Quantum Photonic Computing in Sheffield, with Maksym Sich, Aegiq
Hello there. Thank you for tuning in again. This is episode 24 of the quantum divide. I'm very excited to, to bring this episode out. It's one that's been baking in the oven for a while.
Dan:It's an interview with Maxim Sich, who is the CEO and cofounder of AJIC. AJIC is a UK based quantum technology company. Its origins are in the at the University of Sheffield. Max is the CEO with a couple of his colleagues running the business. They're well known for their products such as Artemis, which is a photonic quantum computer.
Dan:It's a modular setup with integrated photonic circuits. It's reconfigurable and so on. So we'll be touching on that, of course. I just wanted to say also a little bit about Max's background educationally. So he, first of all, did a he did a bachelor's and a master's in applied physics at the University of Kyiv, Ukraine.
Dan:Then he went on to do his PhD in physics at University of Sheffield and just casually happened to do a bachelor's in economics and finance at the London School of Economics at the same time as you do because you've got loads of time to kill when you're doing a PhD. And that kind of sets the scene, really. For Max, it was a great conversation. Hope you enjoy it. Okay.
Dan:Thank you, Max. Thank you for joining us. Let's start as I always do, because I wouldn't want to be unpredictable. Give us a bit of, information on your background, your path into quantum. What's your story?
Max:Thanks, for having me. It's really exciting to to be here. It all began when I started studying physics at school. No no big surprises there. So I did a a degree in applied physics in Kyiv.
Max:So I'm originally from Ukraine, and that involved quite a bit of quantum physics. That's where the interest started, and we just started hearing all the exciting things about terms like quantum dots, quantum wells, and all the exotic stuff being made at the time. So I continued. I then moved to Sheffield where I I did my PhD and also was a research fellow. That's where the UK National Quantum Technology Programme started.
Max:So we've been there from the start.
Steve:And as
Max:you can see, that's pretty deep immersive, thing that happened here in the quantum for me.
Dan:Very nice. So what what drew you to Sheffield? Was it the was it applying for the PhD and getting into the the faculty there?
Max:Yeah. So it really aligned to something I was very interested back when I did my masters. So semiconductor physics with quantum effects on it, And Sheffield is the National Centre of Excellence for Research in Semiconductors in in the UK. So we have National epitaxy Facility. Well, have the Sheffield group led by Morris Skolnick, whom I joined and which is now the core of the Sheffield Quantum Center.
Max:So they've been working on these topics since early nineties. So we recently did a little tracker of all the funding that went into this, all the all the grounds. So it's just little bit short of 100,000,000 that was invested in the whole activity over this over these years, so that's a lot. And National epitaxy Facility, out of that, took about 60,000,000 investment into the equipment, the facility, of course, people who who are working and doing research there. And those are just the major investments that have been made.
Dan:Yeah. Big money. No no doubt. Yeah. You've mentioned semiconductors.
Dan:What about photonics? Let's go on to your company a little bit.
Steve:Oh, wait. Can I inject this before I
Dan:move? Steve, go ahead.
Steve:Yeah. So you had an academic background, it sounds like, but you must have gone into some entrepreneurial something happened. How did you go from academic to entrepreneurship? Was there something that happened in between your PhD, master, or how did that happen?
Max:I've always been interested in entrepreneurship in general. So I also have a degree in economics and finance from LSE. So I did that in parallel. That was my physics studies. So they it was pretty simple.
Max:So one day I just came to a conclusion it's time to change the career path and I did. So I started with a few other things, like cryogenic microscopy, a company I had. When that very quickly went for us to see the potential that the IT that coming out of the Sheffield Group had. And so that that was the beginning of the story, Regik.
Steve:Because it's generally it's the training you get at academic institutes doesn't really prepare you for running a business. I'm just I was thinking it must be there's something missing. How do you go from PhD student to running a company? It's tricky.
Dan:Sounds like the way to do it is to do a master's in finance and economics at the same time. As if doing a PhD on its own wasn't hard enough.
Max:No comment.
Dan:Okay. Yeah. So you mentioned AJ. Let's go on to that. Tell us a bit about how how the company started, and how did you end up focusing on photonics?
Dan:That's the way I see looking at your product, looking at your photonic content computers. There was obviously a path to get there.
Max:Sure. No. So we started EJEC because we knew that photonics is a really important linchpin for the technological map and and ecosystem in general. So it's got multiple users, and it underpins work they can do and a lot of devices and and methodologies they can do and metrology. So sensing, then you're talking about communications, so that's networks and networking as well, and also computing.
Max:So you have these three key pillars which really rely on good, photonics and on good quantum sources of light, if you're talking about quantum real for these things. So that was the the very early, understanding that we had. And in terms of sensors, you might say, hey. We have gas sensors, stuff like doing, for example, and many others are doing different things. Communications, obviously, QKD is there, but there's more to that than just QKD, and there's obviously quantum networking ahead of us and, of course, computing.
