IQMX

Good morning. Welcome to IQM Quantum Computers first earnings call as a public company, following our historic debut on the Nasdaq Global Select Market and Nasdaq Helsinki. Joining me today are Dr. Jan Goetz, our Chief Executive Officer, and Jan Kürschner, our Chief Financial Officer. Before we begin, I'd like to remind everyone that today's discussion may include forward-looking statements, including comments on our product roadmap, customer demand, system deliveries, bookings, revenue timing, operating plans, and outlook. These statements are subject to risks and uncertainties that could cause actual results to differ materially from our current expectations. Please refer to today's earning materials, our public filings with applicable regulatory authorities, and our listing prospectus and subsequent stock exchange releases for more information on those risks. We will also discuss certain non-IFRS measures and reconciliations to the most comparable IFRS measures are provided in our earnings release.

With that, I'll turn the call over to Jan Goetz.

Thank you, Chiara. Thank you to everyone who is joining us for our first earnings call as a public company. Our first earnings call and the underlying transaction of going public has been an important milestone for IQM. We officially entered the public markets on July 2nd under the ticker IQMX. We listed at Global Select Market on Nasdaq New York, followed by a listing on Nasdaq Helsinki. From the announcement in February to trading in July, our go public transaction was completed in just five months. The successful closing of the transaction reflects the execution capability and the organizational strength we have built over the years. Part of building organizational strength, we have also significantly strengthened our board of directors. We welcomed three new independent board members. First, Barbara Venneman, former Global Head of IT at Deloitte and a board member at Vanguard.

Second, Juho Sarvikas, CEO of Nasdaq listed Inseego and former president of Qualcomm North America. Third, Jeff Tuder, an experienced investor and board professional. They join our chairman, Sierk Poetting, COO of Nasdaq listed BioNTech, Alex Doll, co-founder of PGP Corporation and managing partner at Ten Eleven VC, Hannu Martola, president and CEO of Dedicated Technologies, Inc. and experienced board professional, and myself. Our dual listing in particular marked a historic milestone showcasing how technology from Europe can capture global investments to turn quantum computing from a research ambition into real product and customer-ready computing infrastructure. At IQM, we have a differentiated business model. We are leading the deployment of quantum computers into data centers. We have a unique technology approach based on vertically integrated full-stack quantum computers with advanced error correction.

From day one, our founding thesis was to bring working full-stack quantum computers directly into the hands of HPC data centers, connecting quantum computers to AI supercomputers. We have deployed more quantum computers than any other company, with a focus on modular open architectures. Our quantum computers are optimized across the full stack to create a faster lane towards fault-tolerant quantum computing. Our performance across the first half of 2026 shows that we are on the right track with this thesis. Our future success can be measured in three distinct KPIs. First, commercial traction. This KPI becomes visible through our order backlog that converts into revenue over time. Second, product development. We have consistently hit our technology milestones and developed competitive products for a global market. We have a very detailed technology roadmap scaling to millions of qubits. Third, financial strength.

We are well-capitalized by a global investor base, leading to a multi-year runway. These three KPIs are enabled by our distinct business model. We operate our own quantum computer factory and quantum data center, enabling us to manufacture, deploy, and support quantum computers for customers around the world. Our quantum computers are designed for commercial adoption and can be easily integrated into existing data centers. We call this approach production quantum. Quantum computing that institutions own, operate, and grow with it. Before diving into our financial performance and operational update, I want to address our view on the broader quantum industry. Investors are navigating a complex technology with competing modalities, research milestones, and commercial statements. I also want to speak directly to the risk disclosure noted in our listing prospectus about the timeline of large-scale commercial quantum traction.

As a quantum computing pioneer, IQM operates in a highly regulated and naturally cautious environment. Yet our growing commercial traction demonstrates that quantum computing is already delivering real-world value when being integrated into supercomputing data centers. IQM is the commercial leader in full-stack quantum computing based on a differentiated and vertically integrated approach. We have sold more on-premise quantum computers than anyone else, and we distribute quantum computing via the cloud. We have our own chip design tool, our own chip factory, our own assembly line, and our own quantum data center. Running our quantum computers works with a vertically integrated open and modular software stack for real use cases. First, take a first look at these use cases we are exploring as IQM. Looking ahead, we see commercial adoption solidifying across three key application archetypes where our superconducting clock speed forms a distinctive competitive moat. First, quantum simulation.

