IonQ CEO on Taking Quantum Computing Mainstream

[MUSIC] Hello, and welcome to the TechDistructors podcast hosted by Bloomberg Intelligence. In this podcast series, we speak with company executives and management teams about their view on destruction and how it is driving the decision making and strategy. Bloomberg Intelligence is Bloomberg's research arm and covers over 2000 companies globally across multiple asset classes backed by Bloomberg and third party data, supported by nearly 500 research professionals. My name is Jake Silverman, Technology Analysis at Bloomberg Intelligence. I'm joined by my colleague, Conjan Sabani, Senior Technology Analysis at Bloomberg Intelligence. And today, I'm pleased to have Nicola Demasi, CEO of IonQ, one of the leaders in the field of quantum computing as our guest. This is not the first time Nicola has been the CEO of a publicly traded company. This is the first company that's focused on quantum computing if I'm not mistaken. Nicola, welcome to the show. >> Honorable pleasure to be here, gentlemen. I looked forward to a conversation. >> Great, yeah. I was hoping you could start by giving a little bit of background on yourself and how you ended up as a CEO of IonQ. >> Well, I'm a physicist originally. I studied physics undergrad and grad at Cambridge University. My master's degree thesis was on next generation electron beam lithography. And I vividly remember reading Crispin Rose paper from 1995, our founder, who built the world's first quantum logic gate. And I feel like physicist of my generation who were in school at any point in the middle eight, nineties, consider that a shot that was heard around the world in a lot of ways. Because it sort of turned Richard Feynman's postulates from the early 1970s into an empirical reality. And it showed that all of a sudden it was very possible to build something here in the real world that would turn into eventually a quantum computer that was operational, commercial, and solves problems that can't be solved any other way. I'm proud of the fact that I've had a career in both fundamental science and in the public markets for the last 25 years. And so I have been the chairman or CEO or lead director or committee chair on 14 public companies. IonQ is the third one I'm the CEO and chairman of. And it is also the one that I've had a lifelong passion for going back. This point we're actually celebrating the 30th anniversary right now of Dr. Monroe's 1995 demonstration. And I'm proud of the fact that this has been a year of acceleration for IonQ where we've continued to show scientific milestones, commercialization milestones, the very much honor where we got started and demonstrate why IonQ, in our opinion, very much is the leader and is the inter-pound gorilla of the quantum space, no matter how you want to measure it, but it's by revenue or market cap or cash balance or patents or PhDs or employees, we're very proud of the breadth and depths of our very unique quantum platform, which is leading not only in computing, but also in networking quantum networking, quantum cybersecurity and also quantum sensing. I actually took IonQ public with my, one of my blank check vehicles five years ago, and so I had the privilege of being able to diligence the entire quantum sector in 2019. And before we chose IonQ as what we felt was going to be the perpetual leader to do the fact that it got started first, it's able to commercialize first, our machines powered on first, and we've been able to run commercial applications on the public cloud first and longest. And I'm a big believer that, you know, the history of the tech markets is pretty clear, which is those strong ecosystem, virtuous circle feedback loops that get created when you power on first and you commercialize first and you get customers using your software and hardware. And so that's really the backdrop for IonQ success has been, you know, we were here first, we have led every step of the quantum revolution first, we continue to do so, and we believe that we're very much just getting started here. I've always known in my, in my sort of projections the last 10 or 20 years that quantum was going to inflect in the, in the 2020s, and we're in the mid 2020s now and IonQ is very much demonstrating that as the first quantum company in history, you know, to be on track to deliver nine figures of revenue in 2025. Yeah, that's great. I think we want to get into the company in depth a little bit, but just to get started, maybe you could help us define quantum computing, you know, what is the benefit compared to a traditional computer? And then the specific approach that IonQ is taking is, is trapped ions, maybe you could explain a little bit about the background of that as well. So Richard Feynman, you know, from Caltech, Nobel Prize winner in the early 1970s, recognized two things that could be done with quantum entanglement, which is a fundamental principle of quantum mechanics, right? One was you can have perfect communication security. And so, you know, we wrote down these so-called Alice and Bob eavesdropping problems as far back as 50 years ago. And quantum networking is a vibrant piece of the IonQ, you know, company ecosystem investment. We believe we are the far and away leader in quantum key distribution, which is what enables this ultimate unhackable security and the transmission of information. So QKD, we've known for 50 years, you know, and we've demonstrated it now for over a decade with commercial and government customers, is unhackable because you can simply not transmit information