Lab to Market Leadership with Chris Reichhelm

Building Useful Quantum Computers: Constraints, Customers, and the Two Religions of Quantum | Richard Murray

• Deep Tech Leaders • Season 1 • Episode 31

Share episode

What happens when quantum computing startups can’t wait 15 years for fault tolerance? 

Richard Murray, co-founder and CEO of Orca Computing, reveals how his team chose commercial usefulness over technical idealism - and why that decision drives everything from recruitment to product development.

Operating from a University of Oxford spinout with limited resources compared to Google or IBM, Orca faced a choice: follow the same path but years behind and millions of pounds short, or constrain themselves differently. They chose constraints. Starting with just £1.5 million forced creative decisions - building quantum computers with one light source instead of dozens - that ultimately benefited customers through more practical, deployable systems.

Richard describes the ‘two religions’ of quantum computing: those pursuing billion-pound fault-tolerant systems versus those focused on immediate commercial value. His team deliberately chose the latter, applying quantum to generative AI with measurable results: reduced GPU requirements, lower energy consumption, and potentially more creative AI models capable of generating results beyond training data.

The conversation explores operational realities most quantum companies avoid discussing: the talent gap when no industry exists to recruit from, the cultural shock when physicists encounter commercial constraints, and why ‘comfortable with ambiguity’ matters more than technical brilliance. Richard argues that customers don’t know what they want from quantum - and neither do most vendors - making the shift from vague promises to specific value propositions the defining challenge of the next phase.

His most provocative insight: Goldman Sachs apparently defines ‘quantum advantage’ not as an academic milestone but as when a company invests straight off their balance sheet because they can define the business return. That’s the real test.

Essential listening for Deep Tech founders, investors, and anyone navigating the trade-offs between breakthrough innovation and commercial survival.


Let us know what you think...

Learn more about Lab to Market Leadership: https://www.deeptechleaders.com

Follow us on LinkedIn: https://www.linkedin.com/company/deeptechleaders

Podcast Production: Beauxhaus


Richard Murray:

I think one of the things I've observed is that when you get started with a new technology, it may be in a university research lab, you, you might have a sort of technology vision, but often that's completely unconstrained. You know, we wanna build a full to computer, that's it. That, that's the thing that will win us, the Nobel Prize and beyond that doesn't really matter, but we all know that in the world of, you know, real technology and, and business, there are lots of constraints. Um, and I think actually by the way, working within constraints can be a really creative process. But you've gotta learn to do that. And, you know, that constraint might be, well, okay, we can build this thing, but at what cost? At what cost to the customer at, you know, what value will it provide? Um, or any one of a number, like a timeframe, like, we wanna build this thing, but do we want to take 15 years to build it? Do we wanna take 10 years to, or, but, but adding in constraints to your definition of the technology vision you have in order to lead to a. A more sort of outcome driven mission, you know, more, uh, application if you like. Driven mission, I think is, is, is is really important.

Chris Reichhelm:

Welcome to the Lab to Market Leadership podcast. Too many advanced science and engineering companies fail to deliver their innovations from the lab to the market. We are on a mission to change that. My name is Chris Reichhelm and I'm the founder and CEO of Deep Tech leaders. Each week we speak with some of the world's leading entrepreneurs, investors, corporates, and policy makers about what it takes to succeed on the lab to market journey. Join us. What are the trade-offs that deep tech innovators need to make in order to realize their ultimate innovation? That is the topic for today's discussion and to help bring this to life. We're gonna set it this topic against the context of quantum computing, which perfectly illustrates I think, the inherent conflict. What do I mean by this? There's a lot of talk about what the world will look like, about the problems we're gonna be able to solve when we have realized truly fault tolerant, full stack quantum computing capability, that drugs will be discovered in a faster period of time. We'll be able to develop novel, advanced materials, we'll be able to better model more accurately, uh, climate change and so on. If we're a company like let's say Google, which is very active in the field of quantum computing capability with lots of money to throw around and invest, we can almost pursue this task and an academic slash RD type way. Hire the best scientists and engineers, incentivize them, let 'em get on with it. It's not as simple as that, but you know what I mean. But if we're just about any other company, and especially if we're a startup, that's not possible. And why? Because too many stakeholders around the table need to have value developed along the way. We can't wait to get to that point, which may be realizable 10, 15 years away. Maybe it's not that long now, but again, you know what I mean. So we need to demonstrate that we're delivering value along the journey, and in order to deliver value, we've, that means we have to develop commercial value. How do you make a project that's in motion commercializable? My guess that's what this topic is all about. Where are those compromises? Where are those trade-offs? Where are the trade-offs technically, operationally and commercially? To help me get into this today is Richard Murray, who is the co-founder and CEO of Orca Computing. And Orca is a UK headquartered University of Oxford Spin Out, which is developing full stack, full tolerant quantum computers. But they're not doing that right now. They're on a path that may or may not get them there, and so along the way they're gonna be making these trade offs that I've just referenced technically, operationally, commercially. Where is the line drawn for Richard and his team along that path? That's what we're gonna get into today. I'm certainly gonna get a huge amount out of this. I hope you do too. Let's get into it. Richard Murray, thank you so much for joining me.

Richard Murray:

It was great to be here. Thanks for having me.

Chris Reichhelm:

Take me back, take us back, if you will, to when you and your co-founding team decided that it would be a good idea to build quantum computers. What was, what were you guys thinking at the time? What was the vision? What was the, I guess the motivation and what did you, what was the anticipation? What was the expectation? Sorry? What was the expectation? What did you think you would build?

