I'm going to be the quietest MC that you've ever seen. I just have to do three things. I'm going to tell you about what the safe harbor. Bear with me. What the safe harbor language that you should go see at the SEC, what this is telling you. I definitely do suggest you all go look at the documents that we have filed prior to this meeting that have all the content that we'll go through. Just some housekeeping matters at the end. As you're all aware, this transaction was announced in mid to late February 26. On June 5th, we were declared affected by the SEC, which is great. That's about 10 days ago.
Sometime in early July, we hope to have this transaction closed after which IQM's American depositary shares will be traded on the Nasdaq under IQMX and in Finland on the Helsinki Exchange, their common shares will have the same ticker. Before I turn it over to the RAAQ management team and Peter Ort, let me give you a little bit of a synopsis about the safe harbor language. Bear with me here. Specifically, the information discussed today is qualified in its entirety by the Form 8-K that has been filed today by Real Asset Acquisition Corp. and may be accessed on the SEC website, including all the exhibits here. The materials that'll be shared have been filed by Real Asset Acquisition Corp. with the SEC, please review the disclaimers that are included therein and refer to that as the guide for today's material.
Statements made during today's event that are not statements of historical facts constitute forward-looking statements that are subject to risks, uncertainties, and other factors that could cause our actual results to differ from historical results and/or from our forecast. For more information, again, just please refer to the SEC documents on file that have been filed by Real Asset Acquisition Corp. and IQM. Finally, you should carefully consider the risks, uncertainties, and other factors discussed in the SEC filings, and you should not place undue reliance on forward-looking statements which we assume no responsibility for updating. Finally, some housekeeping. We've got a dense agenda here, which is great. I'll go back to it so you can see just how dense.
We're not going to take questions during the presentation, so I urge you to write them down and we will have a 30-minute Q&A period to conclude today's event. Of course, you are welcome to stay for some refreshments on the second floor where IQM management team will be and you can potentially get in some questions you weren't able to get answered in this forum. With that, let me turn it over to Peter Ort to kick it off. Chairman and CEO of Real Asset Acquisition Corp. Thanks, Pete.
Thank you, Mark, and thank you all for coming. Good afternoon. My name is Peter Ort. I am one of the co-sponsors of the RAAQ SPAC. I am joined here with my co-sponsor, Jeff Tuder, who is sitting in the back. In addition to my involvement with RAAQ, I am a partner at Cambium Capital. Cambium is an early-stage venture capital firm focused very narrowly on advanced computing, including quantum computing. My partners, Landon Downs and Dave Moehring, shown here each have about 20 years of experience in the quantum sector as founders, operators, and investors. Dave is sitting in the back. He will be on one of the panels here on the stage later. Worth noting, Dave was the founding CEO of IonQ, that I am sure many of you are familiar with.
I do not know if you saw last night's events at the White House, but they were one of the sponsors which I have to say surprised me. Their logo was right on the mat of the cage. I did not realize that UFC was a demographic for quantum, but you learn something every day. Also, Cambium was one of the founding partners of a dedicated quantum investment firm based in Copenhagen last year called 55 North. And through 55 North, we invested in IQM's Series B round last year. We have known the company for a long time and have long admired the management team and the company. We are delighted to be partnering with IQM as part of this transaction. A few words on why we are so excited about IQM. IQM is widely regarded as the leading quantum company in Europe.
Much more than that, they have emerged really as a top- three global competitor in quantum. Within superconducting itself, we believe they are the leading independent company in that modality. Their achievements are quite remarkable. They are a full-stack company so they control everything from fabrication to deployment. They design and manufacture their own chips in their own proprietary fab. They assemble their own machines, and they design and deploy their own software stack. All this is enabled by an unbelievable team. They have over 300 folks on staff. That includes over 100 Ph.D.s and those folks hail from almost 60 different nationalities. This is a company that is much more again, than a European company. That team has enabled tremendous technical progress. The company was only founded in 2018, so in eight short years, they have really achieved tremendous and hit tremendous technical milestones.
I would note that many of their competitors were founded many, many years before that. For example, D-Wave was founded in 1999. This technical progress that has enabled really significant commercial traction in recent years. This company has delivered more on-prem quantum systems than any other global competitor. A few comments on this slide. The equity value of this deal is $1.9 billion at close. The transaction will bring over $400 million of cash on balance sheet. We have a $146 million common stock PIPE committed. And the deal is targeted to close in early July. Really, the week after next perhaps, or the week after that. Really just in two or three weeks. If we flip to the next slide. This is, to me, a remarkable slide. Maybe one of you can help me understand it given where RAAQ's stock is trading these days.
As you can see on the left, the $1.9 billion valuation I mentioned is the lowest of this group of companies. Yet on the right, we show the fiscal 2025 revenues, where you can see IQM is number two behind IonQ. If you double-click on that revenue, you will note that IQM's revenue is entirely comprised of quantum computing revenue. As I mentioned, IQM is a top- three company, yet this valuation is, we think, quite compelling. You can access that, as I mentioned, through RAAQ today. Worth coming back to this after you hear today's presentations, I think. For all you research analysts and portfolio managers out there, perhaps something to write about and act on. That really concludes my comments. The next slide is a transaction overview. You can see the sources, uses, and how we get to the $1.9 billion.
With that, I will turn it over to the next speaker, which is Jan. I think maybe there's a short video that we will watch before Jan joins, I think. I'm not sure.
The second story takes us to Finland. Jan Goetz, who is among us today, I hope, at least this meeting or maybe tomorrow, is a Munich researcher who moved north to become a professor. Things did not go as planned. In 2018, he co-found IQM, spinning his research out of the university. Here, too, the European Innovation Council believes early on in IQM, supporting it through its growth. Today, IQM employs hundreds across Europe, Asia, and America, selling more physical quantum systems than anyone globally.
All right. We're going to pause for a second. I think we need to change the sound because it's not connecting to the webinar.
We need to change the-
A couple minutes. Yeah.
Okay.
Sorry about that. Your song?
Okay. Are we good to go?
Recording in progress.
Test one, two, three.
Test one, two, three. Do you hear me?
Okay. We're in business. Okay. Good over here.
All right. Very well. At some point, we will have a quantum computer solving these problems. It's a pleasure, actually, to be here today. We just heard the commissioner speak, saying I was a researcher on my way north trying to become a professor. At this stage, actually, I think I couldn't even dream about standing at Nasdaq at some point talking to such a crowd here. I think this was a quite amazing journey over the years. Of course, thanks to the whole team in the background who has been relentlessly working on this. I'm Jan Goetz, one of the co-founders of IQM, and I will tell you a little bit about IQM in more general, about the approach that we are taking, and then we will get also some deep dive later on from Inés about more technical points.
In a nutshell-- Is this thing working? I think this is not working either. Okay. No, it went too far. Okay, here we go. In a nutshell, about IQM, what's really important for us is to build quantum computers at scale, and we call this production quantum. We have been putting a lot of effort, a lot of investment into these capabilities, meaning that we run our own chip factory, we run our own assembly line, our own data center, really bringing quantum to the fingertips of people. We have been building, not only building, actually, quite a lot of systems, we have also been delivering them to customers, selling them to customers, and co-developing them with customers.
The current number is around 23 systems that we have sold, 18 of which are already delivered and working in spec and running 24/7 in data centers, in universities. We do think that this is super important, the ability to deliver. Because quantum is an early industry and there's lots of promises out there, maybe even some hype mixed in. We do think it's super important to put some proof points there to show that companies are actually delivering these systems and running real applications on them. We have been using this actually to grow the revenue. Pete has shown, I think, quite nicely where we stand also vis-à-vis the peers in the market, and the 25 number was around EUR 31 million in revenue.
This combination of having great technology, having strong commercial traction has allowed us then to raise also more than EUR 500 or EUR 545 million on the private side. We closed our Series B round last year. Of course, then typically what happens is you sit down together with the leadership team, together with the board, and discuss, okay, what comes next. Of course, we did see what has happened in the market. In quantum in particular, the window of opportunity was there. We decided that also for us, actually, it would be a great opportunity to take this step and bring this company public. It brings us visibility, brings quality to the business. We really thought that this was the right decision. When I talk about we, I mean the whole team, of course.
I think this is something where we clearly stand out as well is not only the number of people, but it's also the quality of people, really world-class researchers who work on this project, bringing all the technical know-how there to build these machines, to build the algorithms that we run on top of them, to build a software stack in between. We have been bringing best-in-class researchers to work with us and really leading there. We do this from 12 locations. Our main hub is in Finland, where we are headquartered. We have around 200 employees there working especially on the development for the chips, the hardware, the chip production, the assembly of those systems.
We have a larger team as well in Munich, around 100 employees there focusing more on applications and algorithms and how to run these algorithms in a very efficient way on our systems. On top of this, we have smaller teams around the world, mainly focusing on business development and technical sales and sales. In addition to Europe, we have teams here in the U.S., of course, but then also in the Asia-Pacific region. That's what we do as IQM in a nutshell. The way we see quantum and quantum computing in particular is not in an isolated way, but we do think that it's going to be part of the larger compute sector. You have to think about it like in an orchestra. There are many instruments playing in the orchestra, and each instrument plays the part that it's best at.
It's the same in compute. We are going more and more into this hybrid compute where you have different modality, and each modality is doing the job that it's best at. For example, GPUs nowadays are used a lot for AI and LLM training and these kinds of things. You bring a lot of data in parallel through the GPUs. CPUs are still good at logic and coordinating. Quantum are particularly good at those problems which are super hard to solve mathematically. Actually, some of them are so hard to solve that we know today we will never be able to solve them. You can show you would need more transistors in a transistor-based computer than we have atoms in the universe. Physically impossible to build such a machine. This is the real value that quantum brings.
It's not so much about doing a certain application 10% faster, 20% faster, three times faster or cheaper or more accurate. It's really about solving problems where we know today we will never be able to solve them with GPUs and CPUs. This is why we need quantum on top of making it more hardware efficient and faster and all of this. This means it's actually also not only about the applications that you see here that we are aiming to unlock, these are the typical examples in finance, for example, portfolio optimization or financial simulations, in the whole field of logistics where it can be supply chain optimization, network optimization, autonomous driving or the like, or then, of course, in security and defense. This is a sector that we see actually with increasing interest, in particular, given the current geopolitical situation, unfortunately.
Finally, in the chemistry and pharma sector, where it's really about using a quantum computer to simulate quantum physics, for example, in molecules and the like. It's not only about solving these applications faster or cheaper or more accurate, but it's actually about solving problems that we don't even dream about today. Like in AI. When AI algorithms were invented in the 1950s or 1960s of last century, many people didn't think about the ChatGPT and whatever we are doing with them today. It was really hardcore machine learning. Another example is the internet. When the internet was created, it was about sharing scientific knowledge in a decentralized way. This was kind of the idea of the scientists who did this. Nobody thought, or maybe very few people thought about online banking or video calls and these kinds of things.
This is the real potential of quantum, is building a platform for applications that create huge value, huge new markets that we don't even dream about. All of this, as said in connection with AI. There are many applications also in AI that can be accelerated through quantum computers, quantum machine learning, creating synthetic training data, but also the other way around. Quantum computers benefiting from AI. You can use quantum computers to stabilize the machines. You can use AI to stabilize the quantum computers. You can use AI to develop more efficient quantum computers. This is a really nice interplay, actually, between quantum and AI, and it goes in both directions. It will have a similar effect that we have observed in semiconductors, where actually Moore's Law has now been going on for decades.
This was the interplay of new chips being developed and then these more powerful chips being used to develop again the next generation of computers. This is what has been fueling the logic behind Moore's Law. We think that quantum is very similar, and we are just at the beginning. There will be decades of R&D work ahead of us. With each generation that we develop, there will be a new product and new commercial opportunities that will be unlocked by these computers. This is kind of how we see the space, and our role in the space is really to be a platform for ecosystems. It's much more than just providing computers to supercomputing centers or data centers, and it's much more than just providing computing time through the cloud.
