The Quantum Daily interview
Sal Bosman’s CV is eclectic. With over ten years’ experience in high-tech entrepreneurship, his career has also taken in roles as designer, physics teacher, researcher and systems architect. When Sal launched FLO Solutions back in 2006, it was one of the very first Internet of Things companies in the Netherlands (“It completely failed”, he freely admits on his LinkedIn page), but little did he know that he would find himself leading one of the most exciting start-ups in the quantum computing industry a decade later.
Quantum computing itself was first proposed as a concept in the 1980s, but since then it has mostly been an almost esoteric curiosity, of interest only to theoretical physicists. Now, however, it is finally entering the mainstream, with IBM, Google, Intel, and many other tech giants wanting a piece of the action. As Founder and CEO of Delft Circuits, Sal Bosman has plans to play a vital role in making this transition from dream to reality.
Most experts agree that quantum computers will one day outperform even the most powerful traditional computers, but the biggest obstacles to this are, and always have been, practical ones. Giant fridges, superconducting cables, and rare gases are all requirements for the cooling of the quantum device – and that’s before you start thinking about the chips used in a quantum processor. None of these are standard components, many of them are brand new and made to order, and therefore lead times are prohibitively long. Research and Development (R&D) in the quantum computing arena frequently comes up against these pinch points – and this is where suppliers like Delft Circuits come in. Sal’s ambition for the company is to radically reduce these lead times: “If you need a cable today or tomorrow then we are able to provide it. If there’s a problem, we quickly solve it.”
When Delft Circuits pulled out of a potentially lucrative venture capital deal in 2019, many in the industry were left scratching their heads. But, as Sal explains, he’s not in it for the short term. “We want to have control over the company now, but also in six years’ time”, he told us. And one can understand why – Sal’s passion for tech in general, and quantum computing in particular, is clear to see, and he brings this passion to the development of key hardware and components with his small but dynamic team of 20 people. Sal’s enthusiasm is infectious, and it seems he was born to work in the field; “I always imagine reality like a bunch of switches,” he said during our conversation, just one clue that his is a mind that works in a truly unique way.
In this interview, Sal talks about the company’s plans and what he thinks the future holds for Quantum Computing…
The Quantum Daily (TQD): We’ve been aware of Delft Circuits for quite some time, but for those who are less familiar with your work, could you briefly introduce the company and what it is you do?
Sal: Sure. Delft Circuits started in 2016. We are active in the quantum industry, mainly quantum computing and the quantum Internet, but also in applications such as astronomy. Our first product line is focused on cabling, or what we call input-output systems. This product is called Cri/oFlex.
TQD: First of all, perhaps you can explain to us what quantum computing is. For the general audience, let’s say, for people who don’t know anything about it.
Sal: I’ll certainly try. We still don’t really understand quantum mechanics very well. We understand it on the mathematical level, we can predict everything very clearly, but the intuitive picture of what is really happening is still very vague and really not very well understood even now. What we do understand is that nature itself has a fine grain or pixelated quality to it. If you look around, the world looks infinitely precise – just as the screen of your phone also looks like a perfectly continuous thing – but if you go in with a microscope then you start to see pixels. I always imagine reality like a bunch of ‘switches.’
Usually in computing you have resistors or transistors – also like switches, and they are either switched to ‘0’ or ‘1’. We call these ‘bits’. Quantum computing, however, uses fundamental excitations in place of classical bits. A quantum bit, or ‘qubit’, can be in a ‘superposition’ – in other words it can be in multiple combinations of 0 and 1 simultaneously, and this ‘uncertainty’ effectively means it can run through multiple variables in a calculation, instantly.
One way to think of this is as the flip of a coin. With a classical bit, you get heads or tails. In a quantum bit, it is like a coin spinning on the table instead of having a definite position. Quantum computing also uses a strange phenomenon called ‘entanglement’. This is a relationship which causes a pair of qubits to affect each other in a predictable way – even when separated by distance – and increases those positional combinations exponentially.
