Why Control Electronics, Not Qubits, Cap Scale

Summary

Niels Bultink, founder and CEO of Qblox, is interviewed by Yuval Boger. Niels and Yuval talk about quantum control of qubits, scaling up to thousands of qubits, qubit crosstalk, the Dutch quantum ecosystem and much more.

Transcript

Yuval Boger: Hello, Niels, and thank you for joining me today.

Niels Bultink: Thank you. Hello, Yuval, as well. It's great to be here.

Yuval: Great to have you. So who are you and what do you do?

Niels: I'm Niels Bultink. I am CEO and founder of a company called Qblox, which is a company based in the Netherlands.

It's a startup, or perhaps a scale-up company, operating in the field of quantum technology. So we are really in this deep tech, high-tech field. I'm personally coming from a background of physics. I've done a master's and a PhD in experimental physics, trying to get prototype quantum computers to work.

And over the last four to five years, I've been on this journey from a bit of a different angle than an academic one, because I've been growing this company called Qblox. So I'm happy to talk about myself and about this company in this podcast.

Thank you.

Yuval: And what made you start the company? Were you part of an accelerator, or did you just wake up one day and say, "I want to be CEO of a quantum company"? How did it work?

Niels: Well, it started out 35 years ago. No… Actually, the idea of one day starting a company, together with my co-founder, Jules van Oven, started more than 10 years ago, when we were doing our bachelor's education, which is, to be honest, already 17 years ago.

Yes. We always had the idea of one day starting a company in some high-tech field. And I think after our master's, we both went our own way. He went into oil exploration using sensor networks. I myself went into the semiconductor industry. During these careers, we were looking for spin-out opportunities, but we never really found the right subject, the right opportunity.

After a few years in my career, I decided to go back to university, because I was missing the physics, and in particular the quantum technology. And already early on in my PhD, building these quantum computers, I found that this was actually going to be quite a nice opportunity.

A lot of technology was being developed within the university, which is TU Delft, but also at the governmental institute, TNO. A lot of instrumentation was being developed, and it was actually much more engineering than science. The fact that all this engineering was happening to support the science meant to us that many more labs in the world would be needing these types of engineering solutions.

So that's how we found the opportunity, and we decided to take all this engineering work and turn it into a real product for others to use.

Yuval: So what does the product do?

Niels: Yes, what does it do? It's part of the quantum computer stack. If you look at a quantum computer as a whole, it is built up of a lot of layers and components.

Of course, you need a quantum chip, also called a quantum processor, where quantum phenomena are being controlled and used for computation. But next to that, or as a part of that, there's also a whole cryogenic environment, or ultra-high vacuum, that keeps your chip well screened off from the environment.

And then next to that, you need a system to control the quantum devices. The quantum chips themselves are actually a bit of dead material. They're like registers, pieces on which you can carry information. But to actually do that, you need a control system, and that's what Qblox builds.

So we build the system that translates user algorithms. It translates user ideas into the electrical signals that go into the quantum chip. And then, after the algorithm is done, the result is fed back to our control system. So we are sort of the interface layer between the user and the quantum chip, between the classical and the quantum world.

Yuval: Is this aimed primarily at researchers in academic labs, or do you envision manufacturers of superconducting qubits saying, "Hey, we're going to use Qblox in development or even in production"?

Niels: I think that's a very good question, and the answer lies in both, and in the fact that it's in a transition.

The whole quantum technology field is really at an early stage. At the moment, I would say the majority of research and development is happening within academic settings. So right now, about half of our customers, or a bit more than half, are either universities or governmental R&D institutes.

But there is a very important pickup happening right now within industry and the private sector. That means big corporates like Google, IBM, and Intel, but also big startup companies like Rigetti or Pasqal, that are now building these quantum computers in a more corporate or industrial setting.

So our market is gradually shifting from the more academic side toward industry. That's the transition that's happening right now, or one of the transitions.

Yuval: Is the product geared only toward superconducting qubits, or does it support other modalities as well?

Niels: Yes, that's a good question. Our products are quite widely applicable to a plurality of quantum chips, and I think that's also one of the nice opportunities. Instead of having to do everything as a vertically integrated company, trying to do the quantum processors and the electronics and the software on top of it, there's now the opportunity to be a horizontal company, and that is more what Qblox is.

So instead of doing everything for one specific application, we can do this for a much larger market and a much larger set of applications. We support superconducting circuits, quantum dots, trapped ions, and even some photonic systems, which are a bit more exotic but still, in many ways, share a lot of the difficulties that other quantum platforms have.

So that's indeed the case.

Yuval: When I looked at your product a few weeks ago, I saw that it says it scales up to hundreds of qubits. Now, what happens when customers need thousands or tens of thousands of qubits for the ultimate utility?

Niels: Yes. Scalability, and empowering our customers to scale up, is one of the core parts of our company. Indeed, at the moment, people are mostly working with a handful or tens of quantum bits, so the majority of the market needs solutions in that range. Some very exotic customers need 100-qubit systems, or even more than that.

But those examples are very rare. So at the moment, we provide systems that support experiments up to hundreds of qubits for the biggest systems. Indeed, in the coming years, this will grow further, and we are, of course, very actively engineering toward these larger systems.

But at the moment, it's not really yet necessary or desirable to have such large systems, and the biggest reason for that lies not on the quantity axis, but more on the quality axis. To make these bigger quantum systems useful, the fidelity of the operations, the quality of the operations, needs to go up.

Right now, the qubits are not yet good enough to actually make use of large systems. So that's another thing that needs to happen. And that's also a very important area where we distinguish ourselves: providing very low-noise electronics so that the qubits are not disturbed.

