All right, good morning, everyone. It's so exciting to see so many familiar faces here, as well as, hopefully, new friends that we're making for the next decade. We're in for a fantastic show this morning, as I like to call it. You're going to learn about where we've been, where we're going, and most importantly, why we have led this market for the last, I would argue, as our founders sitting in the audience, I would argue we've been leading this market for the better part of 30 years, and we are poised to lead it for the next 30 years. Today's agenda is going to have about half a dozen sections before we take questions. We will then have lunch once we finish the proceedings, and my colleagues will be taking a number of these sections to deep dive on topics that I will introduce shortly.
So let me start with IonQ's proven quantum leadership. And this is not only on the computing side, but also on the networking side. We're going to hear a fair amount about the networking business today. The last decade, we've accomplished a lot, just like we've accomplished a lot in the last 30 years. And we will accomplish even more in the next decade. I like to remind people that we've been thinking about this and working on this longer than anybody else. And you can see it in the timeline. Our machines powered on eight, nine years ago. Our machines were the first on the public clouds. We were there at our IPO. And as you look forward, in the last five years, we've kept up that incredible momentum. We have demonstrated quantum advantage twice this year with partners like NVIDIA, AstraZeneca, and Ansys.
So, what we've been doing the last decade, two decades, three decades, it is challenging. But because of the pathway that Chris Monroe picked for us so long ago, it has really been an engineering challenge that we've been able to prosecute relentlessly. I'm very proud of what we've done the last year and what we've done the last 10 years. Many of you probably saw our press release this morning. We have regulatory approval for our Oxford Ionics transaction. You know that we have large partnerships with key government partners, such as the Air Force Research Lab. You know that we've been building our patent portfolio since IPO. When I met our friends at IonQ five, six years ago, I think we had about 35 patents. Today, it is 1,060. So it's an impressive moat for us, not just on the computing side, but also the networking side.
Chris Ballance is going to talk a little bit later on about the fact that error correction is only something you do when you don't have good hardware. And it's a great phrase of his, and Chris Monroe's, for that matter. We hold all the fidelity records, so we haven't had to do a lot of error correction. So we've been able to have machines that work before everybody else. And this merger with Oxford Ionics has given us a roadmap to continue to not only lead the industry, but extend our lead. You'll hear me say this a few times today, but we believe we have a five-year lead on our technical roadmap over any competitor.
In a moment, you're going to hear from my colleague, Dean Kassmann, who will talk you through the exciting news that we've made great progress on our 100- qubit AQ 64 system, whose prototype in our lab here in College Park is, believe it or not, 36 quadrillion times larger as a computational space than any system IBM has publicly unveiled. I thought that was a typo the first time it came into my inbox, because I've never seen a number that large in my professional career. I've done 14 public companies. That actually is correct. It's very impressive. It's a massive testament to our leadership position and how we continue to prosecute it. Last thing I'll say is we're the best capitalized company in this sector.
And we believe that that capitalization allows us to drive growth factors in computing, networking, and of course, not just on the ground, but up in the heavens, as you've seen us do. So this transaction has been approved, as I mentioned. It is likely to close relatively soon. There is a state visit upcoming for President Trump to the U.K. next week. And we very much hope to be a meaningful part of it. What I love about IonQ's technology is that we are the best solution across every unit economic vector. I think some of you have heard me say this probably on analyst callbacks and investor callbacks. We abbreviate quantum computing and quantum networking as our key business product families, QC and QN. We're very clear that computing revolutions are won by lowering unit economics, driving mass market penetration, and of course, miniaturization and shrinking the footprint.
We are the only company that has a pathway to not only have done that the last decade, but to continue to do that. So for us, QC stands for quantum computing. Pretty much everybody else who's in the quantum computing space and says they're a competitor of ours believes that QC stands for quantum construction. They talk about machines that are the size of football fields that cost billions of dollars and that will need their own small modular nuclear reactors. That is never the future of computing. It never has been. It never will be, particularly in a world where sovereign systems are the name of the game and the name of the day. Everybody wants their own. They want to control it. They want it in their basement. They want it near the front lines, if that's pertinent.
Sovereignty is something which is a positive tailwind for our company. If you think about everything we do here, quantum computing, quantum networking on the ground in the heavens, sovereignty is very much something, sovereign systems that we are delivering in a geopolitical climate which favors this globally. We're lower power. We're lower footprint. We're definitely much lower cost. And you'll hear more about this from Dean in a moment. We have validated that our cost of goods sold is under $30 million for a fully fault tolerant system that has two million physical qubits. There's no one else in the space. It's anywhere near us. They are orders of magnitude more expensive, less powerful. And my favorite part is they won't get there for probably five years after us.
