Modelon AB (publ) (MODEL) Earnings Call Transcript & Summary

Good morning, and welcome to this Live Q following Modelon's Q2 report. Today, we have the CEO, Jan Haglund; and the CFO, Jonas Eborn, with us to present the quarterly results. And after their presentation, a Q&A session will be followed. And I would like to remind you to submit questions throughout the presentation via the chat function below the video window. But with that, I hand it over to the company, and then I will be back for the Q&A session.

Thank you very much, Jessica. So I'm Jan Haglund, CEO of Modelon, and I'm broadcasting live today from our headquarters in Lund, Sweden. I will be presenting the earnings results for the second quarter of 2026, together with our CFO, Jonas Eborn, who will give more financial details. I'll start off by giving the highlights, both financial and operative highlights of the second quarter. Jonas will go more into details of numbers and financials, and I'll be summarizing at the end, also including an outlook and some updated financial targets. There is an opportunity to ask questions as usual at the end. So a quick view on the numbers for the second quarter. Net revenue was SEK 16.6 million, which is a decrease compared with last year. There is some currency effects in that that contribute to the decrease and also a loss of some accounts that we saw in previous quarters, especially in Q3, Q4 last year, and some also smaller accounts in Q1. ARR came in at SEK 45.9 million. It is a slight decrease compared with the previous quarter, but we do see a stabilizing trend for ARR. And with the growing pipeline, we have -- we see good possibilities for growth during the second half of this year in ARR. I'm very pleased to see that we continue to execute on our plan for improved operating income. So operating expenses were down by 35%, which then contributed to a significant improvement in adjusted EBIT, minus SEK 4.4 million in the quarter, which is SEK 5 million better than the corresponding quarter last year and even bigger improvement if you take away some of the nonrecurring items that happened last year. In this quarter, there were no nonrecurring items. From a customer point of view, we have made some significant -- closed some significant contracts in the quarter. One in our area, in Northern Europe with a major Northern European industrial equipment provider. This is a contract that we closed in Q4 last year, but where the customer now has acquired additional licenses for digital twin creation or as it's called in our technical language FMU export. In the quarter, we also have 3 very important deals in Korea, one with KIA Corporation that increased the number of licenses for vehicle dynamics, one with Hanon Systems that have selected us as a platform for system simulation and that grew the number of Modelon Impact licenses. And we also onboarded a new customer in Korea. It's a major Korean power company, that will now start to use our thermal power library. We're very happy about that. I actually had the chance of visiting Korea and Japan during the quarter, and I could feel the enthusiasm and energy in the discussions with our customers. And one topic which was really hot on everyone's agenda was, of course, the use of AI or as it's often referred to, physical AI, where AI technology is used not just for language or text, but actually to attack and solve real industrial problems in the real world. And I think that's where Modelon really has a chance to excel because by combining the power of AI agents, large language models and neural networks with the laws of physics, which is something we have been working with for more than 20 years. That's where the powerful combination lies because as efficient as AI models are for multilingual interaction, problem-solving, analysis and knowledge basis, they are not always efficient in actually solving the laws of physics and the corresponding equations that are needed. That's where we excel. And our strategy is to make this combination inside of our key tool Modelon Impact. So during the quarter, we have actually done several releases of Modelon Impact, where we have added AI functionality, also what's often referred to as agentic AI functionality, where the tool and the AI assistant not only helps you, not only guides you, but actually executes the simulation, executes the analysis and runs the work that you need to be done. And -- what we see here in terms of customer benefits is huge improvements in productivity. 5x is a number often used, and I think there's potential to go even beyond that and shorter time to result, which is a key factor for all of our customers. We also see a faster on-boarding of new people, new users, a quicker learning curve and also significantly faster troubleshooting when you run into issues of problems. This means that technology for system simulation, which is the core of our business, becomes accessible to wider groups of engineers. And those engineers that includes developers and simulation specialists, typically people that are developing new models, people that may be working with code, codevelopment typically in Modelica which is our open language for simulation technology. We also have groups of design engineers that most probably are using models to build entire systems, to configure systems, to configure different experiments and to run advanced simulations. And then also groups of, let's call them, configurators and operators, large groups of engineers that want to parametrize, that want to analyze, that want to do different sweeps and that also want to compare with the real world running digital twins in parallel with simulations. We see that AI technology will help all of these categories through generation of models, through building of systems and through analysis of existing simulation technology versus the real world. So in brief, you can say that our strategy for AI going forward is based on AI-driven physics-based simulation, where we use AI for workflows, for design exploration, analysis and where we supply what's often referred to as the skills for the application domain, for actual system simulation and, of course, the know-how of physics. We combine this with a huge asset that Modelon has built up during decades, which is 1,000 validated simulation models, describing phenomena models in aerospace, in automotive, energy and process, HVAC and refrigeration. And this quarter, we're actually adding one area. We're adding the area of data centers with a brand new library for data centers, for pre-validated models that we have now started to ship to customers. So a quick sneak preview into that, which I believe can be a growth driver for Modelon given the huge build-out of AI data centers and the corresponding challenge for cooling these data centers. And what you see to the left here is a graphic it's directly taking from Modelon Impact, where you have rows of GPUs, GPU racks at the bottom. These are typically transmitting energy of the order of 80 kilowatts, which is huge for that kind of size. That requires a combination of liquid and air cooling, which typically is run by the heart of the system, which is the coolant distribution unit, CDU, where we have them validated models actually replicating different commercial suppliers. This is all run in a data center where also people have to be. That's why you need a computer room air handling, which is what you see to the left. And then all of this heat or excess heat needs to be handled at the top through chillers and cooling towers. So in brief, this is actually a reference model for 1 megawatt data center that is executable and runnable as part of the entire library that we have now built and shipped to customers. What we expect here is a significant improvement of development time for customers that use our library. We see an opportunity for customers to cut OpEx, especially through lower energy use up to 25% would be a reasonable number to get. We see opportunities, and we have testaments from customers that by running system simulations, you get a more stable system. And therefore, you can avoid incidents. And incidents in data centers are extremely expensive, USD 1 million would be the typical cost of a dramatic incident in a data center. And finally, resource efficiency, typically water. Water is a scarce resource in many places, locations where data centers are present. And by running system simulation and digital twins, we have seen big opportunities for savings through better planning and better optimization. Simulation and digital twins of AI data centers is actually the theme of a project where Modelon has been awarded a grant from Vinnova, Sweden's Innovation Agency in a project called Collins. This project then where Modelon will be the coordinator, we will run that together with RISE, Research Institutes of Sweden and several other leading companies, both Swedish and international companies. It's a project that starts in September this year. It's going to run over 36 months. And the focus is AI data center cooling simulation and validated digital twins. And we're very happy about this, and we see that this is going to even further strengthen our position in this important and growing area. So with that, I'll leave it over to Jonas Eborn to summarize the financials for the second quarter.

