Good afternoon. My name is Jukka Tapaninen, Aiforia CEO. Welcome everybody to the first half earnings announcement today. I'm going to give some highlights of what happened in the first six months. Veli-Matti, our CFO, will go through the numbers. I will highlight a little bit our strategy going forward. Let's get started. The first half of the year, it was in many ways, I think, really, really good. We have some new closings on the customer side, and I will cover this in a later part of the presentation. Our order book grew 58% compared to last year's situation. On the revenue side, we were flat, but there are some good reasons why that happened.
All in all, on the sales side, we got a good success, and we were progressing as planned on getting those kind of first clinical customers so that we can use as a reference, and we can scale with those customers and then as a showcase for the others. A significant amount of those clinical agreements done. What was, I think, even more important than winning those customers was the regulation part. As we are dealing mainly now on the clinical side, of course, we have research customers, but the focus where we think that, okay, commercial perspective, the business is growing, is the clinical. There you need a lot of different kind of certifications related to security. We have this medical device category where we have our solutions. We got the IVDR. That's the new regulation for Europe.
It used to be CE-IVD, and it's still CE-IVD, but under the IVDR regulation. We got significant progress on that side because that's actually the important part so that we've been able to build the biggest portfolio of CE-IVD-marked solutions in this marketplace. That gives us a good opportunity to move forward. Of course, we are not alone. This is kind of an interesting marketplace. You have scanner vendors, you have the laboratory information system vendors, image management vendors, AR vendors, and then you have kind of a reselling partner.
We've been moving forward on all of these fronts and strengthening the partnerships and trying to find new ways to the market with those companies. I just wanted to show this because if we look at the history and not going back to the 10 years when the company was set up, but building the core of the technology, getting the research customers. In 2020, we decided to go on the clinical side because of commercial reasons and multiple other reasons. There is a huge need for additional resources and quality improvements. Of course, all of those things that the AI can provide, providing things that the human eye cannot do. In 2020, we decided to go to the clinical market. In 2021, we won the first major deal, which was Mayo Clinic, number one hospital in the world.
That is the best reference you can get in this industry. In 2022, we introduced the first five CE-IVD-marked solutions. Going forward from then, we have been growing the number of clinical closings every year. In 2023, it was three. Last year, it was five closings. In the first half of this year, we closed five. We are on a good track to achieving our targets on that side. Just to name a few, the NHS was a nice deal. The Memorial Hospitals Group in Turkey is important as well. They took five AI models from our portfolio, and it is the biggest private hospital in Turkey. That is a really important one. We have a lot of deals from Italy. There was some acceleration coming from the European Union because there is this COVID recovery fund. So far, there are 15 hospital regions in Italy.
We have closed actually seven today and a couple of really meaningful ones. Veneto region, 5.5 million people. On that particular opportunity, we are going to go to the next phase because solutions are implemented, and now they continue to use it. That is probably the first real case as this business should be evolving going forward. The Lombardy region that we closed earlier this year is really significant. If Veneto was the size of Finland, Lombardy is the size of Sweden. It is about 10 million people in that region, 31 hospitals. We could use Veneto as a reference for Lombardy. The customer said, "Okay, we want to have exactly the same system as Veneto has," because they wanted to get the best practice and guarantee that somebody has been actually implementing this successfully. Multiple others.
I think Fimlab in Finland is a really important deal for us as well. There were multiple lots that they chose Aiforia. They chose solutions for nine different areas, and Aiforia is part of all of those. In a couple of areas, there are competitors offering. As far as I know, Aiforia is the only company that has got actual orders from that public tender. We try to serve our kind of key customer in Finland as best as we can and make the customer happy in that way. My point is that there are multiple good closings. The Paris region, for example, is the biggest hospital group in Europe, 38 hospitals. We have delivered one AI model to a couple of hospitals, but there's a huge upsell opportunity. Even more important than that is that now we are in France. We have delivered something.
