the medical, medical is usually very long-term business. So they look beyond the kind of economic downturns or upturns, they have a very solid long-term plans. And if we look automotive, it's kind of a mixed bag in a way that in China and APAC, our automotive is doing pretty well. And obviously, our customers in Europe and U.S. are a bit suffering. On renewals, we've always had a fairly stable churn. It's relatively low churn. So people that sign up for Qt usually renew and that's, well, the product is good. And usually when people start developing with some products, they don't do kind of a running changes. They go with the one product all the way through the product life cycle. That's a very typical thing. But we do have a very long-lasting customer. So the ones that have started using Qt way back when are still doing that. So that renewal rate hasn't changed. Well, I remember during the spring, there was a lot of discussion that, well, AI is going to replace the developers and AI is going to reduce the number of developers. Well, we haven't seen that, the renewal rates, if we look, people are pretty much renewing what they've been renewing -- what they've been having licenses. So we don't see AI effect on that. We do see people using AI, of course, as a tool helping them, but we haven't seen reductions on the developer license used because of AI. If we look on the -- where do we see usually people renewing less licenses is a situation where there is no new project starting and the project goes on kind of a maintenance mode. So when you are on a maintenance mode and not so on a development mode, then you might need less licenses. But if I look the -- I've been in this chart quite a while. And if I look at the customers we've had in the early days, they're pretty much all the big ones that were there then are still there. And so, in that sense, it's a very stable market. CRA is -- well, that's particularly a European thing because in Europe, we like regulations and laws and these type of things. So the CRA is a -- it's a European thing. And it means that you need to have Docs and kind of a lot of documentation and you need to support your product at least 5 years that it's cybersecure when using Qt 6.12 or beyond, you get that Doc from the Qt and so it helps a lot for our customers on that. So particularly customers, if somebody has been using open source before and thinking that doesn't need a commercial license, this will drive that conversion. And of course, for new customers, it's a good selling point. IAR subscription has been going well. Well, that's kind of expected because in the beginning, we didn't know where it's going to land. But as it started going pretty well, it will continue to go pretty well. When we did the subscription change in Qt, it took us like 3 years and then the majority of the people had changed into subscription. That's my expectation as well on the IAR case. We've done also the integration, has gone forward. All the main things have been integrated. So I think it's fair to say that I don't expect any integration problems anymore. Many things have been integrated. We are working together, and I don't see any major problems over there. So, in that sense, of course, the integration continues, but the -- I don't expect any trouble or problems on that front. So I would say that the integration risk is more or less already behind us in that sense. Of course, when you buy a big company like we did on IAR, there is always a risk that there comes all kinds of integration hassles. But it's been amazingly smooth, I must say, probably the Nordic culture and the closeness helps as well. So, if I -- well, I'm going to talk more about on the future outlook. If we look on the -- what happened during the third quarter, well, of course, there was a bit of a turmoil in oil prices and whatnot, and a bit of a war going. Well, that war started earlier, but during the third quarter, that what's going to happen. Usually, when we have a lot of uncertainty, we do see people shifting from 3-year to 1-year licenses. No major shift now. But if I look at the -- on different product groups, maybe a bit more 1-year license sales than 3-year license sales, but no significant numbers changes over there. So, in that sense, a pretty typical quarter for us, no big surprises. I think, like I said already on the second quarter that we've seen the worst kind of behind us. We're going into right direction, and I see that there is lots of room for improvement in U.S. And our profitability -- obviously, the third quarter is the slowest quarter. The profitability without the one-off costs would have been a bit over 20%. So it's on a good level. I'm already going into the future outlook. So I stop here and talk more about that. But like I said, on the change negotiations or this change we did, we said that we are looking for EUR 20 million cost savings. And that is going to happen. So like I said, we have a few small minor things in Europe to close and then the whole process is closed. So that EUR 20 million cost saving is going to happen. When you look next year, of course, we're going to be investing next year. So don't make an estimation that our cost level will be this year cost level minus EUR 20 million. I mean, business does go forward, but from a current -- this year, current run rate, EUR 20 million will disappear, which means that we announced then that we're looking to downsize roughly 200 people, and that was mainly some in U.S. and -- well, it was U.S. and Europe. Europe was a big part of it. APAC was basically not affected a whole lot at all. And we're going to get into that 200 million number or very close to it when this -- the whole thing is over. So that's where the majority of the EUR 20 million comes from. So more about the future outlook, but now I'll let Ann to go through the numbers in a bit more detail.

