Podcast with Himadri Majumdar, CEO and co-founder of SemiQon, and Janne Lehtinen, CSO

Transcript

Yuval Boger: Hello Himadri, hello Janne. Thank you so much for joining me today.

Himadri Majumdar: Thank you for having us.

Janne Lehtinen: Thank you.

Yuval: So who are you and what do you do?

Himadri: My name is Himadri, Himadri Majumdar, and I’m the chief executive of an up and coming spinoff from VTT called SemiQon, and today we’ll be talking about that. Before I joined SemiQon, I was the Program Manager for quantum technologies at VTT, which is a technical research center in Finland. And in that role, I led various national and international initiatives in quantum tech for the last couple of years, but now full-time in SemiQon, leading the company.

Janne: I’m Janne Lehtinen, Chief Science Officer of SemiQon. I was a research team leader at VTT before joining SemiQon. I’ve a background of roughly 10 years of superconducting quantum technology and lately 5 years of semiconductors. Well, you see where I’m now, so where I’m thinking that the future will be – semiconductors.

Yuval: Excellent. So what modality does SemiQon use?

Janne: We are using silicon spin qubits. Our target is to do hole spin qubits that operate at elevated temperatures. Elevated here means a little bit above 1 kelvin. So it still requires cryostat, but the power budget of the chip that’s available for our qubit systems is significantly higher than in the current systems.

Yuval: And what is the advantage of spin qubits over superconductors or trapped irons or other quantum modalities?

Janne: The advantages are several so let’s start with the first one, and that’s just the physical size of the qubits. So currently if you compare, for example, with the superconductors, you have something which is roughly 200 micron sized qubit. And for semiconductors you can go by almost a thousand factor smaller, so something like 100 to 200 nanometer per qubit in size. And this, of course, enables scalability in a sense that you can pack more of these qubits on your chip. 

Then the second thing is that we have really well established fabrication technology. So, all our traditional hardware that’s based on silicon processing, and now we can leverage this to make high quality components for silicon quantum processors. 

And then, what we already touched a bit, was the operation temperature. The silicon qubits can be operated up to temperature of several kelvins, which then enables monolithic integration with readouts. So we can put control hardware in addition to the chip, and this relaxes a lot of the infrastructure which you need for your QPU. In the end, qubits actually make a QPU that can just have digital in- digital out kind of operation. It functions as real processor, not just qubits with the classical processing hardware outside.

Yuval: Right. So we are talking about quantum IC here, quantum integrated circuits so to speak. Aren’t there other companies that are using spin qubits?

Himadri: Yes, of course. We are not the only company doing it, otherwise you can ask the question, why are you the only company doing it? No, there are of course competition. And the biggest and the most recognized would be Intel trying to do it based on their Horse ridge processors and so on. And then of course there are competition in UK, Australia as well. But how we see it is, we don’t necessarily see them as competition. We see it as potential collaborators because our business model as a company is not to become a full-stack quantum computer maker. We want to focus on the processors and build the best processors that there are.

So if all goes well, we might hold the best processors that there can be, based on semiconducting spin qubits, and the other companies can potentially become our customers or partners as well. What we want to see is, we are the best in the processors and then others who are doing the system integration, they are good at that. So we don’t want to do their job for them. We want to build partnerships, we want to build collaboration and then address the million qubit era, hopefully in the future building the best quantum computers there can be, with our processors. That’s our big goal.

Yuval: I think you mentioned that you’re a spinoff from VTT, or I believe that VTT, and perhaps you while you were working at VTT, made some exciting demonstrations of spin qubits. Could you give us the highlights of what these demonstrations were? What were the achievements that you were able to show?

Janne: Yes. What we did at VTT was establish a fabrication platform for this quantum-dot based device. And we demonstrated that we can have several of these quantum dot arrays embedded in a cryo multiplexers. We monolithichally integrated the quantum dots into a classical CMOS chip.  So that showcases the possibility of extending these types of systems to large scale processor units. And now what we are going to do is to operate these quantum dots as qubits. The quantum dots are pre-stage of the qubit and that’s the current technology level, but what we are aiming is to be able to do solutions at scale and that’s why we take these step-by-step. So first we need to be able to do the best quantum dots and then we can make the best qubits.

Himadri: And we have already demonstrated one of the best possible low charge noise devices.

Janne: Yes, that is there. In terms of quantum dots the charge noises which we measured have been always at elevated temperature since that’s the temperature where we want to operate our qubit in. So the target is to be there and still have full systems.

Yuval: Building quantum hardware takes time. I’m guessing how long you anticipate before you have something to show the market from Semiqon, from the new company

Himadri: Janne can answer from the technical aspect but I can say that what our investors would like to see, and is one of the things that we actually are hoping we are very good at, is to have a very fast design and fabrication cycle, which allows us to do the development quicker and get to the next generation faster and faster. So from our perspective, what we want to do is in a two year timeframe, go beyond the state of the art that exists right now and then have different milestones in terms of tens and hundreds going forward as we continue the journey. Janne, if you want to add some more details to that.

Janne: Yeah, I think that that’s the key message. So what we are doing is to compress the fabrication cycle in two months that typically take, let’s say up to nine months or so. And then we can iterate fast and develop fast these processors and of course the concrete demonstrators that we are aiming, those will be similar to what’s already done. But we do it in a way that there are no hero devices. We do it on wafer scale and characterize the devices on wafer scale to see that we have good yields and we can actually produce the processors on quantities and yield that are sufficient.