Richard Murray:

That is a great question, and if you don't mind, there's a complicated story that I'll unpack a little bit for you because, um, so I'm not the academic behind the company. I. So, in fact, my, my journey, I have a PhD from a long, long time ago, but I was working for the UK government on the quantum program that I helped set, set up. And then, you know, I've been in and out of technology consulting on the commercial side. So I came to this through the lens of being involved in quantum, involved in the policy behind quantum technology, quantum computing. But actually I went around the country and met lots and lots of different academics with lots and lots of good ideas before meeting orcas to academic, uh, co-founders. So Professor Ian Warmsley and Dr. Josh Nunn, they were the ones who'd been working in the lab, uh, to develop this sort of cool technology. Um, but in the middle of those two backgrounds I think comes a really interesting story that relates to what you were describing. So on the academic side, they'd been working a lot on quantum networks, these sort of big optical networking type things. Lots of different things. They can be useful as networks, both quantum and classical, and they can, they, the academic team found that they could also be useful for computing. So when the academic team and met, yeah, we knew each other, but when we sat down and talked about what Orca could be, there was a huge range of different things, some of which were not quantum computing, which we could have done. And, um, in fact, for a long, long time we were sort of working out like, what exactly do we wanna do with all this amazing technology? But, you know, could, can we do all of it? And, and when I came in, I was quite keen that we didn't try and do everything, uh, by the way, to our conversation today, that has been an, an ongoing discussion about how much we do do and what, what we do focus on. But from the very start, we, I tried to sort of set, you know, set the team up so that we weren't gonna do everything. We, we called ourselves orca computing for a very good reason. To keep reminding ourselves that we're trying to do computing, and if we'd have called ourselves and, but this is discussed, if we called ourselves Orca Quantum or Orca Photonics, then we could have been, we could have maybe not had that clarity. So right from the very start, we got together as a team to focus on computing and, and the quantum computing. And, and the reason for that was a few different reasons from the technology side, from the, you know, academic perspective. They saw a great potential to see their research realized. You know, in the, in general where they sort, they sort of came from a tech push point of view. They wanted to see their, their staff made real, made more robust and all of this type of stuff. But from my point of view, and maybe I've always had a more market facing point of view, I, uh, I saw that a lot of existing computers were moving more and more to photonics. So more of the photonic, you know, more of the computers that we rely on having on the inside, more and more platonic components. Also there's this huge overlap with the world of, um, hardware making, machine learning better. So even though I wasn't totally from the market perspective, I, I think I was really keen on what the application of this cool stuff could be for computing. And, you know, I guess the coming together of those two themes landed us squarely in the middle of photonic quantum computing, which is what we do. Um, yeah. But that it, it, it, there is, there are a few different perspectives that I think came together to maybe make us really excited about what we could do in the world of computing with our photo quantum technology

Chris Reichhelm:

and I suppose a certain amount of that kind of ambiguity, if you'd like, that, that it's not always a hundred percent clear, I think with any group of founders. You're not an academic founder, but you are a founder of the company, generally speaking, uh, as its, as its market, face and CEO. Um, but. You know, I wonder if that, you know, how common that is for that. And I wonder if there's a piece in, in the company's history around we want it to do computing, and if it's more quantum related and great, it's more photonic related, great. But we want do computing for particular applications. Would that be fair?

Richard Murray:

Yeah, I, from my perspective, definitely. So, I mean, I definitely saw, you know, this huge sort of almost gaping hole where technology hardware could be better. Um, sorry. Computing hardware could be better. You know, I spent my whole career watching computers evolve. I, I do quite often focus on where I think the biggest market is. And, you know, it's not that, it's not that hard to see that there's a massive interest and market for computers. Even when we started six years ago before the AI boom sort of kicked off. I think that's really important because I think with bots in general, you can do a lot of stuff with bots and I think a lot of bots companies end up in what I sort of call research mode, where they get really excited about serving the research community to build a new microscope that might serve another researcher. And for me, that all sounded just too small. Um, and it's, it, it is sort of like in the, in the world of computing that I see a real, a really exciting need for new technologies, new ways of doing it, and then all the way up to the top because of what we now know that computers use loads of energy that sort of, you know, worrying for all of us how much energy computers are using. Um, also, I always say this, and I always think it's true that computers just aren't, computers just aren't very good at lots of things, so we. I stood up on stage a few days ago in front lots of people and said, just, just think about how bad Siri is. So everyone loves to talk about ai, but let's just be honest with ourselves. How bad, uh, things like Siri or, you know, uh, satellite routing, uh, you know, uh, GPS type systems are and things like this. Computers is, are really not very good and we like to think of them as human, but they're a long, long way from being any type of sort of having the computational power to be able to be more, more natural in the way that they approach things. So, or maybe

Chris Reichhelm:

our expectations have become completely, our expectations are increasing because I think if you compare them to, you know, probably computing standards five years ago, they're much, much better. It's just now our expectations are that much greater.

Richard Murray:

That that's true. Yeah. And we all expect those computers to be in our phone as well as performing this human-like interactions. So, yeah. But I mean, the, the big picture is that I do love the idea of computing. You know, I love the idea of us as humans being able to process information. I, I do really passionately believe that computers could be a whole lot better and a, a very, uh, sort of surface level photonics has always been this technology which had the potential to completely change computing. So people have been talking about photonic computers without the quantum part, but for years now, for decades, and there's never been a moment where that made sense. And now we've got the convergence of, you know, maybe new ways to do computing, photonics with obviously the, the need, urgent need for new types of computers driven by AI and, and our expectations that computers should be better. So it seemed to me like the, sort of the most amazing moment to just apply some very new technology, but the, the need for that new technology has never been more, more pertinent. Yeah. Um, in my view.

Chris Reichhelm:

So what it, so the, the end goal, however. Is still to develop full stack fault tolerant quantum computing capability. Is that right?

Richard Murray:

Yes. Well, on our technology roadmap, that's the, that's the endpoint that we observe. But maybe if I was to unpack that a little bit, I think as a company we've changed quite a lot. So that if, if I was to describe our vision, um, it is to build useful quantum computers and, and those just might happen to be fully fault tolerant, uh, quantum computers or they might not. But the, the thing that now is our sort of north star is it doesn't matter what the technology is. We want to see them being useful being, and you're smiling because I'm sure this is what we'll get into a lot, but there's a, there's a such a drastic difference between being sort of guided by a technology mission and, you know, the, the worrying thing there is that that technology mission might be wrong, you might get it wrong, versus a, I don't know, a more market led mission. And in, in everything related to how the company approaches things. We're trying to, and you know, I shouldn't present it as a, as a completed task, but we're trying to move everyone onto this new mission of just doesn't matter what it is, but what we're trying to do is build useful, commercially useful, uh, you know, valuable quantum computers in whatever technical form that may take. And yeah, that's as, and that's worrying for lots of people.'cause it's so much more ambiguous. It's harder to define it if it's not one very specific technology, but it's like, oh, it's useful. Well, it's useful me, um, we dunno. Let's find out.