Yes. Hello. I'm going to tell you a little about the financial numbers. We had, as you saw, a pretty good Q3 with 27% growth with an exchange rate impact of 26% growth, which was quite positive, and we were quite pleased with it. The distribution licenses grew 18%, which was nice. And as Juha said, we were specifically happy with the development in EMEA in application device licenses and device creation licenses that were. We had a good increase in new sales. But of course, it is the IAR also that affects the Q3 sales as we didn't have IAR at all last year. So the growth that you see in maintenance, 155% in Q3, that is from IAR. And looking at the year-to-date then, the growth was 19.1% with the exchange rate change impact, it was 21.7% at comparable currencies. The distribution licenses year-to-date has declined, but that was expected as we had a very high level of distribution licenses sold last year. And of course, the effect from the new sales in EMEA in device creation licenses and from IAR also is a large part of the growth in the year-to-date numbers. But also, I often get the questions about how the multiyear contracts affect the P&L revenue as it is difficult to understand sometimes because the QT revenues are, to some extent, recognized upfront for multiyear contracts. But looking at the third quarter then and also the year-to-date numbers, there is really no shift to longer contracts to push revenue growth. It's actually a lower percentage of the total revenue share in Q3 and also year-to-date when it comes to those longer 3-year and 5-year multiyear contracts. And the annual recurring revenue then, it grew by 32% at comparable currencies, which, of course, is IAR, like Juha said, but IAR also has a very strong ARR growth during this period as we're moving into subscription revenue, which increases the ARR, but also the QT base business in SQS and Qt also has a nice growth in ARR because we have very strong long contracts and good relationship with our customers and a very stable low churn. And looking at the expenses side, we have done a lot of cost efficiencies and so. But of course, the personnel costs are affected by the reorganization parts, the restructuring. Those are one restructuring costs that are booked in Q3. Those are EUR 5.3 million, and year-to-date, they are EUR 9.7 million. We do estimate that this is the full cost for the restructuring process, even though we haven't fully closed some of the processes yet. We are about 1,000 people at the end of September, and IAR wasn't there last year, and the IAR was over 200 people that were acquired. So even though it is an increase of 80 people, you can see that we have declined the number of employees quite heavily during this period. And also the other costs are actually lower in Q3 than they were last year due to all kinds of synergies and cost efficiencies that we have chosen to do during this period to get a better profitability in the company. It's also a natural effect when you made an acquisition that some of the costs are double. And that hasn't fully been realized yet. There are still entities to merge and still offices to merge. And so we can still have some effects from that moving forward. And the total cost then, they grew by 26% about in Q3 and also year-to-date, about the same. And that restructuring is still ongoing, like Juha said, you shouldn't -- we do expect that we will have that EUR 20 million-plus savings from the actions that we have taken, though we do have certain investments that we're planning to make. So you must have that into consideration when you plan our P&L or analyze that for the future. But then looking at the profitability in Q3, it was EUR 5.7 million, a margin of 11%. But if you adjust for the one-off costs, we actually have an EBITA of EUR 21.4 million, which is quite -- much higher than it was last year. And the EBITA year-to-date was 12.1%. But if you adjust for the one-offs, it's actually in level with last year's 17.5%. So we actually reached 17.9% if we adjust for those one-off costs. So quite a decent quarter that we are pleased with in many ways. Then looking at the Balance sheet. Of course, we have a large part of goodwill and that is kind of a base, say a stable base, not changing from quarter-to-quarter as we don't write goodwill off. But the other intangible assets, which are technology assets from acquisitions and also internal development assets, we write those mostly off over 15 years. The internal development assets that we capitalize, as you might remember, I mentioned to you before, those were EUR 0.2 million in Q3, and this is IAR that still capitalizes some costs in projects, some development costs. But part of the projects have been finalized, and we expect that the rest of them will be finalized during this year. So next year, this will probably not be happening, and we will instead depreciate those assets moving forward as they are final. The trade receivables are around [ 20% ], which is over a rolling 12 revenue, which is quite normal for a software company like us. We have large companies and they require fairly long credit times, but -- so it's a normal and stable level for us. And the ending cash balance was EUR 50.2 million, obviously lower than last year because we purchased IAR, and we are also repaying that debt. So it will not -- I mean, we will have a good liquidity, but it will not increase past last year's level as long as we continue amortizing the bank debt. The interest-bearing debt was EUR 126.7 million, and we still have bank loans of EUR 120 million, and we pay those off every half year with EUR 15 million, so yearly EUR 30 million will be amortized. And on the other receivables and other short-term liabilities, we still have those EUR 5.2 million that we hold to buy the rest of the IAR shares. The arbitration hasn't been finalized yet. I'm hoping it will be at least moving forward. So hopefully, we can close that one soon. Of course, there is also short-term liabilities increasing from the restructuring and letting people go. That is also booked as a debt in the balance sheet. Equity, EUR 213 million is 53.6%. So stable balance sheet. And as we start, I mean, continue amortizing our bank debt, that will also increase over time. And looking at the cash flow, January to September, that was EUR 29.5 million compared to EUR 32.4 million last year. It's basically the same operational cash flow, but we are amortizing the bank loan with EUR 15 million. So that affects the cash flow for a period. And the operational cash flow is affected by the interest that we pay on the bank loan. So that is a short summary on the financials. And with that, I guess I will hand over to you again, Juha, to talk about the outlook. Thank you.