if someone's snooping on the line. And you should come and see those demonstrations that we have running. Actually, we had one of our analysts day, I think, Jake, and I think you were there. But that's sitting in our Geneva office and off in our San Diego office as well. The other postulate that Richard Feynman created, the Crispin row really kicked off empirically 30 years ago as a notion of using entanglement as a superpower for running quantum algorithms that allow unique insights and unique problem solving because of the way this entanglement allows computational states to be operated. You know, it isn't that quantum computers have a higher clock speed than classical or analog computers. That's not why they're powerful, but powerful because of the way they can effectively look at the computational space and the way that quantum algorithms can use entanglement to solve a number of long identified opportunities quickly and efficiently. You know, running shores algorithm and factoring prime numbers has been known about for, you know, 25 or 30 years. The traveling salesperson has been argued about on and off of the last 25 years. And there's, of course, opportunities for perfect simulation of chemistry problems, material science problems, pharmaceutical problems. And into the day, as Richard Feynman used to say, we live in a quantum world. And we have been demonstrating that the world's most successful physical theory, which is quantum mechanics, can be turned into the most powerful means of communicating securely as well as a the most powerful way of driving unique insights to extend artificial intelligence AI, but also to back solve for where materials, pharmaceuticals, biotech needs to go in order to allow humanities, technological progress to plow onwards and even accelerate, right? So you know, there is more than one way to in theory build a quantum computer. But I also think that there is, frankly, a fair amount of misinformation and disinformation that runs around. You know, we have been the most successful quantum computing company because we've been running the longest. We commercialize first and we have the world's most precise qubits, physical qubits. And the challenge to build a quantum computer is how many physical qubits you need to be able to have what's called a algorithmic or logical qubit that you can perform calculations with, right? Because we have the highest fidelity and our physical qubits are the most perfect. And I've known this since 2019 when we diligence the whole landscape. Our ability to build a full gate model, quantum computer that lives up to the expectations of, you know, Richard, five and fifty years ago is simply a lot sooner was a lot less in the requirements around resources, physical scale or energy consumption. We only need, you know, 10, 12, 13 to one as a ratio between our physical to logical qubits. There are other pathways that I'm happy to spend time on. But the other pathways need a thousand, you know, physical qubits for every logical qubit. Some of them require more. They might need 10,000, 100,000 million physical qubits to have one logical qubit. And the issue with those tremendous error rates that other pathways have is they have errors that build up not just within a qubit, but also across their, you know, their machine. And so, you know, everybody else in this ecosystem that is, frankly, in our opinion far behind us, not just temporarily, but also from a sheer unit economics perspective are going to find tremendous scaling challenges when it comes to trying to get to not just hundreds of logical qubits, but getting to thousands, let alone tens of thousands of logical qubits, right? And so what's unique about our approach is that we have these high quality physical qubits and we have the ability to put them on a semiconductor, you know, approach, if you will, that enables us to go farther, cheaper, sooner than everyone else. And so we estimate that we have as much as a five year lead over some of our nearest competitors in the superconducting space. And we think that we have probably a hundred X cost advantage. Those two things in combination are unbelievably powerful as a moat. You know, you're both students of the semiconductor business and students of the technology industry overall. You know, I always like to remind people that Steve Jobs announced the iPhone in January of seven launched in June of seven. And for a man who was known to not lack self confidence, he said, look, you know, we have, I think, a two-year lead over Motorola Blackberry and Nokia in this touch screen technology. He did not have a cost advantage. The iPhone cost more, but they had a two-year lead. And they've turned that two-year lead into a $4 trillion market cap difference. I and Q has, we believe at least a five year lead, temporarily, in building powerful, fall tolerant quantum peers. But we also have probably a hundred X cost advantage. It might turn into a thousand X cost advantage. And that's really no exaggeration because we don't require rare earths. We don't require a year's supply of helium three. You know, I think there's a lot of people who are physicists who have gone down pathways that they thought were going to make sense. And that at the best, at most, they're going to build one machine that will look a lot like the Spruce Goose, if you remember them, and Howard Hughes building the world's largest transport plane. And it'll take off once and they'll film it and it'll go straight into a museum. And there'll be a lot of these museum pathways because people seem