Skip to content

Analyzing the Quantum Hardware Landscape

Overview

In his role as Chief Analyst at Global Quantum Intelligence, David Shaw combines a PhD in physics with years of advising executives in tech, healthcare, energy, and financial services. This understanding of quantum mechanics and the technology life cycle gives him a unique perspective on quantum hardware. In this episode, he explains how many “chips” it will take to build a quantum computer, how far we are from modularity, and why quantum sensing should get more attention.

 

Transcript

Veronica: Hello, and welcome to The Quantum Spin by HKA. I’m Veronica Combs. I’m a writer and an editor here at the agency. I get to talk every day with really smart people working on really fascinating subjects, everything in the Quantum industry, from hardware to software. On our podcast, we focus in on quantum communication, and by that I don’t mean making networks safe from hacking or entangling photons over long distances, but talking about the technology.

How do you explain these complicated concepts to people who don’t have a background in science and engineering but want to understand all the same?

Today, I’m talking to David Shaw, who is the Chief Analyst at Global Quantum Intelligence. David has, I think, the perfect background for talking about quantum technology. He has a PhD in particle physics, but he’s been working in consulting, market analysis, and advisory roles for the last 20 years in healthcare, energy, technology. So thanks for joining us today, David. 

David: Unfortunately, it’s actually more than 20 years. I don’t know if I should be admitting my age at this point. 

Veronica: No, I think, the broader perspective one can have on quantum, the better because it really shows that we’re all in this for the long haul.

 So what is Global Quantum Intelligence? And what do you do there? 

David: Yeah. So there’s a clue in the name. We take a global approach to the sector. We’re all about quantum and quantum technology in its broadest sense and intelligence, and hopefully people will find our input intelligent.

The group came together: myself, André König, Doug Finke. We’d all been working separately in this area for a whole number of years now, Doug, I think  right the way back to, I think, 2015. And, we realized we were doing things that ultimately intersected. So 18 months ago now,  we brought the businesses together, and that’s been great because it’s given us that much enhanced reach. Uh, and also complimentary reach across, across the quantum sector. So we are, we’re focused on everything to do with that quantum technology market or the quantum technology markets, the analysis of those markets, advisory services around those markets.

Veronica: Yes. I remember when Doug made the announcement, was it a Q2B Silicon Valley maybe? But I remember thinking, Oh yeah, that totally makes sense to bring all of you together. 

David: So I think Doug’s one of those guys who’s been to every single one of those Q2B meetings and actually he started out before there was a Q2B meeting.

Veronica: Yes. I refer to him as the OG of quantum computing and I’m not sure he appreciates being called an original gangster, but he does know everything and everyone. And so, we rely on him for guidance and steering every now and then. So your role at the company is to assess the market and spot trends and tell clients what you think is coming or what they should be focusing on.

Is that a fair description? 

David: It is. And I think one of the things that’s a big part of what GQI does is we want to put that analysis in a context which brings, the, real data, real facts about the market, to elucidate what we’re saying. And we do that by building frameworks that help us make the rather disparate inputs that the different things that different players are doing with different perspectives, we deliberately create not just opinions and words about what we think of those players, but we create structures, frameworks that let us analyzesis the market in a way that makes it, more apples to apples and less apples to pears in terms of how one looks at progress and challenges across the sector.

Veronica: One of your recent reports, Outlook Report: Scaling Quantum Hardware, talks about modular computers, but I think we’re still not there, but at least that gives people something to hang on to, oh, that’s what we’re heading towards.

So do you want to tell us how you put that report together? 

David: It’s a good example of how these challenges play out because there’s so many different players out there in the market that are doing different things. And quite apart from some companies are in quantum computing, some companies are in quantum communication, some companies in quantum sensing.

But even within an area like quantum computing, there’s such a range of activities. Range of activities across the hardware software stack, but also range of activities across different time horizons of when they think different important things are going to happen that different companies are pursuing.

And so, on the one hand, the popular language that the market likes to talk about, it naturally evolves quite rapidly, partly for technical reasons, but also to be honest, often for marketing reasons, because companies like to be talking about their ideas in new ways because it’s exciting and it makes better marketing, it means you can do a press release but that can also get in the way of it.

Because it confuses the picture from other people trying to look into the sector. And we try to help by looking at that hardware software stack. And there’s some very unique things about it. You might come to it with an understanding of how the classical computing stack has emerged, but even there, there are challenges because people will have maybe learned their computing in the mid 90s, when we had a conventional desktop PC operating type operating system type stack.