One of the largest leads I've seen in the tech industry in my career, which leads me on to this. I will leave my colleagues to deep dive on it. But 36 quadrillion times larger in the computational space is where IonQ is versus IBM. For those who follow the news, there have been flashy announcements from players like Microsoft and IBM in the last six months. If you read the whole articles, you'll note that they are accurate, usually towards the end of the article, when they will say things at Microsoft like, well, we've only really built one prototype qubit three or four times, and the machine's never turned on. The machine's never turned on, right? Most of our colleagues and founders at IonQ will tell me generously that Microsoft might be where IonQ was in 2001.
But I have other colleagues like Rick Muller, who joined us from IARPA and has worked in the government's quantum programs for, I think, 26 years, who will tell me that they're not even in 2001. This is an astonishing leadership position. Don't listen to the fake news, as I like to call it, that's out there. Listen to the facts, for sure. It's easy to get a splashy press release when you have a $4 trillion market cap. But when you double-click on what people are actually doing in the sector, they are all a long way behind us. And the reason they're a long way behind us is they didn't start anywhere near as early as we did. They didn't pick the right path. And most importantly, they didn't power on early enough.
And, by the way, most people haven't powered on, really, other than IBM or IonQ, for that matter. So, here's where we are today. And it's only going to accelerate henceforth. And this is something that I think most analysts and investors don't always get their head fully wrapped around. The verticality of advancement for our company's computing power only gets steeper and steeper every generation from here. So, as we add logical qubits, because of the double exponential nature of this, you will see that the quadrillions become quintillions and septillions and beyond until we probably run out of nomenclature for it. So, let me spend a couple of minutes talking about the next decade.
As many of you have heard on my last earnings call and most of our earnings calls, I'm a big believer that by forward investing in the world's greatest talent, we are able to not just continue prosecuting our leadership position, but able to very much extend it. So I'm deeply proud of how well we've done on this metric. We've not only added people like Dr. Chris Ballance and Dr. Mihir Bhaskar. Dr. Chris Monroe is spending lots more time with us. And I hope to get as much of his time as he'll give me. We've got Marco Pistoia, who's joined us from JPMorgan, where he previously ran research for the largest bank in the world. Rick Muller, I mentioned.
You'll hear more from him later on today, as you will from Robert Cardillo, who served in multiple presidents' administrations last decade, running the National Geospatial-Intelligence Agency. These fine folks are not only driving technical acceleration for us, but also driving commercialization advantages, and you will have seen earlier in the week that we talked about the launch of the new IonQ Federal Board and our ability to assemble fantastic experts, not only in the engineering and physics space, but also commercially to support driving our nation's national security, as well as national economic security, and for that matter, our nation, the Five Eyes, NATO, and so on, radiates outwards, so what does the next few years look like on the technical roadmap? Well, I mentioned that it gets more and more vertical, and that's absolutely true.
We struggle, honestly, sometimes to think about how we're going to present this in meetings like this, because the double exponential nature is very difficult for people to grok, right? If you think about the compute power as growing, effectively related to the exponent of the number of logical qubits, you can see that things get really powerful by 2027, if not 2028. And they get so powerful that you start to unlock a large number of applications and algorithms that have been postulated, in many cases, for decades. All of these algorithms have tremendous implications for, of course, national security and national economic security, and of course, quantum machine learning and quantum AI, which we'll talk more about later this afternoon. Our acquisition of LightSync was important for us strategically.
And one of the things I am keen to impress upon everyone in the room today is that not only have we won the last decade and will we win the next decade, but we've also already thought about where we're going farther out than that. We're extremely confident that the next five years will be owned by IonQ, just like the last five years will be. But beyond that, into the 2030s, we already have a plan on how we are going to build data center-scale quantum computers and how we'll be able to continue driving growth, not just to 80,000 logical qubits, but beyond. We're not going to stop here, right? There's another five years that come after this where we may end up having as many logical qubits as we have physical qubits, for example, in 2030.
So when players like IBM say the day after we announced this roadmap that they will have 2,000 logical qubits in 2033, so roughly the same number we have in 2028, you can be darn sure that by 2033, we won't have just 80,000 logical qubits, but a bigger number than that. And we're doing that with LightSync's quantum memories and quantum networking. And you can see when I look at the analog of Mellanox and NVIDIA how important those data center-scale value creation events, of course, are for companies. One of the things that runs around our ecosystem and the news is this notion of quantum advantage and when it is happening. We're here proudly to say that IonQ is the first company in the world delivering quantum advantage and quantum commercial advantage. And we're delivering it already right now.