Chris Reichhelm:

Yeah. But I think in, in, especially in deep tech, I think this is relevant. You know, we're gonna come back to all those points you just mentioned because they're, they're, they're very, very relevant and I think they represent maybe compromises or trade-offs is a bad term. I don't mean that in the pejorative sense. I mean it in the, in, in almost a practical sense and. And, and maybe in a sense that forces us to take a look to reevaluate what our real intentions are here. Um, you know, we see this in ai. There's a lot of talk about realizing a GI, uh, putting the safety concerns to one side for a minute, but, uh, with, without necessarily understand and understanding of the pros and cons of, of, uh, of realizing that goal. And I think deep tech innovators and, you know, I think, uh, true fault tolerant quantum computing capability is another one of those goals. We want to do that. I think scientists and engineers within the deep tech communities are particularly, uh, uh, good at focusing on a goal like that. Then just going for it regardless of the benefits. Yeah. It almost to to demonstrate least of all to themselves. That, that it can be realized that it can be done. That they can use their powers of, uh, of, of logic and reasoning and, and, and brilliant to, to accomplish such a, uh, a difficult, uh, a task, a challenge. Mm-hmm. Um, but then of course you rightly bring up, but you know, and you ask the question beyond it, which is but to what end? Yeah. And if we're going to apply such a, an enormous amount of capital and investment, not just in terms of finance, but in terms of human capital as well, and energy and resources. All of the stuff that goes into building all of, and realizing all of these innovations.'cause it's not just money, is it? We know it's not. So, you know, then for what benefit? Is the, is there not something we can do? Is this not, is, you know, should we not be considering this potentially from a different perspective?

Richard Murray:

Yeah, I, I totally agree. Um, I mean, one of the, about that trade off, uh, point as well, because I think one of the things I've observed is that when you get started with a new technology, it may be in a university research lab lab, you, you might have a sort of technology vision, but often that's completely unconstrained. You know, we wanna build a fault tolerant computer. That's it. That, that's the thing that will win us, the Nobel Prize and beyond that doesn't really matter, but we all know that in the world of, you know, real technology and, and business, there are lots of constraints. Um, and I think actually by the way, working within constraints can be a really creative process. But you've gotta learn to do that. And you know, that constraint might be, well, okay, we can build this thing, but at what cost? At what cost to the customer at, you know, what value will it provide? Um, or any one of a number, like a timeframe, like, we wanna build this thing, but do we want to take 15 years to build it? Do we wanna take 10 years to, or, but, but adding in constraints to your definition of the technology vision you have in order to lead to a, a more sort of outcome driven mission, you know, more, uh, application if you like, driven mission, I think is, is, is is really important. And it's, it's so hard for people. One of the things I've noticed employing people who tend to come from a university background is they hate me. Give me telling them, or the market telling them constraints. Like, oh, you could do this, but you are not gonna, because if you do it this way, this is not gonna be, you know, you're not gonna build this thing any less than a billion dollars. So we're not, we're not gonna do that. That fundamentally challenges them, I think because they're driven by the science. They're driven by, you know, building this thing that no one has ever done before. Yeah. Um, so I think that's a really important challenge.

Chris Reichhelm:

Yeah. I, I like the point you raise on, uh, on the, on the piece around constraints and how constraints can actually be, uh, what's required in order to realize an innovation or a pathway to part of the innovation. And it re it reminds me of, uh, of airplanes in the Wright Brothers. I dunno if you know this story, but when they were trying to get the planes off the ground, literally, uh, there was another group, uh, a consortium of wealthy benefactors and individuals who came together and applied a, a, you know, uh, you know, uh, all resource, um, a resource intense, uh, gr group to, to build airplanes and to get 'em off the ground. And they had lots of people and lots of smart engineers and uh, and uh, and lots of capital to be able to do it. And then you had, uh, the Wright brothers. Who had none of that just, and it was, and it was a few of them. And, uh, and the Wright brothers won. Uh, you know, they were the first ones to get, you know, that's why we remember them and not this other group because they had constraints. Yeah. They had capital constraints and they had the human constraints. And so they had to figure out things in a different way. And, uh, I think with deep tech innovators, uh, and deep tech innovation generally, particularly in 2025, we're in an age where, uh, where we, where it's easy to think about the absolutes and the realization of these absolutes and only think in terms of absolutes, but to forget all of the wonderful benefits which may emerge, uh, along the journey, which we haven't quite put a label on, and which may not represent the absolute realization of gold, but may wind up delivering a huge amount of value nonetheless.

Richard Murray:

Absolutely. It's a funny story that when we first got started, I should have said this, when we started talking about the introduction talker, when we first got started, we raised a siege pre-seed round that was 1.5 million pounds. So we got started with a tiny amount of money and we, I remember us sitting in this office and, um, we were talking about what we might do and it was so obvious to us at that point that we weren't gonna build anything substantial with the funding that we had, you know, along the normal path of just, okay, let's just go out and build this massive system. So we sat down around the table, what can we do, uh, with the, you know, money that we have available. And we came, actually, came up with actually a much more resource efficient method of delivering a very first quantum. And that was great for us because that meant we could build, build the system with just one. It was cool because we just had one quantum light source and one detector rather than, you know, tens which other people might need. Um. So it was good for us 'cause we could do something, but then we, we quickly found it was really good for our customers. We could, we could do something for them as well for that same, same amount of money. But this very early on we were constrained and had to be creative to learn to work within that constraints. And, okay, there is a trade off. I don't think you wanna be too over constrained too often, but I think a certain amount of constraints or, or limits to resources at your disposal stop you just being somewhat like, I dunno, idealistic maybe in the way you approach things and do, force you to be much more grounded. And I think that that does actually drive innovation, um, in a, in a useful way and Great. Maybe one day I'll compare us to the Wright Brothers. Yeah,

Chris Reichhelm:

yeah, yeah, yeah. Well, how difficult was it to convince your colleagues to pursue a, a commercial line and, uh, as opposed to the ideal?