Yes, thank you. So well, we haven't changed our guidance that we've given for the year. So the full year guidance is still the same. So the net sales 10% and operating margin 15%, at least. And we're keeping those. If we look on the business, how I see it. Well, the fourth quarter, obviously, is always the busiest quarter. And at this point, we do have -- the pipeline and everything is over there in place. If I look overall, all our product lines that the -- what we do see, obviously, on IAR, we do have a lot of subscription changes. But if we look next year, IAR is going to be growing because of that price change already. And if I look on our other products, we do see an organic growth on all of them. We were actually on this very same room to -- on our shareholders' meeting and on that -- the shareholders' meeting in the spring, there were questions about our profitability, and we said that we're going to have -- we're going to do measures during this year in order to be on the 30 plus/minus profitability next year. And that's where we are definitely solidly going towards. And as we do have organic growth, and we will have a revenue growth also next year, we will be -- when we said that we're going to do a EUR 20 million cost savings, of course, we're going to be also investing next year. And now if we look at where the investments are largely going in this current environment, they are more and more on AI-related investments like the -- now we had a World Summit earlier this week, and we introduced a -- well, it was renamed Clate. It's about our design tooling and design and one of the major ideas over there, obviously, is that the designers can start using AI and doing their designs and later on, it generates QML code right upfront. So we're going to be investing, adding AI agents increasingly amount on our tools so that they are easier to use. Over time, we do see that this will actually kind of hinder a bit that the languages, underlying languages, like on Qt now, it's a C++. But in the future, the languages are going to disappear more on a background when you can use AI on front. to use our frameworks and our tools. So on IAR, this transformation into subscription, I don't see any -- it's going to go as it is going now and going to take a couple more years to happen. The reorganization thing, we do have some minor things in Germany to close down, but I think they're going to be closed. So as -- we think it's pretty much handled. And the -- well, the market challenges are kind of -- when we operate in 3 regions, I mean, APAC, EMEA and U.S.A. and 70 industries, it's kind of very difficult to say like I was that -- on one sentence that what does the market look like. So we're doing pretty well in China. We're doing pretty well in Japan and APAC in general. Automotive is doing there pretty well, as you know. It was few years back, I said that there is going to be a day that we are going to be driving Chinese electric vehicles. And I think that was a pretty accurate prediction. I mean some people still might have a Tesla, but the BYD is coming. But we're well positioned over there. We're doing pretty well over there. And then in automotive in Europe and in U.S., where obviously, our customers are suffering and not only the automotive manufacturers, but the Tier 1 manufacturers as well. Whereas on medical, we're doing pretty well. All around the globe, defense is a very hot topic. I mean you know that around defense industry, there is a lot of money, a lot of start-ups, a lot of investments going on, and we're getting our share of that. And I think that, that's going to continue for many years to come. So saying on one word that how does the market look like, it's very difficult. But in general, if I look at the -- I'm happy how we're doing in Europe overall when I combine all the segments. I'm happy how we're doing in APAC. It doesn't mean that we wouldn't -- there is always room for improvement. And I'm not that happy how we're doing in the U.S. So I think that in the U.S., we can improve quite a bit. So there is a room for better execution. Well, the long-term growth, I think that the -- what I do get challenged is that -- I actually don't get challenged that will there be a screen pretty much on every device? And will there be touch screens and will all the devices be digitalized? That is not -- I don't get challenged on that. I mean everybody agree. I think that's commonly agreed upon. What I do get challenged is that will AI take over this all. And you can -- well, who knows when we look 5 years or 10 years ahead that what AI can do and what cannot do, right? Usually, what we do is that, when a new technology arrives in the beginning, we overestimate the effect. And then on the long term, we underestimate, right? So we usually overestimate what's going to happen in the next 2, 3 years. And then we underestimate that what the world looks like 10 years from now, that's traditionally been. Of course, AI is moving faster than anything else we've seen before. But that's usually how things goes. So, if I start from the kind of our embedded systems that somehow are functional safety or safety critical and they need certifications and whatnot and CRAs and whatnot, the AI is quite far from that. So, if you go with AI, getting into the -- in a situation that you can have certificates for car brakes and whatnot and these type of things, I think it's that we don't see at all. Actually, we see more than the use of AI is limited in these type of environments because it is crucially important that nothing goes wrong, and it is crucially important people understand that these certificates are met and nothing bad happens. Then we have kind of our in between embedded systems. And over there, we see that people are using AI to help their work and whatnot. But we still see that people are pretty conscious on AI usage and basically for 2 reasons. One is that the -- when you do embedded devices, a crucial cost effect is the hardware, the amount of memory the hardware has and the hardware itself. And manufacturers usually want to have as clean and as efficient code as possible so that they can have the cheapest possible memory and the amount of hardware for specific functionality. And AI -- well, AI still has room to improve on that, right? It also means that these products need to be supported a long time. So it makes sense that whatever is built in there, it's somehow structured and architecture is known and it can be moved forward to other developers maintain and whatnot and their