TQD: What sort of advantages does this give, and what are some possible applications?
Sal: Well, we can do reasonably powerful computation using classical bits. But, it’s quite easy to write down a calculation that is basically impossible to compute on today’s computers. For example, let’s say Airbus wants to do some very complex simulations for aircraft. The simulation will always have a limited amount of points, where your memory is limited. Or, to take another example, imagine you’re a multinational chemical company like Dow chemical or BASF or Bayer. They have, let’s say, 30 plants, 200 products, five ingredients and they have to optimize the logistics – these are ‘traveling salesman’ type problems, but the calculations typically bloom; they become exponentially difficult.
Quantum computing in the long run has the ability to solve those types of problems. One can use quantized behavior, superpositions and entanglement, to leverage the computing power of your system – either the memory, or the processing power. Programmed well, it could basically do those exponential calculations instantly.
TQD: So, in layman’s terms, quantum computing is basically a new type of high-performance computing?
Sal: Exactly. In the 80s in the 90s, we used deep processors – an Intel processor is a classic example – and the computation eventually came to normal computers. Then we had the revolution based on the graphical GPUs – all the bitcoin mining is done in that way. Now of course we have machine learning and AI, and currently in this field they are designing specific hardware chips to solve specific problems, in chemistry or biology for example. Quantum computing is in the same league of high-performance computing revolutions – it is still an immature technology, but it could have the biggest impact.
TQD: Where does Delft Circuits come in? How did the company come about?
Sal: First and foremost, I’m a designer. I became a physicist later and then started working on quantum computing. We wanted to start a company in this field, but at the time, all the companies in the quantum sector were building quantum computers. We decided to position ourselves as a hardware supplier for quantum engineers and the quantum industry, where we would solve specific problems – we are very good at recognizing specific issues that we know will be a problem, maybe not now, but in the future. The first things that we’ve focused on are cabling, circuits and electronics components.
TQD: So what problems are you looking at currently, and how will your technology help to solve them?
Sal: Okay, I talked about quantum behavior before, but it’s actually not very apparent – except for in optics, when it is still very quantized. But the point is that such behavior only gets obvious when it’s very cold. If you’re at high temperatures or even room temperature, there is just too much movement. So, when you have quantum technology it has to be cold – really cold, we’re talking about one of the coolest places in the universe – so quantum computers basically have to exist in enormous refrigerators.
In these massive fridges there is a problem with cabling. First of all, the cabling is very expensive, but it’s also very big and bulky, and quite often very fragile. But the biggest problem actually, in cabling, is not conducting the signal but rather avoiding the transfer of a lot of heat. These fridges are very inefficient – it’s just because of physics, a big fridge for a quantum computer can use 30 kW of power, but at the exact point where the quantum processor sits you only have tens of micro-W of cooling power. If you compare it to a financial investment, it’s like investing one trillion dollars and getting a single dollar in return. That is why you want valid technologies to prevent or avoid every single bit of heat that seeps into the cold stage.
Another big and important difference between quantum computing and classical computing is the fact that one cannot copy quantum information. You can move it around or let it interact with other information – but you cannot copy it. I’m not sure if it’s been proven yet, but as far as I know, in all architectures of quantum computing, each quantum bit needs a separate control line – we call it the input-output line. With a modern Intel processor, you can control nine billion transistors with just a thousand 24-pins on the outside of the chip. For quantum computing this is impossible, I believe, because you cannot copy quantum information. As you can imagine, this means the cabling problem becomes even bigger.
TQD: Let’s talk about the business. In the middle of 2019 Delft Circuits walked away from a big VC (venture capital) deal. This was huge news in the industry, so I have to ask – do you have any regrets about this and what has happened since?
Sal: I certainly don’t regret it. At many start-ups that work with venture capital, the founders have the goal of working on the company for five or six years and then selling it to one big corporate or another. We started Delft Circuits because we couldn’t find a job that we liked. On the other hand, we are not really scientists in academia – we love high tech. We love design, engineering and entrepreneurship.