Richard Murray:

Um, funnily enough, I think back then it was very easy 'cause everyone could see the constraints. Here's the money we've got, what are we gonna do? Let's do this. That's great. Um, I think then when you come into more resources, then maybe more of the challenge comes, like, you know, the question will be, well we've, we've now got more funding. We've just secured a bit more funding or something. Why are we still pursuing this constrained model? Why don't we just flip over to the, you know, idealistic model of building quantum meters. And it's funny, it's hard then to explain. Well that's because, uh, that's what, and it actually's not that the market has spoken, it would, it's not as easy to say that we've got all these customers lined up for this meal resource constraint because you have to sell a vision. And that vision is that I believe that within, if we constrain ourselves to build quantum in a more useful way. The, the market is bigger for that, but that's something that you can debate. You know, I can, you can, people can disagree with me and say, well, actually I disagree. I think that actually these much bigger quantum competitors have a much bigger market potential. So I think, I think the challenge comes later when you have the option to pursue a much less constrained approach, maybe a much more scientific approach. We're always having to explain why we believe in listening to customers. As crazy as that sounds, you know, why, but in the world of quantum, you know, like, do our customers really know what quantum of eating should look like? And you know, I think that to an extent they do. Of course they do. They know what a computer looks like. They know how much money they'll spend on a computer, and the fact that it's a quantum computer doesn't change that. Um, so it's, it's an interesting thing that, and by the way, it's, it's, it needs constant reminding. I constantly have to sort of reiterate with my team and also to investors. This is what we believe. Um. We are, we do believe in this, um, more commercial, uh, the commercial reality behind quantum computing rather than, um, yes. A more idealistic version. Yes. Um, I call, I call it a religion. I say I actually, sometimes I say there are two religions in quantum computing. Those that believe in nothing but large scale fault tolerance, billion dollar systems, and we have a different religion that's about more immediate usefulness, commercial relevance, value. And they can align of course, but they are, to a certain extent, quite different. Sometimes

Chris Reichhelm:

it is a different path. It is a different path. Yeah. I love the point you raise about the constraint. When the constraint is lifted through additional resources like finance, suddenly that focus goes away or can get a little, you know, can get a little blurred. We, on the, on one of the recent podcasts we did, we talked about this a lot, how one of the. One of the real challenges for a lot of deep tech companies, especially with fairly young teams, is the pressure of so much capital and how when thing, you know, we start to see, we can often start to see things go wrong. I'm not, by the way, making a case to limit the capital you give to deep tech companies at all. But the, it's, it's how we react to the large s to having that resource. It's the, it's the pressure we feel as humans and our ability to deal with, with, with fewer constraints. And then the creeping back of that idealism and what we might build.

Richard Murray:

Yeah, that's a real pressure. It's so, it's so important to, to me and then maybe the way I do things, um, and, uh, it, it's a sort of strange, wanting to make life slightly harder for myself than, or ourselves as a company that maybe we could do because we want to constrain ourselves and want to. Uh, make sure that we think carefully about what we do to make the most of the, what I mean, even when we've got more resources, make the most of the resources that we have. Um, and, and by the way, I think in this world of quantum computing, there are other companies with almost limitless amounts of resources. You mentioned Google, uh, IBM also privately funded startups have essentially infinity money right now. So you have to think to yourselves, and actually we went through a big think thought process about a year ago where we thought, well, if we, if we also operate in a totally resource unconstrained way, all we're doing is following the same path, but $990 million fewer and three years behind. So we've gotta constrain ourselves in some way. We've gotta think differently, otherwise we're just gonna end up at the same point. But, you know, three years after those other, you know, if we're lucky, three years after that. So we definitively decided to take a different path to be much more commercial. And also to think harder to, in order to constrain ourselves according to the, what we thought the needs of the customers are and should be, um, in order to make ourselves competitive. And it's always quite funny. Like I do, I do think that we are still competitive with those companies that have raised way more money than us because we just had this much more commercial approach. And because we've made life maybe a little hard on ourselves, um, in order to apply these constraints,

Chris Reichhelm:

I I can believe that I can, I can definitely believe that. I, you know, one of the things I've observed with, especially in, in, in the field of quantum where you have a number of different, uh, uh, players who are building, who are trying to build fault tolerant full stack quantum computers is the dec is the ultimate decision that's pending around focus. What are you ultimately really gonna focus on? Because if you're doing. Everything from the quantum development, the chip design to the, to the, uh, the cabling and, and the distillation, refrigeration, cryogenics and the firmware and the electronics and the software and the application layer and that. Oh, it's just so much. There is so much in there. Uh, and it's, it's not cheap. It's definitely not cheap, realizing all of that. And, and so even if you raise a big round and if you have lots of people, that money's gonna go pretty quickly. Yeah. If you're, so you postpone that make buy decision that most other established technology companies have to have to consider. Where is our core focus? We, where can we deliver the most value and what are we gonna rely on partners for in, or, uh, and, uh, in order to, in order to build ultimately a successful company.

Richard Murray:

Yeah, you're right. Actually make buy is one of probably the hardest decisions that I think we have to make because. Mostly driven by the fact that so often, um, it's more natural for the team to approach it if they just make it themselves. Um, so, and it might take more resources to your point, but you've got complete control over, it's easier, right? Um, but the, the bigger picture is, well, no, actually, firstly, it's much more efficient if you do it another way. If you focus on what you do well and you leave everything else to other people. Um, but also if you invest the time and the effort into that. But that investment, by the way, takes a different skillset to outsource a technology to a supplier is a much more diff a much different skillset to being a physicist and building stuff in the lab. Um, so it requires it. So you, you almost by default, fall into the way of working, which is just, well, we need more money 'cause of course we're gonna do it all ourselves rather than, again, more thinking, more efficiently in a more focused way. Well, we don't need to do this. And actually, here's a great company over here and it's gonna take us three months to get that company up to speed. But once they're, and by the way, they can also ramp up production if we ever need them to so much more. I think it's a great example. Make buy 'cause it's just so often seen as a default option. One way of there's any one way of doing it. Yeah. But when you, when you operate in a different way, focus towards, you know, achieving some commercial outcome, you've gotta do things differently. You've gotta think in a different way. Um, and yeah, outsourcing can, yeah. The, the buy side can be much more important.

Chris Reichhelm:

Yes, it can. You, you've mentioned customers a lot, um, and uh, and you had in our earlier conversations too, do you think customers understand the quantum offering? Do they feel, and let me kind of expand on that question a little bit. Do they feel, well this is gonna be like classical computing. And I'm gonna type some stuff into my keyboard here and it's gonna just produce more accurate stuff. Is that what it is? Um, or do they, uh, and, and I guess it's depending on the software, maybe it will be, maybe it will be 'cause software will disturb and that to an extent, or, or do they see it as a truly new computing paradigm?