Max:In our view, photonics is fundamental to any to enabling any type of actually large scale quantum compute in any case. We're talking without, like, deep specifics here on how exactly you interface things, but if you want to build quantum links between computing nodes, photonics is your best bet. So that was the the starting story. We started with a a very fundamental discovery that an engineering solution that came out of the university in terms of very efficiently generating quantum states, was in photonics. And that's something that was missing on the market at the time.
Max:So they can interface devices at telecom wavelengths or you can actually do different wavelengths. And from there on, we went on to quite a heavy product discovery, customer discovery, which led us to conclude that the 2 themes we're currently working on are computing and networking. There's an element of cybersecurity to it.
Dan:Yeah. Makes sense when you when you put it that way, photonics is really central, honestly, to to, I guess, holding quantum state with some kind of asset, the photon, which can be passed around and manipulated in many different ways.
Max:Person think about speed of light. The reason why speed of light is the fastest thing we have is in fact related to the fact that photons retain their properties for really long time. But unless you have, of course, extra impacts and so on, so it's not without challenges to say that. But for long distance transmission, and long distance in quantum means even meters. Right?
Max:That's that's the best bet that you can possibly find.
Dan:So, Max, what products have you got on the market at the moment?
Max:So at the moment, we have our single potent sources as chips, but also as a complete turnkey solutions. So one of the big challenges generally in quantum market is lack of easy to use systems and devices. So people have done a huge ton of work on demonstrating different phenomena, different properties, but it all comes down to being able to have a a simple box with simple interface that you can just plug in into whatever you're doing as an OEM, if you're talking about manufacturing, just even for research. So we started there with, complete turnkey solutions for single photon sources. We are the 1st company really to bring it to the market and to integrate with fiber, which is robust, repeatable, doesn't break with time, doesn't deteriorate.
Max:And that product actually underpins our quantum computers that we've just released earlier this year and sold to the National Quantum Computing Center because the efficiency of this device, of the quantum computer, the Artemis, relies heavily on the source of photons, that we have. And that those 2 are fundamentally there together to have the very high efficiency.
Steve:And those single photon sources are based on quantum dot technology? Correct.
Max:So those are quantum dots in semiconductors. Speaking of dots, you can see how the dots are getting connected with the with the quantum dots that have been very exciting scientific topic, let's say 20, 30 years ago to the time when we actually are bringing them onto production now.
Steve:Okay. That's interesting because I think as far as I know, I don't know much about quantum dots, but I know it's the cutting edge of single photon source. It's very difficult to manufacture 2 quantum dots that have the same properties. Have you experienced
Max:I didn't say it was easy.
Steve:Yeah. Of
Dan:course. Connecting the dots. I like that. Max, I'm gonna use that, store that one in my back pocket.
Max:The quantum dots. Connecting the quantum dots, sis, this is the thing that we do.
Dan:Exactly. Yeah. The next thing I wanted to ask you about was integrated photonics. Could you explain that in layman's terms overall, what it is and how it's different to just photonics?
Max:And
Dan:is it a core part of what you do? And if so, could you describe that in a bit of detail for me?
Max:So when people say integrated, it usually means making things on chip, like a microchip. And the analogy there is, say, it was electronics. So you have integrated electronics, and all of the CPUs, GPUs, all the all that stuff is integrated electronics in that terminology. So you could obviously have separate electronic components as as we had in many devices, that would be just electronics. Same with photonics or optics, which is the same thing.
Max:So imagine you have a telescope that you used to look at the stars and and and the moon. So that one is a photonic device, but it's got lenses. It's big. You can hold it in your hands. Now once you put it on a semiconductor chip with all nano elements, micro elements on it, and you have multiple different functional blocks inside of it, say waveguides, maybe lenses, maybe face shifters, you name it.
Max:But it's one functional block, which is pretty complex, but on a single chip, it becomes an integrated photonics. And the benefit there is exactly the same as with electronics. When we moved away from lamp radios and the huge massive devices onto chips, everything became a lot smaller. And really important element is became a lot more energy efficient. And this is the cornerstone for usability in our real world applications.
Max:So those two things, small and power efficient.
Dan:That's really helpful. Thanks for going back to basics. What would be really useful actually if you could extend beyond that into the the inner workings of your photonic computers. Right? I assume at this point that you're using photons sent through your chips, which are manipulated in different steps to perform gates and so on.
Dan:And then there's a measurement at the end. Right? This is my very simple view.
Max:That's what that's how it works. There's from conceptual functional blocks of a computer, a quantum computer, a photonic quantum computer, it's quite simple. So maybe as a little detour here, there's fundamentally 2 different kinds of, groups of quantum computers that you can make. One is that we know of more commonly, which includes superconducting computers, trapped atoms, trapped ions cold atoms. Sorry.
Max:Those are qubits which are static. That is you created in a particular point in space with, say, having a cold atom in a particular state, and that's your qubit. And you and you test there is to maintain it, perform operations whilst it's there, and then read out the the information. But that is done on single point in space, and you separate it in time. Photonics, on the other hand, because photons are always moving.
Max:So they always go somewhere. You cannot make them stop. You can put them moving in the circles, but nevertheless, they keep moving. So you create your photons, which are effectively your cubits. And it it moves, like, from one end to the other of the set of the compute system, similar to how you actually do it with electronics.
