“Building a quantum computer is something very tangible”
Developing novel qubits in a start-up
Leon Koch has an ambitious goal with his start-up Peak Quantum: to build an error-protected hardware architecture with innovative qubits in order to bring quantum computers into industry-relevant applications. The outstanding entrepreneurial ecosystem in MQV and the excellent research at the Walther Meißner Institute provide him and his co-founders with the perfect starting conditions for this.
By Veronika Früh
Leon Koch enters a hallway that houses some of the laboratories at the Walther Meissner Institute (WMI). Right behind the door, he slips out of his street shoes and takes a pair of slippers from a shelf labeled with his name. For visitors, shoe covers are provided. This is to keep the path to the clean rooms as free of dirt as possible. “I also bought a robot vacuum cleaner here,” Leon says with a grin. It does its rounds independently in the hallway, ensuring that even less dust is brought into the laboratories. When the 31-year-old started as a doctoral student at the WMI around five years ago, not all the rooms were even filled yet. “On my first day here, a cleaner was just polishing everything,” he says. “The first cryostat was delivered a week later, and I was allowed to help set it up right away. It was impressive to experience everything from the beginning.” Nowadays, however, he spends less and less time in the labs – as co-founder and CEO of the young start-up Peak Quantum, his days are filled with a variety of new tasks.
“I really wanted to work on quantum computing,” Leon says. After completing his master's degree in physics at the University of Tübingen, he first tried to convince his professors there to expand their research in the field of superconducting quantum computers. But the incredibly expensive equipment, coupled with competition from Google and IBM, which are researching the same technology, made the project unrealistic – his professors in Tübingen recommended that Leon apply to the WMI. There weren't many labs in this field anyway, so he also applied to IBM in Zurich, the doctoral student recalls: “It's a funny story,” he says, “because both applications ended up with Stefan.” After a brief moment of confusion on Leon's part, it was sealed: Leon became Stefan Filipp's first doctoral student after Filipp moved from Zurich to Munich as head of the new Quantum Computing and Information Processing department. They started together at the WMI five years ago – a time Leon looks back on fondly: “I loved doing my doctorate here!”
At the beginning, despite his great desire to work on quantum computers, he admits that he hardly knew what he was getting himself into: “That was the time when this big Google paper came out, and since I didn't really know much about the field yet, it sounded like things were really taking off. As if it would soon be possible to build systems that deliver real value.” He jumped on the hype bandwagon to a certain extent – and learned relatively quickly that there was still a lot of work to be done. But what excited him about quantum computing in principle remained the same: the promise of what could be achieved with it in the future. “In my opinion, computing power is now really the foundation of progress,” he explains. From friends who work with climate simulations or protein folding, for example, he has learned that the supercomputers they use need many days to calculate the corresponding models. And these are still very coarse-grained, with low resolution and few parameters. “Of course, quantum computers can't solve everything,” he admits. “But I definitely see that they will one day deliver real added value for a large class of problems.” Being able to contribute to this is what excited him so much. And since he had already worked with superconductors, quantum computers based on superconducting qubits were the most obvious next step. “That was and still is the leading...” He pauses briefly, grins, and corrects himself: “It is one of the leading technologies.” Ultimately, the goal is to make a usable system available as soon as possible that enables relevant applications beyond university research, which is often focused on basic research.
Leon Koch, 31
Position
Co-founder of the start-up Peak Quantum
Degree
Physics
As part of the start-up Peak Quantum, Leon and his team are developing an error-protected hardware architecture based on novel qubits. The new technology is expected to reduce the number of physical qubits required by at least one order of magnitude, bringing quantum computers one step closer to the application for relevant problems.
Start-up founding: "Everything just came together at the right time."
That was also the driving force behind the spin-off of Peak Quantum: “In order to put our world-class research into practice, it was simply the next logical step to tackle these last few meters, which also involve a lot of engineering work, with a start-up.” At its core, Peak Quantum is about developing new qubits that could replace transmon qubits, the current standard technology, when it comes to the application of quantum computers. “We have seen that these standard qubits, the transmon qubits, are very error-prone,” explains Leon. In order to calculate truly meaningful algorithms, the error rates would have to be reduced by several orders of magnitude. The current error rate of around 0.1% may sound small at first, but it still needs to be reduced by three to four orders of magnitude in order to calculate truly meaningful algorithms. “Transmon technology is very simple,” the physicist continues. “It's a great platform for learning, for practicing, so to speak. But in my opinion, it won't ultimately be good enough to calculate the desired applications.” Error correction is possible, but the overhead of physical qubits needed to create a relevant number of logical qubits would bring many other problems with it. A cryostat would no longer provide enough space, and multiple cryostats would have to be connected with couplers, which would cause bottlenecks in the data connection and reduce performance. “We said that this had to be done in a fundamentally different way; the problem couldn't be solved with a little optimization in the manufacturing process.”