Driven by heavy infrastructure investments from advanced material science and pharmaceuticals. For example, we have reported an 85% accuracy boost and an 88% reduction in execution time when validating quantum chemistry algorithms on IQM Garnet hardware. Second, optimization. Focused on real-world enterprise optimization across finance, manufacturing, logistics, and transportation. Most recently, IQM and Deutsche Bahn demonstrated a hybrid quantum classical railway scheduling solution using real operational data on today's IQM hardware. Third, quantum machine learning. An explosive cross-industry layer accelerated directly by the growth of artificial intelligence. Speaking of commercial acceleration across these exact verticals, on the 6th of July, we acquired selected assets of Quantistry GmbH, and I want to explain why. Quantistry is a pioneer in cloud-native AI-powered chemical and material simulation. Through this strategic acquisition, we acquired a proprietary software platform, valuable application IP, and a world-class quantum chemistry and machine learning team.

This expands our application capabilities and accelerates the delivery of practical quantum solutions for our customers. Now, let's have a look at the products we are offering as IQM to implement these use cases. We deliver value to our customers through a unique and diversified product portfolio. Our commercial success is based on a three-tier system lineup plus a cloud offering. First, IQM Spark, our flexible on-premises entry point for universities and research laboratories focused on education, training, and early-stage research. Paired with IQM Academy, it helps address the quantum talent gap while expanding the IQM ecosystem. Strategically, IQM Spark allows us to build early platform familiarity. As students, researchers, and developers trained on IQM systems move into industry, we believe Spark can support broader long-term adoption of our technology.

IQM Spark also serves as a springboard for inventing new supporting technologies that will be directly compatible with IQM's quantum computers. Second product is IQM Radiance, our scalable noisy intermediate-scale quantum, or NISQ platform. It has been our core on-premises system for HPC centers, national labs, and enterprise customers. IQM Radiance is designed for direct HPC integration, delivered as an open platform and upgradable with the latest IQM processors. With configurations from five, 20, and 54 qubits today to 150 qubits, IQM Radiance gives customers a clear path to scale quantum capability within their own computing infrastructure. The third product line is IQM Halocene, our next-generation on-premises platform for quantum error correction, designed to bridge the transition from today's NISQ systems toward fault-tolerant quantum computing. It gives customers a platform to learn, research, create IP, and build applications on the critical path to error correction.

This includes compilers, code implementations, AI-enabled error decoders, HPC integration software, calibration tools, and quantum algorithms. These research areas are critical for future fault-tolerant quantum computers and are areas where our customers get to innovate and spin out companies based on their research results. IQM Halocene is also built to explore architectures and software techniques that reduce the physical qubit overhead required for error correction. For IQM, this makes Halocene a strategic platform to deepen customer engagement around the technologies that will define the path to fault tolerance. Both Radiance and Halocene work with our HPC integration service, enabling to operate as SLURM nodes alongside classical CPUs and GPUs already operationalized at supercomputing centers. Fourth, IQM Resonance. This cloud offering extends access to our systems through the cloud. Customers can access IQM Resonance either through our own cloud platform or via AWS.

IQM Resonance positions us for the next phase of cloud adoption. It gives researchers, students, educators, and developers broader access to IQM's quantum computers and IQM Academy. At the same time, it supports advanced capabilities such as pulse level access and quantum error correction capabilities. What all our products have in common is the open and transparent stack. While many in the market offer black box systems, IQM is built around openness, modularity, and deep integration. That's particularly important for highly regulated industries such as financial services, healthcare, government, and critical infrastructure, where security, compliance, and data sovereignty often require on-premises deployment and tight integration with existing computing infrastructure. This open integration philosophy is one of the reasons IQM has become the repeated partner of choice for many of the world's leading supercomputing centers, enterprise customers, and research institutions.

Okay, let's talk a bit about the commercial traction we are creating with this product portfolio. Our differentiated product strategy and roadmap results in a strong commercial success, one of our core strengths. Our commercial scale is increasing continuously. As of the close of the second quarter of 2026, IQM has cumulatively, since inception, sold 26 full-stack quantum computers worldwide, built more than 50 quantum computers, and successfully delivered 17 systems globally, with eight systems being currently in production. These are leading concrete operational metrics that set us apart from pure concept business in the quantum landscape. Perhaps most importantly, our customers include some of the world's most advanced computing centers. Four of the top 10 supercomputing centers in the world utilize IQM full-stack quantum computers, such as the LUMI AI Factory in CSC, a deal we recently closed.

