The computer scientist Michael Lachner graduated from one of the DAAD’s Konrad Zuse Schools of Excellence in AI. His Munich-based start-up Aqarios develops software for quantum computers – a technology many experts don’t expect to be market-ready until 2030. Lachner believes this will happen sooner. 

Supercomputers are coming – and should be with us pretty soon. That’s one thing Michael Lachner is certain about. “In a sense they are in fact already here,” says the young computer scientist during a video call from his office in Munich’s Prinzregentenstraße. Lachner is the CEO of the software start-up Aqarios. The company provides solutions to complex planning and optimisation problems for business and industry – with a particular focus on quantum computing, classical AI and advanced optimisation algorithms. 

Lachner was one of the first to graduate from the Konrad Zuse Schools of Excellence in AI. The schools are one of the ways the German government is fostering international AI talent in Germany. At numerous sites across Germany – including Munich, Dresden and Darmstadt – the Zuse Schools offer excellent master’s and PhD programmes. Lachner heard about the Munich Zuse School relAI, which is jointly run by LMU and TU Munich, in 2022. “It was a perfect fit for me.” He applied – and was accepted.

Insights into the black box

relAI stands for “reliable AI”. For AI to be genuinely reliable, when used for example in industrial production processes, it needs to deliver understandable results. This is something that has long interested Lachner. While he was studying computer science, he focused among other things on developing algorithms which do precisely that: which don’t operate in a black box but give the user the chance to understand how the machine arrived at its results. 

In this context, a new technology can help – one with which his firm is engaging very intensively: quantum computing. In stark contrast to conventional computers, such machines harness the laws of quantum mechanics. This enables them to process not only the numbers zero and one in a particular step, but any number of states in between. This makes the “supercomputers”, as Lachner calls them, extremely good at solving complex problems. Especially when the best solution has to be identified among a vast number of possibilities: logistics routes, production planning, portfolio optimisation in the financial sector. These are areas in which conventional computers quickly reach their limits.

For a long time, quantum computers were virtually impossible to use in practice. The main problem is that they are difficult to build and still prone to computational errors. Currently, the systems are expected to be market-ready by 2030. Lachner believes this will happen more quickly.

His company has developed a software platform that firms can already use to work with algorithms that offer not only quantum computing but also quantum-inspired approaches and purely classical optimisation algorithms. “They can in fact already create additional value today,” says Lachner. Another big advantage is that the platform is completely hardware-independent. This means that a firm can use it today to resolve its optimisation problems using classical algorithms – and then easily update it to a quantum computer in two years’ time. 

A revolution with advance warning

This aspect is particularly important to him: “We want companies to understand today how and where they will be using quantum computing in the future – and how they can prepare their workflows and teams accordingly. So that they don’t make the same mistakes they did with AI.” He explains that artificial intelligence has been discussed for decades, and that the huge potential offered by AI applications then became apparent in the 2010s. However, it was not until ChatGPT was launched in 2022 that many companies woke up and took notice. “As a result, firms suddenly found themselves stuck with outdated processes and no idea of how AI could help them. We’re now nearly four years on, yet most still have problems integrating AI in a really meaningful way.”

When he looks back at his degree course at the Zuse School relAI, it is the high quality of teaching and the many networking events that particularly stick in his mind as positive aspects. “It’s really helpful to be able to discuss the topics one is interested in with clever people.” And this didn’t happen only within the confines of relAI in Munich, as there are two other Zuse Schools with numerous sites in Germany, including Darmstadt, Berlin, Freiburg, Dresden and Leipzig. “Once a year the entire network gets together to share ideas and talk about their work. It was very inspiring to learn about the research currently being done at the other Zuse Schools.”

Munich rather than Silicon Valley

When Lachner talks about Germany as a centre for AI he is optimistic. He’s happy to admit that things didn’t always work out all that well in the past. “Germany was really bad at commercialising its top-flight research.” He thinks it’s a good sign that the German government is investing heavily in quantum computing. “The amount of public money that has been made available in the past years is considerable.” He highlights the various funding programmes at federal and state level, institutions such as the Leibniz Supercomputing Centre and the close cooperation between universities, research centres and companies. “I believe Germany has a real chance of being one of the world’s frontrunners in this area.”

For the moment at least, he’s certainly not considering packing his bags and moving his firm to Silicon Valley.

Klaus Lüber (12 May 2026)