Max:So you have you create your signal somewhere, it goes through the CPU, and then it emerges somewhere in the monitor or is it you know, whatever the output device is. So here, we create them. We route them towards a quantum processing unit, which the routing device is an integrated photonic device, and then the quantum processing unit, again, is an integrated photonic device. And it consists of 100 of elements, which are wave guides, so light guides. So it navigates, makes the photon go in the right direction a little bit like wires for electrons, but they are, like, really tiny.
Max:Then you have devices which made them interact, and change some properties such as phase of the photon. And then on the output, you have photon detectors, to read out so your binary result of the computation. And on that large interferometric and phase shifter array, you can use you can actually encode different types of quantum computation. So you can encode gate based model, That is, you take those gates that you've seen everywhere, like, you know, Hadamard gate, c naught gate, and so on. But you can also encode analog computing modes, which are not digital in that sense, but nevertheless are very powerful in certain applications.
Max:For example, you can do sampling, such as boson sampling, or you can do variational quantum eigensolvers. So all those algorithms, they rely on heavy digital post processing compute. So it's a feedback process that happens. So you do one computation with Quanta, and you verify the results. You adjust your parameters, and you keep going until you arrive into, say, an optimal solution.
Max:And it's an optimal solution by checking it. And, typically, there's a range of kind of NP hard problems as they're known, where it's very hard for a digital computer to suggest the solution, but it's very fast to check that's the correct solution. So you know when it's right or wrong, but it's very hard to offer, the thing. And this is where the quantum computing comes, in power. It can offer come up with that that offer much faster.
Steve:And in the quantum photonic based quantum computer, it's very important to have these cluster states. Oh, let's see. Maybe we don't have to go that far yet, but the noise source in a quantum computer must be different than in one that's a superconducting base or ion trap.
Max:Yeah. 100%.
Steve:So the air correction techniques probably are diff of course, they're different. But what is the actual source of noise in the photonic based quantum computer? Is it the wave guides manipulating the polarization? Or what what's the source of noise, let's say? What do you have to overcome the most in that type of
Max:So the main source of noise and errors, let's put them in photonics, is loss of photons. Because as they travel through whatever medium, unless it's an absolute ideal vacuum, you're gonna experience loss. So whilst they travel, they maintain their properties really well, and even if they change, you can always track it back, quite easily to correct for that change. So it's not exactly a noise then, but losses are the dominant and pretty much the only source of errors that you have to deal with in electronic systems. And that actually is the reason why things like QQD systems have got limited range instead of unlimited.
Max:And, actually, our digital communication systems also have limited range, except there is amplifiers that just bump up the signal all the time, say for, like, transatlantic communication or for undersea cables. It's much harder to do quantum amplification in practice. So, hence, we are observing what we have now, what is limited distances. In fact, a good source of entangled photons, deterministic, or single photon sources, as potentially one of those keys to unlocking effectively amplification. You it's not amplification from physics perspective, but let's say translation of the signal over the longer distances.
Max:So that's also the bit where our networking longer term is we've got an interest in networking longer term.
Steve:Yeah. And that's a focus of the company to produce high quality, air correctable signals, or is it something you don't need immediately?
Max:You you need it, and there's multiple facets, multiple elements to this. So one of the key ones creating states which are very good, high quality, so you have more redundancy in in whatever you're doing after this. But for, yeah, for quantum computation, again, that's the same thing. So you have to deal with and you have to find materials which are the lowest loss. You have to find ways how to interface different chips without introducing loss, and that's virtually impossible.
Max:So you have to minimise that. So there's a number of engineering challenges which are still there in terms of reducing that that kind of error rate. And was that's our focus of our companies to roll out the technologies that will be absolutely leading in terms of their efficiency. Because then we're talking about much smaller devices being very performant, requiring a lot less error correction on top of it, so a lot less computational overhead. And in our view, that's the route towards unlocking, like, proper commercial value.
Max:Because if we're talking a location and a data center, in our view, quantum computers don't have to necessarily be better than your biggest exascale data center. They just need to offer better compute per watt of consumed power than whatever you have, as a GPU cluster or a CPU cluster. It doesn't matter too much. And once we hit that threshold, this is where real value is unlocked.
Dan:There's different compute modalities that you mentioned, the classical and and quantum. They're solving different problems quite often. Right? So it's not it's, I guess it's the power is super important. Sure.
Dan:Because of the huge amount of power consumed by data centers. And that leads me on to the NQCC. So congrats on your award for that.
Max:Thank you. It's been very exciting for us.
Dan:Yeah. I'm sure. Could you like, talk about the process a little bit about achieving that and what it is the requirements are that are being set upon you? Is there anything you can share about the what the NQCC are doing going forward with all the different computing platforms that they're gonna be building?
Max:I guess the the NQCC knows a little bit better and what what they're after there. But our understanding is NQCC the role of NQCC is to become the center of excellence in the UK and actually globally on advancing quantum computation. Therefore, they want to have a full toolkit at their disposal for research purposes, for educational purposes, and actually for commercial purposes. So they're gonna sit there with interfacing and bridging the gaps between industrial users or commercial users, between academic research and simply education for students, and actually doing a little bit of their their own research. Therefore, they wanted to have different types of quantum computers, different modalities so that they could put it there.
