Smartoptics Group AS

Good morning and welcome to Oslo and Hotel Continental, a fairly calm Oslo, clearly affected by winter break and the fact that Norwegians ski this week. Welcome to you in the room who are not skiing and also welcome to all of you online. We are here to present Q4 of 2025, Smart Optics. And also, of course, full year 2025. We will start as we always do and talk a little bit about the highlights of the quarter. Like last quarter, I will let the numbers speak a little bit for themselves. Stefan will come back and talk in great detail about everything that you see on the left of this slide. I will just conclude that it's a fantastic quarter and a fantastic year for smart optics. It's an all-time high revenue quarter. We're seeing demands stable to accelerating. I would say accelerating towards the second half of the quarter. And continuing, we're seeing increased traction with our large accounts, independent of customer type, network operators, neoscalers, cloud providers, and so on and so forth. And we have an absolutely fantastic market in the US. And interestingly enough, we have a Europe that looks very promising and looks very interesting for the coming years. And I will come back and explain that. last but not least from a product standpoint we're growing in all our product areas and and i'm very happy to report that it's a record quarter with super good growth in our business area optical devices which is a business area where we change the leadership where we started to invest about a year ago and and we're seeing like smart optics in the past where we spend our money where we invest we also get growth and good business so it is working We are continuing to invest in the company. I said in Q3 that not investing at this point with this great market ahead of us would be foolish, and I will just reiterate that. We are going to continue to invest. And the question right now is, what do we mean by disciplined investments? Well, we do mean that we're going to invest or rather grow our OPEX slower than we grow our revenue. And we are going to grow with profit. But the most relevant question now is, is this a good time to hit the accelerator to capture the market opportunity that's coming from AI-driven capacity extensions over probably the coming decade? And we will, of course, work hard on that over this year. But it's no doubt investments are continuing to for yet another quarter where we are proving that we are the challenger of challengers in our market space. with one exception in the numbers that has been reported so far to the industry analysts with one exception, which is purely related to hyperscale build out. We are growing far faster than anyone else. So we are the challenger of challengers. I want to talk a little bit about what's happening in the market and explain a little bit where we are doing well. And, of course, it's nearly impossible not to talk about AI at this moment. And I'm going to do that from a standpoint of our three customer segments, enterprises, network operators, and cloud AI, what we also sometimes call the ICP, Internet Content Providers. So what is happening in the market? And the reason why I want to talk about it from all these three standpoints is that all of them are equally relevant from the standpoint of AI. What's happening in the world is that data centers are being built at a pace that we have never seen before. When you're building data centers, you need one thing more than anything else, and that's cheap electricity or low-cost power. That's available in certain spaces and places on the planet. And if you find it, you will build a data center and you will load that data center up with GPU technology, etc. The other thing you need secondarily is to cool that equipment and you need access to water. Having network connectivity to these places, because obviously if you build a data center, much like cloud in the past, you need to connect these data centers to the internet, to the peering points, and to other places in the network. That's easier to solve. You just buy bandwidth, build networks, and the problem is solved. And that is what our customers are doing. These data centers, they can appear in many different locations. We've heard about building in cooler geographies to achieve natural cooling. I recently attended a big conference where a lot of data center companies were present. What else are they talking about? Well, they are talking about building data centers nearly everywhere, so submerged data centers under the sea. We're talking about data centers on decommissioned oil rigs because you have access to wind power, and you have space, and you have water. We're talking about – I was in Tokyo last week. In Japan, they are building dedicated platforms in the – Tokyo Bay where they would place data centers and really any idea is as good as the other ones so that's clearly one element our customers are building new data centers they need to connect these with multiple terabits the biggest projects that we have been looking at this is day one requirements of hundreds of terabits and connecting that to the Internet. That's one application, clearly. Who are doing that? Well, it's affecting. So we are selling equipment and networks to the people who own the GPU clusters, so the Neo scalers and similar. We're selling equipment and networks and services and software. to people who own the data center, who in turn rent out capacity to the people who own the GPUs, etc. And of course, we're selling an awful lot of networks and services and software to the CSPs, the operators of the world who are providing bandwidth to the people who own the GPU factories, etc. So it scales across the whole thing. The last one that I want to talk about enterprise there is the fact what I believe and what many in the market believe is that the whole idea of outsourcing your inference to the cloud, if you're a large enterprise company, is great at the moment, but the belief is that more and more enterprises will start building their own GPU-based AI clusters for their demands, for their inferencing demands. One good example of that is smart optics. Of course, we're an early adopter. Of course, we need to have control over our models. We have already invested in pretty advanced AI servers where we are running our models, where we're expecting our customers to use our softwares and to run our cloud-based services in the future. that will accelerate. So it's going to be across all customer segments. I only talked about connecting the AI data centers to the network now, so kind of north-south traffic for those of you who are uh more knowledgeable about that terminology the other thing with ai is of course the west east traffic so the scale up the scale out the scale across which is really connecting