We are considered as a French company, and the French government is now supporting the adaptation of AI to hospitals. They're giving grants so that the hospitals can do the initial investment on this area. All of those kind of public sector buying portals and all that kind of bureaucracy related to the French market is sorted out. We also set up a legal entity so that we see that market is the next big thing after Italy. All in all, I think Europe is doing quite strong on this AI side. What else? We are a technology company in a new market, and we keep on going, investing so that we stay ahead in the technology.
Currently, our offering is the widest and deepest in that sense so that we can really provide the information that pathologists are needing for their decision-making and provide efficiencies and accuracies doing that and readiness to go to the next level as well. We've been investing on integration with the multiple image management companies and LIS vendors so that it's easier to scale. That's something that you need to do, and it doesn't show as a kind of it's not really visible, but you need to make your solution to work with all the leading vendors in the market. From this slide, I think the biggest point is the last bullet point, which is the Aiforia Create. That's our development platform. We've been renewing that technology. We've been heavily invested on foundation models. We are using the transformers.
We are connecting with the third-party foundation models and getting ahead on that technology. It means basically that we have the current version, the older version, which is a CNN, convolutional neural networks-based kind of AI engines. Now we are bringing on side this foundation model technology. I think what is meaningful here is that, okay, so what does it actually bring? What are the things, nice words, buzzwords, what's happening in the industry? The important thing is that, okay so we can develop models faster and cheaper and with better quality. I think those are kind of three things that come out on our foundation models. Obviously, we are kind of keeping a kind of a tight look on what's happening in the market from a technology perspective so that we can leverage all the latest ones in the market. This is what I already mentioned.
Really, really a big thing. May 2022, Europe changed the regulation. Before that, it was a CE-IVD. Then May 2022, they introduced this new enhanced regulation, IVDR. We started to kind of build our processes, document our processes, doing those audits. It took some time because there were no notified bodies in the beginning, so on and so on. All in all, the good thing is that we have done this kind of groundwork for this regulation. We have been accepted. We have that in place. Under the new regulation, we can bring AI models. We don't need to do all the kind of first steps again. Now we can focus on bringing new AI models in the market. This is the main thing if you think about it. As I said, medical device and CE-IVD.
If you want to sell a clinical market, there are only a few companies who can actually do that. You can have always those discussions so that, yes, somebody can build an algorithm. That's not the solution that you can deliver. No regulation, no kind of maintenance cycles, user experiences, all of those things that a software company needs to have in place. We have that. I think we are ready to scale on that side. I give now word to Veli-Matti to continue with the numbers. I'll come back with the strategy update at the end.
Thanks, Jukka. Good afternoon from my side as well. In the following, I would go through some of the key financials from our first half 2025 report. As mentioned, our revenue growth was 2% to €1.4 million, and most of that came from Europe. Finland represented 14% and North America 25%. The order book growth was 58% to €5.1 million. We invested actually €3.7 million, with total product development as the main part of that, €3.6 million. That's about €1 million more than last year. As Jukka mentioned, there was successful progress in our product development in AI models and also this new AI engine, which is great news for me as well because I have to check out the costs. EBITDA improved by €1.2 million to - €3 million, and EBIT improved by €700,000 to - €5.4 million. We had, on average, 69 employees during the quarter.
That's a pretty big drop from last year. On the other side, we have enjoyed new colleagues already this month that have started. We're going to wait a couple more to start within a month or two. We had €11.9 million cash at the end of the period. The revenue growth was flattish, so pretty much the same levels as the year before. If you look at the figures, you could see that the clinical business growth was approximately 60%. If you add on that the delivery that Jukka mentioned would have happened, let's say, first half, the overall growth would have been something like 35%, and the clinical growth would have been over 100%. The order book growth, as I said, was 58%, and that came from the clinical business. The growth was €2 million or 90% from the year before. Costs, we are pretty tight on costs.
Although the employee cost went down to €3.7 million, the other operating costs went up to €4 million. That's due to the fact that we invested pretty much on the computer power. The cloud costs were up, and we purchased quite a bit of samples to enable the model creation, which also increased the other operating costs compared to last year. Going forward, I'm sure that we can keep cost control in place. Obviously, we are in a growth phase in that sense that we still maintain and have a strong mandate from our investors a couple of months back that we should keep up the progress on the product development that the market requires. This is pretty much from my side. Thank you, and Jukka.