to have forgotten in the broader quantum ecosystem that computing revolutions are driven by three things. Like how much power you get per unit of cost, how much computational power you get per unit of energy, and then how much space you take up, right? And so computing revolutions are about miniaturization, power reduction, cost reduction, per unit of computational power. And we have a pathway that we've known for years and years and years would turn on first, run applications first, and be the cheapest by any definition of sheep, energy consumption, just cost of inputs, you know, et cetera, cost of operating. And so we feel like regardless of whether anyone else has a Spruce Goose Museum future or guaranteed future, if quantum computing is going to be a commercial mass market affair, it's eye-incuse pathway that stands head and shoulders above anybody else, private or public who's trying to follow us. And I think it's safe to say, you know, that investors have taken notice of this fact. I mean, we have raised over $3 billion this year in the two largest fund raises ever for a quantum company in history, you know, first from a billion dollar single institution. And again, from a two billion dollar single institutional raise, we have something like an order of magnitude more revenue than any other public company in the sector. And we have 1100 patents, and we have uniquely built not only the depths and computing, but also the depths and quantum networking to enable us to build quantum data centers, as well as maintain all the patent protection around quantum security, quantum security communications. And we've also added on to our platform to add quantum sensing under the ocean, on land, in the air and up in space. And so when you look at that broad spectrum capability, as we said in our last earnings call, we're able to sell solutions now, full solutions to customers, government commercial, we're wondering about how quantum computers can be networked together, how their information can be kept safe, not just in a world where our adversaries may have quantum computers, they can crack sure as algorithm, but also in a world where classical threats are becoming quite sizable. I mean, there's not a day that goes by when there isn't a horror story in the news of someone who has been hacked. And believe it or not, QKD and quantum security solves that issue today also. So it's unhackable, not just practically speaking, as it is difficult and slow, but it's completely unhackable because it would require a breach of quantum mechanics and the fundamental laws of physics to do so, right? So we're seeing an uptick in large customer interest in our quantum solutions, government interest in our quantum solutions. And we're seeing that customers increasing recognize that I increase advantage is not just the temporal time to market advantage in our computers, but also the costs and the accessibility that we provide, but also that our computers have networking built in. We have been thinking about quantum computing longer than anyone else going back 30 years, but we've also been thinking about quantum networking and the photonic interconnects between our machines longer than anyone else. And so when I hear about other people waking up in the last six months, three months, nine months, to the realization that quantum networking, photonic interconnects are very, very important to making what we call the quantum internet a reality. You know, it puts a smile on our face because we've known that for decades and we've been investing for the last decade to do that, right? So I feel like the market's very much coming to us, Jake and Kunjah. I think I think people are realizing that what we've been saying since our IPO is absolutely the case, which is we're going to lead the computing revolution, Kunjah, we're going to lead the quantum network data center opportunity. Now, we're going to lead the quantum cyber security opportunity. And we're going to connect this together for government and commercial customers in all four or five theaters that matter. And I think you're going to see the same compounding effects in our business that you've seen in many other flywheels where the technology industry always likes to end up looking like a barbell where there's usually a giant. Sometimes there's a giant in a fast follower or slow follower as I like to call them. You know, Nvidia and Broadcom, you know, sort of exhibit that to some extent. Apple and Samsung exhibit that, you know, even in the software space, the really isn't the second place player that matters in the search space, for example, right? And so we're feeling increasingly confident that our ability to control the hardware stack will give us a lot of power in the software stack. We are, we are, you know, students of history, we're going to invest continually, as I said, every year, it's called in the compiler application layer. We are an open structure or open platform. So we want to interoperate as we do already with Google Microsoft Amazon and the cloud. And we have the last four or five years, we're going to be open to everyone because we want to be ubiquitous. And we know that the more people that use our hardware and software and compiler applications, you know, the deeper and broader ecosystem is going to be. But at the same time, we're not going to be allow ourselves to have an MSDOS moment as Intel did, you know, 20, 30 years ago, 30 years ago, at this point. And so we're going to maintain control and make sure that we're able to capture value in a vertically integrated