data centers in the region to optimize where you run your workloads etc we are doing that too um and and that's a market that i believe is going to continue to grow as these data centers continue to grow. So it's a fantastic opportunity out there and it's spanning across all our customer segments and really nearly all of the applications that we are involved in delivering. I want to take you through some of the numbers where the numbers are coming from in the quarter and I will start with a slide that we have used one time per year mainly in Q4 every year We look at the invoiced customers in the year and we draw some conclusions. The conclusions are very similar to every other year in the smart optics history. It is accelerated, yes, but we can see that A certain amount of our business this year, about 15% of our business is coming from brand new customers. We can also see that we have a super good retention of customers. More than 50% of the revenue in 2025 is coming from customers that we had in 2020. we can see that our partner network is expanding. We have more partners who are using our equipment in their bigger solutions to their customers. And overall, we are seeing just an expanding customer base, which is very good for 2026, 2027, 2028 and onwards. When we look at these customer segments that I talked about, enterprise, CSP, and ICP, or data center, cloud, and AI, we can see that two of them are growing much faster than the third, namely enterprise. There is also a geographical aspect to this that I will talk about in the next slide because A lot of the business that we have done in Europe is enterprise related. And those of you who have been following us know that several years ago, we started to talk about traction with larger operators in the U.S. market. So U.S. is a couple of years ahead of Europe there. I will come back to that. But The largest growth we find in our CSP segment, and these are – there's a lot of customers. It's effectively tier two and tier three operators across the world with some dominance in the U.S. We have now several new customers in the year and several existing customers who are continuing to roll out. technology and I talked about availability of power where do you build data centers and where do you need capacity well in the US one place where you have low-cost electricity is Texas so it's not a coincidence that our biggest customer in 2025 is a US regional operator with focus on Texas on the Texas market And we're also seeing tremendous growth in the green segment, the ICPs. That's where we put in everything that has to do with cloud. So that's where you find the neo scalers. That's where you find the cloud operators. That's where you find the content, the gaming, the Internet exchanges and all of that. And, of course, enterprise growing a little bit slower. This is heavily dominated by our classical enterprise business, so storage area networks, disaster recovery networks, data center to data center communication, basically for security purposes. I'm expecting enterprise to pick up as AI inferencing is affecting that market more and more in the future. Looking at our channels, we can see a very healthy growth with our indirect business. The fact that our direct business is growing faster than indirect is, I would say, nearly 100% related to the red on the left, namely the CSP. When the CSP market is growing faster, a lot of CSPs prefer to have a direct relationship with us. and hence our direct business is growing. I'm very happy with this split. We have super good partnerships out there, super good channel really across the world. It's several hundred people who have contracts with us and can use our technology in their solutions. So a very solid base of indirect partners. Now we're coming to the geography. And, of course, America is super impressive. So thank you very much for my team in America, which has grown. By the way, we're investing more in America on the back of this success. So fantastic. EMEA, as I said, The way I look at Europe is that Europe is where America was two years ago. We are now working. We have closed. contracts with larger network operators in Europe. We have delivered networks to larger operators in Europe, and we're sitting on a pipeline and a list of opportunities that is highly developed of at least 10 operators in Europe. Very, very similar to what I talked about in Americas two years ago. So, of course, we're seeing signs of this. The fact that quarter over quarter, Europe has developed in a fantastic way since midpoint of the year. Q4 is a very good quarter for Europe. I'm expecting Europe to pick up and be – it's a very promising market for us over the near term. Asia. It is a little bit more project-driven, so we can see that when the large projects happen, the quarters grow when the large projects don't happen. And with large projects, I'm talking about the 3, 4, 500K projects that we do out there, and apparently Q4 was not one of those quarters. Am I worried about Asia? No, quite the opposite. I think we have great opportunities in the markets today. that we are addressing, so Australia, New Zealand, Japan, Southeast Asia, and so on. And we have recently invested a little bit more into Asia and Africa, I should say. I mean, it's relatively small investments compared to America's, for instance. But we are investing, and the opportunity is there. I just came back Tuesday morning from Japan, where we have opportunities with Japanese customers. We're continuing our IOWN efforts, and we're continuing. the proof of concept with entity document business for the Ion architecture. We are having good dialogues with a bunch of Japanese companies now, but it will take time. Rest assured. products so about a year ago we did the leadership change or rather put in a new leader for business area optical devices we started to renovate and improve the back end of that business and we can see that that has clearly paid off we are in a much better shape to deliver fast and quality to our customers We are continuing that investment. So right now we're building a brand new manufacturing facility for business area optical devices in our brand new facility in Kista, Stockholm. That's being built for us as we speak. So focus there, automation, robotization of that business. It is high volume business. It's several hundred thousand devices that we're shipping every year. So we will invest in that going forward. That's also going to result in a restructuring cost in the first half of the year of about $500,000. But it's all done to capture future growth and capture the momentum that we have. The most important business area still, solutions, software and services, growing in a very nice way. So all good. I would like to hand over to Stefan to take you through some of the financials.