Yeah. Just wanted to point out on the numbers. As Veli-Matti said, on the clinical side, we grew 60%. Overall growth was 2%. The debit here is in the details as well. We decided to go to a clinical market. We decided to build the portfolio for that market and also capture the first big customers. On that roadmap, we've been moving forward very, very, very well. If we look at the profitability on the research side and the clinical side, obviously, we see that the profitability, profit will come from a clinic. Now we have kind of first steps on that side of the business. It seems to be growing quite nicely. This strategy, I've been showing this slide a few times, but still very valid. There's an active problem recognition. Our customers, they know that they need something. They're actively looking for solutions.
That's why the large flagship hospitals are really important. The reference value what they gave, it's really, really important in this phase so that somebody can show what is the best practice, how it will be delivered, what type of components, what type of solutions you need to build a kind of working model. Demand is going up. The cancer rates are going up, aging population, all of those same basic things are existing. Also the capacity, in the different continents and countries, there are no countries that, okay so there would be an overcapacity of pathology. There's a need. That need is growing. Our kind of play or place in this workflow, we are talking about the physical samples that if there's a prediction so that you might have a cancer, so leads to a tissue sample. Tissue sample will be prepared, scanned.
Then it comes to a software piece, which is kind of uploading those images, sharing those images to pathologists, storing. What currently drives the market forward is the AI. That's the area where we are. We are the ones who understand what's on those samples. We can take the information from those huge files because physically small, but we are talking about 2 GB size of images. Then you are looking for really, really tiny details. What Aiforia can do here, we do it much better than, yeah, any of our competitions. They're not just the heat maps. We go deeper. We do a quantitative analysis. We populate the report. We actually give the information that the pathologist needs for their decision-making. That is in hot demand at the moment. That drives the demand for the image management vendors, the software vendors.
That drives the demand for the scanner vendors as well. If you turn it around, the scanner penetration is an important piece as well so that they have the devices so that they can actually have the digital images. It uses an AI. All in all, they are connected to each other. What we have, we have the technology leadership. We use the foundation models, transformers. We use the CNN, which is the latest kind of technology as well, which is actually working and deployed. We are recognized as a leader on this side. We have a scalable business model. Now we have a couple of customers moving from a first phase to the phase that they are actually using the software and paying for the usage. That is a profitable part for us because in the first phase, you need to invest on the selling.
You need to invest on kind of delivering the solution. Once it's there and the customer is ready, you can add more AI models, and you can scale up the usage. That is good for the software company. Interoperability, as I said, we are connecting with multiple systems. We have interoperability agreements and testing done with the multiple scanners and multiple image management systems. The last point, this is something that I would like to kind of everybody focus on as well. It is a certified quality. To be able to operate in the U.S., hosting patient data, you need to have a HIPAA compliance. You need to have multiple different kind of security regulations and certifications in place. You need to have some patents. Now you need to have the IVDR-regulated products and the CE-IVD products under that, meaning that it is a heavy investment.
It is a barrier of entry for this industry for sure because whoever wants to come into this industry, it takes time to get all of those certifications in place. We have it, and we can now start selling in a volume on that side. Just a visualization of different. We have been looking at kind of main kind of areas where we want to part with. I have to say, we have a pretty good offering at the moment. That is the most kind of or widest in the market at the moment. Of course, we have a pipeline of new coming products as well. This is an interesting picture because four years ago, when we did the IPO, we said, okay, we didn't have any CE-IVD those days.
We said, okay, our strategy is that, okay we build a portfolio of solutions that the users, they can cover 80% of the workflow with this because it's not really convenient, maybe not even possible that, okay you kind of pick and choose vendors, five to three vendors with a different user experience, different maintenance cycles, and so on, and try to manage that as a solution. We decided that, okay, we bring all the relevant solutions with the high quality. We aim to 80% coverage by the end of a decade. Currently, we have a 60% coverage. We cover all the kind of major kind of disease areas on pathology. This is something that, okay we as a company, we are really, really proud of. Of course, now working with kind of increasing the number closer to 80. We are well on track on that target.