It's really being the facilitator of a complete new industry and growing with the industry and having the industry growing with us. This actually happens at all levels of the stack. It happens already today at the highest level, if you think about governments and each, let's say, major government of a tech nation nowadays has a quantum strategy and billions are being poured into it, not only here in the U.S., actually in each major country in Europe and Asia and the like. We are a trusted partner for many of those governments. They come to us and say, "Hey, can we work together? Can you help us with real quantum computers that we deploy somewhere to build ecosystems around this?" They come to us because they know we can deliver.
It's about this trust in the technology, trust in the ability of the team to deliver, and being a long-term partner, being there for the long run and building this technology together. The way then these programs are implemented, as said, is by providing advanced compute to certain hubs. These hubs can be large labs like DOE-funded labs here in the U.S. We have been selling to Oak Ridge, for example. It can be other supercomputing centers or research centers around the world. They host the machines, and they build together with us the technology stack. They activate the research community, but also they open up the machines for example, enterprise customers so they can use it either through our cloud, through the AWS cloud that we are offering, or then through the machines that we have deployed. This is then where the applications happen.
All these examples I gave is actually what we also do together with partners on our machines. Just a few examples is we have been doing some battery development with Volkswagen. We have been optimizing production processes in factories with Siemens. We have been doing some fraud detection work with an insurance company and the like. This is then where these enterprise customers start building their know-how and start engaging, building up teams, building up IP, and creating the workflows and integrating the quantum workflows into their existing workflows. This ecosystem thinking is really what drives the value and what ultimately then also drives the return on investment for our investors.
As said, we were lucky enough, actually, to convince a very strong group of also institutional investors who have been backing us in the past, and we are very much looking forward also to collaborating with some of them going forward. This is our view on the world. Quantum is going to be there. Quantum is going to be an essential part of advanced computing, creating complete new markets, complete new industries, and we are a key player when it comes to building this industry, building the ecosystems around it. We have a very distinct strategy, actually, to implement this. This strategy, it works based on three focus areas or three priorities. One of which I already talked a lot about is delivering best-in-class quantum computers and solutions.
This is really the ambition that we have is a world-leading technology that we bring out into the data centers that we make available through the cloud so that people actually can benefit from the advanced compute that comes from it. This is a lot about investing into the technology, investing into the infrastructure that we have to produce these systems, and of course, also the delivery and service capabilities that we have been building up over the years. The second part is then accelerating quantum adoption. This is really about the use cases and make sure that there's buy-in from the community, from the industry, from the customers. Because in the end, we don't want to build museum machines that are somewhere and nobody uses them. We need the active users. We need to develop the use cases and the applications on top of this.
Thirdly, and this is kind of the foundation, is building this ecosystem and doing this together. This ecosystem development, supporting startups, supporting researchers, supporting enterprise customers, and everyone working together to build up the supply chain to make quantum a reality. This is the role that we see, and this is the strategy that we work on. Actually, the rest of the day is structured around these three pillars. I will talk a little bit more about this first one, deliver best-in-class quantum computers and solutions. We will hear a bit more from Inés about the quantum adoption and the error correction and what's behind this. Finally, there will be a panel about the ecosystem building, and then we have some hopefully engaging format to talk about this. I hope you enjoy the rest of the day.
I think there's another video. I hope it doesn't crash the sound again if I click, let's make it an experiment.
I think I must tell you that IQM is the only company to provide the real-time demonstration of Q-Nova's performance two years consecutively, and you are going to be the third year consecutive one this year, Quantum Korea.
Okay, let's talk a little bit about the technology, the computers that we build, and what's behind this. I'm going to show you now a very busy slide about our roadmap. It has been online already for quite some years, I will not go into all the details. The main message is that we have been very consistently delivering on our roadmap over the years. As said, this is super important in these early stages of the industry that you have companies out there who say what they're going to do and then do it over and over again to build the trust with the customers, with the partners, and the whole ecosystem. The structure of the roadmap is so that at the top you see the applications that we think are the ones that will be driving, at least in the near term, the adoption.
It's structured in three fields. The first one is what we call simulation. This is really using quantum computers to simulate the quantum behavior of molecules and materials and the like. In a nutshell, this is what a famous physicist once said, Richard Feynman, when he said that nature is so complicated, damn it, you better use a quantum computer to simulate nature. This is what he was thinking about when he was saying the sentence. He didn't use the word quantum computer in this statement, but I think he was thinking about such a machine that we are building today. The second one is this whole field of optimization and optimization problems. This is of course, very interesting because optimization you find in each and every industry. I don't think there's any industry that doesn't have optimization needs.
This can be very broadly deployed across all fields. Finally, this topic of quantum machine learning, especially in these times that we are living in, where everything is about AI and what can you do, and how much compute power and how much investment is needed for the next data center. It's super important that we develop more efficient solutions to fuel AI and to run these AI algorithms. To run these algorithms, we have been building up over the year actually a super attractive software platform. What makes our software platform special is that it's very open on the one hand side.
I said we co-develop a lot together with our customers, which means we have different modules in there, and if a customer wants to improve something or change something, since many of these modules are open source, actually we are very happy to have the customer do this, or we do this in partnership. It's very transparent in a way that people and customers and users really see what's happening inside the machine. Whereas in other cases, I think there's still a lot of black box happening, and it's not really clear what happens inside the machine. If you want to build trust in a new technology, and if you want to build a developer community on top of this, it is super important that there is actually a certain level of transparency and openness, especially on the software stack.
All of this is fueled underneath by having these best-in-class processors. We have been developing over the years a set of processors where each year, basically the number of qubit grows, the quality of the qubit grows, and also the design of the processor architecture has been improved. This is something that really makes our approach special and Inés will talk much more about this, is bringing these different topologies together to have a world-leading error correction implementation. This is only possible because we have this closed loop of a design tool, a chip factory, and all the testing capabilities to iterate very fast on the new layouts. I said I will not go into all the details of the roadmap. This has been out there for a while. We have been very consistently delivering.
I will just double-click a little bit of what we think are crucial points when it comes to scaling, especially scaling to large-scale systems and running also algorithms at scale so that you reach this commercial usefulness. We have four points basically here, and we start at the bottom of the stack, on the processor side, what's important there. Then we work ourselves up in the stack. We talk a little bit about the control that's necessary about the operating system, so the software part, and then the applications that you can run, and then finally what you can gain actually from optimizing the full stack. Not only being a kind of component provider for something deep within the stack, but actually being able to optimize very holistically the whole quantum computing stack.
This is the rest of the session that I will present is walking ourselves through the different parts of the stack. If we start at the lowest level, really at the processor and at the component level, even on each processor, the basic building blocks in a quantum processor are qubits. This everyone knows about it because you read it every day in the headlines. At least equally important are the coupling elements, the couplers, because this is where the magic happens. This is where the logic is implemented. A qubit is, in a way, just a storage unit, right? It stores zero or one or any superposition in between zero and one. Where the dynamics comes into play and where the logic happens is actually the coupling elements. This is where we can apply two qubit gate, for example, CNOTs and these kind of things.
It's not only important to build very good qubits, which live longer and longer, but it's actually important to really master the coupling element because this is where the error comes from. You often read about gate fidelities and the like, and this is really about the coupler, how well does it work, how fast is it, how accurate it is. This is something that we have been developing since day one, basically. A very unique, very advanced coupling element, which has a few advantages compared to others. One is actually you can turn it off completely. Which sounds very trivial, right? If you turn off the light and it's off, it's off, and you come back later into the room and the light is still off, it's off. Right? Actually, on the chip, it's not so simple.
If you have there a coupling element and you have many things in the neighborhood, you can have interference and you can have crosstalk. This means even if the light is off, there is still some light coming from the side, and actually you can see in the room, which introduces errors. What we have been building is this coupling element which you can tune so that you can compensate for all the stray light, so to say, so you can really turn it off. That's super important because this is what brings you high fidelities. The other thing is that it's very fast. It really operates on a nanosecond regime, and this is one of the key advantages actually also of the superconducting technology as such, is that it's so fast.
If you want to run very long algorithms, actually you will see the difference at some point. If you do a gate in a nanosecond or a millisecond, it is three orders of magnitude. If the whole algorithm runs an hour, it is an hour to optimize your financial portfolio, or 1,000 hours. Right? Nobody then wants to wait 1,000 hours if you can also do it in one hour. This is why it is so important to build these couplers so that they can really run as fast as possible. Then finally, this will become hopefully clear in a minute, is we have been building these couplers so that then on the chip they can extend for long distances, what we call. Over a certain distance, not only to the direct neighborhood, but actually also very far away.
We can do high- quality and very fast gates also over those distance. Why is this important? It is important because it enables advanced error correction. I talked a little bit about the capabilities to do chip design ourselves, to produce these chips and test these chips. This is something that we have been doing very consistently over the years, really experimenting, like in a playground. What can you do? How can you design these chips to become more efficient? We have done a couple of things. One thing is you can think to yourself, I am a qubit, and I want to make a logic in this room, and this means I want to interfere with as many people in the room. In an ideal case, I would have a direct connection to each of you, because then I can share my information directly.
The alternative path would I only have a connection to one person, and this person has the connection to the other one. Whenever I share information, I have to go through someone else and someone else and someone else. This is like these games you play with kids. There is some error. Every time you communicate, someone misunderstands maybe a little bit the information. It would be great if I could talk to everyone in the room directly. This is actually something that we implemented. We call it the star configuration. One qubit can talk to as many qubits as possible in the room. In our case, we have one customer, for example, in the Czech Republic. They have a system from us where a qubit has a connectivity 1 to 24, which is quite unique.
Some of the architectures usually have 1- to- 3, 1- to- 4, maybe 1- to- 6 if you are good. We have managed actually to build a 1- to- 24 connection. High connectivity, that is great. Then also connectivity over long distances. Because if I want to talk to someone in the back of the room, it does not help if my connectivity limit is somewhere here. I need a long-range coupler to be able to talk to someone in the back of the room. We have been developing these long-range couplers so we can not only connect into many directions to many qubits, we can also do this over a longer distance. Then people actually, they have a set of tools to unlock new theories how to implement error correction.
This is the latest release we did on these so-called barbell codes, Inés will talk more about this. That's actually very unique because then we managed to show that you can run error correction in a very efficient way because obviously the less of those long-distance couplers I need, the better. We managed actually to develop this barbell codes so that only every other qubit needs one of those long-range connections. Very hardware-efficient. That's important. On the chip, we managed to really innovate on the way we can design and structure the processes so that the logic can run in the most efficient way. This is one thing, then of course you need to scale it to huge numbers because it's not enough to talk to qubits like in this room.
In the end, you want to talk to millions of qubits, and you want to implement the logic there. This means we need to change things compared to the way we do them today. If you Google quantum computer, often you see a picture of these golden chandeliers. This is how we call it, like a structure which looks a little bit like a chandelier. A lot of gold there. It's actually not full gold, it's only gold-plated. It doesn't make sense to invest in the material as such, but of course the value there is really the design and how you do this. The point is then if you look at these systems, you see a lot of cables. They are quite bulky, for example. You will not have a system where someone connects millions of those cables, maybe even by hand.
It would take years probably for someone to do this. What we have been doing is we have been developing a completely new stack, a stack that gets rid of these cables, because not only is it cumbersome to build these machines, it's also quite expensive, and these cables are not the most cost-efficient way to do this. We have been building a new platform, which we call here IQDB, Integrated Quantum Design Platform, which is demonstrated already for a 1,000-qubit stack and which actually has the potential to go way beyond into the million-qubit regime. This combination, we think, is really powerful. Having best-in-class processors on the component level, on the design structure level, then having the platform to scale to millions of qubits, we think this is really where the value comes from. Okay, now we have a great processor.