frameworks do help. So I think that the artificial -- the AI models will be using frameworks also going forward in that sense. And then the third is that at least at this point of time, the software usually represents a small fraction of the cost of the whole product, not to mention that what does it cost if you have to do product recalls, right? So it's a -- the people are kind of calculating that how much do I want to save on doing software, increase my hardware costs and taking risks that I might have product recalls versus that I save on the cost of the software? So that all kind of sums up to that humans and developers are very much involved in the whole process so far and thinking about the architectures and using frameworks. Then we come to a software that is not so much on embedded anymore. It's type of software that I'll do my own web pages, right? And if I do my own web pages and there are a couple of bucks, I can fix them tonight and then I have a new web page, right? And for that kind of use, AI is very good. So we see it on a low end where there is a -- the risk of something going wrong is limited there if the AI usage is increasing, and that's kind of the world we see now. Going forward, what we see is that if we talk about a product, I think that the -- and there is no timeline to this. But the -- first of all, we do see that the -- over time, we're going to go more on consumption-based pricing. We're going to introduce -- well, we have already some, and we will introduce more in the -- when we go forward. Some people have asked that is there a -- there is a huge risk involved in that. And I don't think there is because when we're going to start introducing consumer-based pricing, we can, of course, do it in a limited market. We can do it on the new customers only and see that how it really goes. We don't have to roll it up. We don't have to go on a consumption-based pricing to all customers globally, existing and new ones. We can launch a consumption-based pricing in a particular country only for new customers and then view that how that usage goes. So it can be done in a way that you don't take a humongous risk doing it. Going forward, I think that we're going to have more and more AI agents on our products and which will help the use of our products. And those agents will also be combined to each other in a way that if you do coding with Qt, you're going to be able to have integrated testing on that in a way that every time you do a code change, it automatically runs through the test on that code, finds bugs and fixes them automatically and then just reports that this would happen. So we're going to have a design tooling that you can write text prompts and build designs. You're going to be able to have agents that help you creating code with Qt framework, testing with our testing tools and then IAR turning that all to machine language. And I think that then using all those, you can use the whole set of products or you can use a subset of products on consume-based. And AI is helping you on all that. And that's kind of the future in the 3 to 5 years from now. Do I see that the -- well, at least now we see that if we do AI helped coding with Qt and it's done by ourselves, it's far more efficient and consumes tokens far less than if you just take a normal language model and put it to do something. So I think that those efficiencies will always be there. I don't think that the efficiencies having very efficient software on embedded devices that have very little bugs to prevent product recalls, that's kind of the fundamental thing on embedded world, and that's not going to go anywhere. So eventually, what that means is that our addressable market will grow because it will enable people with less skills to do more, right? So when this world kind of comes, it means that you and I will be able to do simple demos. We can do that runs on real hardware. We can be able to do product managers, we'll be able to do simple demos and simple coding and whatnot. So if you think nowadays, if a product management has an idea, then they have designers designing something and then it goes to R&D for development, when this world comes more in reality, you can do -- we can do more of that. And it means for Qt a bigger addressable market as we go forward. Well, then another change that we usually don't talk so much, but it's also happening is the way that the processes are run in companies. And that means, for example, that in former days, it was very important that you were visible on Google and you had people that were able to optimize your Google visibility and whatnot. Well, nowadays, that's kind of getting lost. You need to be visible on AI, right? I don't know about you, but at least I'm talking so much -- whatever I want to do, I ask AI. I don't go to Google anymore. It's even better. I can use my native language on AI and it translates and searches and does what it does and make recommendations and whatnot. So a buying process that you need to be visible on somewhere or a buying process that we had previously that people came to read our documentation, that's, of course, changing quite a bit. So one of the first things you need to be, you need to be visible on AI, which means that you need to be able to train the AI that it knows your product, it can recommend you a product. It knows that how to help if there are bugs and whatnot. And we are very well down on that road. We've been doing that a long time now already. So the -- but that's kind of a change we talk a bit less. So of course, we have a Q4 to close. And of course, that's going to be the busiest quarter and the best EBITA. If I look Q4, I'm pretty comfortable with it. If I look towards next year, this year has been a tough year in many respects. We've laid off people. We've done many not so pleasant things and had one-off costs and all kinds of challenges. If I look next year, on a number-wise, obviously, it's going to look substantially better than this year, particularly on the profit line. And I'm looking forward that it's going to be a much more -- well, it's going to be a better year. So I'm hopeful to that. And then it's always this disclaimer that if something terrible doesn't happen. But I think that the -- our new normal is that there is at least 1 or 2 wars starting somewhere and I think that the -- probably fair to say that for the next 2 years that I see forward, there are going to be a lot of surprises. What happens after that, we'll see. Yes, with that, thank you, and I'm sure you have a lot of questions.