Thank you, Magnus, and good morning, all of you. We had a strong quarter. We had the revenue increase 37.7% to 23.2 million compared to 69 last year, and that was mainly driven by high sales in Americas of 13.9 compared to 6.9. The gross margin in Q4 was down to 46.1 compared to 49.0 last year and was impacted by inventory reserves, write-offs and other inventory related one-off items. The underlying margins, however, are very consistent and strong quarter over quarter throughout the year. The full year margin that I think we should focus on was stable at 47.8 for 2025 compared to 48.1 last year. And that serves as a better guide for going forward as an indicator for the margin development. Looking on EBITDA, it was good on 3.6 million compared to 2.4 last year and an increase of 1.2 million. And that split is 2.4 is related to the revenue increase and a minus 1.4 is related to increase in employee benefit expenses that has increased to 5.7 over 4.3 last year. And that's driven by the organizational growth of 7% from 131 to 140 full-time equivalents. And that's including new hires in sales in the US. We have FX impact that worked against us with eight percentage points. We have inflation and increased variable compensation due to the positive development in sales. Other operating expenses is rather flat year over year and quarter over quarter. EBITDA margin increased to 15.3 compared to 14.4 last year. Cash flow from operations was super good. This quarter was at 6.8 million compared to 0.2. last year and mainly driven by the good underlying business and some reductions in inventory. The equity ratio was 54% as a result from the growing balance sheet. We have non-current assets of 9.1 compared to 7.1 last year. Current assets is 39.7 compared to 33.9, and it's mainly inventory and trade receivables. Cash is up compared to last quarter, up to 7.3, but down a little bit from last year, that was 8.0. We have available credit facilities of 75 million NOC, 7.4 million US dollars equivalent. We have a high focus on cash and mainly inventory forecast process that has been improved. And we are doing continued management on trade receivables. Non-current liabilities is only 0.3 compared to 0.8 last year. Current liabilities amounts to 12.8 compared to 10.7 last year. and it's mainly trade payables and tax liabilities and personal related expenses. Deferred revenue up to 12.8 compared to 9 last year. And the increase is related to stable high revenues from business area software and services and growing revenues. The working capital amounts to 14.5 compared to 49 last year and down from 18.4 last quarter. And that's mainly driven by inventory reductions and increased deferred revenues. The inventory is now on 18.7 compared to 12.6 last year, but a little bit dropped from last quarter when it was 20 million flats. And the increase from last year is mainly driven by that we now have longer lead times and the components we have now are more expensive. And the higher inventory level that we set, we talked about that last quarter, that is essential to secure the future growth. in sales. But the improved forecasting process that we have done during the quarter have reduced inventory slightly. But despite the high level, there is a low risk sitting in inventory. Trade receivables decreased to 18.7 compared to 99 last year, down from 19.0 last quarter. We have good collections in Q4, and this quarter the sales was more evenly distributed over the month compared to the average quarter where we actually have a bump in the last month of the quarter. And we don't see any risk in trade receivables at all. Trade payables increased to 5.6 compared to 5.0. And net other short-term liabilities amounts to 17.2 compared to 12.5 last year. And the increase is mainly driven by increase in deferred revenue. And then we also have a little bit of tax liabilities of 1.7, slightly bump up from last year when we had 11.1. Thank you. And back to you, Magnus.