With the Leibniz Supercomputing Centre in Germany and VTT in Finland, we have repeat institutional customers highlighting the performance and reliability of our systems. During the first half of 2026, we expanded into several new markets, including Japan and Spain, through new customer wins and customer engagements. We also delivered new systems now fully operational at the U.S. Department of Energy's Oak Ridge National Laboratory in Tennessee and another customer delivery to CINECA Supercomputing Center in Italy. We also completed installations of our educational quantum computer, IQM Spark, and IQM Radiance R2 systems at universities in Finland, Germany, and Poland. We're also seeing growing demand from the private sector with two recent system sales to enterprise customers. First, we reached a commercial sale where Galaxy, a space company in Poland, purchased a 54-qubit Radiance R3 system for on-premises installation.

Second, we signed an agreement to deliver an IQM Radiance R2 system to TOYO Corporation in Japan, securing our third Asian market footprint. As you can see, we have a global customer base ranging from the U.S. to Europe and all the way to Asia. Our commercial reality is clear. Cloud adoption is growing, but gradually. On-premises installations remain our primary revenue engine today, while cloud and software capabilities expand the reach, usability, and long-term value of our systems. In order to deliver on this commercial success, we have developed the concept of production quantum. This means we invested in industrial strength for manufacturing since day one. It shows superconducting technology because it can be manufactured and scaled to millions of qubits while operating at a much higher clock speed than other modalities. This speed advantage at high processor quality offers a clear commercial advantage for our customers.

Superconducting technology also makes IQM's quantum computers the right platform for data centers. Our quantum computers deliver what data centers actually need: speed, scalability, and stability. It's already proving it. We have been running quantum computers in customer data centers and cloud systems uninterrupted for several months. Not in a lab, not a demo. Real systems, real customers, real workloads. This is where our vertically integrated model really matters. The tight innovation cycle between chip design, chip manufacturing, and system testing allows us to move faster, maintain the highest quality standards, reduce supply chain dependencies, and improve hardware margins. It truly creates a flywheel resulting in the best product for our customers. Those fast innovation cycles translate into tangible results when it comes to IP creation. According to an EPO and OECD ecosystem study, IQM ranks first among European quantum computing companies and patent assets.

At the heart of IQM's business model is our proprietary fab in Espoo, Finland. The first-of-its-kind facility in Europe that delivers constant supply of working quantum processors with world-class performance. We've built it to accelerate our own technology development and manufacture our own quantum computers, not to operate as a commercial foundry. After our initial ramp-up of the factory in 2021, we have invested another EUR 40 million into industrial expansion of our factory. This has doubled our clean room capacity to enable the production of up to 30 full-stack quantum computers per year. At the same time, our ecosystem partnerships with NVIDIA, with AWS through reseller agreement, and with HPE ensure that our quantum computers can integrate directly into existing AI, cloud, and HPC workflows.

For example, at HPE Discover, we announced a collaboration with HPE to integrate IQM's superconducting quantum computers with HPE Cray HPC infrastructure for hybrid enterprise environments. Let's talk a bit about the interplay between quantum and AI. Whether they're deployed on premises or accessed through the cloud, our quantum computers are designed to integrate seamlessly into customers' existing computing environments. We do not view quantum computers as standalone machines. We view them as part of a hybrid compute architecture alongside CPUs and high-performance GPUs for AI. This hybrid reality is where IQM holds an exceptional structural advantage and where we have had strong customer results. For example, most recently, the use case of molecular simulation with Leibniz Supercomputing Centre in June. The hybrid approach also deepens our position at the intersection of quantum computing and AI. Let me be precise about what that means.

Quantum computers are not designed to replace GPUs or process massive data volumes. They are designed to solve highly complex computational problems that can sit inside AI and high-performance computing workflows. A good example is the transaction we announced in early July where IQM was selected to integrate a quantum computer into the LUMI AI Factory, connected to one of the world's leading supercomputers. As AI infrastructure becomes increasingly energy-intensive, quantum computing offers a complementary path. Applying specialized processors to problems where brute force classical scaling may become inefficient. This is how we see quantum developing. Not as a standalone machine, but as a specialized accelerator within the next generation of hybrid AI and HPC infrastructure. This is exactly the direction of our work with NVIDIA, bringing quantum processors into AI and HPC infrastructure, while also using AI to improve the performance and reliability of our quantum computers.