The result was a new, error-protected architecture that reduces error rates at the hardware level. However, manufacturing this new architecture is anything but easy. “We had to do a invest a lot of time to be able to manufacture it at all. We had to develop completely new processes that did not previously exist in our field,” says Leon. Despite all the complexity, however, he has not lost his pragmatism – in the end, these qubits are just coils and capacitors, only much smaller than the oscillating circuits familiar from physics lessons. And the new technology looks very promising: "According to our calculations, we are already saving at least one order of magnitude in physical qubits. This means for example that instead of a 100,000 qubits, we now only need 10,000 for a powerful processor, which can then fit into one refrigerator instead of ten. This also makes the computer ten times cheaper and ten times more energy-efficient," says the physicist. Of course, there are still many questions to be answered, and they are not yet ready to present the world's best quantum computer tomorrow. But the combination of a strong research group, the right infrastructure, excellent manufacturing at their own institute, and the prospect of brand-new clean rooms being built directly opposite would provide the perfect conditions: "Being embedded in such an ecosystem makes it possible to found a start-up that can certainly compete with IBM or Google in the first place. Everything just came together at the right time."
"In the end, it's the people that matter."
Leon laughs and smiles a lot when he talks about his work, which he clearly enjoys, and his enthusiasm is contagious. As soon as he starts talking about his new responsibilities as CEO of Peak Quantum, his grin widens even more. Team leadership is particularly important to him: “In my opinion, the most important thing in the long term is to bring the right people together and offer them a working environment in which they can flourish, develop, and grow. Ultimately, it's the people that matter.” Even as a doctoral student, Leon supervised a large number of master's students. Many of them are now his fellow doctoral students and colleagues—a fact that fills Leon with pride: “It's great to see the next generations growing up and sharing the enthusiasm. It was great to see that we were able to bring a group here to the forefront of science from the ground up.”
With the founding of the start-up and his new responsibilities, Leon was also able to learn a lot of new things. “After five years of doctoral studies, almost nothing that happens on the scientific side shocks you anymore,” he says. “So I found it incredibly exciting to start something new.” However, this is not the physicist's first entrepreneurial experience. At the age of 14, he founded a micro-enterprise selling special substrates for succulents. “It all started when I was given the task of caring for our family cactus,” he says with a laugh, but it's a long story. “Of course, it's on a completely different scale now,” he admits, adding that he didn't hold any investor meetings back then. Leon also had to develop a feel for how investment firms view a technology: “You have to learn what the market needs and communicate a crazy new idea, make what you're actually doing tangible. Even for an investor who may not have a deep background in quantum physics.” In the process of founding the company, Peak Quantum also received help from an experienced partner in the MQV ecosystem: TUM Venture Lab Quantum/Semicon offered valuable support with its various programs. “We were all physicists,” says Leon, “we didn't have anything to do with business at first.”
And yet, the deep tech sector is very open to people who are trying to find their place between being a scientist and a founder: “I'm still allowed to be both,” says Leon. “At the moment, it's not about conjuring up the best sales figures, but also about solving the fundamental problems in order to put quantum computers into practice.” But more and more, he is facing a new challenge: “Letting go! With every new area of responsibility, you naturally have to let go of old ones. At some point, you can no longer sit down and repair the machine or manufacture new chips yourself in order to really focus all your attention on the new things.”
Building quantum hardware: Between wrenches and nanotechnologies
Fortunately, he is currently still allowed to work in the laboratory, and he particularly enjoys the hands-on aspect of the job: “Building a quantum computer like this is something very tangible. When it really comes down to building the hardware.” In everyday life, he may well find himself picking up a wrench because one of the machines needs to be repaired: “You also have access to normal tools and not just electron beam writers or nanomechanical tools.” The production of quantum processors, on the other hand, takes place on the smallest of scales. The layers of superconducting materials, which are layered on silicon substrates as in the semiconductor industry, are around 100 to 200 nanometers thick, a hundred times thinner than a hair. Structures of the same tiny size are then created on top of these layers using etching processes, for example, “which is where all the quantum magic happens.” Of course, you need microscopes to see this, explains Leon, “and there are now microscopes that allow you to zoom in wonderfully and see our metal layer in great detail. You can see that it's a little rough, not perfectly smooth, but of course still orders of magnitude smoother than anything else we know.” And you can also see very clearly when problems arise: “In the worst case, you see ‘oh, one of those nanometer arms has broken off,’ and then the qubit can't work.” For Leon, the manufacture of the processors connects two worlds, so to speak – “fancy, fancy quantum mechanics and, quite simply, a metal arm, another metal arm, maybe a little rust, i.e., an oxide layer, in between.”
If physics becomes too much for Leon at some point, “at around 70 or 80,” he already has a plan for retirement. Then maybe it will be time for his second dream: to build beautiful furniture as a carpenter. “But that will have to wait a while,” says Leon. “I'm going to stay in this field for many years to come and drive it forward. I definitely hope that we can achieve this as a team and that we can pass on this enthusiasm.”
Published 31 October 2025; Interview 5 August 2025