Richard Murray:

Really good. Um, I don't, I don't think they know, I don't think they know, by the way, I was gonna be slightly provocative in saying that. I don't think a lot of quantum companies know the answer to that question either. So it's not like we're helping with that uncertainty. I think. Um, and this, I think describes the promise or maybe the potential promise, but also the opportunity of quantum of eating. It can be both. It can be one or one or the other. It can be more of a, an accelerator to a class, an existing computer, or it can be a completely new thing, unlike anything they've ever seen before. Those customers, they don't know. Right. You know, they don't, I mean, they might have employed some quantum physicists, but they're relying on the vendors to tell them. And I think then the vendors don't really know. You know, there's, like I said, there's different religions, so it doesn't help. They talk to two different vendors and they get two different answers. Um, so I don't think it's as easy as just saying you can just listen to your customer, because I think, and I don't think it's as easy just to say, well, yeah, we would just let this play out and see how it goes. I think it does require, well, the starting point of a conviction that it works in a certain way. Um, and then so in to a certain extent, we go to our customers and we say, here's our vision for quantum computers. And by the way, most of them turn around and say, oh my goodness, this is much more understandable than anything I've ever, I've, I've, I've heard from so many different people, all of this confusing stuff about quantum is half of which I don't understand, half of which doesn't seem to really relate to what I'm trying to do within my business. Finally, someone's actually explained it in terms of a, a value creator or, or a product. Um, so I think that for me is the sort of initial moment where I think, you know, you've got something where you propose something and the customer at least says this makes sense. Um, and then the next part is then to obviously move forward and validate with them to the point of them spending money and committing to you and obviously, you know, scaling operations with, with you and things like that. Yes. Yeah. I dunno if that answered your question. I think my answer is it does. No one really knows.

Chris Reichhelm:

No, no, but that's, but I, but that's my sense, that's my sense. Um, and I think it comes back to these, these, I won't say arbitrary goals that we set ourselves as humans. You know, putting a man on the moon in 1969 was gonna be a good idea, not just through the fact that we, you know, it was of limited value to put a man on the moon. We could, you know, look at space rocks and um, and those sorts of things. And, uh, uh, but it was all of the second and third order considerations that. That, that, uh, extended from that, from putting a man in the moon, all that we learned about science and engineering, that that, you know, that gave way in some ways to this great era of innovation that we've had since. Um, and I see that with Quantum, uh, a bit. And I think one of the benefits of going early with customers is you can have those conversations with them along that journey. I wonder, and, and this, you know, we're gonna kind of glide into the, you know, the, the, the topic of solutions and product market fit and, and products and, you know, what are we actually offering here and what are they gonna be able to do that they couldn't do before? And yeah. You know, how do we measure value? How do they measure value? But when you're talking to an audience of customers and what they're looking, I'm assuming is, you know, they're looking for, you know, some kind of competitive advant, you know, advantage. So look, you know, so actually it would be really helpful for you to discuss Orca computing and where you guys sit in that customer journey. Who are your target customers? Um, what kind of benefit are they getting because I know there's a strong machine learning component to their, uh, and a data distribution piece. So actually before I kind of go down that, yeah. Could you extend you, could you talk a little bit about the solutions that you guys are kind of building and why?

Richard Murray:

Yeah. No problem. In fact, before we do that, do you mind if I just make a sort more general point about all that, that product market Yeah, of course. Market that you mentioned. So, um, I, I, I don't think I've mentioned yet, but actually before, or I worked for the UK government on their, I, I help run their, their quantum program. And I think it's interesting because there is a massive difference between how countries approach innovation. If you've got a big innovation program and okay, you wanna see some succeed, so you just sort of. Where the idea of innovation can be a lot more just related to sort of osmosis. You know, you fund a few things, those things may or may not work. It doesn't matter 'cause that company might sort of not succeed, but then the founders will go on to do something else that might succeed in that world. Just innovation could be a lot more, like, let's say like osmosis. You sort of hope that, you know, eventually the right ideas coalesce on the right sort of i right markets or the right, um, product solutions. I, I don't think that model works at all for startups because, you know, we've only got one company. We've only got a small number of customers that we're hoping to build solutions with. So the idea of not being super focused on what exactly and, and then, you know, product market iterating on that product vision with them super closely. But knowing that, you know, two months wasted going down the road that you don't think will work out is, might be fatal. I think everything can be, needs to be much less like osmosis, much more focused on where you really think the product opportunities lie. And, and, and I'm only highlighting this because I think this highlights the massive gulf of understanding between say, government and policy makers talking about innovation, but quite often without sort of a lot of substance without, without a lot of specifics and potentially university academics as well. And then what, how companies operate. And so, sorry, sorry about that slight setup, but I think that was interesting. Um,

Chris Reichhelm:

we could have a whole hour on that.

Richard Murray:

Yeah. But stop, and lemme because you set up nicely and I should't go through this without tell you more about, ok. So, um, so I mean, firstly orca with, with the way that we build, uh, our quantum meters with photonics, the first thing it allows you to do is build these systems that look much more straightforward than what a lot of people imagine a quantum computer to look like. So. We, our products, they look like normal server rack, so they're sort of black boxes. They roll off the back of a lorry, they get installed into data centers very easily without the picture that some people have of quantum computing is that it's a sort of golden chandelier, a lot of expensive equipment, um, and very different from the normal server like boxes that, you know, customers and data centers expect. So, so that's the first half of what we do. We build these quantum computers, but in a way that looks completely different, much more practical, much more like what existing data centers, for example, are use, are interested in using. But that only gets you so far and because you know, you could install a computer and in sort of computer and inverted commas 'cause no one really knows what one has to do with this thing. And really, in the early days we were largely sort of physic led. We went to customers talking to them about photons and qubits and. Pretty much a hundred percent of our customers had no idea what we were talking about. Um, so we quite quickly realized that unless we had an application view, which actually is also in the way we allow customers to use our systems through the software layer as programming environment, but also a link to applications, um, it's very, very hard for those customers to really understand tangibly what what you do. So actually from sort of the first year onwards, Orca spent a lot, a large, large proportion of our effort investing in applying our systems. And what a great piece of work that came outta that is, um, most firstly the application, the link between quantum and generative ai. So applying quantum to existing generative AI models and actually demonstrating how quantum can make those Gen AI models better. We're just on the journey to demonstrate that. But that alone would be a game changer for the world of quantum video. It would be huge. Um, and then slightly more sort of boring if you like, but also just as important is the, the way that we allow quantum computers to talk and communicate to with existing computers. So it was also a realization that quantum computers don't, they're not a standalone thing. What they need to do is integrate. They need to be side by side with existing computers in a hybrid way. And to do that really well, you can do it in a sort of hand wavy cut type way, in a way that customers don't think is very good. But we, we needed to really invest in, um, close, we call it close coupling, close integration between the quantum and classical computers so that those two systems can work really well. So well in fact that the user doesn't really see that they're two different types of computers, even though behind the hood we know that the quantum system is operating completely different from the classical. We provide that integration layer so that customers can sort of understand what's happening, not see them as too, too different a thing and, and easily operate with those things.

