John Martinis, a distinguished physicist and 2025 Nobel Laureate in Physics, is renowned for his pioneering contributions to superconducting quantum computing. His research has been central to developing high-fidelity qubits and engineering the architectures needed for scalable quantum processors. He previously led Google’s quantum hardware team, where his group achieved the landmark 2019 quantum supremacy experiment — the first demonstration of a quantum computer outperforming the world’s most powerful classical supercomputer on a computational task. In 2022, he co-founded Silicon Valley-based WE where he now serves as CTO and continues to advance next-generation superconducting qubit technology and quantum system design. Martinis was a speaker on a quantum computing panel moderated by The Innovator’s Editor-in-Chief during the World Economic Forum’s annual meeting in Davos in January. He recently agreed to be interviewed by The Innovator on why quantum is here and why it is not.
Q: You were on a panel I moderated for the World Economic Forum in Davos entitled “Why Quantum Is Around The Corner and Why It Is Not.” What is your view on that?
JM: This is a very exciting time. There’s a lot of clever ideas out there. I’m happy that there’s many ways that people are trying to do this, but in my opinion, it’s generally going to be harder to build a quantum computer that people claim it to be. It’s a huge system engineering issue, and there’s many things you need to get right, and in the end, you need to make it manufacturable and cheap, which is its own problem. That’s why, in our company, we’re really thinking about manufacturing as a key component.
Q: Can you elaborate on Qolab’s approach?
JM: If you look at superconducting qubits, which I’ve been working on, and I worked on when I was at Google, we’ve been able to build big systems where everything works fairly well. Other groups are also doing that very successfully. However, if you look at the actual experiments, there is a little tiny chip in the center with all these wires and microwave components and connectors and co-ax around it. That’s great for the basic physics that we’re doing now but our feeling is that if you want to build a million qubits that’s just not going to work. Just look at the cost: it’s$10,000 per qubit, so the cost of these machines is tens of billions of dollars. Our point is we just need to build it in a totally new way to make it scalable and cost efficient. That’s hard to do. We came up with an idea on how to do that, which is basically using semiconductor manufacturing and packaging ideas that we are using right now for chips and especially the GPU [Graphics Processing Unit] systems, and to take that technology, which exists already, and then modify it to make it to work for superconductivity. The basic technology is there, and just by working on the processing and proving everything we think we can put it together and build a cost-effective system. The semiconductor industry really knows how to do this well, so to do that efficiently you need to collaborate with them. We found that by working with a couple of the companies we can do things that are unimaginable in a lab run by physicists. So, we have a plan and we have a paper out that describes that and we’re making progress on that. We think this is the only way to do it. But it’s hard. It’s very different than what people are doing now so we need to invent a lot of new things.
Q: So you are creating a consortium of companies
JM: That’s the idea. At the low end is the quantum hardware. We’re the system integrator, because we understand the quantum mechanics, and we’re working with companies like Applied Materials. We describe what needs to be built, and they know how to put things together.The high end of our consortium is being led by Hewlett Packard Enterprise, because they build supercomputers and clearly,quantum computers are going to be used with a supercomputer. HPE understands how to break up the problems, and they have the infrastructure. There are also other companies involved but that’s the basic structure of it. It’s a little bit more complicated than a supply chain. The hard thing about this is that you need to share in the rewards of building a quantum computer and it is hard to share. We all know that. Most of the large companies working on quantum are what’s called vertically integrated. They try to do everything, so they gain all the advantage and the money. It’s hard for a small company to do this because we don’t have the resources.
Q: I know you are the brains behind this, but what makes you convinced that the superconducting approach is the way to go? There are other competing approaches.