Max:They and after that, they can also maintain that capability so that they could benchmark them, for example, against each other, which is a super important and exciting idea for the sector because nobody's really put together 7 different devices in one room in a public setting or publicly funded setting. And almost, like, offering it as a resource to the world. And this is an incredibly important role that they're playing there. Because the question of how do you compare all these all that zoo of different modalities is a very hard question to answer. Because you we started already a little bit saying, hey.
Max:This is static qubits. This is flying qubits. How do you com compare gate based performance on these 2 if they're so different under the hood? And the answer is that's not trivial. You can come up with numbers, but there's always a star and a law of small print associated with it, what that means.
Max:So what you ought to do is come up with the tests, say, okay. Let's run these different types of algorithms, and we'll see how they perform in reality, like the whole system. It may not necessarily be the qubit problem. It might be actually the control problems or other things, but that's what matters in in the end. It's how well does it perform in when it's fully assembled and it's just thrown a task on it.
Max:And so what are the metrics out of this? How much power it's consumed? How long it took? And how big the problem it could handle? So that all is very interesting things that no one had fully demonstrated, on a side by side comparison.
Max:Yeah. We're super excited. We're part of this because that's how we're gonna form the common knowledge about quantum computing.
Dan:I think they've got a lot of work ahead of them. Right?
Max:A huge amount of work. It's very ambitious. I'm glad they're doing it, but it is will be very challenging for them as well.
Dan:Yeah. Yeah. It's gonna be fascinating to watch from the sidelines.
Max:Well, I mean, maybe maybe we can play with something there once it's up and running.
Dan:Yes. That would be good. That would be good. What are the downsides of, photonic computing? What are the challenges?
Dan:You've mentioned the the noise, but almost every modality has its own idiosyncrasies and challenges because of the way it's using the laws of physics. What is it with photonics that people say neutral atom can be quite slow? I imagine photonics is quite quick because you're traveling at the speed of light, but that must bring other problems as well along along with
Max:it. Yeah. So it can be very fast. It brings the problems of synchronization and actually quite a lot of other devices such as detectors, switches, which are good for quantum are not fast enough yet. One of the bigger challenges, of course, is that it's both.
Max:It's a blessing and a curse in a sense because with photonics, we heavily rely on decays, basically, wave investment that went into telecommunications industry. So all the fiber optic switches, transceivers, and all that ecosystem, They've been investing a lot in the manufacturing base, which is what we're leveraging, quite heavily. But it also means that we're dealing with a very complex process from a supply chain perspective, from manufacturing perspective. It's very demanding, And that's not as easy to manage if you're talking about discovery processes, discovery projects as with others. Me also talking on the limitation of what you can achieve per chip.
Max:Because while photons fly fast, you need to include a lot of if you want to build, say, 1,000 or millions of cubits, you've got to include appropriate amount of functional elements to all those beam splitters, space shifters that I mentioned, and they take space. So one of the potential bottlenecks there is you run out of, space on the semiconductor wafers faster than you would like to. So there's a limit to how many effective cubits or effective devices you can implement for some of the applications. There's also obviously a question, which still sits on quantum state generation for how long you can maintain the status that all the sequential photons are of a good enough quality to build up the computation. Because you need to have multiple qubits in order to do anything sensible, and they have to be in that really perfectly identical state when they when they create it.
Max:And then you can implement really good devices. So you still have things like noise and the decay of it, kind of long term coherence in these devices. So those challenges still remain. To a large extent, they're engineering challenges, so that's a good news for us because we know that fundamentals are pretty good. But there's uncertainty as to how much effort needs to go in the end into those problems before they we can crack on a really large scale compute.
Steve:One question I've always thought about and never got the answer to just probably because I didn't look it up, but how universal is a photonic quantum computer? How do you program photonic quantum computer? So in my mind, you have this solid you know, this chip. It's fixed. You send the photon in.
Steve:It can do one, calculation. Or not not one photon, many photons, and you have this one calculation. Can you modify the chip dynamically, or how much dynamicity how dynamic is the chip? How programmable is it? What's that like?
Max:Depends on how you build the chips, but they're all programmable on the fly. So you send a problem, you encode it. Again, depending on the type of computation you're doing. You're doing gate based, type, or you're doing those in a variational methods. So that implies slightly different logic of operation.
Max:But, normally, you'd repeat the same experiment a couple of times or maybe more than a couple of times to, get rid of the kind of noise and error impacts, and then you need to reprogram it. But it's reprogrammable on the fly, so you don't have to do anything, really. And your interface is the same as with other types of computers. You write code in Python, and you say, hey. This is the gate or that's the function that I want to implement.
Max:Then, of course, there is a low level control software that goes into this, how you actually control every single and tweak every single subcomponent on the chip. But the procedure is pretty straightforward. So you build it, then you have a particular calibration routine for your device so that you know what one of those 1,000 different electrodes that or electronic current electric currents that you or voltages that you applied to chip does to the output. Once you run through this calibration procedure, you're ready to go and program it.