Thank you very much. Like last year, the board intends to propose a dividends of 0.6 NOC per share. So no change there. This is, of course, pending AGM approval later in the year. I want to talk a little bit about how we're viewing our future right now. This is more or less the same slide that we've been using now since we introduced our new long-term ambitions. One major difference today compared to six months ago that we need to focus on our core markets we need to focus on our business because we're doing great there is no doubt about that so continue focus on our two big home markets and of course the initiatives that we have running so you notice that When we talk about new growth drivers, we have now removed M&A from the chart. And the reason why we've done that is we want to convey that we're not actively working with that question at all at the moment. Maybe we will later in the year. We shall see. And maybe opportunities will arise that we choose to pursue and go after. But at the moment, the focus is on what we're doing. And of course, adding new growth drivers, committing to major accounts. Yes, we are doing it. You can see it in our product development. We're developing much more advanced products for release later this year and next year. We are developing the company in all aspects, anything from compliance to procedures and and how we conduct our business. We are going into the new geographies to a very large extent, following our customers. Also, don't forget that it is not going into South America with a completely brand new business. brand-new play it is the same the hyperscalers are building data centers in Mexico to meaning our network operator customers needs to connect those data centers and so the business model is the same so it's all very controlled in a sense and we are investing more in our software automation and AI tools this is both external meaning software's that we will sell to our customers both as part of the SoSmart network orchestration platform and as a la cat software packages that will help them automate their processes. It's all about going after the OPEX of our customers and help them to improve that. And also our internal tools. AI is now something that scales across our whole company. All functions are building a pipeline and a roadmap for our internal tools development team. And I have really good hopes that that's going to dramatically improve the way we conduct our business in 2026. And remember that we are a company that can do this. In order to do this, you need software skills. You need to have an understanding of how you develop these architectures because the architecture is the important thing here rather than the actual coding. And we're in a very, very good position to do that. Looking forward, a great market out there. We maintain our ambition to grow our market share by two to three times in the relevant markets that we have been talking about. And we're continuing to believe and to target the 13% to 16% EBIT range that we have been targeting. discussing before. Now, the question is, of course, how big is this market and how fast is it growing? And we will have to come back with that. We have through 2025 talked about five to six percent growth in the market. Up until 2030, we know that the industry analysts are now working their numbers, working their Excel sheets to figure out how fast is this really growing. I believe we will see a larger number in the first half of 2026. So when we come back in Q1 or Q2, we will be able to share some more projections on that. With that... I'm happy to take questions, and I suggest we start in the room, Per. Any questions in the room?

Thank you. So, Marcus, SCB. I have two questions. I can take them one at a time. So, the first one is on several competitors are flagging now supply constraints into 2026 and long lead times. How did that affect your Q4 and your outlook into the next quarters?

So far it has been mainly positive. We are clearly winning new accounts because some of our competitors, especially the larger ones, are of course super focused on hyperscaler business. We have also seen some of our competitors coming into 2026 pretty much sold out. probably not completely, but to a large extent. That is, of course, positive for small optics. The midsize players are turning to us for help to get deliveries. They need to connect their customers, and we have several wins of that nature. The good thing is that These are not decisions you take lightly. You don't bring in a new supplier for transport networks and optical networks and throw them out a quarter later because you're done and dusted and you can get deliveries from someone else. It is strategic choices. that these people do to bring us in. And my expectation is that we will stay there for a very long time, just as normal. The only difference is that it's faster now.