Thank you, Jan. And we will start looking at the ARR for the quarter. It decreased to SEK 45.9 million quarter-over-quarter decreased by 3% and an annual decrease of 14% versus the second quarter of 2025. The number is reported in constant currency, which means that it's showing the real volume change in our recurring revenues. The decrease versus the previous quarter, 3%, is explained as the tail or the end of the small accounts that we saw losses from in the first quarter. And it's also a combination with smaller decrease in the multi-platform business. Over the year, we see a decrease in Modelon Impact of 21%. This is the accumulated effect from the past 4 quarters including the federal accounts that we reported on in the third quarter last year. If we take a look at the revenue, we have stabilizing revenue, even improving over the first quarter. It's SEK 16.6 million in the second quarter, minus 10% versus 2025, but increasing then over the SEK 15.5 million that we had in -- SEK 15.3 million that we had in the first quarter. Our software revenues were stable, SEK 12.1 million, whereof the recurring revenues are SEK 11.5 million. So there were some nonrecurring revenue in the quarter in Asia. And part of those is, of course, the Korean deal that we had reported on previously also. Our services rebounded to SEK 4.4 million, an increase of 10% versus last year and also an increase over the first quarter. Development costs were SEK 7.1 million minus 32% compared to the 2025, the second quarter. And we can see that the -- we've stayed at the level of SEK 7 million to SEK 7.5 million in development cost for the past 4 quarters, and that's where we expect to continue our development efforts also. Looking at costs in more details. Our OpEx in the quarter were SEK 21.4 million, which is minus 35% versus 2025, primarily, this is from lower personnel costs, which decreased by SEK 6.5 million, excluding nonrecurring costs. We can compare that with the second quarter in 2025, where we had personnel cost of SEK 21.5 million or SEK 26.8 million, including nonrecurring. This also resulted in an adjusted EBIT that improved by SEK 5 million. So we are at minus SEK 4.4 million EBIT versus the minus SEK 9.4 million last year, excluding -- adjusted EBIT, excluding nonrecurring. Also, cash flow improved versus last year. We have a cash flow of minus SEK 7.6 million in the second quarter compared to minus SEK 14.6 million in 2025. And our cash liquidity at the end of June is at around SEK 32 million, excluding any credit guarantees from the bank. And that concludes all of the financial details. So we will continue with a summary and outlook.