JM: If you talk to any physicist and they will say their approach is the best. However, when you build a system, everyone knows it’s not the best that matters, it’s the worst part. And no one talks about the worst part – what they need to fix – except for our company. We say super connecting qubits are great. They work, but they don’t scale. So, let’s talk about the problems of other approaches. Take, for example, neutral atoms and ion traps. Neutral atoms proponents say they are good at scaling up. Okay but the problem is those are very slow. It’s 1000 to maybe 10,000 times slower. You can maybe make up a little bit for that with architecture – people need to show that – but a factor of 1000 is a big deal when you build a computer. You can essentially make 1000 times more money if it’s 1000 times faster and you can solve problems that you can’t now. The other thing is, people point to how neutral atoms are scalable, and people are talking about 1000s of qubits now, and yeah, right now, if you’re talking about the horse race right now, and who’s ahead, that looks interesting. However, if you want to build a general-purpose quantum computer, you need to go to millions of qubits and make optical traps that can do that and connect them together with some optical interface, which is not going to be easy. People are working on it, but it’s not clear what the overhead is on that.
Q: It sounded from our discussion in Davos that you, you agree with Arvind Krishna, the chairman of IBM, that, the physics has been solved, and this is now an engineering problem. So, although it’s hard, it can be done.
JM: It’s not just an engineering problem. There’s still a bunch of materials physics and other things we need to figure out that tends to be kind of fundamental. We’re still building physics models for our qubits, and I would say that’s the hardest part, because you aren’t 100% clear of what’s going on. And then I would also say that if you take what people are doing right now with superconducting qubits, and you take their systems and you say you want to incrementally improve the approach, it will fail. We need to come up with a whole different architecture, which we’re doing. I imagine others are doing the same. So, do you call that an engineering problem? I understand companies want to give that message so that all the stockholders feel good about what they’re doing. And in some sense, it’s true, but I try to be a little bit more forthright about how there’s still some real big problems to solve. I think they can be solved, but we still need to do it, and it’s not easy.
Q: Still, you are relatively confident that by the 2030s we will have a fault tolerant quantum computer?
JM:. I’m going to define fault tolerance for you, which is 10 to minus big number errors. That’s a lot of qubits. People are talking five to 10 years. Everyone is now. You need to realize the reason everyone’s doing that is that there’s a government program and everyone is making bold claims to be part of that program. I’m glad people are thinking this way and we are too. We’ll just have to see who can deliver.
Q: What are the key takeaways for business? What should they be doing to prepare?
JM: The number one thing is to think about Internet security. RSA [a family of public-key cryptosystems widely used for secure data transmission] which is the basis of Internet security, could be broken in five to ten years. And the nice thing is this is not something to panic over, because people have developed quantum resistant protocols. There’s a very serious effort at NIST [The U.S. National Institute Of Standards And Technology.] They’ve developed it, they’ve talked about it, and there’s a bunch of companies that are implementing that. You just need to learn about it, and figure out a plan to implement it. The good news is that people have been thinking about this for a long time. The way I look at it is in the history of spying encryption, all encryption systems have lifetimes so it would be unrealistic to think that RSA has an infinite lifetime. What I’ve also been told by people in the industry is that RSA was developed over time, and there’s a lot of different platforms and ways to do it. I think this is a chance where there could be a more uniform deployment so that in the future if we have to change encryption it will be relatively straightforward.
Q: What excites you the most about the potential of quantum computing.
JM: Being a physicist, I’m naturally drawn to the idea of using a quantum computer to map the quantum mechanics of chemistry. I really like that application, because it is already kind of proven out algorithmically. It would be great if you can do that with molecules to make new materials, for drug discovery and the like. It’s going to make all sorts of things cheaper and better. The example I like to give is we need rare earth magnets for electric car motors and other things. There are worrisome supply chain issues so it would be great if we can build magnets out of less rare earth materials. There are many, many things like that. Maybe quantum computing could enable the mining, or the extraction of the ore, the refining process, to be done more ecologically. We’re entering an era where our materials are harder to get and more of a concern so if we can do better with that supply chain, that can make a huge difference to the world. There are also ideas for how quantum computers can be used for optimization problems. It is a little bit like AI right now, people have a lot of ideas about what we might do with it, but it’s not 100% clear until you build the systems.
This article is content that would normally only be available to subscribers. Become a subscriber to see what you have been missing