way. Yeah, no, it's great. I think you gave a really comprehensive overview of what IonQ is working on. And in an area that you touched on a lot is systems. But, you know, digging into a few topics really, the big headline numbers that a lot of people focus on when they think of quantum computing is qubit counts. You've outlined a roadmap that I think is pretty ambitious. Clearly, you feel pretty confident that you're going to be able to reach it. This pathway, all these different technologies within, you know, botanics that you've been leveraging, is this what helps drive some of those step function improvements in qubit scaling to reach those targets that you've outlined? Because, you know, you've hit 64 algorithmic qubits. You know, there's a path to 256 and 2026. And then ultimately, you know, reaching 80,000 algorithmic qubits and 2 million physical qubits by 2030. Are there any like additional sort of breakthroughs that you think you need to in order to reach these goals? It may be areas like optics, for example. Yeah. So here's what's unique about IonQ. You know, we are the only company in the world that has no breakthroughs required. We have hit our fully fully full tolerant fidelity and announced that in the past three or four months, 49, 99.99 percent fidelity. What is important, and by the way, Jake, it's 80,000 logical qubits, you know, with 2 million physical qubits is what we're targeting in the lab in 2030, which means we could ship it in 2031 to customers. And we also know, and we've had AT Carney crawl our supply chain, you're at our analyst day, we announced a sub $30 million bill of materials in 2025 dollars to deliver that machine. Now, of course, inflation's a real thing. And we expect that we can do even better than that over over time at scale with more manufacturing capacity. But we're very heartened by the fact that our bill of materials doesn't move up very much between the 256, 10,000 qubit system, 100,000 qubit system, million qubit system, and so on. We have proven out all of the scientific elements of our system from the software compiler through to how we run applications, through to the really quite minimal objects at this point in our system relative to the ion trap on a chip architecture that's so unique to I and Q at this point. And I want to point out that not all logical qubits are created equal. And so because of the tremendous fidelity on our physical qubits, our physical qubits are more, you know, have higher fidelity than the logical qubits of pretty much everybody else's pathway. And so the computational power will be unmatched, I think even on our 10,000 qubit system, you know, which should have, you know, hundreds, 500-ish, hundreds, certainly of high-quality logical qubits. I think you're going to see an acceleration of the distance between us and anybody else in the sector, not just in computational power, but also in unit economics. And I think the unit economics really people need to spend more time on this. There are companies that are well-funded who, yes, could build, you know, possibly a, you know, 1,000 logical qubit machine, you know, in the next five or 10 years for $5 billion. But it might cost them a lot more. It might cost them $15 billion. You never know. Our machines are going to cost us tens of millions of dollars all the way through, right? And so you've seen capital markets wake up to that. You've seen governments wake up to the fact we have the largest government contracts out there in $100 million, AFRL, our first research lab contract and so on for quantum, two quantum networks, computers. And people are waking up to the fact that, you know, look, processing time or processing cost or a real item. You know, you've seen a lot of these semiconductor guys not make it into DARPA phase B, presumably because, you know, they don't have a path on costs or supply chain or what else. Our supply chain is all, you know, domestic. And I think no one should be confused about companies whose machine has never turned on being in so-called under-explored paths of quantum computing in the DARPA program. Under-explored if your machine hasn't turned on is exactly what it sounds like, which is, you know, to put this in perspective, Jake, I'm selling the iPhone 17 Pro for $800. There are other people out there who might have good brand names who can offer you an iPhone 1 or an iPhone 5 for $30,000. Everybody else has a mobile phone that hasn't turned on. I mean, let's get real here, right? If I said to you, I'm going to build a mobile phone, but it hasn't turned on yet, let alone run an application, let alone be available for sale. You know, how can you really say that you're in a race when, you know, you're 17 generations behind my phone and never turned on? I mean, you can't really take people seriously if their machine won't turn on for years from now when ours turned on in 2017, right? So, you know, we think this is a two-horse race at best. You know, there'll be one other person that, you know, can sort of stagger along. But I think, you know, capital markets see through this pretty quickly when you have a path to a very expensive spruce goose that will cost between $5 and $50 billion. You know, you're never going to build more than one of those. And so you're not going to really satiate computational needs in the real world of both governments and commercial customers, right? Because at the day, time and cost to run these quantum algorithms is going to matter dramatically, dramatically, right? So there's certainly some trade-offs, right? I think you've