So you didn't see any meaningful impact in Q4? Yes, we did. And the second question is partly related because we see this surge in memory prices and also availability of components. So two-part question there is, how is that affecting the discussions with customers, the end demand projects ramping up? Are they being impacted? And the second part of that is, of course, your own gross margin and how that will affect 2026.

So I think it will affect nearly all our customer dialogues this year. All our customer dialogues this year are, to some extent, going to be related to delivery performance and our ability to deliver. So it's going to be a very important topic to us. Stefan said that we have, in a disciplined way, worked through Q4 with our inventory. And if anything, I'm expecting our inventory to go up now. That's a good thing for us. If we can deliver, we will win business. And there are very long lead times, not only on memories, as you said. Also on optical components, every optical component that sits inside our solutions is typically half a year lead time. So my procurement team is now working with Q3 and Q4 demands. That's of course a difficult task. But to me, it's more problematic if we sandbag that than if we have a slightly higher inventory and a little bit more working capital. That is not a big problem for me. So, it's very important, no doubt. I think the margin impact of this will be marginal. I don't think that's where we should look. We should look at how much inventory we have and, consequently, how much we can deliver. So far, so good. We still have very short lead times. We are operating with four to six week lead times typically, which is extraordinarily good in our industry right now. And of course, our ambition is to maintain that through the year. That's going to be... a real power play in 2026 and 2027, I'm sure. On the other stuff you mentioned in memories, and I can just say that we're good for one and a half year or something like that. So we have secured that position.

That's very good. So just one final follow-up there is on the gross margin. I think you mentioned that you expect that to be sort of flat into 2026, or you should look at the 2025 gross margin, and that's a good guide for 2026, or are there other things that we have to keep in mind? on the gross margin for 26?

No, I think it's a good guidance. So 2024 is very similar to 2025. It's going to fluctuate over the quarters, as Stefan mentioned, but I think it's a good guide for the future now that we will operate in this range, 47 to 50, for the foreseeable future.

Thank you.

Thank you.

Any other questions in the room? Then let's go to the call. So we have Kristoffer Wang Björnsson from D&D Carnegie. Please unmute yourself and ask your question.

Good morning, guys. Can you hear me? We can hear you. Good morning. Great. So just wondering, you had like an interesting comment on the east-west traffic. I think you're primarily and historically your bread and butter has been metro area networks and more like, let's call it north-south. So can you help us understand kind of what's prompting you to to start talking about within data center, east-west traffic and that part of the network?

Well, what's prompting me is that we're doing it with customers. We have a fairly large project in 2025, which is all about that, connecting AI data centers to each other over a geographically dispersed area. So typically what we're talking about here is is probably below 100 kilometers between all data centers so that's that's what prompting it why is this important well you know as as as the knowledge and as as as the ai companies are developing their methods they will need to distribute the workloads across many data centers for several reasons. You know, one thing is the whole model routing or whatever you call it, right? Where do you run a specific request? Where do you have the resources available to run a specific request? And then it's also, of course, a question of scale. How big data centers can you build? Well, there is an upper limit there. Although I have had the pleasure to see one of these data centers, they are huge. That was one of our NeoScaler customers that I had the pleasure to see. They are huge. But there is an upper limit. Then you need to go to a second location and a third location and so on and so forth. And that will up the requirement for data center to data center or front end to front end or however east-west type of traffic. So it's as simple as that. We're doing it.

So it's still kind of coherent plugables for longer distances. So it's between data centers, not kind of within the data center rack to rack.

That's correct. What I'm talking about is between data centers, and it's absolutely coherent technology. In fact, it's our whole product portfolio. Well, maybe with the exception of devices, of course.

And then on the inventory, so it's been kind of steadily growing sequentially over the last couple of quarters, but now it's down, so that downtick in Q4, is that a hint that you're expecting kind of lower top line momentum in Q1 or is it more about just the man pulling so hard that you haven't been able to continue to grow the inventory? Some reflections on that seems a bit cautious in my view.