We've delivered it with Ansys back in March at NVIDIA's GTC event. We delivered it again on June 10 at their GTC event in Paris, and you'll notice the tonal change that Jensen has had in the last nine months. We've driven almost all of that change because they've been a partner of ours, obviously, in making fast progress on things like computational drug design. When you deliver a 20x speedup, and NVIDIA and AstraZeneca, the partners along with AWS, they can't claim that that's not a real thing, right, and turning a month of work into a day is just getting started for us, right? We ran that on our AQ 36 systems, Forte Enterprise. That was really last year's technology. Next year's systems are something like 240 million times more powerful, right? Value that comes from quantum computing and from computing revolutions in general tends to be emergent, right?
So if you think about where NVIDIA was as a business a few years ago, it was just a game company. They made graphics processing units for the PlayStation 2, right? Obviously, they figured out they could do some other stuff with that, right? The same thing is likely to happen here. But because our machines powered on eight years ago, because they were on the cloud five years ago, Dr. Chris Ballance, when we announced our merger, told his team something that also made me smile, which is IonQ's machines have been operating in the public cloud longer than we've been incorporated. And it's a serious point, right?
A lot of people have looked at IonQ's leadership and success and gone, "I'd like to back a venture company and hope that I can get an exit somehow." And the reality is, because we powered on first, because we went to the cloud first, because we delivered early quantum advantage first, we have a software stack and applications that enable us to deliver applications today and last year and the year before and the year before. And we also are able to manufacture our machines and ship them around the world. And we do that all the time. And so the algorithms we're working on will all drive value for our company. There will be a lot more value, I'm pretty sure, that will come from quantum machine learning and quantum AI.
And the sky is the limit on that one, because we use about a tenth the power of classical GPU data center AI setups. I have to believe that due to cost alone, that is going to be extremely compelling, aside from the fact that we probably can't have half the grid going on Microsoft recommissioning Three Mile Island and all that stuff. That's a practical concern, too. But on cost alone, quantum AI is going to be really compelling a few years out. I'm going to close this section mentioning our quantum networking business, which I hope everyone in the room gets to spend a lot of time with Jordan Shapiro and Grégoire Ribordy, who are with us today. Marco Pistoia is also a foremost authority on quantum key distribution. And this is what I'd like you to hopefully all reflect and integrate into your thinking.
Quantum key distribution and quantum networking is not just something for the future. It's something for today, because classical cybersecurity challenges only continue to be nastier, nastier, more pervasive, right? The old adage that there's only two types of cybersecurity issues, right, the ones you know about and the people inside your system that you don't know about, right, continues to be a challenge for corporates and nations alike. There is nobody that wants to get hacked, but there's a lot of hacking going on because people are downloading now to decrypt later, and they're doing that because nation-states know that they will eventually probably have a quantum computer that can decrypt RSA-2048. The beauty of our quantum networking business is that it is not just practically secure. It is theoretically and absolutely secure.
As Marco Pistoia likes to say, "Even in 50 years, you won't be able to crack our quantum networking solutions." And so companies who work with us, you can spot because they are not in the news for data breaches, right? And governments that work with us, same thing. And obviously, banks and telcos are realizing that QKD and quantum networking is vital not only for their business next year and the year after, but their business right now. It's also vital to the future of the quantum internet and vital to the future of quantum network computers. And so what you'll see today is the fantastic synergies amongst our business lines, right? We can not only do QKD on the ground with terrestrial networks that are now much larger in potential reach because of our LightSync acquisition.
We could actually build a nation-state-sized and even international QKD network and quantum network when we want to. You're going to see that we're able to do that from ground to space, space to space, and space to ground. We follow our customers. It's something that banks and telcos worry about, and of course, governments. And we're already the company that I think it's safe to say the only company, actually, that can do quantum computers that are networks together using secure quantum communications. So we are planning for a bold, grand future here of not only selling sovereign systems, but being able to quantum network those sovereign systems, being able to allow quantum communications and the quantum internet. And I look forward to taking questions in a few sections here. But very much, please store them up.
We are building a business which is great in its individual capacities, but even more powerful in its whole. Without further ado, I'm going to pass you over to Dr. Dean Kassmann to talk a little bit more about our quantum computing roadmap.