highlighted scalable manufacturing, you know, build materials costs, very, very important, really strong gate, gate fidelity is a clear path to tens of thousands of logical qubits. What about areas like gate speeds? Are there other, there must be some trade-offs as you think about the ability to scale up? I mean, how do you overcome some of the other, maybe roadblocks that might, you know, get in an eye on Q's way? Yeah, so gate speeds entry in one. People actually don't realize that time to solution is what matters. So time to solution is what drives, you know, real computational unit economic discussions. And time to solution is exponential when it comes to error rates. So if you look at our trapped ions, again, superconductors and neutrons, you can see that our two qubit gate performance is staggeringly ahead of everybody else when it comes to the two qubit gate fidelity. I mean, this is orders of magnitude better, right? Factors of 10, factors of 100. But what's really interesting is when you look at time to solution of an application, because the error has an exponent in it and you can debate with various academics what the exponent is the power two or the power three or worse. You actually hurt yourself exponentially in having poor fidelity, much more so than gate speed helps you. So gate speed might be a factor of 10 better on a superconductor. But if the error rate is a factor of a thousand or at least a hundred worse and you're squaring the error rate, you can rapidly see how your time to solution is going to be a much, much bigger problem. And so this is what is not in the narrative for almost everybody that says they're trying to compete with us in the computing space is they don't talk about the fact that the application is hurt by tens of thousands, possibly hundreds of thousands of times, even though you might have a gate speed that's like a factor of 10 better than us. It'll take you orders of magnitude longer to actually do anything useful, right? Because of the way your hurt and that time solution's conversation. So yeah, I mean, look, I think, you know, honest conversations all send the market and I, you know, an I and Q's direction. I think there are academics that have dreams and whiteboards and they're never going to raise the money and they're never going to build a machine the power is on, generally speaking. And if they do power on, they'll find their time solution is so unbelievably poor that they're never going to be in this race to do anything useful commercially. So shifting over to applications, I think you've made a pretty clear case for where Ion Q is ahead of the pack, but ultimately you have to monetize your systems that you're building, right? There's a lot of different areas that Ion Q and quantum computing systems can have some real advantages. I think you talked about a few of them earlier, some clear ones are, you know, chemical interaction modeling, protein folding, recession forecasting, just to name a couple. They seem like these are pretty great areas that are primed for breakthroughs. What do you think is some of the biggest opportunities long term? Where do you think quantum computing is really going to make a difference long term? Look, I mean, I think the areas we have been active in are where there's going to be massive, massive upside. So I think it takes about three large little qubits for every atom that you want to model, you know, to an extreme level of accuracy with all the energy levels. And so as you see our large little qubit count progressing, you'll see that we can effectively back solve for certain criteria that we want in a drug and a piece of biopharma and a piece of material science, right? And so these areas are going to just lead to an explode, you know, Cambrian explosion, you know, of new technology opportunities. And as I would like to remind people, you know, what we're in the business of doing is turning problems that will take all of the world's fastest supercomputers in the classical world working together somewhere between a century and a millennium, you know, or even possibly a billion years, you know, into days. We are not competing, we're trying to compete with, you know, data center ad network servers and ad servers that are focused on monetizing humans and, you know, microsecond latencies, right? I mean, the GPU world is fantastic at looking at correlations in a black box in a neural network on what humans are clicking on and what humans might want to buy based on prior behavior. And that's why you see six shoes expenditure by Facebook, Google and players like that. And they're getting the revenue results, like they're spending lots of money and they're getting growing revenue because it's really, you know, neural networks are just fantastic correlation engines. They're a black box. They don't know how they actually work, right? Quantum computers are not. We were a deterministic algorithm that only executes on what we ask the computer to do. And so the human partnership with quantum computers will be a strong one all the way through to what we're looking to back solve for in terms of future, you know, let's just say material science, drug discovery, vaccine creation, protein folding, etc. I also think you're going to see a partnership between classical, what we used to be called machine learning. We now call it AI, but it's still black box machine learning. We're going to see a partnership between classical AI and quantum computers because we've shown this year that our quantum computers can help speed up training and also create