For us, independence is really important. As the quantum industry is still in its infancy, we want to have control over the company now, but also in six years’ time. We don’t want to suddenly lose control of our company. This was a confusing issue during the negotiation – I don’t want to go into too much detail but the contract that was provided to us was unclear. Although the investors would have a minority stake, we would lose control, and the VC would be able to sell one hundred percent of the shares of the company. This, for us, was a no-go.
Over the past five years we have never really had a funding problem; we are customer- funded. We landed some good projects, we secured some loans and so on. Now, we are in fact working on another venture capital deal, but these are different parties, different VCs.
TQD: So you’re still interested in working with investors in the future?
Sal: Yes, of course. On one hand we want to stay very focused on sales and making revenues today. Last year we had a couple of weeks with multiple sales, and we were cashflow positive. We are very focused on getting products to the market. We sell worldwide, to China, USA and Europe, to big corporates, large start-ups, universities and national labs.
On the other hand, we do need a lot of machinery. We have a team of 20 people, and the industry is growing. At the moment we are looking at an investment consortium – but we only want investors who can bring something to the table, like contacts and experience, and who are very long-term focused.
TQD: In the meantime, how do you finance your R&D? I understand you acquired public funds…
Sal: Yes. We actually got a big project grant from the European Union which covered our R&D expenses. It’s for small and medium enterprises that are highly innovative and work in the international arena on cutting edge technologies.
TQD: Delft Circuits appears to occupy a unique position in the supply chain for hardware for quantum computers. What else can we expect from Delft Circuits in the coming years – can you share your plans with us?
Sal: I can share parts, but I’m not going to give everything away. Steve Jobs said famously, “you should never talk about what you’re going to do,” so instead, I’ll tell you what we are currently doing. We want to be able to supply the industry with new technology that allows bigger and bigger processors, and right now, we are working on multichannel cables. If the number of qubits increases rapidly, then you cannot simply make your fridge bigger – it doesn’t work like that. What the quantum industry can expect from us is an increase of the channel capacity within the current thermal budget. Our cables are much smaller, so their footprint is way smaller, and we can get many, many more lines or channels in today’s cryogenic fridges.
What’s important for us is short lead-time, great customer service, and ease-of-use of our product. Originally I was an industrial designer so I know how difficult it is to work in a fridge. It’s one thing having electronics, and it’s another knowing where your electronics are and being able to access them – especially with cables, where you often have problems. We are very much focused on making the life of the quantum engineer easier by having excellent design, excellent performance, and good prices. If you need a cable today or tomorrow then we are able to provide it. If there’s a problem, we quickly solve it. Well, this is our goal at least. We are still a growing company and we are not perfect, that’s for sure. But, great products and great service; this is what people should expect from us.
TQD: You’ve worked in the quantum computing industry for nearly a decade, initially in research and now in hardware production. It strikes me that you are ideally placed to talk about strategy for the industry as a whole. What do you think needs to be done by policymakers in terms of turning quantum computing into a reality? After all, those are the people who make decisions and organize budgets…
Sal: First of all, I think the policymakers should continue funding, but think much more strategically about how to fund. There are different types of players: there are the academic groups, the national laboratories, and then we have industry – industry consists of the big tech giants, and smaller start-ups or suppliers. These four types of players make up the quantum industry. Policymakers should be very aware of how you shape the ecosystem and how you leverage the capabilities of each player.
TQD: For example…?
Sal: Okay, today there are academic groups who want to build quantum computers, but they are not really suited to do that. I would encourage the academic groups to do the pioneering work – let them approach how qubits work, or how quantum computers are made or how qubits can operate. There are drawbacks to just funding academic research – it can make the professors rich and starve the industry, and I’m not sure the ecosystem would develop in a very healthy way.