On the infrastructure side, we are combining the IQM Halocene system with NVIDIA's NVQLink platform to enable tighter QPU, GPU integration. On the operation side, we've worked with NVIDIA on AI-driven parallel qubit calibration using the NVIDIA Ising open family of AI models, helping automate system tuning and improve system uptime. We also enable customers and partners to push the boundaries of quantum AI. For example, a team at University College London published results in Science Advances this June showing how an IQM quantum computer improved the reliability of machine learning models. In this case, it was about predicting complex physical systems such as turbulent fluid flows, blood flow in the human body, and atmospheric dynamics. These are chaotic systems that have historically defeated conventional AI.

We have also designed a multi-scale biomolecular design pipeline combining quantum circuits with GPU-accelerated classical supercomputing alongside NVIDIA, University College London and Leibniz Supercomputing Centre. As generative AI and large model workflows hit classical scaling boundaries, our customers are actively exploring quantum machine learning and synthetic data generation. By integrating our quantum computers directly with GPU-based clusters at Maya supercomputing nodes, we are helping pull forward real demands, creating a structural framework where quantum accelerates classical AI pipelines. The integration of quantum and AI will be driven by our long-term technology roadmap. Let's have a look at what's happening on the technology side for IQM. We are extremely focused on implementing hardware-efficient error correction codes in our large-scale quantum computers.

Over the last couple of months, we first introduced tile codes and immediately further refined them with the barbell codes, a novel family of QLDPC error correction codes customized for the IQM Constellation topology. Based on our published analysis, barbell codes show a strong improvement in efficiency, either up to 1,000 times lower logical error rates at the same physical qubit budget, or a comparable logical performance with up to 8 times fewer physical qubits. Independent MIT-led research on hardware-aware QLDPC implementation has also highlighted IQM's earlier tile code work, reinforcing the relevance of our code hardware co-design strategy. Barbell codes build on this direction by further reducing hardware complexity and physical qubit overhead, supporting our broader approach of co-designing codes, hardware topology, and system architecture on the path to fault tolerance.

Ultimately, the efficiency of the error correction codes depends on the performance of the processor technology itself. Our technology doesn't just produce theoretical codes or chip designs. Our technology produces excellent quantum processors. As reported in a peer-reviewed paper published in PRX Quantum, we demonstrated a quantum processor simultaneously exceeding 99.9% fidelity for single qubit gates, two qubit gates and readouts. Alongside our error correction and processor developments, we are expanding the software layer for our quantum computers. For example, we launched IQM Pulla, our unified pulse level compilation engine, giving developers greater transparency and control while making IQM's products attractive to a broader customer base. IQM QAOA, IQM Benchmarks, Qubit Selector, dynamical decoupling, and classical feedforward tools are examples of our active expansion into software, providing tooling for applications and algorithms developers.

To execute our competitive and ambitious roadmap, we have significantly strengthened our executive team over the last six months. We have appointed Dr. Craig Ciesla as Chief Technology Officer and promoted Dr. Inés de Vega to Chief Scientist to drive vertically integrated product engineering. Earlier this year, we have appointed Dr. Søren Hein as Chief Operating Officer, overseeing fab, production, and delivery of our advanced quantum computers. Together, the team will execute our roadmap toward fault-tolerant quantum computing. Reaching that goal means scaling from today's systems to thousands, and ultimately millions of logical working qubits. That's why we are investing heavily in miniaturization and cost-efficient technologies that will make quantum computers practical to build and operate at scale. In short, the first half of 2026 has been a transformative period of execution. We have proven that a global quantum computing champion from Europe can successfully tap elite global public markets.

We are very excited about what is still to come. I will now hand the call over to Jan Kürschner to discuss our financial results and outlook.

Thank you, Jan. As this marks our inaugural financial address to the public markets, I want to ground our investors in the specific evaluation framework needed to accurately assess IQM's financial model. Our business operates in a high-value, project-driven market where annual revenue is derived from a limited number of large contracts. Revenue visibility is influenced by customer approval cycles, site acceptance processes, and project implementation schedules, resulting in inherent variability in the timing of order intake and revenue recognition. This can have a significant impact on quarterly revenue, often making quarterly results look much more volatile than the underlying business activity. For this reason, we guide our investment community to track a comprehensive set of commercial metrics: Order intake within the period, revenue recognized within the period, and order backlog balance at the end of a period. Let's review our reported figures for the first half of 2026.

Let's start by looking at our order book and our order backlog. Our order backlog is still exceptionally strong. We opened 2026 with an order backlog of EUR 67.3 million. In the first six months of 2026, we added a further EUR 10.6 million to the backlog as order intake, whilst recognizing EUR 8.9 million of revenue. As a result, we ended the half year with an order backlog of EUR 69.1 million. Since the end of Q2, the order backlog has grown by another EUR 33 million, bringing the total order backlog to over EUR 102 million.