Good morning, everybody. Thank you for joining us today. I'm super glad to see a lot of familiar faces, as well as new faces in the audience. I want to start this morning to dig into a little bit of our technology and talk about our products, and so a big differentiator of IonQ is the fact, as Niccolo said, we have many generations of commercial systems that we've been putting out into the market for some time, right? We've been building them to be enterprise-grade, scalable, high-performant, and solve customer applications. Now, we've been doing this since 2019. Our first inaugural system was Harmony. That was followed by our Aria systems. That was the first system that we got on all three major cloud platforms, so that meant software development, integration work, and being able to actually start running 24/7. That was followed by Forte.
Forte became and has become our major workhorse in terms of our early application development. A lot of the early results that we've published have been on our early Forte Enterprise systems or, sorry, Forte systems. Forte Enterprise we released. That is our data center compatible system. We have multiple instances of Forte Enterprise that we're currently building. We have instances that have been deployed that are currently running customer jobs. A lot of the current research that you see published on IonQ systems is on Forte and Forte Enterprise. Now, we have Tempo, which is in active development. Those systems have been being built for the last year, the design, engineering, all that work. The Tempo system is a 100-qubit system, 99.9% fidelity, capable of running AQ 64 benchmarks. That system is going to be deployed next year.
Now, the Tempo system underneath the hood is supporting mid-circuit execution, dynamic or mid-circuit measurement, dynamic circuit execution. That are needed capabilities to do quantum error correction demonstrations and all of the underlying hardware. The Tempo system also has a move from our past ytterbium species to now barium. That barium gives us larger uptimes. It allows us to do many different kinds of atomic protocols that, at the end of the day, regardless of underlying physics, result in more performance for the customer, a higher set of nines. And so the overall laser system development and everything else that's going into Tempo has been a step change in our underlying technology. We have our Oxford Ionics 256-qubit system right around the corner. That has been in play for a while, right? We've already started integration work. We'll talk a little bit about some of that a little bit later.
But we've been doing this for a while. The muscle and energy that we've put into it is significant. Now, that energy and muscle, to give you the best analog, if you were to think about other players in the market that are producing and designing the software stack, the hardware stack, if any of you have ever, many of you have taken a plane to get here, how many of you would actually take a vendor that's the first plane they've ever built, and you're going to get on it for the very first time to do something useful, fly somewhere? None of you would do it. It's just not realistic. You're going to trust the plane that is flying every day, has been put through the wringer, because flying is learning, right? We've been doing learning for a large amount of time.
And so our systems over time have been used. We've been kicking the wheels. We've been changing the underlying technology, but we've been delivering across that time frame. So I want to talk a little bit about the software stack as well. So in addition to the hardware development, we have been driving the software from the ground up to the FPGAs, the system-on-chip software, all the way up to our cloud integrations, right? That's compilers, circuit optimization, our control and scheduling. All of that software stack is built in-house with amazing capabilities. We have back-office capabilities for users, account keys, all of the different infrastructure that's required to be able to run a 24/7 operation. Now, that also includes applications. All of the infrastructure required for being able to do session management, being able to drive hybrid workflows is all there as well.
Now, that and my colleagues will a little bit later talk about networking feeds into our networking capabilities. And so we have been driving the software and the hardware for a significant amount of time, and we're going to be continuing to leverage that as we move into the future. I want to shift for just a second into some of the benchmarking work and kind of give you a quick update. If you remember, we have been anchoring the IonQ roadmap against a benchmark called algorithmic qubits for a while. That benchmark is intended to measure the number of useful qubits that you can use in the system. Now, it is based off of work done by the Quantum Economic Development Consortium, QEDC. It is comprised of six different circuits across three different large families that include optimization, simulation, and machine learning.
And so the measure of the benchmark is intended to be able to kind of measure the width or the number of 2-qubit gates you can execute in applications, as well as the number of qubits that are actually useful. And it's a measure of hardware and software. And so we have prototype results from one of our development systems for Tempo, and we are able to demonstrate AQ 64 in that development system. We still have additional work to go to fully check the box, but we have very promising early results. The plot there on the left is an example of those results. Blue dots represent the ability to not only execute the circuit family, but get meaningful results out of it.
If there were red dots or orange dots, that means that the outputs from the quantum computer, one, either are not accurate or are close to the equivalent of noise. And so I want to do a little bit of a comparison. Over time, as we've run comparison benchmarks across the world, IBM right now, on their 150 physical qubit system, is seeing roughly eight to nine usable qubits. This is the 36 quadrillion times more useful computational space to actually do work, right? And this is on development hardware at the moment. It's going to be even more exciting next year as we continue to drive fidelity and drive hardware. Now, we know that AQ is a benchmark that's been out in the market for a while.