Chris Reichhelm:

And so how will the customer know that they've got quantum capability? Will they be able to process more data or perform more Yeah. Complex tasks.

Richard Murray:

So the short answer is that in the performance that they, they're able to achieve. So I mean, we do benchmark our systems to show a performance uplift. You know, there's no point in working with these quantum systems unless they can be shown to be better. And in this world, the customers we speak to, they, they are all in on very, very detailed technical benchmarks of machine learning performance and things like that. So, so one answer would be that they just, they see, uh, uh, an improvement in their model. And the, the types of results that we've started to see are things like you can perform the same operation with fewer GPUs, so fewer spend less money with nvidia, which of course is useful for lots of people. Um, you, that also correlates to a potential very significant energy saving, and this is the one that maybe really gets me outta bed knowing that current AI models consume a, a, an unsustainable amount of energy in a way that should really worry us if we're thinking about using AI so, you know, sustainably in, in the future. So the idea of reducing the energy consumption, those models, not only helps of course with carbon footprint, also makes them economic. The final really core result, which we're still validating, but if, you know, if we demonstrate this we're, we would be a huge shift in the quantum market is an ability for those AI models to be, to be more expressive, to, to start to predict results that fall outside of the training data. So normal AI models are based on just lots and lots of training data. That's how they work. But they're normally quite constrained to working within the sort of more or less, the sort of limits of that training data. As soon as you start to ask it, ask for data that pulls outside that training data, it struggles. So say for example, say in the world of using AI to predict new molecules, which is where we spend a lot of our effort, the model will be very, very good at giving you a mole molecule or a new drug that looks very similar to an existing drug, finds it much harder to predict a completely new drug and create new molecule that no one has ever seen before. Um, in that world of completely new molecules, completely different types of data, we see that our quantum system can really boost the performance way away from the existing training data set. If even a little bit of that ends up being true, uh, you know, of course we've validated, but that, that is huge. And that's that really then you can start to think of your AI system as much more, um, creative, which we think AI is creative, but reality is not so much. So yeah, if even a hint of that is true, that could be enormously powerful in seeing AI evolve into much more creative tasks. Um, now I'm not going to go too far down that road because all the, the results, what we've seen are things like molecules, things like, um, industrial processes, quite constrained things. I'm not gonna talk about AI become, becoming cognitive or anything, but, but really tangible industrial results that come outta making AI slightly more creative, slightly more expressive, different from the training data. And, um, that makes me really excited and, and all the while building on this huge backbone of quantum of eating, which we all know will keep evolving, keep getting better. Time after time, time, but now linked to an application that makes a huge difference. If, if you can, if you can land it.

Chris Reichhelm:

Yeah. It, it's, um, I mean it's very exciting to hear you talk about that and the, the, the benchmarks you guys are currently going for. It reminds me a little bit of the late nineties, early two thousands when Oracle used to run this big campaign around their servers and the performance of their servers, and it was a very, very effective marketing campaign. It was generally in the Economist, and they had a white backdrop and red lettering. Oracle servers run X percentage faster than IBM's or someone else's and use less energy. I don't think they ever said and use less energy. No one talked about it in those days, but, but they, it was that, it, it was the clear benefit. It was a clear, it was. It was a clear message, it was gonna be useful for engineers who were looking for efficiency, um, or were looking for just performance. It was all about the performance of their servers and, uh, and, but it, it got the message home. And I think that's something that the, I think customers can relate to. If they can say, we can do your job more efficiently, we can make your AI or your machine learning platform, um, much more creative or more able to process complex tasks more accurately. That's a, I mean, that's a, as you rightly say, that's a game changer.

Richard Murray:

Yeah. And it's funny in quantum computing, because I think the last five years of quantum computing, it's been that long. First Quantum have it sort of merged about five years ago, the last five years of quantum computing, quantum has survived on the basis of. The, the applications are groundbreaking. We can invent new molecules and all that, but without being very specific and yeah, about it for being promoted as the, the first change, uh, or paradigm shift in computing that has been, you know, many decades. But I think now users of the customers are much more sophisticated. We've seen Goldman Sachs, we've seen, you know, BP invest quite heavily in, in their own quantum research team. So they know that they need to land specific applications, which are of above and beyond just the sort of generalities of, oh, well maybe it could, you know, do this, but don't ask me exactly when you'll be able to do this or, or what exactly it can do. So I do think it's not just orca, but I think in general, the whole of the quantum of market is shifting to be much more specific and needing to answer much more specific questions. Industry to customers on what exact fine I get. They can do lots of stuff, but what specifically can it do? Yeah. And, and when and how much. Yeah. And so I, I think that in general the quantum industry is about, is you going through a pretty exciting shift. And obviously there are lots of publicly traded quantum stocks at the moment and they're having to answer this question as well. And I, so I think it's a sort of exciting transition from it as a exciting new era of science to Well great. I get that. But what are the products,

Chris Reichhelm:

and one question on that. You mentioned Goldman and BP starting to, uh, you know, hire their own quantum scientists and engineers. Are they doing it to enable further, enable their AI capability? Or are they doing it kind of as a standalone operation to see where that goes?