Steve:Okay. So it's not about just selecting the path within the chip to program it. It's also there's some external magnetic fields.
Max:Yeah. So part of the programming is how is the path of the photon or or not so much the path of the photon. So we don't really control the task per se. What you program is the impact on the photons on the chip. And then it's got some probabilistic nature to how the pass changes.
Max:Now you don't necessarily need to know what it is, but what matters is the output measurement that you achieved was that input or change in in in the voltage. And that's the actually, the difference between digital compute and quantum is that with digital, you actually can know precisely the position of your bit or the electric signal at every point in time. In quantum, you don't need to know this, but you treat it more a black box of of a kind, some properties of that black box. And so you have, say, a 1,000 inputs, control inputs, 20 or 32, in our case, quantum inputs, and they quote a number of outputs that you measure. And the calibration consists of ensure no.
Max:Given the known input, if I apply a particular set of control parameters, what's the output? What happens inside is it's different to how it operates, with a digital signal. But it's it nevertheless, it that that's the actual nature of quantum compute. With gate based operation, again, that you dedicated a particular area of the chip to implement a gate, and the gates themselves are sometimes probabilistic. For example, if we're talking about NIST type implementations, that the gate is gonna work once and 2 times or once in 4 times or 1 in 10 times, depending on what type of gate it is.
Max:So we know when it worked, and we count that result, when it worked. For more advanced or tolerant type compute, those can be removed. But then, you it is removed because you're measuring errors in in almost real time and correcting for them.
Dan:I love the fact it sounds like you're preprogramming the the chips or chip, and then just letting the photons go and letting them, letting them do their thing. It just, and you're trusting in the way that, that they're gonna take particular paths and have some particular behavior impact affected upon them. The one one outstanding question I've got is in order to perform calculations and gates and so on, you need multiple cubit source cubits to come together for manipulation or to operate a gate on them together so that they interfere with each other and then and so on. I assume that's all done through very precise timing mechanisms. Is you know, how do you ensure that cubit x and cubit y get to the right place on the chip at a certain point in time to interfere with each other in the way that you need them as a gate?
Dan:And what kind of what kind of accuracy of time is required for that?
Max:A very high accuracy as the as the answer. So you're talking sub picoseconds, actually femtoseconds even in terms of what what you need to be doing. So that again comes from developing a manufacturing process where you build up device integrated chips, which have got exactly precise path lines, exactly precise changes to how they impact photons. And once you can manufacture this, yeah, you you can do it. And with photons, because they travel very fast, a tiny delay or offset of delay is actually a microscopic distance.
Max:So, actually, correcting for this, those, like, external devices or larger devices is quite feasible. So you can reach extremely high precision without resolving to manipulating individual atoms on the chip, but you're rather talking about things like of the lengths of 100 of microns or even millimeters. Now that's already not even a none nano devices. That's bigger things. And to create those photons, we use right now, we use one single device, which is a single photon source.
Max:So it definitely generates identical photons out of it. And we convert that time series into blocks of cubits that come into the compute device. So it can demultiplex it from a single stream onto a parallel stream or multiples, and they come at just batches, basically. And it's all driven by a very stable you'd call it a clock, but it's just a very good pulse laser, which is the, in fact, the most stable clocks out there. So if you dive into industrial standards for, like, stability and you might have come across things like frequency comms and so on.
Max:So these are just very stable clocks, which are made on optical lasers.
Dan:Got it. Yeah. I'm boiling it down to quote, unquote simple calculations of speed, distance, and time. And obviously, it's not that simple, but, yeah, that's a good way of looking at it.
Steve:So we spent a lot of time so far discussing the computing aspects of photonics. There's also the path of networking, either networking devices or large scale networking, quantum devices. How does that, domain fit into your roadmap?
Max:So to build large computers, eventually, you'll have to have a way to link them natively with a quantum link. So if you have multiple QPUs, in order to effectively have a large compute system, not multiples that just operate in parallel, there has to be an entanglement link between them. So for us, that's not the thing that we'll be doing separately. It is just embedded in the process of how we make things. So, from from that perspective, it's not a different domain or paradigm for us altogether.
Max:It's just part of the, fundamentals of the design decisions that we've made. So that's that's an important one. But the kind of intermediate solutions that that we looked at and we worked quite a bit in the communication space, it's also considering we're now looking how we can help people adopt networks which are good for quantum. So that there is a lot of activity and interest in that area as well. And, yeah, you might have seen we've signed a memorandum of understanding with Honeywell.
Max:So we've done some work in the past on modeling, for example, quantum networks. And our view, if this technology is going to quantum technologies are going to make a difference, they have to operate everywhere. So they have to operate deep under the sea, obviously, on the ground, in your laptop, in your data center, on in air, in space, as digital does. Our digital technology works just fine on Mars recently. Quantum has to have the same ability.
Max:And for us, it's how do we unlock that keep that interest? How do we unlock measurable sort of KPIs that come out of, for for this? And that's why we started working with Honeywell. As they say, there's there isn't a thing that flies on Earth. It doesn't have a Honeywell component in it.