Let's go over our reported revenues. Q2 and H1 revenues respectively were EUR 6.7 million and EUR 8.9 million. These revenues are generated from quantum computing only and primarily from sales of on-premise systems, including the associated service and maintenance. Revenue from hardware and services was EUR 6.4 million for the second quarter and EUR 8.3 million for the half year.

Total revenue also includes revenue from cloud-based usage of IQM's quantum computers, as well as co-development projects, again, from quantum computing related only. Our gross margin stands at 46% for the second quarter. Our reported operating loss was EUR 30.9 million for the second quarter and EUR 60.5 million for the first half of 2026, reflecting the significant investments we are making in R&D. The product roadmap we have outlined earlier is world-leading and delivery against it is well on track. I would like to provide essential context regarding our expense base. Our reported operating losses include one-time expenditures necessary to achieve our public listings at Nasdaq and Nasdaq Helsinki. We invested in legal and regulatory compliance as well as in the necessary dual listing requirements.

Whereas many of these expenses will be recognized in Q3 with the closing of the transaction, EUR 9.9 million is included in our H1 general and admin expenses. The listing has transformed our capital position. IQM's cash position was at EUR 309.4 million immediately following the listing on July 2nd. This substantial capitalization reduces near-term dilution or liquidity risk, provides a runway to Q2 2028, and gives us a robust balance sheet to fund our strategic milestones.

As we look forward to the rest of the year, we are confirming our full year 2026 guidance. Our full-year order intake target is EUR 65 million-EUR 75 million. Our full year 2026 revenue target is EUR 42 million-EUR 47 million. Given that our first half revenue is a smaller fraction of our full year guidance, I want to reinforce our core cadence principles to remove uncertainty around our forward estimates.

Our 2026 revenue model is structurally second half and Q4 heavy. This concentration is an outcome of our project timelines. The largest part of our 2026 revenue recognition is scheduled for Q4, tied directly to the scheduled delivery and customer acceptance milestones of our first 150-qubit system. Investors should expect three structural seasonality factors. First, extended backlog conversion. As our systems scale up into higher qubit classes, the timeline spanning from the point of order intake to revenue recognition naturally extends to 1.5 to two years. This reflects physical site preparation, high-vacuum testing, and tailored on-premise installations and calibration. Second, European seasonal slowdown. Commercial procurement and site access is typically moderate during the European summer months in Q3, leading to a natural clustering of final installations and revenue realizations in the fourth quarter. Third, nonlinear order inflow.

Core institutional and sovereign funding cycles move in irregular blocks rather than in predictable linear quarterly increments. The surge in backlog growth is entirely normal, followed by periods of operational installation. From a capital allocation standpoint, our priorities are clear and unwavering. We are deploying our cash directly into assets that further increase our competitiveness and value proposition, accelerating our physical customer delivery schedules, achieving the Helsinki error correction milestones, expanding internal fab capacity, and deepening our cloud software ecosystem. We are also scanning for opportunistic M&A that truly will enhance our solutions. We are completely focused on becoming an enduring, profitable market leader in global quantum infrastructure for the long term. With that, I will hand the call back to Jan Goetz for closing remarks.

Thank you, Jan. We remain optimistic about the potential of quantum computing, something we have put in numbers in our latest edition of the State of Quantum report. We just released the fourth version of this annual industry report, indicating a six-fold contract market surge since 2021, raising up to $2.6 billion in deals value. The report shows that 46% of buyers now mandate on-premises architecture. We are a category leader in data center deployments, which is globally recognized. For example, we are named a major player in the IDC MarketScape report, which is the Worldwide Quantum Computing 2026 Vendor Assessment for our distinctive on-premises full ownership deployment model. As we close our first earnings address as a public corporation, I want to leave our shareholders with three definite takeaways.

First, IQM is operating from a position of verified commercial scale. With 26 systems sold, 17 delivered, and a comfortable cash position, we have established an operational baseline that no one else in quantum can point to today. We serve anchor customers that command strategic importance for the future of global sovereign computing infrastructure. Second, our competitive differentiation is sharp and defensible. The combination of our proprietary chip fab manufacturing, a dedicated superconducting architecture optimized for error correction and robustness, and an open, deep on-premise integration model gives us a unique moat against both restrictive cloud-only providers and alternative, less proven modalities. Third, we are executing an infrastructure-focused operating playbook. Quantum computing is transitioning from scientific excellence to an industrial asset class.

