Quantum Computer Feasible in 5-10 Years

27 July 2026

How far are quantum computers from practical applications? And what challenges must be overcome for them to reach their full potential? Nobel laureate John M. Martinis discusses these topics in an interview during his visit to this year's PGI Day at Forschungszentrum Jülich.

John M. Martinis is a pioneer of quantum computing. His research helped lay the foundations for today's superconducting qubits. He later led Google's quantum-hardware team and played a key role in its landmark quantum-supremacy experiment. Today, he is co-founder and Chief Technology Officer of Qolab, where he is working on scalable quantum computers.

Martinis came to Jülich to give a keynote at the Peter Grünberg Institute's (PGI) annual PGI Day after a visit to Croatia, where he was honoured by the president for his scientific achievements and Croatian heritage. In Jülich, he also met long-time collaborators Rami Barends, Frank Wilhelm-Mauch, and Mohammad Ansari.

In the interview, Martinis explains why a general-purpose quantum computer could be built within five to ten years and reflects on his career in academia, at Google, and in the startup world.

Professor Martinis, one question that interests many people is: When will we have quantum computers that are really useful for practical problems?

The quantum computers that people are building today are kind of at a scientific scale to test things. The real idea is that useful computers will be built - general-purpose quantum computers with 100,000 to 1,000,000 qubits. Right now, we're significantly below that.

I used to avoid answering the question, but now I answer it. This is partly because there have been some significant theoretical developments in the field. There is also good funding, and many people, including my own company, claim to know how to build a quantum computer. So, I think a general-purpose quantum computer could be made in 5 to 10 years. However, it could take longer because scientists and physicists have not fully grasped the system engineering and the difficulty of building a quantum computer.

One reason I say 5 to 10 years is that when you build this general-purpose quantum computer, it will be powerful enough to crack present-day encryption, including RSA encryption and elliptic curve cryptography, which is used for a lot of hardware. It's used for Bitcoin.

There are some serious developments going on that could greatly impact computer security.

It's important to be ready for all that. The issue is that there are some serious developments going on that could greatly impact computer security.

I talk about this because I think it's important for people to start thinking about improving internet security. There are protocols and ways to do this. People are aware of them and are working on them. In fact, I was at the White House where President Trump asked the Department of Commerce and other government agencies to develop these protocols by 2030. So, things are happening, which is good. It's very important that people do that.

In your keynote at Forschungszentrum Jülich, you talk about the development of quantum computers over time. If we compare today's quantum computers with the early days of classical computers, where are we now?

It's a bit nuanced, but I would say that, in terms of regular computers, we're around 1960. At least in the '60s, these computers were powerful enough to solve problems. Of course, compared to what we're doing now, they were very rudimentary. However, I think the technology today is kind of like what we had in the '60s. We've invented it. People have put together systems.

In my talk, I argue that we need to go from these superconducting quantum computers with lots of wires to real integrated circuits. We have a plan to do that, and that's a lot of work. It's not easy to do.

From an outside perspective, quantum computing can sometimes appear overhyped. Is it possible that they will never surpass conventional computers?

People have talked about quantum computing for many decades now, it's been a long time. Quantum computers are much more difficult to build than classical computers. And classical computers keep on getting bigger and better - just look at what GPUs can do today. In that sense, the target is constantly moving.

What speaks in favour of quantum computing is the quantum supremacy experiment we did at Google. It showed that a quantum computer could be powerful and that there is no new fundamental physics that would stop you. Sure, there's a lot of engineering you have to do. But the theory and the experiments look good.

Have there been developments in the past years that particularly surprised you?

Yes, in the last year or two, there had been developments in the quantum error codes. People have new ideas on how to assemble physical qubits into a logical qubit where you don't need as many qubits to build a big quantum computer. I still don't understand all that. There still has to be a demonstration, and I have some concerns about whether it's as easy as people are making it out to be. But it's looking really interesting. Even if that's not the case, there are straightforward improvements people can make to the surface code, which is kind of the most basic way to build a quantum computer.

What is currently the biggest bottleneck in your view?

I think the big problem is microfabrication and fabricating the devices. Right now, people can make good qubits, and they can make systems with 50, 100, 200 qubits. But you have to make them more reliable. You have to make them with fewer dropouts.

It's really the question of how to make manufacturable qubits. You can make a single CMOS transistor. But in the end, you have to make millions or billions, or I guess now almost trillions of transistors. And that takes a lot of effort to figure out how to do that. So that's what I've identified, and our company has identified as a big issue, and we have some ideas on how to do that.

You have worked in several very different environments: at a national laboratory, a university, at Google, and now at your own startup, QoLab. What have you found to be the main differences between these settings?

I'm very happy that I worked in all of these places because, at the time, they were perfect for me. At the national lab, I could focus on the physics of quantum devices. I enjoyed being in the lab. I enjoyed understanding in detail what was going on. I then went to a university where you could have a lot of students. I had a lot of autonomy in deciding what to do. It was really a time to push forward the technology. Google was great because we had the resources to build a large team and hire people. But, in the end, there was a lot of politics, and things were going in weird directions.

In a startup, however, you get to define your own direction and decide what you want to do. Of course, you have to raise money, do fundraising, and the like.

I've enjoyed learning about the different ways of doing science. For my career, it's actually worked out quite well. I was really happy teaching at the university in Santa Barbara. Everything was great. However, if I wanted to build a useful quantum computer, I'd have to go beyond university research. That's kind of why I considered working at Google. So, I'm not going to say one model is better than the other. They're just different. People have to decide based on their personalities and desires.

Interview: Tobias Schlößer

Video of John M. Martinis' Keynote (Duration: 56 min.)