Himadri: And in that respect, one thing that your audience might be interested in is that most of the spin qubit devices based on CMOS are very much dependent on the fab. All the companies who are trying it either are doing it using existing large fabs or doing it in university fabs. But what we have as an advantage is we are operating out of this infrastructure or clean room, which is pilot line based here at VTT. We have full access and a full operational capacity in that clean room. That allows us to not just do early level prototype and devices, but go to the pilot scale already by ourselves. So that’s a huge advantage in terms of delivering the targets quicker as well.

Yuval: Does that mean you anticipate that the pilots or the initial products will be manufactured at VTT and only later when you go to even larger scale you will engage an external fab

Himadri: In a way just to correct that, so the fab that we are talking about is accessible to users like us. So we will be the one doing production in the fab even though the fab is owned by VTT and the universities. We will be the operating company in the fab. And yes, then you are correct that we will be doing the pilot manufacturing and the first generation of products in that fab, and then when it scales and the volume goes higher, we will have tech transfer and partnerships accordingly.

Yuval: Some hardware companies are building application-specific computers. They say well we can design the chip to fit a particular type of optimization problem or maybe we do something for machine learning and so on. Do you envision your machine to be general-purpose universal quantum or do you envision it to be tailored to a specific type of application?

Janne: Well, the target is general purpose computer, but of course there are specific things which we would address with the spin qubits better. So it’s a really broad question to ask, but the target is on the general purpose computer side..

Himadri: And we can also quite easily say, being on the early stage of the company, we are quite open to collaboration. So within your audience, if there are university and research groups who want to try out a certain kind of device, then we are open to those kind of suggestions and options. But like Janne said, our target is for general purpose computing, but we also know we will have different milestones going forward in the early stages, for sure.

Yuval: If you can share how large in terms of number of qubits do you expect the initial device to be? How fast is the cycle time or the clock rate? Any specs you can share on your initial goals?

Janne: The size will be two to six qubits at first, since the state of the art is currently six, so that’s already there. It will be really fast to reach that level. But then of course we cannot much say about progress beyond that. So getting to some microsecond coherence times for the first year in those qubits.

Yuval: How large is the company and how large do you expect it to become in the next couple of years?

Himadri: Based on this strategy of being a very core quantum processor company, we started with the small team of core founders and employees. So there will be five of us to start with, and then in the next couple of years we will most likely grow by one or two employees more, depending on the strategic employees that we will hire. And further beyond that, we won’t be growing exponentially because we know we have a task to fulfill in terms of developing our designs and doing the demonstration of the first tens of qubits in a processor. But after that, depending on how we grow in terms of our revenue and how we grow in terms of our client demand, we have a scope to grow. We don’t intend to grow unless there is a very clear commercial path ahead of us as well. We are very mindful of our burn rate, as it’s called.

Yuval: That sounds like an incredibly capital efficient company in terms of the number of employees, how are you funded?

Himadri: We have started operations and, by the time when your listeners will be hearing this podcast, we will be announcing that we have some early stage pre-seed investment from our investors, private investors and some strategic angel investors. That would be the starting point for us. But we are of course going to have a lot of public private partnership projects based out of public funds in Europe as well where we will be participating and this is how we are going to grow. As I said in the beginning, our goal is to grow in a collaborative manner. So we will work together with other stakeholders in various public funded projects and concentrate on our product development and then grow from there. That’s the goal.

Yuval: What can you share with me about the quantum ecosystem in Finland? Why is Finland a unique place to build a quantum computing company?

Himadri: Yeah, it goes to the core of our existence basically. Finland has been very strong, and you can see it from the companies like Bluefors and  IQM for example, from the fundamental science for decades. And this has led to those companies, Bluefors  and IQM to be successful right now. And it’s the same with the technology that Janne was talking about that we have developed at VTT. It all boils down to years of research and also critical infrastructures, having access to critical infrastructure. So all these factors come together to build this scope for us to be in the right place at the right time. And this allows us to also leverage the ecosystem. Having Bluefors close by, providing the best cryogenic solutions in the world, having IQM who have now developed and built a business and having sort of partnerships with them and discussions with them is important. Also companies like Quantrolox, for example, who are providing specific solutions and can potentially be one of our close partners. All these ecosystem players actually help us and we hopefully will be helping them as well going forward. But it’s very, very important that the research ecosystem, the education ecosystem, and the infrastructure is here in Finland to allow us to do those things.

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

Himadri: Oh, great question. I let Janne answer first while I get some time to think.

Janne: I guess it would be nice to meet (Erwin) Schrodinger if it would be possible.

Yuval: And why so why Schrodinger?

Janne: Well that’s an even better question. He’s the person who I’ve had to deal with most during my studies, I guess.

Himadri: I would actually love to meet Professor Feynman, Richard Feynman. He’s one of the people who motivated me to study physics actually. And, who knows, if I meet him I would even find some solutions for making better spin qubits, knowing how great he is.

Yuval: Excellent. So I wish you lots of luck in your journey and thank you so much for joining me today.

Himadri: Thank you, Yuval

Janne: Thank you.

Yuval Boger is the chief marketing officer for QuEra, a leader in neutral atom quantum computers. Known as the “Superposition Guy” as well as the original “Qubit Guy,” he can be reached on LinkedIn or at this email.