We've talked early in the Q1 earnings call about being able to replace the benchmark, not actually replace, but augment the benchmark with additional items that we're hearing from customers. And so you're going to hear later about the importance of fidelity, the importance of physical qubit count, logical qubit count, error correction rates. And so those physical benchmarks, we're going to start to, I would say, solidify and communicate more broadly. We're also investing in application benchmarks. You'll hear about that as well. The importance of the application benchmarks are really focused on specific vertical industries, chemistry, logistics. And so you'll hear a little bit about that later. And those metrics are going to be things about what customers care about, which is energy usage, cost to solution, accuracy. And so you'll hear about that from our lead of product a little bit later on.
Now, I want to quickly, before I end, talk about roadmap. Now, we published a technical roadmap. It is the most aggressive and forward-looking roadmap that is out there. We have an ability to deliver by the end of the decade 2 million physical qubits. As Niccolo said, it is literally five years ahead of everybody else. Next year, we'll have 256. The year after, 10,000, right? That physical qubit count, logical qubit counts, and the error correction rates are better than any other competitor that we have out there. So I just want to leave you with the fact that if you compare the numbers published by our competitors and others, we are right now literally five years ahead. The early results that we have are showing that we have a system capable of a 36 quadrillion, I would say, more advanced computational space.
We have the best unit economics. I'm going to hand things over to Dr. Chris Ballance from Oxford Ionics to kind of double-click and go deeper into the unit economics and some of the technical details that are underpinning our roadmap. Welcome, Chris Ballance, to the stage.
Thank you, Dean. I'm Dr. Chris Ballance. I'm the co-founder and CEO of Oxford Ionics. Dean's just told you how ambitious our roadmap is. I'm going to tell you the underpinning technology behind it that allows us to achieve this, allows us to achieve it faster than anyone else, and allows us to achieve it with better unit economics. There are two key ingredients to scale quantum computers: size and performance. Size is the number of qubits in the system. This is a necessary but not sufficient condition to do useful things with quantum computers. Every time you add a qubit, you increase the computational state space. This is where you get this exponential increase in powerful. That's not enough. You also need qubit performance to unlock this potential.
If we look at the map of the market, we see that the big differentiator across the market right now is the performance of the qubits. This is where there are orders of magnitude in difference between different players. Also, what we see if we look at this map of the market is that all of the highest qubit performance players use trapped ions. This is because they're natural qubits. You don't have to manufacture them. And this means that you get better qubit performance, potentially, and much longer coherence times. That makes life a lot easier. The other thing you can see from this is that Oxford Ionics has qubit performance that's substantially better than anything else out there using our electronic qubit control. Last summer, we demonstrated a qubit fidelity of 99.97% on the most stringent test, the 2-qubit gate fidelity.
And as Niccolo said, this is a one plus one equals 10 kind of deal. And I think this explains in a visual metaphor why that is. Taking Oxford Ionics' electronic qubit control technology, where we have these incredibly high qubit performance, and merging this into IonQ's compute platforms allows us to scale faster using the best qubit technology out there, way ahead of everyone else in the market. But why does qubit performance matter so much? So it matters because it allows us to deliver value to customers sooner. And it matters in two different time scales. The first time scale is right now. Improving qubit performance right now allows us to achieve more with the platforms we have. This is why, as Dean was talking about, the AQ 64 work. This is why, despite IBM having more qubits, they can't run anything like this size of problems.
They have the qubits. They don't have the qubit performance. But it also matters in the long term as we start using quantum error correction. So this allows us then to achieve and unlock value faster. And let me explain why. So quantum error correction is a software solution to errors. It allows us to take lots of lower performance qubits and make fewer high performance qubits. But this is not a free lunch. It's a software solution to our hardware problem. If you build better hardware, you need to do less error correction, which means you need way less overhead. If you look at the long-term scaling of an industry standard process, what you see is that as you add the number of qubits, the computational power increases exponentially.
But if you have higher performance qubits, if you have lower error rates, and if you have good error correction architectures, you can unlock a lot more compute power with the same size scale of device. Because we have fewer errors to correct, we can get far more useful qubits out. And to put this in context, let's look at our 10,000 physical qubit device. This is our 2027 roadmap target. If you take those 10,000 physical qubits and assume they're in a competitor's device with best of the rest qubit performance, and you assume they use industry-leading quantum error correction, those 10,000 physical qubits go to something like 41 logical qubits. If you take 10,000 physical qubits in our platform, where we have much better qubit performance and architectures that allow us to very efficiently quantum error correct, that can take us to 8,000, sorry, 800 logical qubits.