But then the high performance computing, the HPC stack has already moved significantly beyond there in terms of how one thinks about it. And it’s there. That people think about where the quantum stack will have its application but equally, in reality, the quantum stack is back at that very early days still of the computing industry where you had a series of different computing hardware modalities and just the emerging bones of how those systems are processed and the quantum stack quite naturally has its own challenges because these quantum devices are distinctly different to, the challenges that it takes to put forward a digital device.

The quantum plane is a different focus in its own right as we seek to keep the qubits protected from noise in their environment. But then that gives us a challenge of how we’ll add the control plane on top of those devices to manipulate those qubits. And that’s something that we’ve forgotten about in conventional processing because those activities are just buried in the same chip.

Right. Yes. They’re not important. And then we have these other differences  in how we look at our quantum devices. Some people want to see how far can I push small-scale or intermediate-scale devices that are still noisy, so-called noisy intermediate scale quantum devices. And that has its own priorities of which parts of the stack I have to try to make work.

But other players are still pursuing that type of strategy, other players are saying no, we believe we have to focus onward onto error-corrected devices, which add additional layers to the stack. And so all of these different technologies also get mixed up with the different time horizons that different players are seeking to address.

And that’s one of the things that GQI’s frameworks seeks to bring some order to. Where in the stack are you talking about? So where are the challenges and what time horizon are you trying to address this technology into and therefore, what has to be got ready first before your bit really can work.

And that’s the sense in which GQI provides frameworks that help that synthesis. And, scalable quantum hardware is one of our, most recent reports that elucidates things in that challenge. And one of the things that it specifically seeks to bring to the fore is this notion of, when I was trying to see how far I can go and I still might want to see how far I can go with one individual device.

That’s great. But at some point I’ve run into this, well, how big can my individual device really get? And some of the large scale applications we want to run and that we think of a lot of promise for the value they’ll offer back from, from quantum technology. They need rather large scale systems.

And one of the paths that tends to force you down is: Modular scaling.  Can I take my  independent device modules and coherently network them together to give me that large-scale device. And that’s what scalable quantum hardware really examines as a challenge because it’s a new window to look at the challenges these different plays face.

Veronica: So I’m curious to get your context around a challenge I had the other day. A colleague was talking about quantum chips. How do quantum chips compare to GPUs or CPUs? And I said, well, and so I went immediately to modularity, right? Because like, you know, You can crank out GPUs and CPUs and, it’s just not, it’s the same idea, but we’re not in the same place on the timeline that you mentioned.

So I don’t really think about them as quantum chips. Do you think about them that way? 

David: This is a fascinating one, actually, because it’s tempting  to want to jump to saying, right, well, so what size quantum chip? But I actually, I think one of the, one of the insights you get when you look at what this really means for quantum computing system is that it’s not just one chip.

Literally, it’s not just one chip. It’s either the chip where superconducting circuits are fabricated to make superconducting circuit qubits, or it could be a chip which is really an ion trap.

But ultimately that’s a chip, a QCCD approach to iron trap is a chip, where I’ve got lots of electrodes that let me trap my ions. And so one chip is that qubit plane. But then, actually, to control those qubits, I’ve got to wire a lot of stuff in. Sometimes it’s literally wiring, sometimes it’s guiding laser light to control the qubits.

Other modalities have their own emphasis points, but actually that control plane is typically another chip. You have to think about, all right, so how am I gonna fabricate that? How am I gonna route the signals? How am I gonna bond those two chips together so that it doesn’t thermally overload the environment that the qubits are having to live in? So that it doesn’t introduce more crosstalk than I can handle?  It’s a significant issue because we’re so used to the digital world where all of those things can just neatly live on the same substrate; and the digital world where that digital technology means that I have the power of the digital 1 and 0 gives me an intrinsic error correction.

I can simply boot the circuit back to the 1 or 0 whenever I need to, to keep it trundling forwards. But I can’t do that with qubits because qubits, their superposition of 1 and 0, is intrinsically an analog system. So those control lines that I’m trying to bond between my control plane chip and my qubit chip, they’re analog signals.

 And typically analog signals. And that complicates, first of all, I have to these two chips and I’ve got to bond them to it. And then actually, it’s not just two chips. Because the other challenge that you walk into when you start to say, I’m going to scale these systems is, well, I’m going to do this thing called error correction, which sounds great when you wave your hands about it.