data centers, a data set, sorry, for training of classical LLMs. You're going to see, of course, defense intelligence usage. You're going to see supply chain optimization and topological data analysis coming to the fore everywhere where you have hours or at least half an hour to run something as opposed to a microsecond latency. And you're going to see new, frankly, insights generated by virtue of the fact that quantum algorithms and our ability to look at problems through a different light generally creates new insights that you can't generate through GPUs and LLMs. So I always like to remind us that look, your mobile phone today is using LG again, still has a CPU in it. But all the smart stuff and all the hard work is being done by GPUs. And the same is going to happen with our QPUs, our quantum processing unit says, we're going to add the next level of growth to the overall compute ecosystem. You're still going to have CPUs doing calculator work and excel. You're still going to have GPUs looking at correlation engines and microsecond latency. But QPUs are going to add insights that won't ever be reached, even with every computer ever made known to man working on certain quadratic speedup and problem sets. And so it's an exciting new world. And look, I mean, what I'm excited about is 10 years ago, Nvidia was just a game company. They made graphic processing units for the PlayStation. And no one thought there would be a lot more to it than that. And look at the progress they've made in the past just four years, right? The next four years for I and Q is going to have much more progress than the last four years for I and Q. Because our computers make doubly exponential progress every moment we add a larger little cube that we double the compute space. So it's not, this is what people miss. It's not Moore's law where every two years, you know, you'll double the compute space. For us, it's every larger little qubit, right? So going from 36 algorithmic qubits last year and 24 to 64 this year is 260 million times bigger compute space. And this is really important in the competitive race as well, Jake and Kunjah, because what people don't realize is if we have a thousand larger little qubits, when our competitors only have 100, we're not 10 times more powerful. We're two to the power 900 more powerful. And by the way, we've stopped naming the numbers after about two to the power 20. No one really knows what to call those things, quadrillion, subtillion and you know, no decade million or whatever it is, you know, kind of thing, right? And so people need to realize that, you know, the competitive annex that we have in being just a few years ahead is truly staggering, right? It's a difference between Goldman Sachs and Morgan Stanley proprietary trading teams making no money. It's not you make less money. You make no money. If the other team arbitrage is away everything, you know, if you are Boeing versus Lockheed competing for the next stealth bomber, whatever, 27. And one of you is using our newest machine. The other one is two to the power 900 less capable. You know, you're going to lose a hundred billion dollar bid. If you're a vaccine is a year late, you have a huge patent disadvantage and you have a huge revenue disadvantage and huge brand disadvantage, right? So we see ourselves as creating the gold standard in every single part of every little bit of the market, if you will, of applied science. Every bit of applied science, we revolutionize by our computers. And it's happening soon and the people think and because of the double exponential nature, it's like a freight train that's coming that the human brain really struggles to process because they say the human brain is bad at exponentials, don't understand Moore's law, but we're really bad at double exponentials. So that's computing, right? Networking is just going to become table stakes. Everybody is going to need our quantum network solutions. I mean, there's going to be a moment when fall tolerant machines, hopefully for my IQ first, but our adversaries may get there, you know, not that far behind us. Who knows? They're investing and they're also unfortunately stealing, you know, successfully in some places. This is going to be a moment when, you know, everyone the world has to buy wants to buy, you know, our QKD solutions and we have all the patents around that. Similarly, our, you know, our GPS solutions for the future are gravimators and universal sensors. I mean, you know, the quantum sensing is actually, ironically, the most well understood part of what I increase platform does because everybody uses GPS on their phone, pretty frequently, right? But the future of quantum sensing is very much embedded into people's expectations of the future of communications and the future of humanity. We expect that self-driving cars will have more accurate, you know, position, navigation and time. And people expect that, you know, our quantum computer is they can already do early signal detection and change detection with satellite imagery and 3D object detection. You know, we're really actually empowering more elements of the future than people realize and more elements, by the way, of even the sustainable earth than people recognize. You know, we're very much a defense intelligence center organization. And we have a large team at iQ federal of, you know, security-cleared personnel like General Jay Raymond, who founded the Space Force and was the only person to serve as a four star in two different services for him, air and space since 1947. The Air Force has