Many, many qubits, high-accuracy, high-speed. We still need to control it in the end. There needs to be a control system behind it. This is, again, something where we have been putting a lot of effort and a lot of design work also on the engineering side in building very, very efficient control systems. Not only on the cost side, if you compare, we have our own control electronics compared to what you could buy in the market, it's actually extremely cost-efficient when we do this ourself. It's also state-of-the-art because we know exactly what specifications we need for our systems. Whereas if you buy something off-the-shelf, it's designed so that it can run any quantum computer. There's lots of flexibility, which is great, but it comes at a cost. The cost is often performance.
We can optimize the performance of our control system really for the specifications that we need. It's full-stack optimized. It's really not only optimized towards the processor that sits underneath, it's also optimized towards the software stack that comes from the top, the real-time logic, and all the compilers that runs on top of it. You see one of those systems here in the picture. This is, I think, to control a 54-qubit system, and it looks very nice. It's very efficient, cost-efficient, energy efficient, and this is how we can control our quantum computers really in a world-leading way. We still work with partners on alternatives. There are several reasons why we do this. One way is actually to leverage also effects in the industry to help them support on the production side, make it more cost efficient, but also de-risking.
We are working very heavily on a second sourcing strategy. Every component that we use, we want to make sure there's a second source available so we don't run into any supply chain risks. Of course, if others are developing things, we can distribute the innovation tasks more broadly through the industry. This is our approach here on the control side of things. If we go higher up in the stack, actually we enter the software part. This is the interface. The control is the interface between the quantum world and the software, the programming world. Here, it's really about building a platform. Building a platform for developers so that you don't need a quantum Ph.D. anymore to run a quantum computer.
You really want to create abstraction layers so that any developer out there can actually use the system and create applications and algorithms. This is what we do in this very open, transparent, and modular way. This is why we can actually work with many partners. You see some selected partners here on this slide. In the background, actually, we work with many, many other companies and research institutions, university, to develop this stack. It is about implementing very high efficient error correction codes, real-time decoding. It's about compilers that sit there. There's so many things that need to be done. This is why we think we do have a real advantage that we also do this together with our customers. The customers, they know which problems they want to solve. They know the bottlenecks. They know really the pain points.
Just two examples here on the results side is we work, for example, with Oak Ridge, a big U.S. national lab. For example, also with Munich Quantum Valley, where we have installed a few systems and many others on this software platform in order to enable research, to enable education, also to co-create IP, for example. This is the platform that we are building on a quantum operating system. We can build on top of this. The way it was done so far is what you see here on the left-hand side. I will not go into detail. The main message is that this looks quite messy and complex. It's really hard to understand, even for me, this diagram. This is what's happening today. We want to connect quantum computers with GPUs and CPUs, maybe TPUs and the like.
This means each of those approaches has already a stack. People are trying now to combine all of these stacks somehow and merge them. It's getting quite challenging because you have all of these parallel stacks being developed in HPC, the quantum workflow, and the like. What we are doing actually now is simplifying it, like a one-size-fits-all kind of a solution. Make it as simple as we can. Have one quantum operating system, one unified compiler on top of this. We are working here with an open source project, which is called Qrisp. It comes from one of the Fraunhofer institutes in Germany. Really one stack for all. This makes it way easier for the developer community to develop on top of this. It makes it way easier also for the application stack that you can run.
If you think about, let's say, an institution in finance, they don't want to understand all the nitty-gritty details of the stack and how to bring now the job to the GPU and the other job to the quantum processing unit. They just want to throw their problem in there and have it solved. This is what this is about, is really building one stack, one coherent stack down the line where whatever is using it and for whatever reason, you just throw the problem in and it gets solved in the best possible way. This is about the applications. Then finally, just one word about the full stack and the power of a full stack and having the knowledge on it. There's lots of stuff to be optimized.
What you see here on the left-hand side, this is just a graph representation of individual tasks that need to be solved if you want to tune up and stabilize a quantum computer. Probably if you double click on any of those boxes, you can extend again into many, many things that need to be done. This means if you have a company that only solves a certain part of the problem somewhere in the stack, it's really hard to optimize for the global best solution. This is where the value comes from being this full- stack vertically integrated company because we have visibility into all the bottlenecks, into all the problems that need to be solved, and we can really optimize in the end the full solution. This is where the performance comes from on the system level.
What this means is really systems that work. Systems that work reliably, at scale, and in the long run. What you see on the right-hand side, this is error rates of systems over months. These systems are running 24/7 in data centers on a noisy, shaky floor, hot environment, next to the GPU and CPU racks. It just runs. Actually, also you see there the power of full-stack optimization, because on the right-hand side where we achieve a two- times improvement, we didn't do anything on the hardware. This is just fixing some stuff in the software stack. Again, this is the power of having the full visibility on the full- stack solution. It's not only building systems that work, it's also improving on the full- stack side. The way to do this is actually to own it.
This is what we have been doing since the beginning. We have been investing into our own fab, our own design tool for the chips, taping out. We have been just investing, again, into expanding the capabilities of our fab then having the assembly line right next to it, where we do the full- stack testing as well. The chips, the processors come out of the fab fully packaged in the back end of the fab. We bring them to the assembly line and install them in a full stack system, and we get the feedback right away. We mount it, we measure, does it work, does it not work, what can be improved? There's the next iteration on the design cycle. This is super- important, accelerating the innovation cycle and moving faster. It was mentioned, I think, by Pete in the very beginning.
We have been moving very fast also compared to others. We started a bit late, but I think we have been catching up very heavily, and now we are providing leading-edge systems to the world, either as deliveries on- premise or through our own data center that we are running. This one, the picture that you see here is from our data center in Munich, where we host our cloud. What this results in is a fleet of quantum computers around the world. The number I mentioned already, 23 systems that we sold, 18 are already delivered. We have been delivering all over the world. You see in most of the major tech countries around the world, you find one of our machines actually in the top supercomputing centers.
We have been delivering to four out of the top 10 supercomputers globally, I don't think there's any other company which comes close to this number. We have been delivering in Asia- Pacific region, in Japan, in Korea, in Taiwan. We've been here in the U.S. then, of course, in Europe. What we see, this is more recently, is we see a paradigm shift in the customers. What we have announced for the last two deals that we made was actually sales to enterprise customers. Companies buying quantum computers because they see the value in owning the machine. You get direct access, you get the IP that you create when running on there. The two customers, one was a space company, Galaxy in Poland, the other one was a tech automotive testing company, TOYO in Japan.
This is a real paradigm shift that we see, not only selling into these DOE labs or big supercomputing labs in Europe, but actually selling real systems to enterprise customers. We always set them in the cloud, we do strongly believe also in the cloud business. This is where we see a lot of enterprise customers on their journey to quantum readiness, testing use cases, validating use cases, creating IP. Obviously, the investment can be much smaller if you just do it through the cloud. We do see long term also the cloud still to ramp up. What's new is that we see these customers now also actually willing to buy real systems. All right. This brings me to the end of my presentation and another little video.
Good afternoon to everybody. I'm Inés de Vega. I'm Vice President of Quantum Solutions, and it's my pleasure to present you our path to accelerate quantum adoption. We are approaching a very important point in the quantum computing field, which is this tipping point that is going to happen around 2030, where we will find that quantum computers will start to unleash early use cases where we find quantum advantage. This is because of the advances that we are already observing that will happen also in the next few years in terms of hardware performance and also algorithm performance and performance of error correction and mitigation routines. When it comes to hardware performance, what you can see here is an average number of logical qubits that is predicted by quantum computing companies as published in their roadmaps.
You can see how towards 2030, we will be approaching around more than 100 logical qubits. You can see also the curve corresponding to the roadmap that Jan was presenting in the slide, the roadmap that we published a couple of years ago. You can see that we have a very aggressive innovation slope that I will try to explain a bit more about today, which is due to the combination between high performance quantum error correction codes and a hardware that is adapted to them that is showing high connectivity between the qubits. What we will see in the next few years is first, that the first use cases will be on solving electronic simulation problems with hybrid workflows, where we are combining quantum computers with classical computers. Quantum computers will be solving a kernel within a larger workflow.
It's not until we have more than 1,000 logical qubits that we will be able to see this resolution of electronic simulation problems, which are, by the way, relevant in chemistry and materials, with a full quantum solution. When we have of the order of 100,000 up to even more than a million logical qubits, we will see how these optimization problems that Jan was describing before will start to be resolved with advantage. We are talking about problems such as rail scheduling. For instance, if there is a train that is failing, that is having some problem, this cascades down into trains, platforms, crew, and passenger disruptions. For instance, Deutsche Bahn has published in 2024 a report where they are reporting that they are paying around EUR 200 million per year in passenger compensations.
Another example is UPS, this logistics and parcel distribution company here in the U.S., that have a software for optimization called ORION, that has also enabled to optimize the routes of the trucks, and therefore saving hundreds of millions of dollars per year. What we are talking about here is how quantum computers can actually produce improvement in these optimization problems, and therefore the impact is going to be quite significant. By the way, this curve that you see here is quite interesting also. This is a resource estimation of the number of logical qubits that are required to break RSA-2048 key. As you can see, this is really quite significant that the number of logical qubits that are needed has been decreasing in the theoretical studies from hundreds of thousands of logical qubits into just a little bit more than 1,000.
This is why it's pretty encouraging, the advances both in algorithms and also in the hardware roadmaps. How are we expecting to find quantum advantage? The fact is that it will come in different flavors. The first one is in the speedup, that is kind of like the best- known flavor of quantum advantage. It will also come from precision, which is very important, for instance, for chemistry, for achieving chemical precision in complex electronic- related problems. We might also find advantage in energy to solution. This is something to watch out. It might become quite important. Also, this is something surprising for me at least, is that there is also a chance to find advantage in the sense of finding a different kind of solution for generative machine learning.
Because classical computers, when solving generative machine learning problems, they have some bias, and quantum computers may also have bias, but a different one, and therefore coming up with different kinds of solutions. By the way, this kind of quantum generative machine learning is going to be based on IQP circuits that are a class of supremacy circuits, so they have already brought advantage. The nice thing about them is that they are trainable. Let me see if I'm able to change the slide. Oh, I was too fast. There we go. Sorry. Okay. We are observing a transition from an old paradigm that was all about counting the number of physical qubits and also performing the supremacy experiments, which are very impressive, and scaling up the quantum computers.
The new paradigm is about number of logical qubits, logical error rates, and then tackling, in the future, industry relevant applications in order to try and test advantage. The interesting thing is that we are talking about logical qubits. What does it mean is that we need to use these quantum error correction codes that Jan was already describing, and they are based on basically encoding redundantly the information in many physical qubits. For a single logical qubit, you need to use multiple physical qubits. Of course, it depends on the quality of the quantum error correction code that you will be able to have a better encoding of the information, that you will be able to encode more information with fewer physical qubits as a backbone.
What is interesting for us in this talk is that I will be describing to you how we are defining our leadership in terms of three different blocks. The first one is to have a credible quantum error correction demonstration, and this is all going to be able to demonstrate that we are having working quantum error correction code in practice. What I mean with that is that we need to be able to prove that we implement the so-called quantum error correction cycles that are running all the time during the implementation of a fault-tolerant algorithm. These error correction cycles are going to be composed of a number of elements. The first of them is composed of syndrome measurement. This needs to happen all the time.
Syndrome qubits are those qubits that they are going to be observing if there has been an error basically during the algorithm run. Then you need to decode whether there has been an error or not. You need to have very fast decoders that happen in real time during this quantum error correction cycle. Of course, you need to calibrate, you need to produce some feedback in the qubits, and you need to have a software that orchestrates all these processes within this quantum error correction cycle. Just to give you an idea of how fast this has to go, it needs to last around one microsecond because you need to have several quantum error correction cycles per logical operation. This is quite phenomenal. The second aspect that I'm going to be describing is to have this scalable path to quantum advantage.