But actually, it’s mathematically very complicated. This point of how do I design error correcting codes and fault tolerant gate sets on top of those error correcting codes, and then how do I administer that in real time is a big challenge. And in particular, it needs an activity which is generally referred to as decoding.

And decoding is, reading the error syndromes that I detect, that the things that I’m reading out from, they’re happening down at the qubit plane, but I’m using that control plane to read them out. And I’ve got these syndromes that are telling me whether or not an error has occurred. And I have to run, in real time, this decoding task to decide how to correct the errors.

Now, the thing is, this is a really quite unusual processing challenge because it requires low latency. It requires a deterministic low latency. I have to be sure that my system will run and always deliver the result on time. And it has to be massively parallel because I’m doing this across a whole dispersed group of qubits in my system.

So the decoding task, many people would argue, is quite a unique processing challenge in its own right and will require its own specialist processing. So people that see this element of the industry, they say, well, actually, we really should be worrying about that decoding chip. So that’s the third chip that I’ve got to have in my stack, even before I start to talk about software bits. And physically, these are all separate chips. 

Veronica: Right, right. 

David: And that’s why it’s an additional challenge to start saying, so what is just the quantum chip that I’ve got to connect together? And actually, I’ve got to work out how do I bundle those three types of chips into some coherent module that can be scaled out.

Veronica: And, I can’t make one for you, you know, that’s the other part I think, right? Like, I can’t just put one together and you could take it back to your place and run all your computations. 

I noticed in your report about scalable quantum hardware, you mentioned roadmaps and it’s such a powerful communication tool to say, this is our plan and we might not get there, but this is our plan.

And is it just so intimidating to put times around this stuff? Because people seem so resistant to timelines. 

Why are roadmaps so scary?  

David: I think it’s interesting to frame that in the context of who’s the audience for the roadmap. 

And actually there are, I think there’s importantly two subtly distinct ones.

On the one hand, you’ve got governments who are trying to stimulate their ecosystems to build out quantum. And it’s about trying to make sure the development activity happens or that they at least have an important part of the value chain happening in their ecosystem and that their users in their wider industrial ecosystem are positioned to get the advantages. And there the road map, you can understand, it’s more about specific cycles of innovation being followed through in that wider ecosystem. And some things will fall by the wayside. Some things will be developed and be the basis for the next cycle of innovation.

I think that’s  important when you’re talking about a governmental program or wider aspiration to develop quantum technology. I think that mindset is important. 

Another big audience for roadmaps is, of course, investors and a lot of the companies and it’s different. I mean, a big tech company, it’s a bit different to a startup because on the one hand, it’s about a roadmap, but ultimately about how that division inside the large company contributes, at what point it starts to contribute to a variation of the wider enterprise.

But for a smaller startup, it’s more directly about communicating to the investors about this is how we are planning to progress. And some of those companies are, to be honest, they’re more when they have to be in the private conversation with investors, they’re more forthcoming about what number they’ll put on top of that stage because they have to be in and the investor is in the one hand thinking, What is it that I need to hold management to account for?

If you want to be investing in my next round. I want to see that you hit this thing you promised you were going to hit, and equally the investors are often thinking, and at what stage of development is that potentially an exit for myself? Because these investors, they often are specializing in a particular life cycle stage of a business that then, they are wanting to be able to pass on to other pools of capital at an appropriate point.

And I think whenever one talks about this notion of different eras or roadmap stages and dates, I think you have to understand that sometimes we’re listening to some government slides, sometimes we’re listening to a tech major, sometimes we’re listening to  a genuine startup, and it’s a bit different.

Veronica: Yes. Yes. That’s a good point. Who are you talking to and what do they need to hear? How much do they need to hear? So, we’ve talked about chips and hardware and that’s certainly making progress, but a long way to go until we get to truly modular systems. But one area that is moving along much more quickly is at least, I don’t know, some weeks it is and some weeks it isn’t, is quantum networking.

And I know that you were at an event at The Hague recently. And I really liked your LinkedIn post about quantum networks, get excited and innovate. And again, from a communications point of view, people get that right. Like a super secure internet, like no one can snoop on your traffic and the keys are always secure.

And so, what is your take on the state of quantum networks at the moment? 

David: This is a very intricate one in its own right, because I think one of the things that to realize whenever you hear someone talking about quantum networks and getting excited about quantum networks is really that there have been different communities within the inside wider quantum where this concept has meant something different to them.