So this is 20x better on the same hardware platform. So this allows us to unlock value, unlock much more value with the same physical platform. And equivalently, this allows us to unlock value earlier because we unlock certain thresholds of compute power years earlier on our roadmap. And now I can say that we've broken our own record. In a device like this, which you can see out in the lobby afterwards, we previously demonstrated 99.97% 2-qubit gate fidelity. In our new platform, we've shown that we can achieve over 99.99% 2-qubit gate fidelity. So better qubit performance is better. I hope that I've convinced you of that. But this is a particularly special threshold since this is a threshold that all of our long-term scaling roadmap is based off.
Now we've achieved 99.99% 2-qubit gate fidelity, which is the most stringent measure of performance in these computational systems. We don't have to make any further improvements to this benchmark all the way to scaling out to millions of qubits. Everyone else in the industry is still doing basic science, still doing R&D, trying to achieve our long-term roadmap targets. We have now delivered that, and the technology behind all of this is electronic qubit control. This is our secret sauce, so what this allows us to do is control qubits using electronics rather than lasers, and this is good because it allows us to integrate electronics into chips, and with that electronics integrating into chips, we have an incredibly scalable platform. We can use a standard semiconductor supply chain. Because electronics allows us to get incredibly well-tolerant signals, this allows us to then get these world-leading fidelities.
At the same time, because we can efficiently parallelize things on individual chips that we build out, this allows us to run our computations fully parallelized. And what this allows us to do is get the fastest time to solution of any platform like this. And finally, because we're leveraging the might of the standard semiconductor supply chain, this is mass manufacturable. We're already producing these devices on standard semiconductor fabs alongside chips for laptops, cell phones, high-performance computers with our manufacturing partner Infineon. We produce multiple generations of devices on standard semiconductor fabs. And it's this combination of using the silicon in existing standard semiconductor supply chain that allows us to scale so fast. And it's because we can scale quantum computing the semiconductor way. We build out individual unit cells, elements on chip that allow us to control our qubits.
And then we can build out larger and larger platforms by tiling out these elements. So we can scale by replication rather than by reinvention. And when you put this all together, we can do the same thing. We can just crank the handle and scale all the way from these 256- qubit platforms that are on our technology roadmap for next year all the way out to millions of qubits. But of course, having the best qubit performance, having the most scalable technology platform, that's necessary but not sufficient for winning the race. You also need to have, as Niccolo said, good unit economics. We now have the best-in-class economics across the quantum industry.
If we look at the BOM estimates we have, the bill of materials of parts costs, what we see is that for our 256- qubit product, we expect to use low single-digit millions of dollars of parts. And this doesn't increase fast. As we go out to 2 million physical qubits, we can do this with low tens of millions of dollars in parts. This is in comparison to our competitors who need to use incredibly precise foundry fabrication, non-standard. You need to use incredible cooling power and building things the size of warehouses. The estimates there are billions of dollars. Let me say it again. For low tens of millions of dollars, we can build platforms that competitors' parts costs are measured in the billions.
This combination of cheaper performance, scalability, and unit economics is what's going to allow IonQ to maintain its market-leading position across all epochs of the quantum computing revolution. Thank you. I'm now going to hand over to Dr. Christopher Monroe.
Good morning. I'm Christopher Monroe. I'm a co-founder of IonQ. I want to spend a few minutes talking about scale, expounding upon what we heard from Dean, Niccolo, and Chris Ballance. I want to talk from a very low level and also a high level. At the low level, IonQ technology, the transistor, if you will, of our quantum computers are based on individual atoms. Because an atom of barium-133 is the same in New York as it is in London, the recipes for scale are already at hand. We don't manufacture our qubits. They're given to us by nature. It comes down to, I don't want to simplify and say it's just engineering. But I've been in this field for 30 years to de-risk the physics of our qubit. We never think about the physics of the qubit. It's there.
It's a question of scale and how you engineer. I think when we founded IonQ, the physics had been thoroughly de-risked. But I will admit, when IonQ started about 10 years ago, we didn't know exactly the technology that would bear to allow us to scale. I think we saw one example of a technology that Chris Ballance presented in Oxford Ionics that will allow us to continue to scale. But I think one thing that's very important, and this we learned in classical non-quantum computing, is that to scale something, it needs to be modular. That's probably true of any complex system. Everything that's complex needs to be modular. We don't make a football field chip and call that a computer. We make a bunch of little chiplets and link them together in, say, a data center.