Jan has already mentioned a few elements. What we need here with the scalability is actually that we have a reasonable price per logical qubit. Very strongly connected to it is that we are able to encode logical qubits with as few as possible physical qubits. That's why the quality of the quantum error correction code, the efficiency, is going to be very relevant, and I'm going to explain to you how we are going to be approaching that. The second element of scalability is related to be able to have literally scalable quantum computers because, as Jan was showing you, as you add more physical qubits, because you need to encode the logical qubits, you're going to be having more electronic components, more cablings.
As Jan was describing before, we are working very hard in actually narrowing, diminishing the size of these elements so that the quantum computers are still fitting in an HPC center. Of course, the third element of scalability has to do with having a sustainable funding, sustainable investment in the roadmaps. The third key of our leadership is going to be about our ability to enable system-level integration of our quantum computers in industry environments. This is all going to revolve around having a hybrid computer stack and having this open architecture that also was described before by Jan. Let me now continue with the first block, which is about this quantum error correction roadmap, that we are structuring in three different phases.
The years are kind of indicative of the moment when we have achieved these kinds of innovations that I'm going to describe in the moment. The phase one is about implementing surface and color codes for memory qubits and early fault-tolerant gates. We are talking about Clifford gates, which are, for instance, CNOT, phase gates, Hadamard gates. They are not yet forming a universal gate set, but they are core elements towards fault-tolerance. What we have already achieved, for instance, in our labs, is that we have halved in a year the logical error rate. Perhaps what is most important here is that we are building this real-time quantum error correction software controller stack that enables the implementation of these fast quantum error correction cycles that I was describing before.
This is happening now in our labs, of course, this real-time quantum error correction controller stack is going to be used also in the next phases, which I will describe in a minute. Just to mention a last thing here with respect to phase one, is that we will be also delivering products. The first product within this phase is going to be called IQM Halocene. We already have released an announcement around it, and it's going to be an open platform that will enable our customers to co-create with us quantum error correction experiments and different kinds of applications. It's going to be seemingly integrated in HPC centers, and it's going to be, of course, usable to advance in hybrid workflows. The second phase and the third one are perhaps the most exciting ones.
They are the ones that are responsible of this acceleration in the roadmap that you have observed in the first slide. Why are we excited about phase two is because we will be able to scale up these constellation architectures that Jan Goetz was presenting, where you can find these strong connections between the qubits because of the presence of the central element. What we are going to do is we are still going to implement memory qubits and Clifford gates, but with a class of quantum error correction code that is called quantum LDPC, that is much performant than surface code. Just to give you an idea, a single patch of physical qubits can encode many more logical qubits than surface code. Surface code can only encode one logical qubit in a patch, and quantum LDPC code can encode several of them.
It is in phase three when we are adding two different elements that will actually produce even stronger acceleration of the roadmap. First of all, it will be a hardware-related element, it is that we will add to this constellation architecture, these long-range couplers that will connect syndrome qubits, some of the syndrome qubits, at long distances. Why this is relevant is because actually, thanks to this, we will be able to implement much higher- performant quantum LDPC codes, specifically these barbell codes, which we have published last week in a paper that you can already check. I think the reference is down here. Basically, these barbell codes are highly efficient. They can, for instance, produce a logical error rate that is around three orders of magnitude lower than with surface code with the same conditions, and they have a higher encoding rate.
Let me maybe say a couple of words more about barbell code. This is a very busy slide. This is very high- technology slide. We are very excited about it. This plot that you can see here comes from an MIT study, what they did was to actually represent several instances of quantum LDPC codes, because quantum LDPC is a family of codes. What you can observe here is basically these codes, when you represent them in terms of the quantum error correction efficiency, which is again related to this encoding rate and this logical protection that they can enable for a certain budget of physical qubits versus the hardware requirements. This is very important because this is related to whether you can manufacture a quantum computer that can actually run these codes with this efficiency.
It's not only important to have a code that is very efficient, it is also important that you are able to manufacture a quantum computer that can run this code. With hardware requirements, we are talking about, for instance, the number of extra routing layers that you need to add to your chip, the number of extra long-range connectors, the bump bonds, and through-silicon vias that you need to add. All these elements are the ones that make the fabrication process more complex and therefore can bring delays and more problems in terms of manufacturability. In that sense, we believe we are very well positioned because if you look at the barbell codes, they are located in a region which is pretty low in hardware requirements.
We are talking about, again, this constellation architecture plus a single long-range coupler per plaquette, which is pretty good and easy. Well, not easy, but let's say realistically manufacturable, whereas at the same time, it's the most efficient of the codes that exist with these hardware requirements. Of course, this is just the beginning. In the next few years, we expect that we will still be able to come up with new families of codes that are having the same kind of hardware requirements, but higher efficiency. Okay. Coming now to the last block here. What is important is to build an open architecture to support hybrid workflows. It's all about how to talk with the quantum computers, how to offer access to the quantum computers.
In this regard, what we are building is this hybrid platform that is a combination between two different models of access to quantum computers. The first one is this hosted cloud and marketplace that is very good for easy discovery and for onboarding customers that do not want to spend a lot of resources to basically experience quantum computers. Here we are offering curated, ready-to-use software tools, and this is the best for standard workflows and ecosystem visibility. When it comes to industry customers, it is much better to have also this additional way of integrating quantum computers, which is based on having an external SDK and API integration. Here the workflows can stay in the customer. That can be, for instance, an HPC center, a national laboratory, an industry, or a partner environment.
Here what we are offering is that QPU is therefore accessed through being plugged into already existing software stacks. This is, of course, the best foundation for enterprise and advanced customers that are going to be working with us towards adoption. The combination between these two concepts is what we call this hybrid platform that is really hosted across internal and external kinds of executions. We are talking about a shared API layer for cloud, SDK, and ecosystem access, and a single architecture that avoids duplicated platform stacks. This is, of course, very good for different kinds of customers from different kinds of origins, including cloud users or enterprise customers, high-performance supercomputing centers, or national labs. What this platform will do, it will enable all these different kinds of tools that come from both the customer and our own tools.
This is a combination between these partner software libraries that may be relevant, for instance, for their relevant workflows, error mitigation and correction tools that we are offering so that we can provide for efficient implementation of the quantum algorithms, the algorithms themselves, of course, benchmarking routines that are going to be very important to compare the hybrid solution with the full classical solution in order to be able to understand the quantum advantage metrics, and also sector-specific workflows. What you can see here basically is that even if the year of quantum value is around 2030, actually, the adoption is something that doesn't happen overnight. It doesn't happen overnight from the moment there is awareness to the moment that it is possible to integrate a quantum solution into an already existing workflow.
There needs to come a preparation stage, and if this preparation stage is lasting several months, it means that we need to start now to support our customers in this adoption. Actually, this preparation stage is going to be basically articulated in two different kinds of processes. First, we need to build a usable product, such as a hybrid quantum- HPC workflow for battery material or catalyst simulation, and this typically takes six to nine months. The other process that happens during this preparation stage is actually to build the customer base, which also takes time. We are talking about, for instance, identifying early users, running pilots or test beds, integrating into existing workflows, validating value, and also moving from experimentation towards repeated usage. This is very important.
Here what you can see is a curve that gives an estimation, of course, it's only indicative, of the percentage of enterprise customers that are adopting. We are currently, of course, at very initial stages, and you see here that the slope is going to be very slow. After the year of quantum, the D-Day, we are going to be seeing that this slope is going to be increasing. What is important is that those customers that are already working with us in this preparation stage during the next months to come, will be able to extract the quantum value immediately once the computational advantage is there. This is going to be, therefore, very important that some customers are already joining us in this adoption process. How do we plan to structure this adoption process with our customers?
You can see here that we are thinking about three different horizons. The first one is all about building enterprise users that actually, basically enterprise user story, that builds on customer needs to actually identify the use cases that they care about, and actually also the bottlenecks within these use cases so that we can think about the ways that quantum computer will actually support in resolving these bottlenecks. This is also, of course, about getting feedback from the customers and then improving our processes for the next iteration. The second part is about targeting fewer industry use cases, then thinking about really, or even implementing the embedding of quantum solutions into already existing workflows. We're talking about workflows, for instance, for drug discovery or material discovery, and we are going to be doing this with application partners.
In the third horizon, we are talking about a moment when we will be approaching this computational advantage, and we will be having success stories coming from horizon 2 that we can then integrate these hybrid workflows that result from them in these hybrid platforms that I was describing before. This will, of course, allow us to have a broader reach in the community of customers. Just to give you an idea of the activities that we have when it comes to horizon 1, here you see some of the use cases that Jan was already anticipating across the three different algorithm work lines that we are working on. The first one is simulation, which is related to simulating quantum materials and quantum molecules that exhibit strong entanglement, strong quantum properties, and these are very relevant for problems like, for instance, battery simulation.
This is, for instance, a project that we have with Volkswagen, where we are working on simulating the catalyst region of batteries, which turn out to be owing basically its properties to the strong entanglement that is being created in this region. These properties are also related to the properties of the batteries, for instance, durability and time to charge. In optimization, we are also working in a number of use cases. I would like to emphasize, for instance, this rolling stop optimization coming from Deutsche Bahn, which of course they care a lot about. These different use cases we are tackling with a software library that we have built around some of the best quantum algorithms that exist for optimization. There is an element of repeatability of tools here when it comes to optimization. For simulation, we have multiple tools.
This is more complex because there is no single- solution- fits- it all. We have a number of tools that we are also using to build a library. Finally, when it comes to quantum machine learning, we have worked with Siemens in a number of use cases. Some of them you can see here. Currently, the star project that we have is this very exciting quantum generative machine learning use case that I was describing before, which is based on these supremacy circuits. Finally, I'd like to say a few words about what defines a good partner for enterprises to adopt quantum solutions and why we are actually meeting these requirements. The first one is all about having a reliable and affordable high-performance system in order to support the state-of-the-art development of use cases at the enterprise level.
In this regard, we have shown continual innovation and execution on our roadmap. I think this is something we are very proud of. We are also leading on number of on-premises quantum system deliveries globally, and we have a very high system uptime, which is, of course, very relevant to have stability. The second aspect that customers care about is that there is best-in-class integration to minimize integration and migration efforts. For this, it is very important to have seamless workflow and HPC integration, which we have been showing for many years. Also this open, non-black box approach, which is compatible with multiple platforms that I was showing before.
Finally, customers are looking for companies that have a track record of co-creation because they want to build relationships based on trust. In this regard, we have had a strong track record of co-creating with customers in all these multiple use cases I was showing in the previous slide. There is a high satisfaction of post-sales services that is backed by customer testimonials. With this, I would like to finish this presentation, and I thank you for your attention.
The reason we chose to work with IQM is because their roadmap is more planned and organized compared to other companies, which makes promotion easier for us. This includes their integration capabilities, application team, and service team. These aspects are all considerations for us. Also, they are relatively friendly, making it easier for us to work with them.
Well, thanks everyone for joining. We've heard from Jan and Inés about the first two pillars in our strategic priorities, delivering best-in-class quantum system solutions and accelerate quantum adoption. We're going to focus on the third pillar, which is to support quantum ecosystem development. Some people might ask, "Why is this such a core strategic priority for IQM?" That is because, as IQM, we believe in order to reach quantum advantage, solo effort is not the way to go. What it requires is an industry-wide evolution, kind of matures in both commercial and technology perspective. We believe this is the only thing that kind of attracts more investments and resources into quantum computing. Secondly, it also can attract the brightest talent to work in the quantum computing industry.
Both of them can create a resilient supply chain that kind of push our industry forward. IQM has always been positioning ourselves as the nucleus to build this quantum ecosystem to drive the global effects. From this, what we have provided is, A, open and modular system, so in order for our partners to join our system, to build innovation on top of it. Secondly, it is the global delivery capabilities of our both on-prem and cloud systems. This actually provides the access for partners to build the innovation on top of our systems. Thirdly, what this total in general is going to push our innovation because of co-creation history. You have seen from both Jan and Inés presentation about how we have been demonstrating those three points.