Max:And they have, obviously, huge interest in advancing space communications, optical communications. And here, our expertise in understanding how light propagates, how we talked about losses, so all different mechanism of losses. What impact it has on the quality of the signal came, came to use because on the other hand, they have an extensive set of technologies for optical communications and free space. So for us, it's a great synergy because we can bring some of our technology, even chip technology to them to implement the more global networks. And even in log business decision making, because a lot of the challenges currently that we're seeing, they sit on the fact that it's very hard to come up with a ROI type judgment for deploying quantum technology.
Max:So, usually, it's very qualitative in sense that, yes, we need a better security, and then it's gonna cost something. But how do you make it into a commercial operational system? Or how do how do you fit into your business process, is a very hard question. And so that's actually where we focus a lot there is unlocking that understanding in 1st place. How do you use things?
Max:And one of the things that came out of this is the, like, our ATLAS toolkit. For example, you can go and model different types of satellite networks. So we did, for example, a case study of how do you build a resilient quantum network or, you know, key native bases in Europe. So what's gonna take and that you can do it with this system. Then you can couple it with Honeywell's metrology systems for optical cons, and you're gonna get, like, real time data for a particular location.
Max:So you can actually estimate what it what type of devices do you need to put there, what how many, what kind of satellites, what kind of orbits you can use, and attach numbers to this in terms of budget that that you're looking at. And that's the first step for unlocking proper business value in communications, really.
Dan:That is super fascinating, the combination of those two sets of technologies and perspectives. Right? Looking at the atmosphere and and the weather and building that into the potentially grand to space optic communications, which is optical communications, which is does seem to be a hot kind of area of development at the moment in lots of different areas. I know some of the LEO constellations are testing it, and there are thousands of inter satellite links already using lasers. It's just obviously a lot more complicated with phase shift and weather and all kinds of other atmospheric issues that you get when you go to the ground.
Dan:So, yeah, that's fascinating to hear. Yeah.
Max:Yeah. And exactly. Because with all those effects, you still need to make it find a way for, say, global quantum network to be operational. Because without that, there is no real scale that we can achieve, and satellites are a critical element there. So they will work sooner than we'll figure out how to make that under the water is my strong conviction.
Max:So that's why we have to look into those areas and find and try to find synergies where we can bring that to life.
Steve:Mhmm. I think there was a paper a few months back at submarine to submarine, QKD, a 100 meters apart. So I agree. We've seen a lot better than 100 meters in satellite already.
Max:Yeah. And, QKD is just the one aspect of this. Right? So it's the more mature technology and clearly, it's got its own value proposition there. So we're we're happy to help with that.
Max:But longer term, of course, we're looking What I want to see is, you know, how do we build a quantum data centers in different places? How do we share information in a quantum way? Because entanglement distribution is the way to go. And you can make an argument that the reason why quantum computation is so much faster in some ways than digital compute, quantum communication can be also faster than digital communication with same amount of, say, cubits versus bits, just by pure virtual storing more information and in less number of of programs. Now that's a little bit futuristic, but I think we need to see that there's a path towards that.
Max:And part of, say, what we're doing is our all our devices are compatible with telecom wavelengths, so old fiber comms that we currently have. So we can put our devices onto the same very same networks. And we also see that there's a strong shift in optical comms to kinda can get onto that same communication spectrum. So then it becomes cross technology compatible, and you can really build out those things. And it actually is closer than we might think.
Steve:It's funny that you say the faster rate on the quantum. So my studies in PhD and masters was about classical communication over quantum channels. And something that is interesting, I know there's theoretical results and even some experimental results that show you can do it. But the, the main challenge is being able to measure highly entangled quantum states. It's a very challenging problem.
Steve:It's very interesting to see. Okay. Someone is actually thinking in a commercial setting, classical over quantum, You can even do quantum over quantum, of course, but the classical messaging can be faster. You can maybe not in terms of speed of light, but in terms of rate per transmission or bits per transmission. Yeah.
Steve:Yeah.
Max:And all those things, they really said on qualifying and quantifying the quality of the channel. If you don't have that input, that doesn't it's hard to do anything else.
Dan:Daft question moment. I was reading about. Yes. Which is let me try and explain it. It's using a multidimensional mathematical space, Hilbert space, to model more information onto a single qubit than you would using normal three-dimensional kind of mathematics.
Dan:Is that what you meant by it by by increasing the amount of data per transmission or jumping the gum?
Max:No. I think we're still talking, we were still talking cubits. And I think QDETs is a very interesting concept, but it still needs some time to learn to something because we're struggling to deal with cubits. We pretty much have a binary state that exists in the superposition itself have a continuum of, outcomes in those two coordinates. Now if you have multiple energy transition or multiple variables within a qubit, they can exist in that superposition and also in in some way in superposition between themselves.
Max:Indeed come up with this come to this, multidimensional space. Now you struggle to read out and control the binary thing. And my view is shouldn't really those, higher dimensional things are exponentially harder to physically make. So that's my take on it. But the in terms of information, it's similar to the notion that you can, encode in the same tangled, states as a little bit larger potentially than just that kind of binary.