For example, the traditional quantum communications community, they’ve tended to focus first on cyber security type applications. They wanted to talk about extending the range of, cyber security applications and headline ones that get discussed are QKD, quantum key distribution, also QRNG, but one can rise to the challenge of extending the range of QKD or enhancing the robustness of the security promise around it, and from that community, there’s often a desire to talk about the quantum internetas a way of motivating both that there’s some things we can do now, if we can extend the range, if we can get computers to help extend the range for entanglement, if we can introduce more advanced protocols from the quantum communications. 

Then there’s another community that’s always said, well, it’s going to be difficult to build one monolithic quantum device with one monolithic chip that’s doing all its stuff. So we’re going to have to go to modular scaling, and by modular scaling I mean modules with a quantum coherent interconnection between the modules. 

That’s a form of networking, that is quantum networking, and it’s formalized, in fact, in terms of one of the conceptual approaches to quantum computing is referred to as distributed quantum computing,  which could be a series of modules in the data center, but they’re performing their computation across the modules.

And that’s a different type of quantum networking. And those people don’t really tend to want to talk about the quantum internet so much because for them, it starts off at a lower scale. It’s about these short distance interconnects first, edge-to=edge coupling, local connections, maybe inside a cryostat, or short connections from fridge to fridge inside the data center. 

And that’s, now, one of the cross points here, of course, is that a lot of the underlying technology could have a role in that short distance networking, but it also could very much have a role in that longer distance, enabling that longer distance network and quantum repeaters, etc.In some sense, a quantum repeater, you can realize a quantum repeater to extend the range of a quantum network just by being a very small-scale quantum computer. 

So there’s a touch point in terms of some of the enabling technologies that join these worlds back up again. And it’s very exciting at the moment because this emerging latest generation of technology speaks to exactly that space. And so you’ve got suddenly what were two parallel communities actually having a lot of crossover again at the moment. 

And I speak about two communities, but actually there’s a third of course, because there’s the quantum sensing folk as well. And unfortunately people don’t speak to the quantum sensing folk as much as they should, which is a real shame ’cause there’s lots of great stuff happening in quantum sensing already. Right. And there’s some great connection points in the long run back to networked quantum sensors because if I can start to coherently network my quantum sensors, I get some very powerful enhancements about what they can achieve again.

Both in terms of their fundamental sensitivity and the scaling of that sensitivity, but also in terms of how machine learning of the outputs of quantum sensors can work, where we already have a very interesting demonstration from Google and Caltech, which showed that an exponential speed up by how quickly, how effectively I could, learn the outputs of coherently networked quantum sensors.

Veronica:  I do mention sensing when people are like, what’s happening now? Because I mean, militaries around the world are getting into that. And also when you say, well, sometimes GPS craps out and they say, what do you mean, I can’t rely on GPS?

I said, well, you know, that’s why people are thinking about quantum sensing. 

David: It’s a subtle point and I refer, actually refer people to another one of our reports, which is Quantum Sensing Outlook, because I think it’s one of the things that’s missed in the popular debate about quantum technologies is that this acceleration in quantum technologies started with military connected programs in the U.S. 

And it started off in clocks and sensing, timing and sensing, because it was DARPA related. Projects that really put atomic clocks on the map, enabling what we now have as GPS or GNSS, you know, satellite-enabled navigation. An important spinoffs from that work, with what got people looking at some of these platforms, could actually be used as generic qubit platforms, which have applications in computing and communications.

What I think we can sometimes miss, that some of these technologies, particularly some of the sensing technologies, are really established in the military already.

This new generation of quantum sensing it’s giving us a new generation of quantum sensors that we can integrate into a much more deployable platform.

Cerca Magnetics is an example of a company pursuing that and  that’s the interesting trade off that’s now going backwards and forwards between what’s, you know, almost all of these quantum technologies are dual use in some sense. Militaries across like minded countries are pursuing getting the most out of those technologies.

But we also want to make sure that we get the civil benefits of this technology as well. 

Veronica: Yes, I remember reading about a pilot program that the British government funded.

I think it was last year and one of the projects that got approved was  fluid dynamics and looking for ways to improve flood modeling. And I know that’s one of the most complicated things to model, the movement of water. But obviously with climate change, we need to know, we need better ways to predict flooding.

And so that was a very, that’s again, something people can understand, right? Like the weather’s kind of crazy now and I don’t want my house to flood. So how do we get better at predicting this? 

David: And I think it’s a useful example. I like that particular program because I think there can be a danger that one leads into this.