Now I'm going to use two specific examples to illustrate how we have been driving this. One of them is from Finland. What you can see is in 2018, there were only three quantum computing companies in Finland. In 2025, there were 24 quantum computing companies across every part of the stack and generated 72 patents. With this momentum, we're expecting this to generate 2,300 jobs and to pull in EUR 1.4 billion investments. This is going to make Finland a major quantum ecosystem.
Similarly, we have also seen similar effects in Germany. Back in 2018, there were only five quantum computing companies. In 2025, again, over 20 quantum computing companies focus on various parts of the stack, with over 300 patents generated. We're expecting this will drive Germany into an ecosystem to generate over 1,000 jobs and then pull in EUR 450 million investments. Of all of those new quantum computing companies that you've seen in 2025, four of them are actually spin-offs from direct collaboration with IQM, and two of which have already become trusted partners and value creation partners for us, for our supply chain. That is the effects and the importance of quantum ecosystem.
Without further ado, I would like to invite my panelists to come and join us to discuss the importance and how we can drive the ecosystem forward from various parts of the industry.
IQM had a very specialty to say we are going to have a product which is for high schools, or for universities, or universities of applied science, to really train and put that into education part. To really train yourself from an engineering point of view, from a quantum informatics or a physicist's point of view. All of this is a small solution. I consider the Spark as a small solution compared to the multi-level qubit machines which are in the portfolio. This has been an idea where we said, we do need entrance point for the clients. We already serve networks and HPCs and servers and software licenses and so on.
After the customer testimonial, we're now proceeding with a panel. I want to invite my panelists to give themselves a short introduction before we start the questions. We should start from Suhare. Yes.
Can you hear me? Okay. Hi, everyone. My name is Suhare. I am responsible for global business development for quantum computing at NVIDIA, I've been in the quantum computing industry for about 10 years now. I did my Ph.D. in physics and then spent some time in management consulting, that's when I entered quantum computing and have held various roles across product ecosystem, business development, partnerships, and so on.
Hi. I'm David Moehring. Pete mentioned me earlier very briefly, I've been in quantum computing for 25 years. The first eight years, full-time being trained in quantum computing hardware in academia. The next seven years, I was in the government lab and also deploying money as a program manager in the United States federal government and the intelligence community. For the last 10 years, in the venture capital side, the first two of which I was the founding CEO and the first employee at IonQ. The last eight years, I've been at Cambium Capital and also supporting the previously mentioned 55 North Quantum Fund.
Thanks. I'm Blair. I look after strategy and corporate development for IQM. I think I'm the newest to quantum computing of this panel, I've spent over a decade in investment banking, looking at a variety of digital infrastructure transactions and sort of similar technologies. Yeah, great to be here.
Zia Mohammad, nice to meet everyone. I've spent the past 10 years working in product management, initially commercializing IBM Watson in the early days of AI, for those who remember that. In my experience, I've spent five years across Amazon and IBM in quantum computing.
Yes. I think we have a panel that kind of represents different parts of the industry, hopefully it's going to give us a comprehensive view of the ecosystem. Before we start asking specific questions, I think it would be good to go around the house so everyone introduce what is your definition of the quantum ecosystem. What are the key components from your perspective that should be in there?
I can get started. When I think about the quantum ecosystem, it is what is required to extract value from quantum computing. It has several layers. The QPU, of course, is the most fundamental, the substrate, the heart, you could say, of the ecosystem. There's the rest of the compute that's required to make the QPU work and extract the value. There's the QPU, there's the rest of the classical compute required, and the interconnects and AI and everything else that goes there. I think about the programming layer. This is where developers are able to use the quantum computer to extract that value that we've talked about. Applications and algorithms are the why. Right?
The why behind everything that we're doing ultimately is to unlock these large problems that are inaccessible at this time, the algorithms are the ones that enable us to unlock application areas. There's one other layer that I think about that isn't maybe talked about as much in the context of the technology stack, and that's the workforce. The workforce, meaning the physicists that are building the hardware, the developers, the theorists that are discovering new algorithms, the enterprise developers, the domain scientists, the full stack of the talent that's required to actually build something useful and extract value and translate that to commercial value. The full stack of everything that's required, no one single layer will extract value, all of these together, I think, is how we'll get to the maximum value from quantum computing.
Dave?
Thank you. My answer is, not surprisingly, very similar. I look at it, some of this will sort of repeat or rhyme with what you just said with regards to how I see it. It's really around the hardware, the software, and the end users are the three main parts of the ecosystem. We heard Jan earlier talk about a lot how IQM kind of brings all these together in a very important way. I think one other piece to the ecosystem touched on is the enabling technology piece of the ecosystem. Not just the people, but a lot of the technology that's needed. What's interesting about it is, especially back when I was in the government, but also in the early days of a lot of these companies, you can brute- force your way with off-the-shelf components, just what is given there.
The enabling technologies are what's going to be needed to scale out to thousands, millions of qubits, and logical qubits. Jan also touched on that with the, I forget the four-letter acronym, but the scaling approach that they're doing. There's a lot of very important work that needs to be done there.
I think I sort of echo everything that's been said, but of course, I look at this from the view of the ecosystem being anything that's concerned with the research, development, the distribution, and the usage of quantum technologies globally. This spans from enterprise, academia, you name it, and all the quantum technologies and companies that are around today and to come.
I think the one thing I'll add rather than repeating the hardware-software foresight is I think there's a policy element as well, right? Especially with any emerging technology. You initially saw this in the days of AI, where now you have governments and large corporations reacting to this technology that's came out. I think the one element in this ecosystem as well is a lot of the investment that's coming in, yes, it's coming from corporations, but it's also coming from governments, national institutions trying to invest in quantum computing. Trying to make sure that they're acting in the interests of the technology, making sure the policies that they're working on is something else to be considered in the ecosystem perspective.
Okay. I think hopefully that kind of gives us a comprehensive picture of what the ecosystem's like. Just following on Zia's point, actually, Dave, because you mentioned your experience with the U.S. government, and then you've always been allocating capital to quantum and quantum-related technologies, et cetera. What is your point on policy, and especially when it comes to sovereign systems?
Yeah, for sure. Obviously, sovereign systems are an important part of the ecosystem, and I think one of the most important pieces is about the learning that you can do with them. What you see, especially as a supplier of quantum systems, what we see with the companies we've invested in, and especially talking with IQM, is if you have systems that you can only access through the cloud, and no offense to my panel member, but what's interesting about it is it obviously gets distribution to a great number of people. As you're trying to improve your system, you don't have the kind of direct feedback from the end user, from the folks that are trying to be the most creative with the device.
What you can get with a sovereign system is, one, you can get your hands on it, and you can get direct access, especially with the open ecosystem that IQM provides. What's interesting from IQM's perspective is that you can also get feedback directly from them. They can tell you what they're doing within their own ecosystem in ways that you can't get feedback through a cloud provider, or if you just toss it over the fence and say, "Use this." One of the things I find most important about deploying systems is the ability to aggregate information from all of the different kind of users, be it sovereigns. When I say sovereigns, I don't just mean countries or nation-states. I mean even the commercial users that were discussed earlier.
Yeah. I think that is very true, is we collect information and we kind of feed the information when it comes to policy makers, like what Zia was talking about. We can feed those information to them and help them make more educated or informative decisions, et cetera.
I'll just follow on and say that when I was in the government, it was an explicit part of my job to fund outside of the U.S. because all of the best researchers in the world are not just in the United States. Even as an intelligence community officer, funding outside the U.S. was important because if you want to learn, you need to aggregate information. That's exceedingly important, not just for the government, but also for an independent company.
Good. Suhare, as we all know, NVIDIA has done a tremendous job of building the developer community, and now, in a sense, dominating the AI community. It would be good if you can express from your experience, for example, what are the core elements within the quantum ecosystem that needs to develop in the next few years in order to mature, and what can the quantum ecosystem learn from the AI?
I'm happy to talk about this. Even before joining NVIDIA, I've looked at the evolution of the GPU as the closest analog to the evolution of the QPU that I could find, and this is many years ago. Now that I'm at NVIDIA and I've learned a lot more about the history, GPUs were used for gaming and graphics, and that was it. The big unlock happened because of CUDA. I think similarly, the programming framework is very important for quantum computing. developers a unified platform that allows developers to be able to work across different types of hardware, or we talked about a heterogeneous structure with CPUs, GPUs, and QPUs.
The ability to intuitively and without having to be a quantum physicist, run on a quantum computer, I think is how we unlock this value and expand the reach of quantum computing to a very wide developer base. When you ask what's important, I think the developer is at the center. The hardware is very, very important, and all of the algorithm progress is very, very important. Scaling is just as much a making it accessible to the developers that can extract the value problem as much as it is a logical qubit problem, which I completely agree with also. I think investing in the developers and that unified programming platform is a big one. The second one I'll say is, especially now with the advances in AI, we are starting to see acceleration to quantum computing from AI.
Two months ago in April, NVIDIA released the first mod AI family of open models for calibration and error correction. AI, and IQM was a big part of that as an early tester and a partner and collaborator in releasing that. We're starting to see more and more AI playing a role, and I think that's another critical part of the ecosystem that we need to continue to invest on and focus on. The last piece is, I think, again, making sure that the workforce is getting the access and the attention so that they can develop the algorithms and the applications that are required to realize the value. I think along with developing the applications, there needs to be a standard, a benchmarking standard of what is quantum advantage and what is valuable.
A community- ecosystem-built standard general consensus on what's considered valuable there, I think is also important.
I think that makes a lot of sense and is kind of echoing a lot of the components that NSF describes and what our system wants to offer for the developers to access on, and to the kind of philosophy why we do that. Zia here, another question from AWS. AWS has been very successful providing computing resources to developers. What do you think is the core benefit of doing that? How does that help the ecosystem to develop?
I think there's a couple of things that come to mind specifically, too. The first is, going back to that workforce development and education. AWS, as a cloud provider, has a lot of workforce development programs. I think it was about two, maybe three years ago, we launched Amazon Braket Knowledge Badge. Get exposure to quantum computing. How do you learn about quantum computing? Providing that not only to universities and startups that we work with, but also for larger enterprises, these traditional enterprises that dominate the Fortune 500. How is quantum computing going to work for them? How is it going to impact them? I think that first element is that workforce development and training, like Suhare was mentioning. The second is, what about the people in those large companies who are on AWS that are already familiar with quantum technologies?
We have partners at AWS in all different industries. All different domains. The chemical domains, the financial domains. If you already have experts at your industry or in your industry who are working on maybe Monte Carlo simulations or chemical modeling, how can you work with AWS, who has partners in the hardware and software space in quantum computing, to actually build those use cases, come up with those proof points, et cetera? Outside of just access to these developers, I think it's also trying to bridge the gap between what people are doing today and the methods or techniques that they're using today, and potentially using AWS's expertise to introduce new ones as well.
From your experience, what are the kind of core components for the developers or for the users are looking for when they want to access to those materials or when they want to choose what system they want to develop on?
Yeah, I think flexibility is one of the biggest ones. The Amazon Braket service, for example, provides access to IQM computers. Being able to either have access on a pay-as-you-go model, being able to reserve that time for dedicated access, it's really around that flexibility.
Dave, from your experience, you have a lot of exposure to different kind of quantum companies. Not just full-stack, but also components, et cetera. How do you think IQM is positioned when it comes to the ecosystem development, considering both on-prem and cloud system development that we have heard?