Felix Henriksson from Nordea. I have 3 questions. I'll start off with the distribution license growth in Q3. It got back to positive territory at 18% year-on-year. Was this mainly timing? Or is there something to extrapolate for the coming quarters?

No, it's timing. So the guidance we've been giving before that it's flattish or small decline is still there. So the thing is that we did report the distribution license revenue early on, on a yearly basis. And the reason is the fact that it really goes up and down, right? So on a yearly level, you can kind of see a bit of a trend, but the quarterly fluctuations are really big. I mean quarterly fluctuations like last year, the biggest ones on quarter-on-quarter were like EUR 10 million. So it goes up and down. So yes, unfortunately, I can't promise yet another 18% on distribution licenses, no. So it's like the previous guidance that it's flattish, small decline.

Fair enough. And I appreciate the disclosure on that. Then on IAR, I think in Q2, you stated that around 68% of the customers had been converted into the subscription model. Can you update us on that figure? And you also mentioned that it takes still 2 years to complete the migration. Sounds lengthy given how far you are currently. So a bit of color on that.

Yes. On that particular figure, I need help from Ann because I don't have it. But on the -- before you -- I'll give you a little time. We had it on Qt and we had it -- when we did it on Qt, it kind of -- in the beginning, it spiked and then it kind of slowed down. And eventually, I mean, there are customers that will never convert. I mean they don't know versions and whatnot. So my expectation is that it will slow down. So I think that a few more years is -- and then we are where we are, and then I think that it's pretty much done. But yes, so what was the percentage on third quarter?

Well, it's not -- it is above 50%, but the 68% is not the 68% of the customer base, it was 68% of the bookings at that time. And IAR has a large amount of customers who don't buy anything every quarter. And so there's still a large portion of the customers that will come and want to buy perpetual from us and so -- and then we will convert them. So it's not 68% of the customer base, but more of the bookings there. So that's the number we've been talking about. But that's still positive and more than 50% continuing.

Got it. Good clarification. And then finally, on the acute licensing model change towards a consumption-based model. I think in the last quarter, you were still a bit hesitant to comment about the potential revenue impacts of that. So can you disclose anything more about that? Because logic you're thinking it sounds like there might be some element of disruption given that you're moving from an upfront based 3-year license booking model towards a more stable one?