Richard Murray:

Um, hard to tell. I mean, I think, um, historically they've done it as a general capability that they knew they needed to build up. Um. Mostly looking in general at Quantum and what can it do, I think now almost. And so, and I've worked in the world of sort of industry, uh, innovation and, you know, research divisions that are quite separate from the main body of the business. I think now that those, those quantum teams that have been set up maybe on the peripheries of large companies doing their innovation piece, but, you know, let us know and we'll work out where it applies to the business. I think that is starting to shift to be research teams that are much more centrally located within the business. Um, so I think historically it's been less linked to say AI or the specific applications, but I think that's changed. Yeah. Um, actually one, one great example is I spoke to some, someone who I won't name, but it's a great quote that they said that said that they, in, within their team, they labeled, um, quantum Advantage to mean when a big company invests in quantum computing straight off their balance sheet. They, because they're able to define, you know, the business return that can be derived, that, that is the definition of quantum Advantage. Not some more academic label. But, uh, you know, there's a lot, lot of debate in the industry about what Quantum Advantage means, and this person apparently stepped forward and said, well, it means someone investing in a off their balance sheet. Um, of course, we all know what that means. Absolutely. And how significant that would be.

Chris Reichhelm:

Absolutely. Absolutely. That's very good. Um, let me, uh, and let's talk about something else. Um, the journey you make as an organization requires obviously, certain skills, experience, and talent, and which, and those skills, experience, and talents change, uh, over time. Um, how, and yet talent is, you know, such talent, the type of talent we're talking about, it's at a premium, uh, it's scarce and what is available is at a premium. So how do you think about. Team, team, uh, that recruitment of skills, experience, and talent, the way that's organized, um, what's your, you know, what works for you guys?

Richard Murray:

Yeah, I mean, this is really at the heart of, I think being successful in Quantum is, is the team. Um, the, the funny thing about Quantum, which I think a lot of people miss from the outside, is because of the fact that quantum has never been an industry, there has never been an existing industry for the development of quantum computers. That means that almost certainly you cannot recruit people with a lot of industry experience in, in Quantum because it hasn't existed before. So it is not the so easiest to say, oh, I'm gonna go for the top 10 quantum companies and pick up someone with 10 years experience working on a quantum computer. Import you, I mean, you might find one person, but in, in general, that's not a talent sourcing. Scheme strategy that works. So that then means that either you go for the people who know a lot about quantum computing, but typically they'll be coming from a more research background. So they'll have to learn and adapt to learn how to operate in a more commercial environment. Or you have to go for people with a lot of industry experience, but then find the people who are capable enough of translating that knowledge into quantum. And the question of, well, those people, how, how much, how can, how deep can they go into quantum if they, if they haven't studied it before? And, and maybe there are outliers, people who've studied quantum before, but, you know, and then steps into industry bit, bit like myself. But there is a, a talent gap, uh, and a pretty big one, or people who know quantum, who also know how a company operates, especially at the more senior levels. Really, really challenging. That gets only even more hard for more product related skills. The more, you know, when we really look at the frontiers of product development or quantum computers, you're really talking about, you know, like probably a Venn diagram of one, well, a very small number of people. So I think, um, the answer is twofold, really finding great people who are versatile enough, um, dare I say, it's sort of creative and intelligent enough, but versatile enough to take on a new role that's different from what they did before. Step into a new way of doing it. Embracing new technology or, or, or, or embrace or if they're coming from research background, a bit of a weird, much more, uh, constrained way of, of working than they've been used to. So you've gotta find the right sort of feedstock, if you will. And then in inside the company, I think there's a lot of, you know, um, I think training is too shallow a word, but, um. Up all these words aren't significant enough to really describe the job of really working hard to make, make sure that your team members understand culturally what you're trying to do. That's maybe the hardest piece. You know, making a researcher understand the world of, uh, products and revenue is, is, is a big, is a big old cultural shift and a big, um, thing I think sometimes for them to understand and also just in terms of the, the actual skills that they have to be able to step into this and continue to evolve and develop as people, um, into this completely different world of working. Yeah. Often. Yeah. Um, but, but it's, it's hard. It's really hard.

Chris Reichhelm:

Your, your point around. So I wanna touch on a few things you mentioned there, but your, uh, your comment that training or development is too light a word. I forget how you phrased it, but you, it, it's too, almost pejorative a term. To use for what has to go on it. It's funny how over the, I guess last few decades we've come to see training and development has almost a weak, weak terms for what has to happen. Hmm. And when the whole point of, especially of these kinds of environments, for anyone, quite frankly, for anyone, when you're operating at the limit of what's possible scientifically and technically so on, uh, um, is, is to develop. We have to develop, we're always developing, but we see that as almost a, some kind of, you know, it's got weakness around the term and around the connotation. Mm-hmm. Especially in an area like quantum. And you outlined it beautifully there in just the, you know, the conflict. There aren't a lot of people who've scaled quantum compute. No one has done it. Okay. Yeah. And, uh, inside big companies or small companies, everyone's doing it for the first time. If that's your kind of starting point, then what are you going for? Yeah. And uh, and you know, all right, so, okay, we're all doing this for the first time, so then what are the skills we need? Yeah. Give us a better chance of being able to do it. And we're all gonna have to learn, we're all gonna have to develop. Um, you know, do, do you think that young scientists, especially coming, you know, with, with outstanding Providence, those have emerged from really first class, uh, institutions, universities, and so on. Do you think they get that? Do you think they get that the journey a It has, it's gonna be tough. It has to be tough. You're gonna feel like, I don't know, you're back in school again because there's so much to learn. Your learning hasn't stopped or peaked because you got a PhD in whatever, from wherever. It's actually, you know, you're gonna carry on this journey and you're gonna feel done again. Do you think they get that,