And we also provide many tools to fine-tune the qubits and their operations. That's what we're doing as well. It's to empower the systems to be scaled in a meaningful way. That's the goal.

Yuval: Let's assume I have these low-noise electronics. Obviously, the performance of the system also depends on the particular timing and sequence of the pulses, and so on.

One could think about error correction algorithms or noise reduction. Do you also provide the firmware, or the software, that drives and achieves better fidelity, or do you leave that entirely to your customers?

Niels: I guess the answer is a bit of both. To really understand how the software stack works: there are quite a few layers in the software, and layers in the firmware, or even what you can call hardware or gateware, that need to be addressed to make it a useful system.

Where we are good, and where we define our role, is in the lower layers of this stack. Anything to do with tuning up individual pulses and gates, executing pulses, having high-quality signals, being able to tune in a very fast way, that's where our real core competence lies.

But higher up in the stack, there are also very important parts that our partners are working on. If you think of making very good Hamiltonian models of the pulses and the chips, to really understand the interaction at the physics level, that is more where our partners come into play.

So we especially like working with other suppliers, other parts of the stack, for that. That's how we see it.

Yuval: And how about calibration? I think quantum computers sometimes run calibration routines every now and then to measure certain parameters that then impact the control signals, crosstalk between qubits, and so on.

How do you address those?

Niels: Yeah, indeed. When it comes to crosstalk in the chips, it is very important to have a very good understanding of the physics of your chip. And I think the task of understanding the physics lies mostly with our customers, or with other partners that work on this more software-oriented layer.

But being able to execute the protocols in a very fast and efficient way, that is where we see our part of the problem. That's where we are always benchmarking ourselves, for example on specific protocols to measure crosstalk and then be able to correct it.

How fast can we actually do this? How well can we parallelize the execution of these protocols? That's where our solution lies in these problems.

Yuval: Tell me, please, a little bit about the company. How large are you? What kind of funding have you raised, if any? What are your plans for the coming 12 to 18 months?

Niels: Yeah. Qblox is a bit of an odd company in this whole field. That's because Qblox does not rely on any venture capital funding. We have been able to grow the company in about three years to a team of now a bit over 50 people, in a very lean way, and really aimed at doing business very early on.

That has allowed us to scale in a very fast way, but still mostly on our own legs. That is the situation so far. But for us, it's key to keep growing at this very fast pace of roughly doubling to tripling in size each year. We hope to maintain that, still on our own legs, in the coming years. But if not, we will look for other ways to finance and grow the company, because it is key to keep pushing the boundaries when it comes to how good our solutions are, how much of an edge they have on the competition, and how well they can solve the problems for our customers.

That's the most important part.

Yuval: As far as European quantum computing companies go, there are plenty of companies in France, a lot in the UK, and a few in Germany. I'm less familiar with the Netherlands quantum ecosystem. Is there anything you can tell me about the Netherlands quantum ecosystem?

Are you collaborating with other local companies? Is there government support, or did you just happen to start the company there?

Niels: It actually surprises me a little bit that the Dutch ecosystem is new to you, because I think it's quite well known as one of the vibrant places in Europe in particular, but also in the world.

There has been a leading institute called QuTech, partly TU Delft and partly the governmental institute TNO, for more than 10 years already, with multiple hundreds of scientists working there. And over the last five years, more than eight companies have spun out of it.

As for the network with other companies, I'm sitting right here in a building that is filled with different quantum startups. It's really aimed at quantum startups. I think the ecosystem, and the way of working together between companies in Delft but also in the wider Dutch ecosystem, is very good and also quite well known.

In this building, there is a company called QuantWare that develops and provides superconducting quantum chips. They are actually one of the few companies that can supply you with a quantum processor. Then there are also companies working on software layers, like Orange Quantum Systems.

And there is a company working on cryogenic infrastructure called Delft Circuits. These are just a few examples of the startups. Other than that, there are quite big corporates like Microsoft and Intel who have on-campus activities in Delft, really employing large R&D groups.

So this whole ecosystem of governmental institutes, big corporates, and startup companies is a very fruitful ecosystem, where they really strengthen each other. I think it's a very good place for us to be, and there's a lot of growth in it.

This is the first quantum building, but there are already two new buildings on campus that are entirely dedicated to quantum technology. One is going to be for Qblox itself, and the other will be a new, more ecosystem-oriented building. So that's, in a nutshell, the story about Delft and also about the Netherlands.

Yuval: As we get close to the end of our conversation today, I wanted to ask you a hypothetical question. If you could have dinner with any of the quantum greats, dead or alive, who would that person be?

Niels: Oof, that's a very good question. I think, well, quantum or scientist, it would be very interesting to have dinner with Marie Curie, for being a very early innovator.

I would like to hear her thoughts on quantum chemistry and quantum applications. But also more recent people, like Chad Rigetti. I was actually at a dinner with him a week ago. He's one of the early entrepreneurs in this field. I'd like to hear how they see the field, how they think it has evolved, having been there as an entrepreneur early on, and in which direction it should go.

So that's just off the top of my head.

Yuval: Excellent. How can people get in touch with you to learn more about your work?

Niels: Yeah, they can contact me, of course. Please head to our website, qblox.com, and you will find the right contact details to get in touch, whether it's for recruitment, since we're looking for many new people as well.

We have many vacancies. Or whether it's to use our systems as a customer, or as a partner, we are always open to collaborations. That's what we're here for. So, yes.

Yuval: Excellent. Niels, thank you so much for joining me today.

Niels: Thank you, Yuval. It has been great to be here.