Thank you, Jonas. So I'll just summarize the quarter in 2 slides. First, just reiterate from last quarter that we are executing on our plan for turning around ARR. I'm quite happy with the new, more efficient structure we have, both organizational structure and cost structure, which means that our focus is entirely now on building ARR growth, where we see positive signs and where we have a good possibility to come back to growth in the second half of the year. We're going to do that through a more efficient setup of sales, through new functionality of AI, where that will generate both opportunities with existing customers and an opportunity to win with new customers since I believe that we are ahead of competition in many of these areas. And finally, we see that our library asset continues to attract new customers and now with the addition of the data center library, we enter a new market, which is growing fast across different geographies, especially in North America, which is already Modelon's largest market where we have significant presence and good local competence. We also continue to execute on our plan for improved cash flow and operating profit. And we've seen operating results improve actually consistently over the last 6 quarters. And what we though see is that this pace of improvement is not enough to reach cash flow positive during 2026. That is why we're revising that part of our financial targets and ambition. We're now seeing that Q1 2028 is a realistic target for fully breaking even from a cash flow positive point of view. Having said that, there are effects that could further accelerate this. For example, multiyear deals or additional research grants, and we will of course, feverishly continue to work on those opportunities. So that's why our new revised financial targets are a continued ambition for recurring revenue growth above 20%. And I believe that how we're positioned now with AI functionality and towards growing data center markets, that is a realistic target with free cash flow positive from Q1 2028, so that's a revision compared to our previous ambition and to have long-term operating profit margin above 20%, with a very healthy gross margin, we have high leverage from new deals and new contracts. So long term, that should be a very realistic target for Modelon as a software company. With that said, I thank everyone for listening and leave it back to Jessica for questions and answers.

Thank you very much, both Jan and Jonas for that presentation. So let's start off the Q&A session with the focus on the service revenues that actually returned to growth this quarter, and it was up 10% year-over-year and 33% quarter-over-quarter. Is this driven by a one-off project? Or is it more a structural shift in demand? And can we expect to see a continuation of this growth in the second half of the year?

Yes. Thank you, Jessica, for the question about services. Services is an important part of our offering, and we see services as an enabler also for software deals. I mean our focus is still as a product company and will continue to be. But we've seen more customers coming in with different kind of service project where they need our help for onboarding, for building new models and entering new areas. So the growth that we saw during the second quarter comes from several service projects, not one single large project. And going forward, we see a good pipeline -- actually a growing pipeline of service opportunities.

And as it was mentioned during the presentation, ARR fell for the fifth consecutive quarter. I'm curious to know what concrete signs if it's pipeline conversions, bookings, customer conversations, give you confidence that we will see a turnaround and return to growth in the second half of the year?

It's true that ARR has been falling now during a number of quarters and that has been due to several effects. During last year, it was entirely dominated by the loss of a few federal -- large federal accounts for us due to budget restrictions in U.S. federal business. We've also, during this year, seen a trend where small accounts, single user accounts, of which Modelon have had quite a lot, have a tendency to be less stable. Perhaps it's a person leaving the company. It could be a temporary project. While we see that accounts with more users, more customers and several projects, they are not only happy, but more stable also. So I mean, our strategy right now is on coming back to ARR growth. That's very much our focus. And we see a growing pipeline. We see a growing pipeline, both for services and for software. We also see that we're well positioned now with new functionality. AI is something that all customers want, both our existing customers and potential targets or prospects. And I believe that with the new functionality that we have released both during the first and second quarter this year, we are in a very good competitive position to combine AI with a physics base. Adding that to that, the data center market, where we have actually a customer since before, based on library components that we've had since several years, but what we've done now is to facilitate for customers to use our assets, thanks to creating ready-to-use models, ready-to-use reference systems. So I believe that, that will attract more customers, and we have a good and healthy pipeline for that, albeit with a lot of competition because it's a very attractive market. But I think these things together give us good possibilities to come back to ARR growth during the second half of this year.

And you highlighted 3 new Korean customers this quarter. KIA, Hanon Systems and a power company. How meaningful are these wins individually? And do you see Korea and the APAC region becoming a bigger growth driver going forward?