Max:And so allow you to to put more information in it. Or you can keep it purely in, in a quantum state, and that would be immune to interception just by definition of it.
Dan:Yeah. So sending coherent states across the network basically in a superposition.
Max:Yeah. So there's a there's a range of different ways you can build those to make them work, even to make good Cool.
Dan:Thank you. I'll forget the cubits for now then. Stick the cubits. Yeah.
Max:Yeah. Safer.
Dan:Okay. Yeah. The Honeywell collaboration sounds very interesting. I'm looking forward to seeing, maybe I don't know whether you're able to share at some point in the future, the outcomes of that at a high level, that'd be really interesting to talk about.
Max:Definitely, there will be something. I think when the time comes, we'll make those announcements, of course, but, there there is a lot of interesting going on.
Dan:Did you wanna talk about your 7 point plan? I picked it up from your, I think it's from your website.
Max:Yeah. We did a submission to parliamentary inquiry in Quantum Technologies last year. So they were asking about basically anything that government has has done in terms of spend on quantum. Is it a range of questions that they had? But on top of that, we put the 7 point plan that kinda summarize the review in a more condensed and focused way of key challenges out there and what needs to be changed.
Max:And so those are probably 7 more important areas that we felt the government needs to be aware of. And not everything they it's easy to address for them or cannot be addressed directly, but indirectly.
Dan:Yeah, I guess that was a little time ago now and I guess what's happened since is the announcements about the centers of doctoral CDTs, which are gonna be stood up.
Max:Yeah. So they also announced the next 10 year ongoing program. Now that is still lacking detail to it, but nevertheless, the commitment is there, and, we can be pretty certain about that. The logic in broad strokes is also very much understood. So they were it's not to say that the government was not doing anything, quite the opposite.
Max:And this is more towards saying, hey, this is the areas where we also need to pay attention because the ambition they would become put us, you know, behind superpower. And I quite like it one way or the other. How do we turn the UK's economy into a more innovative and something that still drives innovation on a global scale. So that's how you can survive as a smaller country in a global context, but still maintain that weight, in a global context. And those points were there.
Max:And the viewpoint on CDTs, Centrist's PD doctoral training, it's a great thing. We we like it. So we've supported a couple of those. So CDTs are mainly focused on so that they get PhD students going through them. In the past, the challenge for CDTs was sometimes employability of PhD students because you create quite a big number of highly trained educated people.
Max:Doctorate is is no small feat, but there's not enough academic jobs for them. I mean, after PhD, you're really qualified to become a researcher and progress your academic career. Now there's a huge, you know, mismatch in terms of how many PhDs you create and, like, how many researches you do. So they need to go somewhere else. And this time around, I think there was a huge focus on industrial careers, after PhD, and, essentially, that's an advanced training to take more challenging roles, within private sector after you graduate, and that's part of the deal there.
Max:So I think that was a really good and welcome thing. So I'm very much supportive of of that idea. And Part of the thing that still needs to happen is kind of general, awareness amongst, broader picking up quantum technology in their school programs, understanding what it is. Okay. You're not gonna be solving the hardest quantum problems, but you're not solving the hardest computing problems either when you're learning to code.
Max:So a comparison. So let me just, look it up, but Apple has got 300,000,000,000, in revenues last year and had have had about a 100,000 workforce. So a good quantum technology industry would have to be better than that. And it also tells you the scope and the scale of how many people will have to be employed in the sector longer term. So you're talking 100 of thousands of people who will have to be some of them would have to be really deep into, like, hardcore cutting edge scientific development.
Max:Some of them will have to be and the majority of them will have to be focused on engineering products. And there will be a lot of other jobs which will be dealing directly with quantum technologies that people need to understand how they work on broad strokes, the implications, the position in the sector. And you're talking, you know, lots of people who are doing nontechnical, as in product development work. So talking about negotiating contract, sales, business development, customer success, but even production, manufacturing. So you need to be aware of what you're doing in order to do good quality work.
Max:And so all that needs to happen, and if you don't have that as part of the curriculum, and so it was like, you know, conversation at the kitchen table with your kids that, hey, this this quantum thing, It's not going to be easy for us to scale.
Dan:Yeah. You're absolutely right. There's a whole ecosystem of functions in a in a company and also companies, in a country, you know, in order to support an industry like that. And that's why I liked UK universities must take a longer term approach to commercialization. That's an interesting one because, yeah, you said the highly trained PhD students, They're either going to an academic role or they get shuffled into industry somehow.
Dan:But, actually, what about the long term commercialization development of things which is developed inside the university or partner universities or so on? Right? Do you see anything happening there in in changes? And, oh, you know, obviously in the UK, there's bodies like innovate UK and everything, which is the core of that. Keen to hear your thoughts.
Max:So there's still a huge amount of things that need to happen to universities or as as a whole, UK and actually many other countries, as well, to enable that transition to the knowledge base economy. So if we take a state step back a little bit from quantum technology specifically, the big market shift right now that's happening is moving away from companies I mean, it's not completely away, but that companies that are purely kind of digital type companies. Think of, Google when they started. It was a purely digital company that enabled better search. It was massively better search.