I think it's two things, Kristoffer. One thing is that when we started Q4 on the 1st of October, our ambition was to optimize our inventory, of course. And I think at least I have kind of changed my mind on that topic. It is not that important anymore. The other reason is that – so, meaning we worked with inventory optimization through the quarter. The other thing is, of course, that Q4 is a very big quarter, and we have delivered an awful lot of products in Q4. Hence, inventory goes down before we can backfill it. All right. Thanks. Thank you.

Okay, then we have Øystein Lodgaard from ABG on the call as well. Øystein, if you unmute yourself.

Good morning, and congrats on the strong numbers. I have a couple of questions. Maybe we can start with, you mentioned on the call some new product launches, more advanced products than what you have in your product portfolio currently. Can you say what kind of products these are, what kind of use cases they are aimed at? Is that more data centers or road dams for telcos?

Certainly. So if we look at the journey that we've been on for a number of years, moving from being basically... three, four, five years ago, a point-to-point data center network provider into developing a RODAM portfolio, aiming first at the metro networks, then going after the regional networks. So one natural evolution for us now is to continue that journey, meaning improve the performance of our RODAM family, to be able to do longer reach and also in the other direction, more capacity. So introduce a broader spectrum, what's referred to as L-band technology as an example. So more of that. If we look at our transponders, Maxponders, the Layer 1 products in our family, more advanced versions of that. We have just released our 800 gig transponder now, which I think will become a high runner. We will develop more of those and more advanced versions of our transponders, Maxponders, more multi-service type products. And, yeah, and then, of course, continuing our efforts at the edge where we still see a lot of opportunities in the sort of low-cost, both rodent-based and kind of active-passive type products. So it's across the board. And last but not least, our software platform, which today – I consider to be phenomenal. It is really becoming a mature tool for our customers to really take advantage of. We have a lot more to do on the automation AI side. It's not necessarily AI, but certainly automation, data collection, multivendor. There is a lot of development that's ongoing there that I have seen. really strong belief in the near future.

Perfect. Thank you very much. And one other question on the near-term growth or kind of the growth momentum going into 2026, and it's a two-part question. One is we've seen, like it was mentioned there previously, several of your competitors have very long lead times. That's kind of a more – something that has come up relatively recently. Does that mean that, I guess there's big opportunities like you mentioned in customers switching providers to you because you are able to actually deliver on a short notice. Is that still ahead of us or did you get much boost from that in Q4?

We did get boost from that in Q4 for sure. It was not, I wouldn't say it was... um instrumental in any way but you know these are typically mid-size operators those are typically the ones that suffer when when this scenario happens and it also happens to be a sweet spot customer type for the products we have, the kind of company we are, the responsiveness that we can offer them. It has happened. There are probably a handful of new accounts that we have won through Q4. But the world is big and there are thousands of these and it will continue, I'm sure.

And do you still have some large customers now still ramping up as new customers ramping up volumes with you, for instance, in ESKs, et cetera, or are we kind of all those big customer wins at full pace already in Q4?

No, I would say nearly all of them are ramping up, right? So it takes time to grow quickly. accounts, you know, most of the accounts that we have been talking about in the past, you know, Crown Castle, WIN, all of those type of accounts that we have publicly talked about, I mean, there is still a huge growth possibility in all of those accounts as we qualify ourselves for new applications, as we become relevant for new type of network scenarios, and, of course, as they grow their business. which they are on the back of data centers rolling out left, right, and center.

Okay. Thank you very much. That was all my questions.

Thank you so much.

Perfect. We have a question on the portal as well from Eving Vangsnes. You aim for two to three time market share for 2030. What market growth do you expect in the same period?

Yeah, I talked about that just a few minutes ago. I think it's a little bit early to say. So for now, I think five to six percent. And we will come back in the first half of the year as our friends at Signal AI develop their models, because that's what we're relying on for that purpose. So we're going to come back to that. It will not be lower. Perfect. Thank you. That was the last question. Then thank you very much. Have a good skiing holiday to all of you Norwegians. And talk to you again in a quarter. Bye-bye.