Our focus for the rest of 2026 is pure operational discipline, hitting our delivery targets, converting our premium backlog, and driving repeatable, transparent execution for our customers and public shareholders. I want to express my deepest gratitude to our customers, global teams in Espoo and Munich and around the world, our premier research and sovereign partners, and our public shareholders for their trust. We have built a strong foundation, and we are well-positioned to lead the next era of high-performance computing. Operator, let's open the line for questions.

Thank you. We will now begin the question and answer session. If you have dialed in and would like to ask a question, please press star one on your telephone keypad to raise your hand and join the queue. If you would like to withdraw your question, simply press star one again. If you are called upon to ask your question and are listening via speakerphone in your device, please speak up your handset to ensure that your phone is not on mute when asking your question. Again, press star one to join the queue. Our first question comes from the line of Tanu Chauhan with Rosenblatt Securities. Your line is open.

Hey, guys. Thanks for taking my question. Congrats on a strong quarter. This is Tanu on behalf of John McPeake. Just two questions here. My first question is, you described 26 systems as sold or subscribed. What portion is subscription here versus outright sale? My second question is on error correction. Your published roadmap has four to 36 logical qubits in 2027 at a 10⁻⁵ logical error rate. Does that rung depend on the barbell codes working? That's it. Thanks, guys. Can you hear me, guys?

Yes. Thanks, Tanu, for the question. Just on the 26 systems, all of these are sold to customers, it's all final sales here.

Okay, got it.

I think Jan will take the technical question.

Yeah. Thanks for the question about the roadmap and the error correction code. As you can see in our roadmap, we will actually merge the topology of our processors going forward. Indeed, the error rates that we give there, they are including novel codes, QLDPC codes. You can see also there's a range in the qubit count, and it depends then on the exact QLDPC code that you choose where you land. For example, we have published tile codes recently and we have published barbell codes, and they have different efficiencies. It depends a little bit on the exact error correction implementation where you land in terms of qubit and logical error rate.

Okay. Thank you. Thank you, Jan.

Our next question comes from the line of Tyler Anderson with Craig-Hallum Capital Group. Your line is open.

Hi, everyone. This is Tyler Anderson on for Richard Shannon. Thank you for taking my questions, and welcome to the public markets. I was wondering, with your HPE collaboration, is this for the tight coupling of your computer and HPC environments? If it is, could you just expand upon that and broadly speak about where superconductors are in this today as an industry? Thank you.

Yes. Thanks for the question. Indeed, it's about the development of hybrid approaches. When I talk about hybrid approach here, I mean the quantum computer working together with a supercomputer. As we are focusing on quantum computing and we are not building supercomputers itself, we need partners there. This is, for example, what we do with HPE. Developing the integration between high-performance computing and quantum so that in an actual supercomputing center or data center, they can work hand in hand and then solve hybrid jobs where part of the problem is solved on a quantum computer and other parts are solved on a conventional supercomputer.

With that work, are you the owner of the IP for that orchestration between the quantum and classical, or is that something that HPE is keeping, just noting that they're doing this kind of work with several people?

At the moment, it's mainly still happening on the software side of things, and we do have interfaces there, and these interfaces are, most of the time, they're done in an open-source framework anyway. We are working on our side on the quantum part, and our partners work on the supercomputing part, and then there is an open interface in between.

Thank you. I appreciate it.

Next question comes from the line of Craig Ellis with B. Riley Securities. Your line is open.

Thank you for taking the question, team, and congratulations on the transition to public markets and the commercial and technology development success. The first question I wanted to ask is regarding the very nice calendar 2026 revenue guide of EUR 42 million-EUR 47 million. The question is this, inside of what you indicated would be more of a fourth quarter weighted revenue recognition period, to what extent are we looking at units that would be a larger number of lower ASP units versus a smaller number of higher ASP units? Can you help us understand what the unit implications are inside of the EUR 38 million-EUR 33 million second half guide?

Thank you, Craig, for the question. It's a bit of a mix of both. We are working on the deployment of one or two smaller systems in the second half, as well as we keep building up on the larger systems that are to be delivered. Just to give a little background here on the revenue recognition. Whenever we have projects that are less than six months in time, we recognize revenue at a point in time. With the delivery, this applies to the systems up to 54 qubits. Anything above that is then done over time within reaching milestones in the progress of the project. There is this mixture of these two approaches.