created and we've got Robert Cudillo, also, who we used to run the National Geospatial Intelligence Agency on our board. But despite the fact that, you know, we don't set out to do it, we actually help more elements of the sustainability model for planet earth than you might imagine. We're getting cars and trucks off the road by working on better battery technologies. And we've talked about that we kind of licensed our IPO. We're working on, you know, autonomous driving and logistics with Einride and Sweden. We're working on obviously getting trucks off the road. And we're working also in monitoring gas leaks, you know, through satellite imagery. Carbon sequestration catalysts are the kinds of things you can do when you have tens of thousands of large, little qubits. Nobody else is going to have that by the end of the decade because we're far ahead of everybody else in terms of unit accomplished time. And so, you know, inadvertently, we always say, I always say to iQ is, whatever your passion is, you get to come to work at iQ, scratching that ish and impacting the positive way that passion from logistics to material science, to defense intelligence, to pharma, to sustainability because every aspect of applied science is positively impacted by iQ. Nikola, you mentioned quite a bit about the relationship with the GPU and how the CPUs will have their own place in the future, along with CPU and GPUs. I want to just very quickly a segment to that, a couple of things there. So, you're right, a lot of your peers are embracing the fact that it's not a competition between a GPU and GPU, even though currently GPUs are trying to sort of solve the same problems. How do you see this role checking out? Because, like, if we be honest, right, that was the same thesis before, but today GPUs are eating CPUs lunches, right? So, based on that, if we extrapolate 10 years out the road, a GPUs could be taking most of the wallet share. So, how does it do? And today, Nvidia is partnering with a lot of your peers, you are self-included. They're trying to be the sort of the networking provider because a lot of quantum companies are struggling with their, you guys have a strong solution there. So, just very quickly talk about how do you see this relationship going forward? You know, look, I think the trillion dollar market cap companies have a fairly transparent strategy for Google, Microsoft, Amazon, Nvidia. You know, they're, hey, you know, the hoop waiting for there to be a clear wing at winner and hoping they're going to swoop down and buy that, you know, that winner, both to defend their cloud and also defend the future of their compute model. You are spot on that, you know, GPUs initially expanded the compute market, but now they are more efficiently doing things that CPUs used to do. And there's going to come a day where GPUs will do more cost effectively or certainly more space effectively in your phone and your data center, what CPUs used to do. I and Q's, and I really say I and Q's because it is simply not the case that the word QPU can be generalized. I mean, look, Graphcore used to say that they were going to give Nvidia a run for their money. I don't think anybody will seriously say that Graphcore is a threat for Nvidia or Broadcom, the GPU space. I mean, they're, they're a part of history at this point. As a venture back startup that had a dream, there was once again, five years behind, you know, the venture fallacy, and I think about this a lot because of my career in the public markets and the venture fallacy is mostly caused by startups not having audacious ideas, but having the idea that they can catch up to better funded businesses that are well ahead of them on the technical road map. And so if you look in the public markets, why does the number one player always trade at a premium? The typically has a one or two year lead in the software and hardware space. It's because two years later, it's much more likely nine to 10 times the two year lead becomes a three year lead, not the two year lead becomes a one year lead and then becomes a zero lead. That almost never happens. Right. In software and hardware because of the ecosystem network effects and feedback loops, right? And so that's why I said, like, look, when I delegenced this space in 2019, I could have taken almost anything public, I think. I mean, I had $645 million in our back. Everyone would have taken that money in 2019, you know, leading quantum company, even divisions of company, I think, found would have found that interesting. And I picked INQ because I knew they commercialized first. I mean, you commercialized first, you just get that advantage, first algorithm, first on the cloud, first on all three clouds. As I always said, everybody, look, you know, we've been all three clouds for last four or five years. If you're not on the cloud, particularly if you're owned by someone who runs a cloud, why is that? Well, obviously, it's because your machine doesn't run, not reliably, not sustainably, if at all. If your machine hasn't turned on, obviously, it's not on the cloud, right? You know, let's, let's stop the nonsense here, right? In the sense that these leads tend to increase in time because capital flows to the winning company. It doesn't decrease. There isn't, there isn't a moment where things converge, you know, our machine, we said is 36 quadrillion times larger compute space than what we think is our next biggest competitor. We said that the analyst A.J. you were at and the crazy part is that number is going up in the next year or two or three or