Well, it's too early to say. But of course, we try -- we will introduce a model that our current customers won't get a huge price increase or a huge price reduction to that matter, right? Because I think that the -- our position in the market is that the -- our products are really state-of-the-art, right? They are really good. But they are not the cheapest in the market either. So I don't think that we have a price elasticity really to increase a whole lot of our pricing. So on a developer, if -- and now I think that the -- if we look overall, the cost levels where we are, how much we charge, even we do go in a consumption-based modeling, it's the -- I'm not expecting to have a huge revenue increase through that, but I don't want to see any decrease either. So it's going to be modeled through that. I think it's more a where the world is going that the people are using -- developers are using AI to boost their work, which means that there is more usage of Qt framework, and we think that fair enough to have that. One equation over there, obviously, is that the -- let's see how it goes. But as you know that earlier on, people were saying that we're going to start using AI and we're going to get a lot of efficiencies. Well, yes, sure, if you use a tool that's free of charge, but those days are gone, right? Now people are paying for tokens. And I know that on many companies, the cost of using AI is becoming visible that, "hey, this is how much." I know that there are companies that they are looking at the -- what are we doing and are we really using AI for something meaningful because it does have a cost attached to it. And if you look at the AI companies and how much revenue they need to cover their investments, we might come to a point that the Indian coder is still the cheapest one, right? So we don't know, right? And we need to view this -- from our customer point of view, they want to develop software and it has a total cost. And so we can't increase a whole lot of our costs in that environment. But -- we're looking for a revenue neutral, but we're also looking for that we go into that consumption base because AI is being used to produce something throughout our frameworks and tools, and we do have quite a bit of tools. So we're going to -- it's not only Qt framework, but also the -- our test tools and whatnot. And so we need to look at the whole equation. But when we do get into that, they are very well integrated and AI is helping on using that. We think it's a pretty -- well, we think it's a winning combination. It's definitely something that the -- I don't know that the whole other companies have something like that. So I see, for example, some of the testing companies that are our competitors, they are trying to implement tools for developers. We already have 1.5 million developers on Qt framework. So they're trying to move into that, "Hey, it's not only a testing tool, but you can also -- something for developers to develop." So this combination on consumption-based offering when there is a -- it's a very compelling offering when we get there. And a very unique one. Let's see that if we can add any value. From a customer point of view, what's the added value is that, "Hey, well, I get a whole set, and I can do the whole thing and I don't have to think about it, and it works very well together." So that's where the benefit comes from.

Waltteri Rossi from Danske Bank. I'd first like to ask about IAR. I know you don't disclose this, so that's why I'm asking. What was IAR's base revenue level in 2025? And how much has it declined in sales this year?

Well, the IAR, you can -- if you look at the historical data on IAR, it was on a fairly slow growth, and it was on a bit over EUR 40 million. So I think it was EUR 42 million roughly, and it was on 5% growth year-on-year. So it was a bit over EUR 40 million. And so of course, there was not a whole lot of changes when we acquired it last year. The subscription change did start, but it was on a very early phases, and I think that we've accelerated it this year. So the revenue decline has really come this year. And like we discussed already offline, we don't disclose the BU numbers, so we don't disclose the BU numbers. But we've given -- we were discussing early on that it's, well, high single-digit millions, let's put it that way. Yes.

You mean decline has been higher.

Yes, decline, yes. So between 5% and 10%.

Year-to-date?

Overall, I'd estimate that to be this year, not year-to-date, but to give you some idea. So I think that some analysts were estimating that the decline can be even high as 30%. And I think that -- well, that was a very high end. So not that bad, but substantial.

Okay. That's helpful. Then second question relating to the license maturity mix. You mentioned earlier that I think 1-year licenses are going a bit more than 3-year licenses. So to me, that actually sounds that the split would be close to 50-50 right now of sales. So can you give any indication on the split between 1- and 3-year licenses? Is it more like 50% or 30%, the 3-year licenses?

Well, that number I don't have in my head, but I knew that you're going to be asking this. So I was just looking at how much we sold on third quarter 1 and 3-year licenses. And of course, we do have multiple products and the changes were very little, but on 1-year licenses, it was a bit more bit more than on the 3-year licenses, but no substantial change really. I mean, I didn't calculate the million. I was just looking at the different product lines. So a bit more 1-year license, but no big change in the -- no meaningful change in that shift really.

All right. And final one from me. You have now publicly stated your plans to change the pricing model at some point. Have you seen any kind of -- from customers, any buying -- changes in the buying behavior that they would be buying now the licenses with the old terms?