The large tech giants themselves are traditionally not that great at innovation, which is why they usually acquire successful start-ups. However, in our field they play currently a driving role. Then we have the supplier group, which I think is quite interesting, obviously. These are the fridge builders and electronics suppliers – like us, making cryogenic components. You also have quite a few start-ups focusing on quantum circuits themselves, which is very interesting; we will have to see what comes out of this. I will say this is one of the most dynamic, non- conservative groups out there, and I expect there will be a lot of dark horses with surprisingly good results.
Finally you have those national laboratories (some are our clients), for example CERN in Europe, LLNL in USA, CNRS in France, Fraunhofer in Germany, TNO in the Netherlands, IMEC in Belgium – IMEC has been the cockpit of the semiconductor industry for forty years. Personally, I think policymakers should put these national laboratories in the driving seat, because they have more leverage and a stronger punch. They are a little bit more organized and have the ability to keep employees longer – and you need that permanent staff for massive projects. They can also run and coordinate co-development projects, and they are major initial customers for suppliers.
TQD: Is there anything else people should be aware of in terms of the way the industry is developing?
Sal: Well, the quantum industry is developing in such a direction that is beginning to create a value chain. In another industry for example, we all know Apple makes the iPhone, but there are thousands of components in an iPhone, and the components are sourced from many companies. One company I really like is called SkyWorks, based in New England, which produces the microphones and speakers for the iPhone along with many more components. Last time I checked, this company was worth about $20 billion. This is the sort of company that makes up the ecosystem of the telecom industry.
Eventually, the quantum industry will develop in a similar direction. One single company could never build a full quantum computer: it needs algorithm developers, software developers, electronics guys, cleanroom people, cryogenics people. These are not just very different areas of expertise, but also very different types of technologies that require a different type of character and a different type of culture. No industry has all that under one umbrella – it is usual in the high-tech industry for companies to diversify and specialize.
TQD: Thanks so much for speaking to us today. Before we leave you, I’d like to return briefly to the future of quantum computing. What progress is being made, how do you envision the technology being used in the future?
Sal: As I said, the technology is still premature, but in the near term, we expect that much of the progress will be in material science. For example, superconductivity, especially high temperature superconductivity, is still computationally very difficult to calculate numerically using a normal computer. However, in the quantum system that is now available with 50-60 qubits, we can already beat todays computers for super specific tasks. Who knows what we could do with the kilo-qubits system. Probably, we can simulate what is happening in the superconductors. Going forward, those kinds of applications are likely to result in a lot of interesting and new results in material science.
TQD: Looking further into the future, do you see quantum computing changing our daily lives?
Sal: Almost certainly in some way – but in what way is an almost impossible question to answer. I can only give you a historical analogy. When humanity invented the printing press, the only thing we could imagine printing was the Bible. When William of Orange started to conquer Britain, they brought two printing presses and it was used as a strategic weapon in propaganda. A few years later, one of the most printed documents were erotic short stories!
A similar thing happened with the normal computer. When we built the first electronics-based computer, the first calculations were done for nuclear reactions and atomic weapon research, and it was very difficult to think beyond this. The CEO of IBM, Thomas John Watson, was once asked about the world market for mainframe computers, and he said he believed that something like five computers would serve the whole world. He was only thinking of certain high-performance computing problems, which at that time were seen as useful. However, when Lotus started developing spreadsheets, computers suddenly came into the corporate community. In the past, if you wanted to change one variable in a business plan then the secretary had to go away and calculate for a whole day, by hand or by using a simple calculator. Suddenly people started to realize, wow! I can make a spreadsheet and I can do the calculations in real time. From there on you start to get into gaming – and now of course we have social media.
I would say the same can happen with quantum computing. Right now, we are thinking about quantum materials, chemistry, cryptography… In the future, quantum computation could be useful for many industries. For example, Hollywood accounts for a big fraction of the high-performance computing in the world, making calculations for movies using very powerful computers, so there is a clear need for greater processing power there.