Max:Or things like WhatsApp or Instagram. So that that would call them digital companies who sit on on this technology, towards more varied deep tech. And varied deep tech means nuclear fusion, batteries, climate tech even, to sound like carbon capture, for example. So those types of things that it's a suddenly an incredibly broad funnel of completely different types of technologies, which use very different manufacturing techniques and so on. And those are emerging en masse out of the university research, be it fundamental science, be it engineering science.
Max:But, because the trend over, say, last couple of decades was that the larger corporations were slimming down on their research, whilst the research budgets in the countries were going up in a proportion. So a lot more interesting things were happening at the universities and at academic So there's a lot more of this and, that's fueled also by the realisation by venture capital markets. So, you know, which is also stimulated by the governments who say, hey. As a government, as LPs, we want you to invest in future technologies, which are not, say, next dating apps, but something that's more fundamental to well-being of the country. So there's a turn towards being deep tech, and the word becomes more and more up there.
Max:And university is suddenly facing the challenge that before that, they were primarily educational institutions and doing research. And that that was it. So pretty much we're just writing scientific papers, articles, and so on. So now they're facing that the challenge that they're now suddenly in the venture capital business, and they're completely not prepared for this. Now they do their best.
Max:They have all the best intentions, but the majority of stakeholders at universities are very ignorant of how markets generally work as a thing. We you'll find all over the UK that universities are able to from their financial governance. So they think about how many students we're getting, how many how much money we get in, what are the expenses, and they view IP as one of those revenue streams that just comes in. And if you view it as a revenue stream on a 5 year horizon, you're completely missing out the point of value creation that new businesses do to the local economy. So for example, you won't find a single university, in my view, here in the UK, whose vice chancellor would be sitting there and thinking, hey.
Max:We're going to invest or create this couple of startups that are are going to turn into major successes, like major successes and you're talking 1,000,000,000 plus, they're gonna create an incredible value for the local economy. Chances are they're gonna create a massive endowment funds for the universities upon exits and things like this, but it also creates a culture of, say, entrepreneurship. It creates a culture of seeking value creation rather than seeking a job to get you to your pension. And that is something it's a paradigm shift that needs to happen, and it's nowhere near. And it was then the university technology transfer offices.
Max:So some people do understand this, but it's a different language they we talk even though it's all English.
Steve:Do you think it's, kind of, the way it's it's I guess it's a problem of the structure now because how do you measure the success of a university might be how many of your students were employed after they graduated. And if it goes into something risky, maybe they're scared.
Max:There are use cases. There are use cases where it does work really well, and we all know them very well. You look at don't you don't have to look far. Look at the MIT, look at Berkeley, even look at things like University of Maryland. So where there's a good deal on the table, things start to happen.
Max:And they do not compromise on the employability of their students at all. Because if there are a bunch of extremely exciting, interesting companies that emerged out of the university, where are these students gonna go toward to look for work?
Dan:Yep.
Max:She's at the doorstep.
Steve:It's also problematic. Sometimes these students that they want to do something, but they need a job and that they end up being in the sector they don't want to be in. And then they're kind of stuck there for decades. A lot of physicists go to finance, for example, but it's probably not where they want to be, but what else can you do sometimes? So if you don't have that ecosystem of entrepreneurship, go on your own, but there's also levels of risk that people can accept, I think.
Steve:And that's tricky. You can't really make it unless you have stability.
Max:Well, it's a lot risky if you're the only startup in the country versus you're part of, the bigger movements, so you're actually quite a bit more de risked in the Vibran ecosystem, because you have more capital. It's a more competitive market. We have multiple buyers, multiple suppliers of all different things. You're talking accountants, you're talking lawyers, financial facilities, talking banks that are understanding what risks are. If you're a start up or if you're a a growth company, there's more funds, there's more experience, because you can lean on people who's done it before and you don't make the basic mistakes that, once you know it, it's obvious.
Max:So this is the great value of those ecosystems, and universities can be absolute catalysts of this. And in my view, they're completely failing this.
Dan:Really interesting viewpoint, Max. I I love it. Yeah. And, of course, it pumps up the whole service economy of the whole country. Right?
Dan:If it's because most startups don't have a role for every single thing that they do. People have to wear many hats and therefore they have to rely on external services. But, hey, I'm we could probably do a separate podcast on that.
Max:Absolutely. Yeah. There's there's there's a lot to to it actually.
Dan:Really nice. Okay. Well, listen, I'm gonna wrap up. Fantastic talking to you. Thanks very much.
Dan:I'll be watching with great focus on your next steps and any announcements you make. And I'm looking forward to, chatting to you again at some point. Okay.
Max:It was, great chatting.
Dan:Great. Bye for now. I'd like to take this moment to thank you for listening to the podcast. Quantum networking is such a broad domain, especially considering the breadth of quantum physics and quantum computing all as an undercurrent easily to get sucked into. So much is still in the research realm, which can make it really tough for a curious IT guy to know where to start.
Dan:So hit subscribe or follow me on your podcast platform, and I'll do my best to bring you more prevalent topics in the world of quantum networking. Spread the word. It would really help us out.
Creators and Guests