That's very helpful. The follow-up question somewhat relates to that. The company indicated a one and a half to two year timeline from order intake to revenue recognition. Can you just discuss the extent to which that timeline is weighted more by manufacturing time versus customer customization and acceptance, and how do the costs and RevRec work through the timeline?

The one and a half to two years for completion, that's for the bigger systems we have. As I said, there's some systems that can be delivered even quicker, within six months. Since we're going towards the bigger machines anyway. When we start, the main components are our electronics. Whenever we purchase these and we start assembling them for the clients, this will then hit both revenue and the COGS. There's always, let's say, up to three months of on-premise at the customer build of the computer, and then customization, and then customer acceptance test.

Very helpful. Thanks, team. Good luck.

Thank you.

Next question comes from the line of Mike Harrison with Rothschild & Co. Your line is open.

Hi, guys. Hope you're doing well. Thanks for the call. Thanks for taking my question. If I take your guidance of EUR 65 million-EUR 75 million for FY 2026, that's roughly flat on what you had in FY 2025, but about double of what you had in 2024. On a kind of one to two-year basis, what's the natural run rate of growth you'd expect to see there?

Well, thanks for the question. Obviously, we've not given any guidance on that yet. As we are moving towards bigger systems, of course, the individual volume of a contract will be bigger as such. We, of course, anticipate that to grow, but we're not guiding on any growth rate here. Not yet.

Just if I can ask a quick follow-up there, please. I'm assuming that we should be seeing some of the variance in the gross margin kind of flatten out as you transition to bigger systems. I'm assuming that the higher ASPs go hand in hand with higher gross margins. Is that a fair way of evaluating how the product roadmap ties into the financial evolution?

I think that's a fair way of looking at it, yes.

Fine. Okay. Thanks very much.

Thank you.

Next question comes from the line of Ryan Choi with Bank of America. Your line is open.

Hi, guys. Thanks for taking my question, and congrats on your first print and some solid backlog numbers. I guess my question is on the EUR 69 million of backlog at the end of your reported quarter. How should we think about the mix across sovereign or government-funded customers versus academic institutions versus private enterprises today? Any color on how we should think about margins and sales cycle differences would be helpful.

Yeah. Well, thanks, Ryan, for your question. Currently, the customers are mainly from scientific and government institutions. However, as we have reported, there's our first two commercial customers, one in Poland, one in Japan. We do see a lot more interest building up in commercial customers and them getting quantum ready. However, there's still kind of reluctance to buy because they're waiting for the bigger systems in order to have commercial applications run on these systems. We of course expect the mix to go towards commercial customers in the near future. As it is currently, two of the customers are commercial customers and the rest is from academia.

Got it. Very helpful. Thank you.

Next question comes from the line of Felix Henriksson with Nordea. Your line is open.

Hi, guys. A couple of questions from me. I'll take them one by one. One is on the order trends for 2026. I know you guys touched already on RevRec and the seasonality on that. On the orders, it seems like in Q1 this year, or you almost got zero orders Q2, then it was up year-on-year, and the full year guidance implies a pretty meaningful step up for the second half of the year. Can you just elaborate a bit why the slower start for the year and how is the visibility into the second half order intake beyond the LUMI contract that you've already signed?

Yes. Well, thank you, Felix, on this. Our order intake doesn't really follow any linear quarterly rules or anything. It's really a one-time event each. We've guided on the order intake of the, you mentioned the LUMI, so that's another EUR 33 million, and that already brings us to EUR 50 million as of now in order intake for the year. The guidance we're giving is on order intake EUR 65 million-EUR 75 million is on the cautious side, but we are positive that we will definitely be making that. A lot of orders are still coming from the public markets, or a lot of sales coming from public markets, we are following tenders here, and the rules are made by each country individually. Unfortunately, they don't follow our calendar expectations or requirements. I think that's the main reason here.

Okay. Fair enough. Another technical question. Earlier this year, we've seen some of the earlier movers in superconducting modality, like Google and IBM diversifying into neutral atoms and spin qubits, respectively. I just want to hear your thoughts around that and how that sort of plays out into your conviction about the future of the superconducting technology as such. Thanks.

Yeah. Thanks for raising the question. We are of course, very much convinced of the superconducting technology because its core strengths, like we can deploy working full stack quantum computers today. The clock speed is very attractive, and I don't think anyone has moved away from the superconducting technology. There is a strong belief that the quantum industry and the quantum sector as a whole will grow. You can also do other things with quantum systems like sensing and communication, which we are not so much into. We are really focusing on quantum computing, and this is where the superconducting technology is really strong at, and this is why we keep focused on the superconducting technology.