Richard Murray:

um. I think my honest answer would probably be no. I, I could be more charitable in saying it depends. Um, it depends on a few characteristics, but I think it, unfortunately, I think sort of university teaches you that the physics is the hardest thing and the most important thing and everything else is just easy. And you, you can do it and what's the big deal? I could do that if I really wanted to, but it's not, it's not, it's not something that's really that hard. So I don't really need to try, try very much. Um, I think, I think I important to say that I think there are exceptions. I think that you have to sometimes try and spot the people who are less technically brilliant, who maybe have it. That's not the, the, the most important criteria. It is people who, and I could use a, a bunch of vague words that, um, that we use, which defy detailed description, but things like they're comfortable with ambiguity, um, maybe even more of a stretch. Their personality is one that is. Learning and putting themselves in difficult positions outside their comfort zone in order to learn. I think those types of characteristics, um, to be honest, unfortunately, are sort of sometimes trained outta people as they do a PhD. As a PhD, you, you are supposed to be the expert. You are not supposed to be uncomfortable if you don't know something or explain that you don't know something. There's a lot of, um, sort of maybe arrogance or expectation that you just don't sign up. Don't recognize those things as important. But I think if you can spot people like that, and we've got, you know, some of our best people are ones who, who really step into that. Okay. I, I've seen, and I've heard a bunch of stuff that I don't understand. I'm not gonna dismiss it and just be like, oh, I'm gonna go home and have, but I'm gonna really step into that and really try and understand and pick apart what happened today, why these events played out, why I didn't, you know, why, why there were so many iterations on this proposal that I was trying to write and things like that. Um, in order to. Build themselves into better people. Um, maybe coming back to one of the comments you made earlier, and I think it's important, the, the training bit isn't the hard part. Like we recruit the most intelligent people you can possibly imagine. They get stuff immediately. They need to know, they need to understand why they need to learn that stuff. That that's the harder part, the, the sort of the step into the training as if, if you like the explanation as to why you need to know this. Not because as soon as you, as soon as they know, they need to know it, they can learn it. You know, they can, they can sort of find, you know, they can do a training course like the rest of us, but it's, it's sort of, it's the wrapper around the training, the cultural piece, if you will, that I find even more important to, to explain to them.

Chris Reichhelm:

But is some of that training or development, you know, you are a very commercial organization and we're gonna build products for customers that they can use. So this is less about, again, you know, getting back to the original point, this is less about going for the end goal. Now it's a journey to get there. We may turn off at some point, we may find something even more interesting as we go, but now we're gonna build products for customers that they can use. And so we need you, depending on the role, of course, we need you to get out and get in front of customers, start having conversations and uh, and hear what they have to say. And, uh, and, you know, not take it 100% as as gospel. It never is. You've gotta, you know, thread the needle or in between the lines, however you wanna describe it. But, but there's all this other work out in the market that where, you know, and where, and identifying or kind of figuring out where that market is going, what that customer is gonna need now and likely to need in the future. How are, how what we're doing might align with that, how it could align that, and then getting to that point and being wrong and then coming back and doing it again and being wrong and, you know, until we get it right. Um, that's, that can be very difficult.

Richard Murray:

Yes. Yeah. For,

Chris Reichhelm:

you know, for individuals,

Richard Murray:

you know, for all of us, I, I definitely agree and really the essence of this is really something that's hard to learn. You reading people, reading situations, so for example, your customers saying one thing, they really need this one thing. You may might think that they're wrong, but working through that, in a world that's full of ambiguity, there are very few certainties. Yeah, there are very few wrong answers, but there are probably answers that are more right than wrong. Um, this is a world that's really, I mean, let's be honest, it's really hard to navigate. All of us spend our whole careers learning how to do this, how to read situations, how to adapt to c to accommodate certain scenarios and, and feedback and things like this. And I think what's bring, what's the most challenging thing about quantum computing is it's, it's so much more, it has to be so much more accelerated because, you know, the need for those people is now, you can't wait for people to sort of spend their whole career to figure out how to, how to read a customer signal or how to interpret something that someone said, which is a bit ambiguous. Yeah. Um, yeah.

Chris Reichhelm:

So what are the key characteristics or features? Can you describe, you know, what are the key characteristics or features of talent when you're, you know, you know, getting away from the technical for a minute, you know, because I appreciate that could change depending on the role, but can, you know, what does an arc of computing person look like?

Richard Murray:

Um, we often, by the way, I'll say a bunch of stuff, which my recruitment team always push back. We can't say that it's too ambiguous. It's like, what do you really mean by this? But, uh, we quite often use the phrase comfortable with ambiguity. Uh, as, and again, I'm always, there's always pushback on what, how do you test for that? What does that really mean? But comfortable with ambiguity. Also, I think, related to that very strong sort of problem solving skills. So you notice the point of ambiguity, you know, experience, track, record of finding maybe unique, maybe resource constrained solutions to those track changing problems. Um, our, and, and, and I mean, and then slightly different from that, within a different set of sales, sort of commercial awareness. I mean, that, that actually comes third really, if you can have those two things first. And then also commercial awareness. You know, it's rare, but, you know, we'd be doing really well, but some sort of person who at least is willing or able to understand, um, the, the commercial environment, the commercial side of, for example, what a customer will be looking to go through in order to say purchase one of our systems. Yeah, so I'd say those three probably in, yeah, various different orders, probably summarize. Outside of just, we need a quantum PhD or someone who understands quantum optics. Yeah. Some of the things that we, we strive to, to get from, from our team members, um, in, in what we do. Yeah.

Chris Reichhelm:

Yeah. Richard, this has been hugely enjoyable. Um, thank you so much for, for spending this hour with, with me and us. Um, I think the, where we, where we draw that line, uh, when we're on that lab to market journey, trying to get to that ultimate destination is, is, uh, is really important. Resources. What I've taken away is this idea that resources are going to, uh, have a huge impact on how much, on our adherence to that goal. Uh, and that's a good thing. Um, you know, these constraints are a good thing, uh, and they, and they can lead us to other destinations. I imagine that, you know, the, uh. Communication with the different stakeholders along that journey. The one investor thinks she's signing up for, because I thought we were building this. Oh. And now it looks like we're turning off and we're gonna to do this, but hold on, my portfolio's gonna be outta whack 'cause the valuation changes and, and so on. You know, those kinds of things. Um, I imagine that kind of communication is really important as part of that journey, but that's, that is the nature of a deep tech journey. I think.

Richard Murray:

I totally agree. And by the way, it's what makes this whole journey that much more fascinating. That's it. I mean, personally, I get a real kick out trying to solve these problems that we talked about today. Yeah. It's not a technical problem. It's a human problem and crack it and motivate your teams to work in the right way. And, and that's the answer to being well part, a big part of the answer to be successful. Yeah.

Chris Reichhelm:

Richard, thank you so much again. Great talking to you.

Richard Murray:

Thank you and you.

Chris Reichhelm:

You've been listening to the Lab to Market Leadership Podcast, brought to you by Deep Tech Leaders. This podcast has been produced by bowhouse. You can find out more about us on LinkedIn, Spotify, apple, or wherever you get your podcasts. Casts.

Head Shot

Chris Reichhelm

Host
Head Shot

Tamar Zak-Collins

Producer