No. No. I mean the -- our customers are like Daimler and LG and Fortune 500 companies and they -- very long relationship with our customers. So they -- and if you look on the products that where we're at, I mean, of course, we do negotiate with sourcing organizations and what -- and of course, they are tough negotiations and always that can you give more discount than this and that. But the very long-lasting customer relationships, the users are actually deciding that what they want to use and then the sourcing organization is negotiating. And if we don't do anything drastically stupid, they're not going to start changing. I mean -- and then if you look -- well, we talk too much about cars, but let's talk about medical devices, for example. Our cost in the overall whole product, it's like this big, right? If you think that companies, they want to use the best-of-breed products to build their software, and if they can squeeze the software development, tooling cost and testing cost, they are optimizing the size of the overall cost. If you think car, for example, I mean, yes, there is a software cost, but it's -- our part of that software cost is like this much. And so we're not the first line to start squeezing. It's more important for those customers that they -- for example, in Europe, they fulfill the CRA requirements and the software works and they are not recalls and it looks cool and be optimized. So that's kind of the position where we are. And then we -- so I don't see -- the users are happy with the products. They are not changing. So then we have the -- and most companies are very big. So the pricing discussions are with the sourcing organizers. Their target, by the way, is always that, lower the cost by 5%. So it's a funny discussion. They want to have more cheaper, and we want to -- yes, so that's how it works. But yes, yes, absolutely no. And we're going to move very cautiously on that consumption-based pricing models as we go. We have had customers that are using AI and they've been asking that, "Hey, well, how do you want to charge this," right?

It's Matti Riikonen, DNB Carnegie. A couple of questions. Firstly, you already discussed a bit about the difference between ARR and organic growth. But could you just quantify that how much the maturity mix in your growth is affecting your top line growth? And how much that has done in the first 3 quarters of this year? I'm just trying to understand that what kind of change in the numbers that we see is really affected by the maturity mix.

Thank you, Matti. That was a very specific and very articulated question. So would Ann want to answer.

Maturity mix, you mean 3-year and 5-year contracts then. How much of those are -- I can say that, as I said when I was up on the stage also about the financials that the multiyear contracts are in Q3 and as a share of the total in Q3, they are lower, the multiyear contracts than they were last year. So they have a lower effect than they did last year. So the growth does not come from that. And it is actually the same for the year-to-date, but much smaller change. So it's more close to the same share of the total revenue. Does that clarify it?

Yes, but for exact, we'd actually probably look into...

All right. And no major changes in the quarters either, I assume. So Q3 has been kind of favorable in terms of -- or not as favorable as is earlier.

Not as favorable, exactly.

But then in the first half, you had quarters that had the positive impact on top line growth.

You mean Q1 and Q2?

Yes.

I believe in Q2, we had a higher degree of the multiyear contracts. We communicated that, right, than we did in Q1. So there was some effect there in that quarter in the growth, but not in Q3. I would say, quite the opposite.

All right. Then the second question is that normally, when you have prepared yourself for the next year, you have been growing your resources. Now you also say that you need more resources in '27, so you need to invest. But of course, in this environment where you have just sacked a lot of people that might be a bit difficult. So when can you start investing in growth more? Is that going to be in first half or Q1 next year? Or how do you plan to do that?

We are -- I think that -- well, we are hiring already, right? And it's a bit of a -- well, of course, we do need replacements. People are leaving voluntarily as well. And we do need increased number of people, specifically in APAC. We do see -- we are hiring that. And then on a growing scale, we need more AI capabilities and not only on R&D, but also on the business side that how do we use AI in our own processes to improve them. And then on the R&D side, if you think that what I was describing that we go in the consumption-based model kind of a platform where all our products are and we build AI agents on top of that, it's a -- we need new skills on that type of development, and we are doing those hires as we speak. So in that sense, it's a bit of a shift. And I think that next year on many companies, you're going to be seeing more of that, that there are change negotiations where -- and at the same time, same companies are hiring.

All right. Then you talk about the importance of regulated customer industries for you so that many customers are actually in those kind of environments. What is the revenue share of your customers actually being in regulated industries?