This backs sort of what cloud computing is today. Now, IonQ would not be a company if we did not have a vision for a modular way to scale our quantum computers. While I've been in the field for 30 years, about 20 years ago, we figured out a way to do this using photons, optical connections between quantum computers, very similar to what you see in modern data centers. Once again, this was very researchy in 2007 or 2006. The research labs at universities were really the state of the art, my lab, and also, incidentally, that of Chris Ballance at Oxford. We sort of knew that this was the vision forward. We didn't know exactly the technology that would come to bear to make optical interconnects fast and high fidelity.
As a wonderful next example of new technology coming to bear, IonQ merged with LightSync in order to hasten the speed of optical interconnects and allow us to scale in a modular fashion. One of the stars in that field is Mihir Bhaskar, who I'll introduce next. He'll tell you a little bit of the story on how we hope to and how we will optically scale our quantum computer chiplets. So, Mihir.
Thank you, Chris. So by way of introduction, I'm Mihir Bhaskar. I lead IonQ's distributed computing efforts. And before IonQ, I was CEO and co-founder of a startup company called LightSync. And before LightSync, I, for three years, founded and led AWS's quantum networking efforts. And I want to take the analogy that Chris pointed out a little bit further because when I was at AWS, I saw this happen firsthand. In between 2020 and 2023, the data center was being overturned so that the new hardware that was going in could support the AI revolution. And we all know the company that was really behind a lot of those developments, which was NVIDIA. And as Niccolo mentioned in his opening remarks, their core technology was a GPU, a different type of computer. It was good at solving problems that were hard for CPUs. It was being used by gamers.
It was being used for image processing tasks. It was being used by researchers at the time running small-scale, early AI workloads, and Jensen had the foresight to realize that he had to invest in a technology and platform that allowed for mass proliferation of GPUs at data center scale, and so in 2019, NVIDIA acquired a company called Mellanox, and Mellanox pioneered a standard for networking called InfiniBand, which was a replacement for Ethernet connectivity, which is just too slow to connect GPUs together in a data center, and so fast forwarding today, we can see that's been a very accretive acquisition for NVIDIA.
It was really the thesis and the inspiration for me and my co-founders to found the company LightSync, to take the ideas that Chris Monroe had pioneered in the early 2000s about how to connect quantum computers and how to build new technology to make those connections fast enough to actually scale out in the data center. And so the parallels coming together between IonQ and LightSync really align with that of NVIDIA and Mellanox. So I want to get a little bit more technical about how this actually works. How do we actually build faster interconnects? And to do that, we need to think the way that Dr. Monroe originally conceived of doing this, which was by taking two of these ion traps, having them send very, very weak optical signals known as single photons from each trap, and combining these signals at an intermediate detector.
If both of the photons arrive at the detector, then you've established a connection. In the real world, you have massive amounts of losses in optical systems, which throttle your overall networking speeds. That's why it makes it very hard to build a quantum interconnect that's fast enough to actually connect quantum computers together. The technology that we've pioneered at LightSync and have brought to IonQ is based on quantum memory. It's very simple. It is a buffer that allows each quantum computer to communicate with the network independently. It gets rid of this challenge of having to have multiple photons arriving from multiple quantum computers at the same time. With the quantum memories that we're bringing to IonQ, we expect we should be able to speed up interconnect rates by up to 50x.
So this is the type of generational leap that we see going from a technology like Ethernet to InfiniBand, and is more than enough to begin connecting quantum computers together in a data center. And so it's all great to have these figures in PowerPoint. I do want to emphasize here to those of you in the audience at Analyst Day, this is a real technology. You can see what our chips look like outside in the hallway. It's a technology that me and my team have been building for many years. We've built multiple iterations of it. And we've already shown that this technology works, that when you put this inside of a quantum interconnect channel, you can speed up the rate of this interconnect. So the technology is validated.
And if you get a chance to go look at the technology, what you'll see is a unique confluence of proprietary materials, integrated photonics, optical packaging technologies that are based on over a decade of research that my team has pioneered that exists only at IonQ, nowhere else in any other quantum or non-quantum company. And so to take the 10,000-foot view of how does this bring us closer to actually building quantum data centers, speed is the name of the game. It's the most important thing for connecting GPUs or CPUs in a data center to scale them out. The interconnect needs to be fast enough to distribute computations across the cluster. And so with our quantum memory technology and with the systems that Dr. Monroe talked about, we actually expect to have the speeds already to start building these first distributed systems.
