How Europe’s IQM became a quantum computing soloist - diginomica

While the world’s focus is on the AI market, more interesting developments are taking place at a sub-atomic level - or at least, that’s the view of Jan Goetz, co-founder and CEO of IQM Quantum Computers.Unlike some of its lab-based competitors, Finland-based IQM sought to become a viable commercial operation quickly, completing its $1.8 billion dual listing on Nasdaq and Nasdaq Helsiki in July. This made it the first European quantum company to float on a public exchange. In general, it has held its value since then.This particular unicorn claims to be global leader in on-premises quantum computer deliveries over the past year – 26 or more systems have been delivered, Goetz tells me, including to four of the world's top ten supercomputing centers. Partnerships with NVIDIA, AWS, and the US Department of Energy's Oak Ridge National Laboratory all add to its list of achievements.Impressive stuff, though for context we must acknowledge that the small number of systems shipped to date, while large for this sector, serves to illustrate how early in the quantum uptake cycle we are. Goetz acknowledges: Of course, it’s not a huge number of sales overall if you compare it to smartphones. But we are in the very early days, and it is more than anyone else has sold.Goetz shares the IQM story so far:Like many quantum founders, I'm a scientist. All IQM’s founders are scientists, and we were working at Aalto University [in Espoo], which is the leading tech research university in Finland. At the end of 2018 we had small prototype quantum computers working in the lab. And our technology was on par with what we had seen from the US and other places.But we had also seen large funding rounds in the US: companies like Rigetti and IonQ had already raised hundreds of millions of dollars. So, we concluded that if we wanted to keep our leadership position on the technology side, then we couldn’t scale within the university. And it's not the task of universities to scale technology and build products, anyway. It's all about the science.So, the funding available on the academic side didn't match our ambition anymore. We knew we had strong technology, and we knew we had a strong team. So, the only way to scale was to start a company. Which we did, and we received our first investment in 2019, which was the largest seed funding round in Finland at that time.That pushed the ambition level up from day one, he says: We said, 'OK, if we're going to do this, we don't want to just build another small start-up from Europe. We want to build something big that's globally relevant and which can compete on a global scale'.To that end, a year ago, the company announced a €275 million ($315 million) Series B round led by US cybersecurity investment specialist Ten Eleven Ventures and Tesi, Finland's national investment fund. Less than a year later, IQM became a unicorn on the public markets, a significant player in quantum hardware, shipping enterprise-ready computers based on its superconducting technology.IQM’s early focus on the enterprise is interesting, but begs the questions of whether most enterprises are at the stage where they see real utility in quantum systems. Don’t most regard it as an exploratory space in which few commercial applications have yet emerged? Goetz says:We very much see it as a journey. As a company, our journey is determined by our technology roadmap [towards fully fault-tolerant quantum computers by 2030]. In that sense, it is like other semi-conductor industries, where Moore's law has been active for decades. We put out a new processor generation every year and integrate it into a full-stack computer. But it’s also a journey for our customers. It is not that there will a day when quantum suddenly works at enterprise scale. We call it quantum readiness, and we are taking the journey towards that together with our customers.As to what that mean practically, Goetz explains: The customers learn from us how to run a quantum computer, how to integrate it into their processes, what problems to solve, what kind of teams they need, and what the IP situation is – all those kinds of questions. But at the same time, we are learning from, and with, our customers. And that is important from a product perspective. Plus, we are making them money by selling them the machines, and we are creating revenue for them.These are very much questions for IQM too, he explains:How many quantum computers can we build? How many can we deploy? Do we have a delivery team? Do we have a service team? Plus, there is quality management, finance, marketing, and account management. As a company, we grow together with our customer base. That’s the right approach: to build the industry, and then grow with the industry.As to those sales numbers, Goetz reveals: Just today we announced that we sold the first quantum computer in South America to a customer in Brazil, expanding our global footprint. We have sold quite a few systems in Europe, and to customers in Korea, Japan, Taiwan, and the US – to one of the biggest supercomputing centers. Plus, we have sold to a space company in Poland and to an automotive and testing company in Japan. It is not just on-premises quantum computers, he notes, pointing out that IQM also offers on-demand quantum processing in the cloud. However, the company’s website makes it clear that the focus is tight on hardware sales, stating: Capability is built, not rented. Quantum computing is becoming foundational infrastructure. Access gives you experiments. Ownership gives you capability. This is Production Quantum: full-stack computers that institutions install, operate, and grow with, keeping the IP and value they create. Quantum computers that you own.So, on the face of it – with a new-generation processor released every year – IQM is offloading a lot of risk onto its customers. Why invest in an on-premises computer that might rapidly become obsolete, when a customer could simply follow a seamless upgrade path in the cloud? Goetz's pitch: It depends on the customer and their risk appetite. We sell a lot to data centres, and to supercomputing centres, and their goal is to host and run computers and make them available to their end users. For them, it's obviously super important to understand what it means to run a quantum computer and how to deploy it. So, the cloud is not an option for them. They must own the machines.There are enterprise customers who say, ‘We just want to go through the cloud. We want the applications and the upgrades.’ But there are others who are interested in on-prem deployments. One reason is the sovereignty aspect: the data security, and where the data is handled.As we have seen on diginomica this year, digital and data sovereignty, and resilience to transatlantic shocks, are boardroom topics in many enterprises. Goetz shares his perspective here: Some use cases are sensitive. When we talk about codebreaking or cybersecurity, obviously where your data is processed, and who has access to it is important. And if you think about drug discovery, the biggest asset that pharma companies have is the models of their molecules. They don't want that information to leak anywhere, because it has huge value. But it's also the availability, right? In the current situation, some people might fear that they can’t get access to the technology anymore. If someone decides, ‘We are not going to sell anymore to Europe’, then there's no alternative.For me, this is what sovereignty is all about: we need to come from a position of strength. All the GPUs are basically coming from US providers, and there's a risk that someone might say they're not going to sell those [into Europe] in the future.”Goetz continues:Plus, there is a group of customers who want the machines to run projects with their own clients. If you're running a computational load from a bare financial perspective, it often makes sense to host the system yourself. We see the future as being a mix.In a recent presentation, IQM described the quantum processor, the QPU, as the soloist, the CPU as the conductor, and the GPU as the orchestra. Unpicking that analogy, Goetz says: A GPU is very good at handling a lot of data in parallel, but quantum computers are not good at handling that. Instead, they are good at problems which are super complex from a mathematics point of view, and very hard to solve, but not necessarily huge. If you think about molecular simulation, you cannot simulate that exactly on a supercomputer. So, the problem space might be small, but it's very complex. And this is what the soloist in an orchestra does: play the part that is not very loud, but it is very complex. And the orchestra does the parallel data handling. We work closely with GPU providers [such as NVIDIA] to integrate and build this orchestra. That’s hybrid computing. If you have a big problem, but part of that problem is complex, there’s an orchestration layer that separates those parts. So, the number crunching you push to the GPUs, and the very complex part you move to the QPU.For IQM at least, the melody is strong and easy to follow. In time, more enterprises will add their voices to it and strike the right note. Image credit - Pixabay We'll send you our top articles (and no marketing spam), no more than once a week.