Got it. Thank you, and congrats on the first print as public company.

Next question comes from the line of Waltteri Rossi with Danske Bank. Your line is open.

Hi, team. Congrats on the print. I have two questions. First, when should we expect to see more meaningful cloud revenues as percentage of sales? What is your long-term plan with the cloud? The second one is, if you can clarify why in the Finnish report, you mentioned that the current cash position will be enough for at least 12 months. But in the English version, you have until Q2 2028. If you can just clarify the difference. Thank you.

All right. Thank you. I will take the first question, our CFO will take the second question. On the cloud business, the way we see it is that cloud is actually a very attractive entry point for our customers who want to get first experience with quantum computers before they maybe make a larger investment into a system. It is also actually a very interesting solution if you are just focused on running algorithms for use cases. At the moment, we make the major part of our business through the on-premise deployment, but we do expect that the relative cloud revenue will grow, it will grow as the commercial applications become more feasible. Our quantum computers become more powerful over time. As the computing power grows, more use cases will be unlocked.

This is then also when we expect more cloud revenue to come because users then use it more like in production mode for their own solutions. Whereas at the moment, the deployments that we make, they are going into big data centers, supercomputing centers, which are also serving the research market, for example, where a lot of the workloads are being done at the moment. Your second question on the cash runway. It is a rather technical issue here. In Finland, we are required to make a statement ongoing concern, that is by international definition, 12 months from the time of issuance of the financials. That is why we say that, because it is customary to do so in Finland. Nevertheless, the same applies, of course, to both markets that we are well-financed into Q2 2028.

Great. Thank you.

Our next question comes from the line of Tyler Anderson with Craig-Hallum Capital Group. Your line is open.

Hi, guys. Thank you for taking my follow-up. The system at LUMI, there was language that there was an upgrade. Is this already written into the contract and signed? Is this something that you would expect to be able to do fault-tolerance computing, or is this more of a QEC demonstrator, or is it both? Just want to get an idea as to the size and capabilities of this upgrade. Thank you.

Well, thanks for the question. Overall, the road to fault tolerance obviously is a journey, and it's not so black and white that all of a sudden everything works. You can test, for example, you can test certain combinations of gate sets already with smaller machines, the logical error rate will not be very low. You can just use a subset of gates and use more physical qubits and get a better logical error rate. From this perspective, I think a good way to look at it is as an experimental platform for error correction, where you can explore the different regimes of error correction and either run fundamental operations with rather high quality or run more logic operations with lower fidelities.

Perfect. That makes sense. Since you've become public, how have conversations shifted for you, both in America and also in Europe?

I mean, overall, we see IQM as a global business. We are operating globally. We have been doing business, of course, in Europe, where we started, also in the Asia-Pacific region, in Taiwan, Japan, Korea, and now of course also in the U.S., with our deployment at Oak Ridge National Lab. Going public of course provides more visibility and more credibility. We are very happy about this. I think it has brought the conversation to a next level also when it comes to people having trust in our business. From this perspective, I think it was a very positive development for us.

Awesome. Thank you.

Next question comes from the line of Craig Ellis with B. Riley Securities. Your line is open.

Yeah. Thank you for taking the follow-up question. I wanted to follow up on the comments throughout today's call on the Oak Ridge National Labs placement. The reason for the question is that lab is particularly prominent with respect to advanced compute and high-performance compute. The question is, what are the workloads that you see the system engaging with? To what extent are you able to get feedback on performance and utilize that as you tune systems and engage with other potential customers to expand placements, either with government labs or commercial entities? Thank you.

Yeah, thanks for the question. Of course, Oak Ridge indeed is one of the most advanced supercomputing centers in the world, and we are very happy that it's not just any customer, but really a very reputable customer also when it comes to quantum expertise. The quantum team at Oak Ridge is super strong. Of course, in the end, it's their system now and they decide which use cases they will run on it. They have a broad spectrum on the research side for material science, also optimization problems and other things. We do get, of course, continuous feedback on the performance of the system. Usually, when we sell systems, there's always also a service component to it, and we have a team that can then also help monitoring the performance and making sure that the performance of the systems stays at a high level.

Thank you.

There are no further questions at this time. Ladies and gentlemen, that concludes today's call. Thank you all for joining. You may now disconnect.