That I don't know. I would have to -- because it's not so straightforward. We do have -- if you take -- well, we talk so much about cars that I met some investors that they thought that we are only in the automotive industry, and I decided that I'm not going to talk about cars, but I'll talk cars. So I mean on cars, you do have functionalities that are not safety critical at all, right? If you think back seat entertainment, right? And then you have brakes. So, to kind of -- that's why I'm saying that I'm a bit -- and on defense say the same thing that you do have applications and functionalities on defense that are not safety critical and then you do. So it's not so that you can say one industry, we should kind of look into their functions. But that's a very good question, and that's something that we try to -- we need to look and see that the -- where we would get that. IAR is very much in that business, and we have an acute framework quite a bit on that business. And then our testing is more broad on all kinds. But sorry, I can't give you an answer to that.

All right. Then could you just briefly describe your momentum in your different business segments or product segments like Qt legacy, then you have the testing and quality assurance and then you have IAR. You have already touched on the IAR, which seems to be doing fairly well, considering that there is reductions to be made. But in Qt legacy and TQA, how would you characterize the momentum in those?

Sounds so bad to say Qt legacy.

Okay. Let's define it as everything else than testing and quality assurance and other acquired businesses.

Yes. Well, I'll try to describe you a bit. So before I was saying that the -- if you look at successful frameworks, tools, they are EUR 200 million plus/minus something. I mean, the ones that have been really successful. And then you have quite a lot of businesses that are EUR 40 million, EUR 50 million, right? So you can say that Qt has been already among very, very successful. I mean, we've been able to create a user community of 1.5 million. Many big companies have been trying to do something like that and failed. So it's a great success. Qt being where it is, it's been a great success. But the thing is that there are whatever, 25 million, 30 million developers in the world and 5 million, 6 million C++ developers. And of course, that's the limiting factor on that how big Qt can become, right? Then we became this business model that we have the distribution licenses, right, which is price per unit and that kind of gives you another alternative growth, which we've been using, and that will continue to grow over the course of time. But of course, we were coming to a point that the growth will slow down. And I said that over time, maybe Qt can be like a EUR 300 million business and where do we grow from there. And then the universe send us AI. And now AI actually means that the -- as we go further develop this, the more people will be able to do coding and using Qt framework through AI. You don't have to be a deep tech C++ coder. You can be a developer. And so, our addressable market in that sense is growing. And well, same goes with all our products basically. But so if you look at the old Qt or Qt itself, if we can keep a Qt on a 10%, give or take, organic growth, that's what I'm looking for on a longer term, right? And that's -- then if you look -- you asked that how would I rate the products? Well, Squish and testing, that is something that I've said before that the -- I would say that, over there, you're going to be looking double-digit growth starting with the tool. And that's going to -- originally, we said that our target with Qt is a EUR 100 million business and a lot of people were smiling at us at that time. Maybe they were thinking that our target is too low. And well, we got beyond that. I see Squish -- I've said before that I see Squish like Qt 2.0. So Squish has a bigger addressable market. If you look at the -- everybody needs automated testing. It works on multiple languages. It's very well integrated into Qt framework. So if you're doing something with Qt, it makes all the sense in the world to do the testing with Squish and use also Coco for code coverage. Axivion itself is one of the best tools for static code testing and the architecture verification is a very unique offering on that. So it's very easy to see that it's going to grow into EUR 100 million business. And the growth now in the early days will be that 20-plus/minus-something-percentage on the coming years. Well, on IAR, it's on its own right in its own segment, the -- with this price increase on the next couple of years, you're going to be looking like 20%-plus/minus-something growth. So that's kind of give you an idea. And really, where do we need to spend the time with IAR in the next couple of years while it's growing through this price conversion is the fact that what product development we do over there, where do we find new customer segments and a bigger addressable market for IAR to continue that organic growth after that. And so -- that's basically it. Would I call Qt legacy because it's percentage-wise growing less? Well, not necessarily because it is bigger than all these others, but these others are on the earlier phase of their path. So if you think the really old way of looking at the product life cycle, obviously, the Qt growth is on percentage-wise slowing down and these newcomers are faster. If you look absolute euros, then the story is, of course, it kind of evens out. And what we're looking for, the design tooling, yes, it's interesting to see. I actually don't expect huge revenue numbers out of it. I kind of see it more like an integrated tool, making our offering more complete so that you can start from design and then you move on into developing the things, whatever you're doing. But let's see the -- it might be a product of its own right, and we're going to -- people are buying only that, but that's -- I don't -- let's see, but I'm not counting on it. I'm counting on it more that they are early on, people can start this. They can do the design, they can see the result on target hardware and then they can make it actually workable code. That's the add value, of course. Only 6 minutes over time. Thank you so much. Thank you very much. Looking forward to seeing you after fourth quarter.