There’s a certain attitude that says, what are the real problems that we have in our business? And there is a bit of a danger in that because it’s not that we have quantum algorithms that give speed ups for everything.

It just isn’t like that. But you have to usefully ask. What are the algorithmic areas? What are the problem classes where we think we can get a strong benefit from quantum and where do we think they map into business use cases?

And that one’s particularly interesting because, there’s long been a sense and it’s almost, it was Feynman’s founding intuition about quantum computing that quantum computers should be good for simulating quantum systems. 

And that remains true, but there’s an interesting further add on to that, because Schrodinger’s equation is at the heart of how a quantum computer operates, but it’s a partial differential equation. Any system which relies ultimately on solving a system of partial differential equations, you can get that intuition that potentially a quantum computer is going to have applications there.

And that’s why some of these fluid dynamic simulations, an interesting example of where we’ve got another class of simulation problems where we expect quantum to have a very strong advantage.

Veronica: So I know that you are Global Quantum Intelligence and not the UK quantum intelligence, but I do like to highlight all the things going on around the world. And give each region of the world its due. Is there anything going on in the UK ecosystem right now that you’d like to call out or highlight? 

David: Well, I think, one of the things just to emphasize about that, you know, we pride ourselves in Global Quantum Intelligence about having that global perspective. Our job is not to argue for any one ecosystem around the world. 

I’m based here in the UK, , but I’m not part of the UK program. GQI is not a mouthpiece for the UK program. 

I think there’s a number of pieces that I think are very strong in the UK program. I think the UK program has always had a handle on this important aspect of modularity in quantum computing, the notion of the approaches to distributed quantum computing, they go back a long way in the UK.

Some of the key ideas came out of the Oxford academic group, and the UK’s first phase of its phase one of its National Quantum Technology Program, specifically sought to take this forward. 

Another related strength, I think, of the UK program, in its built up, , not just its hardware capability via the UKRI (UK Research and Innovation) investments, via the NQCC (National Quantum Computing Centre), it’s also now put on the NQCC’s software lab in Edinburgh. And interestingly, it takes an unusually broad view of the challenge of quantum algorithms and quantum software because it includes.

A very strong crossover into the world of quantum communications and distributed quantum networks. It’s led by Elham Kashefi and she was one of the inventors of blind quantum computing, which is a very important crossover, application between the opportunities we can get with, early quantum computing and early quantum security.

The other thing that I think is worth highlighting about the UK program is it’s always had a very strong emphasis on sensing.

It’s been right to champion that because partly it’s championing getting techniques out into civil society that we knew really should be there because people have been hinting about this stuff in the military field for some time. And it therefore generates interesting opportunities for early success.

Veronica: There is so much going on around the world and it’s really exciting time to be in the industry. Do you have anything you’re keeping an eye on for the rest of the year or anything that we should watch out for? 

David: In terms of underlying progress on quantum hardware and on quantum algorithms, the progress is already exceptionally strong. Over the last year, we’ve got a lot of very exciting announcements already. This year, I think we’re actually gonna see some more.

IBM for the last 18 months now talked up where they expect to get to with their Heron Revision 2 processor. And I think it’s going to be very interesting to see, then, what they can get when they deliver on that. I think it’s also interesting to, you know, for many years now, GQI has been pointing out that, you can’t just look at the current leading majors in this area.

There’s very interesting startups coming along with second generation approaches in their particular modality. And I think we’re going to see some very arresting announcements of where people have managed to get to with that next generation of companies as well sets up a very interesting, competitive field, both in terms of people that will say, well, what can we do now with these machines, but also in terms of people saying, and what is that roadmap and that timeline to get to the larger machines that we all know we need, we want in the end, for the most impactful application.

Veronica: Right, right. Well, never a dull moment. I guess that’s the short version. So thank you so much for your time today, David. It’s really been great to talk to you. I always appreciate the chance to compare notes with someone else tracking the industry. And, thanks for sharing all your observations.

David: Thank you very much.

Thanks for joining us for another episode of The Quantum spin by HKA. You can find all episodes on our website, hkamarcom.com. Of course you can find us in all your favorite podcast platforms as well. Follow us on LinkedIn under HKA marketing communications and find us on X, formerly known as Twitter. at HKA_PR.

If you have an idea for a guest, or if you’d like to be on the podcast yourself, you can reach me on LinkedIn, Veronica Combs, or you can go to our website and share your suggestion via the contact us page. Thanks for listening. Talk to you soon.