Coming back to the growth rate, 2% versus 60%. I pick up the 60% because we've been focusing on getting these flagship customers in place and growing the clinical business because that gives us the opportunity to scale up big time. In 2021, we got the Mayo Clinic, and then we built those solutions. Three closings on 2023. This is, by the way, not all the closings. These are the ones that we've been focusing on because we also gave one target out so that we want to have 15 kind of customers where there's a potential ARR over $0.5 million. Annual recurring revenue over $0.5 million. It doesn't happen in the first year. There's an opportunity to build up those customer relations in that level. Last year, five closings. Really interesting one, Fimlab I mentioned. Castille and León, interesting one. That's something so that fantastic deal.
It's a big region in Spain, first good referenceable customers together with the Swedish image management vendor, Sectra. It's a fantastic deal. The challenge there was that they wanted to have a local implementation. Now we are delivering that. It will happen in Q3. That was impacting a little bit on the revenue that we could book in the first half of a year. We could have actually kind of do a little bit better on a RevRex. The main point is that, okay it will be now delivered. We provide a fantastic solution to the end customer. Those are the closings from last year. This year, five meaningful closings in the first half of a year. You can see that France is really picking up. A couple of hospitals from Paris and we won the Nantes region.
Curie Institute is interesting, a really good reference, and they have already announced that they are using Aiforia. We got the deal, so that's why it's in the list. Really kind of nice closings. Lombardia is, of course, commercially also a big one, moving forward. So far, every tender that we've been participating in, we have been winning. In some cases, they've been choosing one or two other vendors on the same tender, but we haven't lost any. That's good to keep in mind as well. Going forward, I know that there are questions coming, like how do we accelerate, when do we actually see the revenue growth happening. We had a big strategy exercise a couple of months ago, and the outcome is that we see the opportunity in Europe. The scanner penetration is 30% in Europe. Sounds low, but it's actually much better than elsewhere in the world.
We see these public fundings speeding up the adaptation, and that's why, as a European company, it's natural that we want to capture that opportunity and accelerate the business in Europe. We will grow the clinical offering, so there will be more CE-IVD-marked solutions coming up. We'll look at the other kinds of opportunities, how we can improve the user experience close to our core and domain. We've been obviously planning all of these different areas, how much revenue and what type of growth will come from those areas. This is at a high level, short term. Long term, this is obviously truly a global business. We have a good presence in the U.S., so we want to capitalize on that and start investing again in the U.S. going forward. We have some interesting ones. I mentioned that when you talk about the AI, currently what we offer is efficiency.
You can do much more with the same resources, you can do it with better quality, you avoid mistakes. The other thing is that now the prognostic models, what we built with the Mayo Clinic, so that you can start predicting the patient outcome, combining the baseline data, combining the AI-driven data. This is something that builds up the value of AI. First, the efficiency piece, but now the additional value are coming from things that you can't do really with the human eye. We have also kind of been talking with a couple of pharma customers. By the way, I didn't mention that the Orion was a nice win from Finland on that side where we sell Aiforia study solutions. That's in good shape. Pharma here means that you talk about the companion diagnostics.
We've been approached proactively by a couple of very large pharma companies that are thinking that, okay, when they are building a kind of medicine or kind of that you need to kind of combine the AI. AI needs to kind of drive certain results, and based on that result, you can then give certain treatment or certain medicine to the patient. Companion diagnostic type of things, but those are long term. That's why I put it in 2030, even though some of those discussions already started. This is a really interesting marketplace, and there's a good opportunity to make it a really nice business. That's it from my side. Now there's room for some questions. I know that Antti is already ready.