four. It's not coming down. It's not going 36 quadrillion becomes 36 billion or a million. It's actually to go up to septillion and octillion the next time we have this analyst day, right? And so it's a divergence between our prospects and everybody else's. And look, I think, look, there will be a day when our QP use becomes so cost effective at certain problems that simply the energy savings alone because our machines are very energy effective, right? To, you know, to train an LLM, the classical space, you know, you need your own nuclear power plant and a football field size data room, you know, our machine that takes a billion GPUs to simulate, presently plugs into a wall socket and is the size of a commercial refrigerator. Now, you know, what we like to remind people is coming around the corner is our 256 qubit system, our 256 qubit system to fully simulate the power of its entanglement would require one billion terawatts of GPUs to fully simulate it, not approximate fully simulate. Planet Earth only makes one terawatt a year. So you're not going to be able to simulate even the 256 qubit system and what we can do. And it's going up from there to 10,000, 100,000 a million, right? So it is a brave new world where there's one thing I'm most confident on, which is that the human mind is fantastic at finding new uses for new computing power and new computing paradigms. And I want to highlight the fact that we're not only thinking out to 2030 on how we can put a million or two million qubits on a single chip. We're thinking about data centers that are bigger than that that will use our quantum networking and our patents IP and lead there to connect our quantum chips together. So there's a lot of ways to win on iincuse architecture and customers are seeing that and that's why they're, you know, they're frankly swarming our direction as is talent in the marketplace. We hold all of the world records in both fidelity for two qubit gates as well as effectively quantum networking speed and transmission speeds. And all of this matters is you want to make sure that your quantum computer is a quantum network and not rate limited by the speed of the quantum network and the transmit data because you want the quantum network to be as fast as the chip, right, and communicating across it, right? So we see the full glory of our architecture not just to a million or two million qubits, but how do we get to 10 million? Right? I mean, someday I hope that we will have a machine that will have a million logical qubits, not just 10,000 or 100,000 logical qubits, right? So it's staggering opportunities and staggering partnerships, yes, with, you know, cloud players and with, you know, Nvidia and Broadcom. All of them are going to realize they need us and my job is always put it is to make the partnerships with us both as easy and accessible, but also as profitable and as expensive for our shareholders as possible throughout this journey. One last one from us, is this a winner take all market? Is there room for multiple winners here? So I am hard pressed to find a piece of the overall tech landscape where there is not a winner takes most. I mean, from search bars to GPUs, it is hard pressed. You know, you see a couple markets where, you know, the number one player will have a 70% market share. The number two player is 15, you know, or 20. And then there's like 10 people fighting over the last 10%. Sometimes you see that in some markets, but, but being first and building that customer ecosystem, I'm building that software and hardware stack tends to relentlessly drive scale, because as I always remind everyone, you know, when you have a 50% market share and you want to grow 10%. It means you got to grow the market and take 5% off and off the other guys. And so you get this hollowing out effect where there might be some tiny companies that can survive in, you know, super niches. That tends to happen. And then the tends to be the lion share, the supermajority tends to be owned by one, maybe two, but mostly one, you know, player. And, you know, that's not simply a common on what our ambition and expectation is in what we're relentlessly driving towards, but it's just sort of simply the way the survival of fittest has worked in the tech markets for the last 20, 30, 40, 50 years. You know, Uniteconomics are going to determine that market share. How much computing power for how many dollars with what energy consumption, what supply chain constraints, right? And so I always say, look, I think we're going to build the world's most powerful quantum computers, quantum networks and quantum sensors first. And I think we're going to do it at the only price points that can be allowed mass market adoption. And that's going to build us the most vibrant, you know, customer ecosystem and application ecosystem. And we are excitedly racing towards that and capital markets are supporting us to do that because they already see a staggering lead today. Well, Nikolo, this was an incredibly interesting conversation. I'm excited to see what Ioncube can continue to do from here and how the company continues to progress. Thank you for joining us today. It was a pleasure hosting you and to our listeners, thank you for tuning in. If you like the episode, please subscribe and leave a review and be sure to check back to your conversations with leading disruptors in the tech landscape. If you want to learn more about our research, including deep dives and topics like AI, check out our work on the Bloomberg terminal at BIGO. This is your host, Jake Silderman signing off.