To double-click on what I said earlier, it's about learning. Quantum computing is still a developing ecosystem as a whole. Trying to do anything within just your own little world, you can chase down some kind of dead-end paths. Working with especially the software developers and the developer community, making sure you're engaged with all of these just across the board is very important. I think, obviously, I'm excited about how they're positioned. They're working with several other of our portfolio companies across both Cambium and 55 North. I get to see it firsthand how well they engage, how well they start to think about the future of the company. Not just what they're making today, but what needs to be made in the next generation after that and beyond.
I guess that gives the fellow investors also another hint that IQM will help to drive the other companies in your portfolio forward as well, if you're investing in IQM in this case. Jokes aside, Suhare, what's also from your perspective, how do you see IQM in the quantum ecosystem?
I think everything that Jan and Inés talked about, the modular and open platform is key. As NVIDIA, we work with everyone in quantum computing, and IQM has been a big part of our development roadmap for quantum computing because it's so easy to collaborate. IQM exposing hardware to open interfaces puts us in a position to be able to build better tools, help us improve CUDA-Q through our integration, and get feedback both ways. That, I think, is one of the key reasons why IQM is positioned to do really well. The second one, Zia touched on this, is flexibility for access. The ability to have on-prem deployments as well as cloud access is very important. Different customer segments have different needs.
A budget-conscious customer would probably want something like a pay-as-you-go, or someone just getting started may want something like pay-as-you-go, or in places where there's data sovereignty or security concerns, or a reason to co-locate with your supercomputer because it's for error correction research, the ability to deploy on-prem, I think that allows IQM to address the entire customer range and meet those needs. I think those things combined, the flexibility, the modularity, position IQM quite well.
And Blair . From IQM perspective, how do you think we will continue to drive to support the ecosystem from more of a corporate development perspective?
Yeah, I think IQM's philosophy with corporate development is not to be like this M&A powerhouse. I think, at least from our perspective, we look at the most efficient and the best ways to deliver on our strategy. You mentioned these three priorities, deliver best-in-class quantum computers, accelerate quantum adoption, and enable the ecosystem, which is what this is all about. We look agnostic across structures. We look at opportunities from IP licensing to standard partnerships, to joint ventures, to M&A. We don't really have a specific structure in mind when we think of corporate development. We're really looking for opportunities to help accelerate quantum adoption and work together with the ecosystem, yeah, really find anything that helps drive forward our roadmap faster so we can get the technology out to the world better, more broadly, anything that helps stimulate this ecosystem.
Okay. To wrap up our panel, I think is important, let's go around the house again and think we can talk about what other one or two concrete things that your part of the ecosystem could do in order to drive the growth of the quantum ecosystem.
Okay. I have three. I'll be quick. I think the developer engagement is extremely important. For NVIDIA, what that would mean-- NVIDIA's role in the ecosystem, if this isn't clear, we think of ourselves as the connective tissue. We don't make the QPU, we don't make quantum hardware, but we do have CUDA- Q and NVQLink, which is the interconnect for GPUs and QPUs algorithms and so on. For us, it's continuing to make CUDA- Q accessible and perform it for developers. The second piece is AI tools for quantum. We just started with Eisert. It's an open model, so it's enabling the entire ecosystem to use it, but also to contribute to it as well. We want to continue to do that in supporting an open ecosystem for quantum.
I've talked about workforce a lot, so there's CUDA-Q Academic, a program tailored towards universities to help train the next generation of quantum programmers, scientists, researchers, developers, and so on.
Okay.
Before I answer the last question, I'll double-click on something Suhare just said around open ecosystem. I can't emphasize how important I think the open ecosystem is within the IQM as a company and the way they deploy their machines. It's because, again, this is how you learn. I know of a lot of the other companies that when they deploy a system, it's like take what you get, and maybe you get some feedback, but usually the feedback is, "Why don't you allow me to do this?" As opposed to, "Here's how I use your machine, and here's how I would like to use your machine." You can grow together as a customer-company relationship. I find that exceedingly important.
To answer your question about going forward, obviously, I mentioned it earlier, I think the enabling technology is very important. I say this is not unique to quantum computing. This is also true for classical computing. There's still a lot of work also in the classical computing world around packaging, about movement of data, memory technologies, and everything in the classical computing world. This is even more true in the quantum computing world. There's a lot that needs to be improved with regards to no longer having this beautiful- looking, but way not to do it chandelier approach with how you would scale a machine. It's important to work on these cabling technologies that Jan talked about earlier in communication and control systems technologies and cryogenic system technologies. There's a lot of work to be done here in the ecosystem.
This is where we're putting our money.
I think we need to form more intimate partnerships with the ecosystem and the other several tens and hundreds of even quantum companies out there. Form deeper connections, more intimate collaborations, and really focus on economies of scale and stopping this fragmentation that we see, particularly in Europe. Of course, it's a benefit that we've got so much great IP and teams around Europe alone. If we can find a more structured way to bring this all together and really work together, I think we can deliver better results as a community as opposed to all trying to do our own thing. Yeah.
Yeah, just echoing two of the previous points, one around access to anyone through the cloud and the scale that the cloud provides. The second is bridging those connections to those established industries, those regulated industries that are primarily always looking for ways to do things in a more efficient way.
Okay. Well, thank you everyone. I think that concludes our panel, and we'll join by the next session. Thank you.
This is one of the main aspects we were looking for. We want to have this down to pulse level control. For the teaching, for getting out most of it, the machine needs to be transparent. We need to be able to have readouts and accesses on all the levels.
I think this was one of the essential parts of our tender and the selection of the Spark machine, for example.
Hey. Well, thank you, and good afternoon to everyone. I'm Jan Kuerschner. I'm the CFO of the company. When we were setting up the schedule for this session, it was like, okay, audited financials, I will just need 45 minutes of Jan divided attention for this. We thought, "Hey, wait a minute. We've published all the financials already." You've probably all studied those before. We just go through the highlights of this. Whenever I'm asked, what does your new company do where I work, it's like, well, we are IQM. We build quantum computers, and we sell quantum computers, and we are a very scientifically driven company with a big workforce of very intelligent people who drive a tech roadmap. That's my elevator pitch, and that only takes three stories to go. What do I want to talk to you about?
As I said, we build and we sell quantum computers, you can see over here we are very proud of this, that we are delivering to customers, we actually do have revenues, we managed to increase these revenues from 2024 to 2025 by 91%. EUR 31 million was our number, EUR 31 million that is. I cannot make a statement like how is this going to be in the future, but I can ask you a question. Wouldn't we all like this growth to continue like that? Revenue is one thing, especially as a smaller and younger company like us, now hitting the capital market and being at more scrutiny on quarterly results.
We have this one metric, of course, revenue is as it is, but we believe that our order backlog is more a metric that is to judge our company, that is something that we would like to promote and want to have investors understand. We're very proud that we had EUR 67 million at the beginning of the year to start this year. Given the nature of our business, most of this will be transferred into revenue, and some of that will be slipping over into the next year. Not to forget that there will be additional sales that either hit this or the following years. It's been mentioned before, but we are very proud of these numbers. We've sold 23 systems already, of which we've delivered 18 so far.
Five are currently in the making, and among our customers are four of the top 10 global supercomputing centers. This is where our customer base is, but I'll come to this in a second. Behind these 18 machines already delivered is not just that we happens to be able to deliver and delivery was according to specs and our customers were happy. It is really there's logistics behind that. We have teams who go out to the world and at customer premises can install the computers, make them work, and with the customers together, work on the specifications. No other company in the quantum world can do that and can perform. When the machines grow bigger and when we have a higher number of sales, our team can deliver even when this scales up.
We've mentioned some of these customers before, but it sums up what Inés told about the companies in the world getting quantum- ready. We see that we have two commercial customers already who started on that path of becoming a quantum ready company. We've heard about the NVIDIA Corporation, and we're very proud of that.
Then we picked out another customer, the Leibniz Supercomputing Centre in Munich. They've just received our 54-qubit machine beginning of the year. They are a repeated customer, and they've also ordered a 150-qubit system that will be delivered at the end of the year, which is then our second delivery of a machine of this kind. That's something we're also very proud of. The revenue development we covered, but we're also very proud of this development when it comes to the gross profit and the margin that has increased. Again, we're talking about still a rather small number of machines to be delivered every year. However, as we're scaling up, we see some commercial advantage for us here, and we are very positive that this will continue like that.
Then when it comes to how do we spend our money on these functional costs, I think it's important to also look at the very last line, and it gives our employee and workforce number. We had 333 people at the end of last year. We have just hit over 400, and we'll be getting close to 500 at the end of the year. We are scaling up in all aspects of our company, in all departments and all fields. The sales and marketing team has increased, and all sales are local, even though our business is international, but our sales are local and where the customers are, we are as well. We put people on the ground who speak the language and who are in the same country as our customers, and that pays off, and we continue to do so.
G&A, we scaled up the team. We professionalized in many, many aspects, legal, finance, IT, people and culture, in order to be ready as a public company. Then most important, and come to this with the next slide as well, is the R&D. As I said, look at the workforce and also the improvement. 50% of our operating expenses is salaries for our staff, and that's something that we take very seriously, and we also need to be competitive with our salaries and packages so that we attract the best talent that is out there. Here you can see that the growth in R&D. You can see the growth in R&D expense from last year, 2024 to 2025. Here we're looking at four milestones when it comes to innovation. We heard before we need to reach really good gate fidelities.
The Halocene platform with a fully error correcting will be launched next year, and this is then going to be our core product, and availability is as of second half of next year. Mentioned also a couple of times, but this is really something that is important for our company. We are virtually integrated, and we can use all these momentums and all the benefits from that. Then the last one is the IP environment, and we're very proud to have so many patents in our name and be among the top filers for patents in Europe. What's the takeaway here? We talked about the revenues. This is something that we're proud of. We're not just an R&D company. We are building computers, and we are selling these.
We will be fully funded after this de-SPAC transaction, and this will get us way on our roadmap into 2028 without the trust, and we believe even further with the trust. As we spoke before, we spend wisely our money, but we concentrate on R&D, and we do everything necessary to fulfill our tech roadmap here. Then, I touched on virtual integration a couple of times, so I'm going to skip this over here. Just a couple of takeaways here. We started well-funded into this year, and we already reached quite some fundings in our history. We successfully finished a Series B funding at the end of last year, over EUR 275 million. That brings us to this EUR 545 million cash that we raised so far. The last number, again, we want to emphasize on that.
We do spend a lot on R&D, and this is the core focus of our company. Again, we are IQM. We build quantum computers, we sell them, and we are very strong on R&D. If you go out there and keep that in mind, then we are the best investment currently that is out there for quantum. Thank you.
Thank you. Now I think it's for our Q&A session. May I invite Jan, Inés, Jeff, and Jan back on stage? Yeah, whatever you said. Sorry. Cool. Now we're taking questions. If you put your hands up, we will have Ciara on one side, we'll have Heather on the other side, kind of passing the mic to you.
Hi, Quinn Bolton with Needham. Thank you for the very informative session. Wanted to start off with, you guys have highlighted the 23 systems sold at 818 delivered, maybe spend a minute on the quantum- computing- as- a- service side of the business. How do you see that playing a role as the industry develops? A follow-up question. You've sold quantum computers to four of the top 10 supercomputing labs around the world. Do you think that there's an upgrade cycle that they will come back? I think LRZ has already done that. How quickly might they come back? Is that every two, three years where they may want a higher qubit system? Thank you.
Let me take this one. The first one, actually, even for selling the systems, often our cloud service is super helpful because customers, before they make a larger investment into a system, they do validate the technology through the cloud. We use it for this purpose quite a lot, to showcase our technology and to attract customers, even for the on-premise systems. The way we see the world is that at the moment, and I think Inés explained this very well, is you can run real applications, you're still limited somewhat in the problem size that you can solve, right? It's more about readiness and educating your own teams, your own workforce. You have seen the kind of adoption curve that Inés has been showing growing over the years. I think this is also very reflective over how we see the cloud development.