Hey, Antti Laura from Inderes. Let's start from the top line and then go a bit further into the financials. First of all, on the revenue, of course, it's been on a relatively flat level for the last two years. We've seen that the preclinical revenue has shrunk underneath, and the clinical side has been taking over. Do you think this will be a recurring trend that the preclinical will be going sort of down, or do you think there's room to bring that up again?
Yeah, relative terms, it will go down for sure. In absolute terms, I think it will stay more or less in the same level. Our focus is really on building up the clinical business. I have to say one thing so that we have an offering, Aiforia Create, which is the fantastic development tool mostly used in research. The same customers who are buying on the clinical side, they want to also buy the kind of research tools. It's something that makes us really strong on a clinical sales situation because we can combine the offering. The focus is clearly on the clinical side.
Right. When you say the preclinical, you expect it to be roughly on the same level in absolute terms. Is there a sort of a cyclical element that drives that up and down between?
Not really. I mean, I think it's quite stable going forward. It's kind of easy to predict because it's a type of a license play or SaaS play in that sense. You pocket the license fees upfront, and then you recognize them throughout the license period. It's pretty stable. Previously, when we tried to find business anywhere, obviously, let's say in the U.S., it was easier to get those research preclinical type of our clients. You were capable of attracting some of those which did some projects going forward. Now, since we are more concentrating on the clinical side, that kind of dropped out from the baseline revenue base. It's stable. Going forward, as Jukka mentioned, the research part includes the pharma companies. I think we have a great chance to take that business as well going forward.
Everything we have done has some meaning. If you think about 10 years ago when we started, we started to build a platform so that you can build these AI models. We brought the first kind of version of Aiforia Create out in 2017. We started to get the research customers because it's a fantastic tool for them. The outcome is that we got the 5,000 users from all the major universities and the research groups. We got a lot of scientific documents out of that, proof points that the technology works. In 2020, we started using that technology to build the offering to a clinical side. We don't want to give away the research business. We are not, it will go there. It's not something that you can scale up big time. The clinical side, yes.
On the clinical side, were there any projects or customers where there were some timing effects, some deals and revenue getting booked to, let's say, H2, for example, instead of H1?
I think we are quite positive on an annual level. I mentioned the Spanish deal that the customer wanted to have because the tender rules were clearly saying that you need to deliver on-premise. That was the first on-premise delivery for us because we are cloud native and everything is in the cloud. In this case, we, of course, wanted to do what the customer asked. It took a couple of months longer. It should be now kind of delivered in Q3. That was roughly about a $0.5 million deal and a big part of that money we could have actually recognized in the first half.
Okay, thanks. That's helpful. On the order book, of course, it has risen from the last year levels. Compared to year end last year, it actually has come down slightly. It was roughly €5.2 million then, now €5.1 million. Can you explain this dynamic a little bit? Is there some sort of contract rotation getting booked into revenue and not rebooked into order book? What's the dynamic there?
That's pretty much as you explained. The big drop, I would say, a couple of hundred thousand, comes from the order book from the research side. The clinical side has been flat or growing a bit, not as much as we wished. There were some negotiations going on, which hopefully would close up this half. It's a bit tricky to kind of forecast how they will be seen. The pipeline is quite nice.
Yeah, I know that there's a little bit of a disconnect so that you see 2% number. You asked before if I'm happy with the results on the first half, absolutely, I'm happy. If you think about the kind of clinical growth, 60%, then a couple of deals kind of moving to a second half RevRex. We could have had quite nice numbers. I think the clinical side is absolutely kind of on target on that perspective. On the cost and overall, where are you investing as a company? I saw your employee count was coming down a little bit in the spring, and now you recruited back again. Have you done some shift in priorities on where you have?
Yes, that's correct. Obviously, there's going to be a rotation going forward. We'll be increasing the number of sales and number of support type of personnel. On the other hand, we maintain the high level of investment in product development. That's kind of like the change. There has been a rotation on the personnel. Obviously, you want to guarantee that you get the best possible people back. Whenever you recruit, it's not happening like that. It's kind of like a McDonald's type of recruitment going forward.
Yeah.
Sorry.
As we are using the latest technologies in software development, if you use all of those tools, you don't necessarily need so much more resources. You can allocate those a little bit differently now because the AI is helping to take out certain things. That's helpful. Also, on the cost side, you mentioned that you've been buying some samples for the R&D purposes, and that was one cost item in H1. How do you see that area going forward? Of course, you are already, if you said, around 60% of coverage in the pathology workflow end of this year. How do you see that area moving forward? Do you still have needs?
Yeah, definitely, there's going to be needs. Going forward, you need to have a certain level of investments on those samples. We are a data company, so obviously, we have to have a certain type of approach to acquiring data, whether it's samples or whether it's some other type of resources. Data is kind of like our core. It's like the Tempus Paige deals. One of the arguments why they did that was the data acquiring. Definitely, we're going to, we have to just justify what's the right level for us compared to the top line growth.
That's good. On the U.S. market, of course, on your strategy, you said that the focus would more be on Europe in the coming years and then U.S. more 2030 and onwards. You've been thinking of the FDA application as one element. Should I not read into this too much and say that has moved forward several years? Are you still thinking actively about the FDA?
Yeah, we've been dealing with the FDA. We did the first application, and we got the feedback. We have a clear idea of how we are going to do it. Whether it's happening this year or next year, I'm not going to share that information. It could take some time, but it's definitely on our strategy to do it.
Good. Thanks. That's all from me.
OK, thank you, audience here in Helsinki. Now we will take a couple of questions from our online audience. There are a couple of questions regarding the order book. It was roughly the same level as it was at the end of last year. Is it a fair assumption that the latter half of the year will be stronger in regards to order book development?
I can't give you any forecast for that. Obviously, that's the target and aim for us, that it's going to grow.
What I can say is that, okay, we have a healthy pipeline of business. It seems that the market is really opening up. Whether we capture all of those opportunities, we'll see that. Currently, it looks really, really promising.
Thank you. Further with the order book, revenue grew by 2%. Why didn't the order book turn as a revenue during the first half?
Yeah, there's a few things. Veli-Matti, you may comment more. From my perspective, two big things are Lombardia region. We closed that in January, and the delivery hasn't started. As we are kind of moving forward with the IFRS, we cannot book the revenue until we start delivering something. We expect that the deliveries for that particular deal will start in the second half of the year. I already mentioned a couple of times about the Castille and León case in Spain. That's the same thing. That's happening right now as we speak. That will be a revenue for the second half of the year. What comes to RevRex, I think we pretty much know all the deals that we are going to recognize because those have been already sold. Now it's in a delivery phase. There are a couple of bigger customers.
We need to deliver, and the customer accepts the solution, and then we can book it as a revenue. Going forward, it's a pipeline and all of those open cases because it's not likely that we close something now and we deliver before the end of the year and recognize as a revenue. We have a good visibility on the second half.
Thank you. Further regarding the order book, how do you recognize the orders in the order book? Is that the total value of the contracts throughout the whole contract length?
That's correct. That's the right interpretation.
Thank you. That was all today from our online audience. Any further questions from the audience here in Helsinki? Okay, thank you.
I would comment on the last question also that obviously, the person is interested in what's the duration of our order book. It's a bit less than two years on average. That might clear up the understanding how we could approximately recognize those orders.
My closing words, three points. It's kind of a fashion to say three points. First of all, keep in mind that we started a journey in the clinical side in 2020. We have been able to build the biggest and best portfolio in that market with the regulatory approvals. That's the first thing. The second thing is that we have actually won real customers, really kind of leading hospitals from the U.S., from Europe, from the main markets. We've been starting, and some delivered already, and some started to be delivered. That's going strong. The next phase, third point, now it's time to kind of scale up. That happens with the different partnership and kind of strengthening our own sales efforts, but also kind of finding nice new ways to the market through the partnership and the channel as well. I think we are in a good position.
The last point, keep in mind the 60% growth on the clinical because don't mix up with the 2% overall because that's the kind of a piece of the kind of preclinical, which is not growing so fast. Where we are focusing also, that's growing really nicely.
Thank you.
Thank you.