We are in this ramp-up phase. It's a lot about education. It's a lot about understanding what works, how does it work? The real boost, I think, in the adoption, especially on the cloud, will come once the commercial use cases get unlocked. On the business model for the selling into the supercomputing centers, indeed, this is the way the industry works in terms of upgrades. If you look at the conventional supercomputing industry, these big supercomputing centers, they do upgrade their systems maybe every three, four years is a typical upgrade cycle for supercomputers. In quantum, actually, we sometimes see even faster pace because the technology is developing so fast. There was one example, the Leibniz-Rechenzentrum in Germany. They started with a 20- qubit system. Now they're running on 54 qubits, later this year, they will get a 150- qubit system.
Also in Finland, we have already to a single customer deployed three systems. I think this is a very typical behavior that we see in the industry. On top of this, usually there is also a service component, which is more recurring in nature, right? Again, in the supercomputing industry, typical values for service contracts are 10%-15% of the total deal size. We do think that this is reflective also for the quantum deals that we are making.
We'll give one, too, on this side, if that's okay.
Thanks. Sami Sarkamies, Danske Bank Markets. I have a question on the stack. You're doing the full stack. You have quite a good amount of revenues already today. I think you're intending to grow 100% also this year. Where is the value in the stack today? Is that mostly in the hardware? Do you see that evolving going forward? How is it reflected in your employee base that what are most of your employees doing?
Maybe I can start answering, and then Inés can continue on the stack, the higher level, and also since you're leading large tech teams, you can tell about how they're distributed. I can talk maybe more on the system and chip side. We do innovate a lot on the chip itself. We have our design tool. We have the chip factory. The rest of the system, many of those components we buy from third party providers, and we just integrate. There's value in the system knowhow, but then not so much anymore in all the different components. On the hardware side, we see most of the value actually in the processor and how it's manufactured and how it's designed. Going high up in the stack and for the workforce, maybe Inés might want to talk about it.
Yes, of course. Currently our algorithms and application teams, we are around a little bit more than 40 people, and we plan to grow more towards the next year, and of course, in the next years as this adoption takes off more significantly. As we are progressing, it will be more and more important to support our customers, not only at the level of using the hardware for error correction, error mitigation, this kind of research and development tools, but we will need to support them to actually develop this adoption pathway that I was describing in my talk. We are starting with around 40 people, and this is a bit of, I would say, around 30% of the rest of the technology stack. It's kind of relatively small proportion, but we're ramping up.
Maybe one gentleman in the front.
Thanks for hosting the session. It's Craig Ellis at B. Riley Securities. I wanted to ask a question on the enterprise engagement action plan, the question was how you plan to scale up going forward and what the gating factors are. Is it actually getting the right people to the region, since I think I heard that we're going to sell locally, or is it just awareness on the enterprise side since quantum is relatively new, even though we have had some partners like NVIDIA make some important announcements that have, I think, raised enterprise awareness. How do we think about the specific steps and what investments are needed to make them to the extent that it's go-to-market salespeople, et cetera? Thank you.
I think I'm going to start, then I will give the floor to my colleagues. When it comes to your question, I think it contains multiple elements. One of them is indeed engagement, because there is a lot of education to be done, there is a lot of exploration because, for instance, I was showing in one of my slides a number of use cases that are also quite interesting for us to learn about. Not only because it allows us to learn from customers, what are their bottlenecks, for instance, what do they care about, what are their KPIs, but also they provide to us their data sets. One thing is that, let's say, this complexity theory is telling you optimization is going to be polynomially faster than quantum algorithms, optimization are going to be bringing polynomial speed up.
The other thing is really the instantiation in specific examples of this generic mathematical statement. What I mean here exactly is that depending on the specific instance of a problem, quantum computers may bring different levels of advantage, or even not bring advantage at all, but just another solution. We are very interested in exploring data sets from different industry customers and then providing them the feedback, it's kind of like a back-and-forth process of education on both sides. Maybe I give the floor to Jan, too.
You asked on the necessary investments, of course, a lot of that is for the production of the hardware and chip development. Currently we are a fully integrated company, we do most of the parts ourselves. However, we're always looking out there, is there any better way and also more efficient and cheaper way to do things. Currently, for our tech roadmap for the 150-qubit machine, it's all in-house. We are on a path to the 1,000-qubit machine that is mainly in-house, unless there is any major advantage we see with partnering with someone else. Further to that, it's always a step-by-step development. The 5,000- qubit development is currently also planned to be in-house, whatever comes beyond that, we can't really know because we first need to get to that step, then we can make further decisions.
We don't see any other development from our competition either. We all need to face the same challenges and need to get there in order to see that. However, if you look in other tech and computer markets, there is the development towards not being fully integrated all the way to the end. This is probably the path, we're very far into the future now, even though it's just a couple of years.
Maybe if I can build a bit on this aspect of how important it is to build the trust and also the common expertise. We need to build a bridge between the quantum experts, the experts in quantum algorithms, and the domain experts in different industry environments. This is how it's very important to build this relationship during this adoption phase, that at the end of the day, we can build these pilot lines and do the testing and the validation.
Thanks.
Hi, Liam Pharr, Bank of America. Thanks for taking my questions. I was wondering if you could start by discussing your pricing strategy. Obviously, you're not getting into specifics, but do you look at it as kind of a flat pricing across a certain qubit number count? Is it variable based on the configuration for a customer, or is it based off of the end value that you're creating for specific customers based off of their specific use case or application overall?
I can start, and then you can give more specifics. It of course, very much depends on the product. We have also seen in the videos and on the slides, we have this IQM Spark product, which is an educational quantum computer. There we kind of have a fixed price. It's a bit under EUR 1 million. On the other side, we have cloud, right? We're distributing also through AWS, and then these are the typical pricing models that you see there. We have these advanced big quantum computers that go into the large data centers. Here, that's typically a longer process. You engage with a customer, and they have a certain budget, and it's a tender, and the question is, how do you create an offer that fits into the budget and that it's compelling?
There's not a list price or anything like this, but this is how the dynamics of these deals work. Maybe you can talk a little bit more about how this then translates into numbers.
In general, since we are mainly talking tenders so far with these public customers, it's all public knowledge afterwards. Having a pricing strategy that differentiates largely between customers is not a good one to follow because the next customer will know how much we sold it for. We of course, try to aim at somewhat of the same price. However, as Jan mentioned, every customer is slightly different, so there's other requirements that they have technically. They might require other services, so there is a difference in price, but that's within a normal range. Going further, of course, we do have prices in mind, but we also need to see what the competition does. Far, we believe that we are very competitive with our prices, so we do manage to get them agreed with customers. But yeah, we will need to see.
Just having a price and then multiply that by the quantums will not work. Just if you make the math from a 150-qubit system to 1 million, it would be nice, but it doesn't work that way. There needs to be some scaling effects on the production side as well. Prices will not just be multipliers.
Makes sense. I guess on that cost side for these higher qubit systems for superconducting, how do you see the cost of the manufacturing of these systems scaling with all the cryogenics required and all of the kind of footprint needs for superconducting versus other modalities? Do you think that the on-prem model that you guys currently have with superconducting may not be as sustainable as other modalities, given that they have potentially lower costs at higher qubit counts? Or how do you see that trade-off progressing as you scale qubit counts to thousands and millions in the next couple of years?
Yeah, obviously there needs to be some scaling factors. You cannot just multiply the number of qubits with the current price. There is a lot of room for improvement on cost per qubit. Maybe that's the metric that you have to look at. I already mentioned the way we currently build the systems with these cables, and it's just not the way it's going to be in the future. For example, if you use, instead of these bulky microwave cables, if you use fiber optics, you can produce fiber optics at much lower price, and also they have a much higher physical bandwidth. You get much more signal through a single line, so you can multiplex a lot. Just to give you one example there. Cryogenics is actually not the cost driver at all. If you look at the systems, this is one thing.
Also keep in mind that nowadays, I think many, if not most of the modalities actually require some cryogenics somewhere in the systems if you think about photonics and others. This is actually, I'm not worried about at all about cryogenics. We just buy those cryostats off the shelf. The main cost drivers are really in the system itself, like the connectivity and then the electronic control, where also you have huge potential to improve. At the moment, we use very advanced FPGA boards, which are quite pricey. If you, for example, would replace them with ASICs, again, you could have a huge jump. There are many kind of screws we can use to bring the cost per qubit down over the system generations going forward.
Any other questions?
Thank you. John McPeake at Rosenblatt Securities. Thanks for doing it, guys. Seems like you guys are leading the qLDPC charge in superconducting. This may be for Inés. How confident are you that you can get non-Clifford gates to work with logical qubits from qLDPC codes?
We are very confident because, well, currently our strategy is to follow similar concepts as the one that exists already for surface code and other different kinds of codes, including lattice surgery and T- gate factories, for instance. We follow the same concepts, of course, or similar concepts, but we want to exploit the improvements that quantum LDPC code offer. Like for instance, what I was commenting before, the fact that we have several logical qubits per patch that can then be leveraged to build these kinds of gates. We have a very relevant collaboration with a professor from Germany, from the University of Mainz, Professor Jens Eberhardt, that is supporting us quite a lot in developing this roadmap to achieve or tackle all the different elements in the quantum error correction stack.
In this regard, our plan is actually to build our workforce also in academia around Professor Eberhardt, and this is, of course, going to bring a lot of significant ideas and workforce and specialized, let's say, expertise in the area of quantum LDPC.
Thank you.
I'll give this side a chance.
Thanks. Kingsley Crane at Canaccord. I think it would be fair to say that you're a global company with global ambition, certainly a European heritage, and that's been a differentiator for you. European sovereign demand has been instrumental in your success. Just would love to hear more about your plans to grow that reach globally. Of course, you've had some early success in the U.S. and Asia. Why that distribution moat might not become
a ceiling longer term. Then just thoughts on if benchmarking initiatives like DARPA could help play a role in that. Thanks.
Yeah. I think, indeed, we have a certain heritage, and this is where we come from, but we do have this global ambition. We have shown, I think, that if you have competitive products in the market, you can sell around the world, even in very competitive environments like here in the U.S. The ambition level clearly has not stopped at this one deal we did with Oak Ridge. I think it is a very good example of what you can do if you have advanced quantum computers available. This is clearly not the expectation level that we stop there.
I think Jan mentioned it well, that usually the way we start engaging is by building up a local sales force. We have been starting in the U.S., but also there, we are still growing the team. We do think that there are effects from scaling the team to scaling the business. Again, this is, of course, only one part of the business, is selling the systems. We do believe in the cloud as well. Of course, there you have different distribution channels and different scaling effects. It depends, I think, very much on the use case that you look at if you have more sovereign needs or if you just need the compute, for example, and can scale through the cloud. From this perspective, we do have the ambition to, in the end, scale globally.
We start now in the very active tech markets, like in the Asia- Pacific region. We have started in countries like Japan, Korea, Taiwan. These are all countries which are historically very strong in compute, advanced compute, semiconductors. Then, of course, here in the U.S., which if you look at the two business models, selling into computing centers, data centers, and selling for the cloud, are obviously key markets. This is really our ambition level to be a leading company also here.
Any other questions?
This is Tyler Anderson from Craig-Hallum. Thank you for taking my questions. You mentioned being able to turn off the coupler fully. Are you able to do this for dynamic decoupling as you're doing circuits? Does your software enable users to go down to the gate or the pulse level?
The way these tunable couplers work in turning them off is you can design them such that they have an efficient capacitance, which can be positive or negative. It has a zero crossing, so to say, independent of how the environment is structured. If you change the environment, the zero crossing would just be at a different operating level. You always have the zero crossing, you can always turn it off. The pulse level control, yes, we do offer pulse- level control. I think this is one of the key points of having this open and a modular stack that you can control the quantum computers all the way down to this level of the hardware.
Can I just add one more? When do you expect to have the architecture in your hardware that supports the barbell code? You had mentioned the need for custom decoders. Is this something that you would develop yourself in-house, or is this something that you would leverage third parties for the decoder?
Sorry, can you repeat the last part?
Yeah. You mentioned needing a custom decoder for the new qLDPC code that you had created. Would you develop this decoder in-house, or would you leverage third parties?
Oh, right. Decoders. We are developing in-house, and of course, we also have partners with whom we are working on multiple kinds of solutions. We are also working with partners like Riverlane on these kind of solutions. Of course, we have to come up with solutions that are specific to our own error correction codes, specifically, for instance, the barbell code.
Would you say that splits between when you're doing the grid architecture and the star architecture, that there could be two different decoders?
The baseline we expect to be the same or similar, but we need to do specific adaptations to specific quantum error correction codes.
Hey, guys. Thank you so much for the time. Ryan Choi from Bank of America. My question is on your top-line visibility. Your three customer sets, academic, sovereign, and enterprises. Do you see broad-based visibility across all three of the customer sets, or is there any particular, is it enterprise you're seeing more visibility in? How should we think about that customer mix as we move towards truly unlocking commercial value in quantum computers? Just the last one there. As your existing tech stack matures, is there anything else you need that you don't have today? I know Blair mentioned earlier that you guys aren't going to lean into M&A heavily, but just curious from that perspective. Thank you.
I'll start with the customer base and then hopefully you jump in for the technology. Yeah, we see the first switch towards commercial customers, and we believe especially banks like your employer, would need their own quantum computer in their own data center because a quantum computer will be a complement to the computers as we know them nowadays. We are selling to all these supercomputing centers, and they are on our customer list from the commercial side, of course. Others might just need computation time and maybe not own the hardware or have the privacy of their own system. These are the ideal customers for our cloud, which they might run our own systems, but they might also run them on AWS and where our machine sits.
Academia will always be out there because talent needs to be educated and the look and feel, at least that's what I hear from my physics colleagues, is important to them. There will always be machines on-premise at universities to teach the future scientists.
Yeah. On the roadmap, I think the roadmap that we show, there are no gaps in there where we would say we need a scientific miracle or breakthrough or anything like this. Some of the elements obviously are lower TRL level than others, and we need proper engineering and proper development. This roadmap is really based on solutions that we know today exist, and they have been shown somewhere in a lab and need to be advanced. Of course, if a scientific breakthrough comes, we are happy to take it, and we are happy to take acceleration in the roadmap. The way the roadmap is designed is really based on solutions where we know they exist, and we need to push the engineering further to increase the technology readiness level to bring it to a real product. No miracles are needed there.
Hi. Felix Henriksson, Nordea. My question is on the technology debate between superconducting and trapped ions. I think one of the advantages for your superconducting methodology in the past has been that it can utilize existing semicon fabs. Also, we've seen the likes of IonQ also moving their products from lasers to a semiconductor-based roadmap. I am just wondering, does that change the narrative there and the scalability advantage of superconducting?
I think what others do doesn't affect, obviously, our roadmap and how we do things. We do see scaling effects in the industry. Generally, what we have seen recently, announcements from companies like GlobalFoundries also getting into the space. Applied Materials, I think, is quite active as well in quantum, and many other semiconductor players as well. Of course, these effects you might also see with other related technologies. I think being able to use semiconductor processes is one advantage of the superconducting technology, but it's not the only one, and probably also not the one that makes the only difference. Some examples I gave is really about the speed of operation, for example, and in making sure you can clock in a very fast pace, the quality, and that it works today. There are several reasons why customers choose superconducting technology today.
For the customer, of course, it doesn't matter so much if you use the standard semiconductor process or some other processes. For the customer, what matters is that they get a solution or a system that works, that is deployed in time and in spec, and using standard processes is just one way to get there.
Any other questions?
Thank you for today. It is Sujay Tata. I am from Stifel. I have two questions. First one, you guys talked about the importance of hybrid solving or heterogeneous compute. I think when Suhare was up there, talked about the CPUs, QPUs, and GPUs working together. I am just wondering, where do you guys get the GPUs and CPUs from? As you scale, is that a margin implication or a pass-through cost to a customer? That is my first question.
Usually when we sell into a big computing center, they are already running hardware. There is already a set, like servers full of GPUs and CPUs, and we integrate in there. Of course, in principle, if a new customer comes and say they want to buy a completely new setup, we would need to partner with someone who produces GPUs. Obviously, we do not develop and build GPUs. In most of the cases, there is already existing hardware running in the data centers where we sell into.
Got it. Thank you. My second question is around the co-developed ecosystem nature you guys were talking about. Specifically, you guys cited a number where you had a number of companies spin- off using your technology. I am just wondering, how are these intellectual property boundaries structured? Do these spin-offs retain any exclusive IP for the application layer of the software? What does IQM's longer-term revenue model look in terms of those equity stakes, potentially? Thank you.
That is a very good question. Of course, because this is a bit always the caveat that exists with co-development, right? Of course, we always take care of doing IP agreements where it is very clear which area is for each company or organization that is within this agreement. We do have some very critical parts of our stack where we are very, let us say, we keep our IP. When it comes to, for instance, applications, algorithms, we have a more open policy, which enables us to really do these co-developments with customers.
Cool. Another question?
Hi. Waltteri Rossi from Danske Bank Markets. Thank you for the presentations. About the commercial timeline, and in terms of technology, what would you say are still the key things that need to be achieved in order to start seeing the end-use applications in a bigger way, if you can summarize?
I can start. Yeah, you can also chip in. The overall way we see this is that there's not going to be suddenly a day X when all applications are possible and before nothing works, but it is more gradual. In the way, I think, looking at what happened with the GPUs is a good example. It was mentioned earlier that initially they were created for video games and graphics and the like, and over the time, they became more and more powerful and more and more applications were unlocked. This is also the way we see it with quantum.
If you look at our roadmap, this is the way it is displayed there, is that if you take the simulation part, for example, you could think about it this way, that as the hardware progresses over time, the molecules that you can simulate, for example, they become bigger and bigger and more and more complex. I don't know, maybe Inés can talk a bit more.
Yes, exactly. When I was saying, for instance, the day, or the quantum day, it's actually not a day. It's going to be, let's say, a set of results that the community will be having. For instance, we will find out that there is a workflow that has shown that they have been able to, for instance, describe a certain catalyst that is key for fabricating fertilizers, for instance, called FeMoco, which is very complex because it has more than 1,000 electronic orbitals, and before it was just solvable with classical solvers that were not so precise. We will find out these kind of headlines appearing more and more. It will not be a day, but let's say a sequence of announcements of this kind.
For that reason, I think it's kind of a moving target, because to be very honest, classical methods are also improving. There will be a moment, I think this moment will be close to when I was in my slide, I was depicting this moment when we have full quantum solutions. This will be a moment when we are able to simply solve all the orbitals of all the molecules and materials that we want to discover. This will be exactly the time when we will stop competing with classical solvers, because we are talking about problems with a size that not even with a powerful classical computer you can solve them. This will come later on in the progress. It will be really progressive, I think, like Jan is mentioning.
Another question?
I wanted to follow up on the quantum error correction, the low-density parity- check codes. Your competitors in the ion- trap and neutral- atom modalities will say they may have an advantage because of all-to-all connectivity. Looking at superconducting modalities, the long-range couplers seem to be the key technology to enable those more efficient codes. Can you just sort of talk about anything you've seen from other competitors in superconducting, where they may be on long-range couplers? Do you feel like you have a lead? Most importantly, are you able to patent some of those technologies that you've developed for long-range couplers?
I think the closest company or competitor company that is working on quantum LDPC is, of course, IBM, and they are using a very advanced quantum LDPC code that is called Gross code, that have very high performance. The price to pay is that they are very hard to manufacture. For instance, one aspect that we have sorted out is that we do not require what is called periodic boundary conditions in our QPUs, meaning that we don't need to connect qubits in the borders of the chip with each other. We don't need to close down the connections. This really implies that there will be less errors in the manufacturing process. It will be more streamlined. We have been able, for instance, to solve this specific problem that, in the case of our competitors, they are not solving.
Of course, this solution is IP- protected. I don't know if this was one of your comments. Yes, we are IP- protecting all of these kinds of solutions that are kind of reducing the complexity when it comes to the manufacturing.
Next question? No? In that case, I think we successfully conclude our Q&A session. Thank you.
All right. Well, in the spirit of quantum optimization, we're actually running ahead of schedule, which is great. I will just do a quick recap on the timeline for going public. I know that Pete mentioned some comments around this at the beginning of the presentation, main thing to know is we declared our F-4 effective last week, which would indicate that we're looking at a very soon or short timeline to finish up the transaction. We cannot communicate the exact date, I think a lot of you have seen these processes go before, you can maybe imagine that it'll be the next couple of weeks, for which we will be listed on the U.S. Nasdaq, hence why we are in the close vicinity of the building. Yeah, we're super excited to be going public, of course.
We do feel that we're in an inflection point in the quantum capital markets. I think Jan mentioned that we're starting to see the entrance of a new customer wave, namely commercial customers adopting hardware, which is, of course, an exciting development for us. For a number of reasons, we believe that now is the right time for IQM. Of course, what we're most excited about is the various benefits that being public brings, not only for being a European company, being able to get more visibility and build more transparency and confidence with customers. Also the access to capital, of course, is the perhaps obvious one, and diversification of the investor base, using our currency as a way to attract and retain talent, of course. Embracing employee ownership and then, of course, not at least our corporate strategy and our M&A agenda as well.
Just a little bit of a board update. Of course, I think we've had many questions over the last number of years on what our governance structure looks like, and maybe there's some familiar faces with here on the SPAC deck. As of today, our board of directors consists of five individuals. You see here on the right, Alex Doll, Hannu Martola, Jan, of course, our CEO, who you heard from today, Sierk Poetting , who's our chairman and also Chief Operating Officer of BioNTech. Last week we announced the appointment of Barbara Venneman, who's a great addition to the board. Following completion of the SPAC transaction, we'll appoint two new board members, of course, subject to the deal completing, but one being Juho Sarvikas and Jeff Tuder. There's a bit of an update there.
One thing I wanted to close on before we can go down to the second floor for some drinks for those who want to join us, is a little bit of exclusive insight into a report that we've run now for the fourth year, which is the State of Quantum Report. This is an independent research paper that we publish together with The Quantum Insider and OpenOcean to explore the various developments in the quantum ecosystem. I think you, majority of this audience here as research analysts, but also many of you investors may find particularly useful as you start to build up your thesis around quantum. A few just takeaways here you've got up on the screen, of course, is the market momentum. Of course, contracts increasing six times since 2021. Of course, we're really seeing the commercial developments here in the industry.
Two, of course, is very highly relevant for IQM, 46% of buyers planning on investing in infrastructure and hardware as opposed to the 24% increase that we're seeing investing into cloud. Then, of course, the amount of capital that's flowing into the space as well. I won't bore you with different stats right now. I think there's a report that I'm going to encourage you all to take your phones out and download from the QR code here to get some early insights into our Quantum Insider, which has not yet been published. It's not insider information, so I'm okay to say this, but you guys will get the first peek into this report. Yes and we can also send this around after today's session as well. Yeah, this concludes the IQM's inaugural Capital Markets Day that we've called it.
Of course, we're not yet public, but we wanted to make sure we had the opportunity to get some early insights into IQM's developments and various technological focus points and strategy for you guys as part of the quantum ecosystem which we consider you all part of to really understand what we're doing here. Yeah, I hope you'll all join us for drinks now down on the second floor. We do have an open bar for those if you want to spend some time with us. Of course, we'll be around to continue with your questions and some more private remarks to the extent you'd like to. Yes second floor, we'd love to see you all down there. I'll be there. Most of the team here you saw today will be there. Yeah, thanks for all for coming. Thank you for the Blue Shirt Team panelists.
Yeah, IQMers. Yeah. Thanks very much.