And so I'm very proud and excited to be joining IonQ at this time because I've been in the field of quantum computing and networking for a long time. And we're proud to be the only company that's pushing both computing and networking, as we'll hear about next. But these technologies come together in the data center. Quantum computers and networks really synergize. It is one big network. It is one big computer when you're building a data center.
And so with the ability to build and link quantum computers at a fast enough speed, I think coming together between IonQ and LightSync is really the NVIDIA Mellanox moment that's going to allow us to take the quantum computing technology we heard about in the previous sections and enable it to scale and proliferate really at the large scales that are needed to ensure that we're not just leaders today with the performance that we have, but in every era of quantum computing and networking in the future. And with that, I mentioned that all of these technologies come together, quantum computing and networking in the data center. We'll hear about how we're excited about our quantum networking roadmap in the next segment.
Thank you, Mihir. Good morning. My name is Jordan Shapiro. I'm the President and General Manager of the Quantum Networking Division at IonQ. We are so excited to share a bit of where we are today and the Quantum Networking Division and our vision for the future. To start, I'm going to turn it over to Grégoire Ribordy.
Good morning, everyone. As Jordan said, my name is Grégoire Ribordy. I've been in the field of quantum technologies for 28 years when I started my PhD at the University of Geneva in Switzerland, and I co-founded IDQ, ID Quantique, in 2001. ID Quantique was acquired by IonQ in the first half of this year, and I'm now very proud to bring the solutions, the quantum security solutions that we've been developing with IonQ to a broader market and to make the cyber world a safer place. There's almost not a day without a major data breach. Personal information, government, corporate information is exposed. Of course, there are several very different attack vectors behind these data breaches. But one thing is certain today, information cannot be secure without cryptography. You all use cryptography many times a day.
When you connect to a website using HTTPS, cryptography is working in the background to protect information. It guarantees data confidentiality and integrity. Now, as we see from these breaches, the traditional security is not working. We need to move to a new paradigm in terms of security. Things are about to get worse. For decades, the security of our national security systems, of our global economy, has relied on one assumption, the fact that the math that we use in cryptography is too hard to break. However, with advances in the field of quantum computing, this is about to change. With progress in quantum computing, a sufficiently powerful quantum computer will be able to break the cryptography that is used today, and this represents an existential threat to our global economy, to our national security.
Try to imagine for a second a world where state secrets are exposed, online payments not possible anymore, or power grids remote-controlled by hackers in foreign countries. It's not exactly clear when such a powerful quantum computer will be available, but it may be as close as a few years, and it's now the time to plan and implement a major quantum-safe migration. It's all the more important that data could be vulnerable to a so-called harvest now, decrypt later type of attack. Hackers could be intercepting encrypted data today, storing it until a quantum computer is available, sufficiently powerful, and then decrypting information. One solution to address this threat is to rely on physics instead of mathematics. Of course, with mathematics, calculations get faster and faster, but the laws of physics are set in stone.
Quantum key distribution is about using a fundamental principle of quantum physics, the fact that a measurement causes a perturbation to protect information on networks. On this slide, I'll try to explain the principle of quantum key distribution. Alice, so there's this convention in cryptography that Alice is talking to Bob, and Eve is trying to eavesdrop. Alice sends information in quantum using a stream of individual light particles, so-called photons. And she encodes the bit values on these photons. Bob will record these photons using single photon detectors, special components. And if Eve, in the middle, is trying to intercept this stream, she can't do that without leaving a perturbation. Alice and Bob can then detect this perturbation.
It's also important to understand that we don't use this technology to directly protect information because if we did that, we would know that we've lost information after losing it. And that's obviously not the goal. Instead, we use it to send a key, a sequence of random bits, and then we use this key to encrypt information that needs to be protected. We heard already today that quantum is now, and to prove this, we've brought a demo that is in the back of the room, and I really encourage you to come see the demo after this show. On this demo, we'll show you two things. First, the fact that classical communication can be hacked without leaving any trace. On this screen, you see two PCs on the left and on the right. One PC sends a video stream to the other PC.
In the middle, we'll show you that if we put an eavesdropper that taps the optical fiber, this eavesdropper can collect the video stream without any detection. Now, in a quantum world, things are different. We're going to zoom now on the monitoring screen of the QKD system. On this monitoring screen, you see on the top left a measurement of the disturbance of the photon stream. It's low. It's a few percent now when we start. If Eve starts to eavesdrop on the communication, a perturbation will be caused, as you can see now, and an alarm will be triggered. Alice and Bob, the emitter and the receiver, will be informed of an attack and will be able to react. Of course, when the interception stops, the system goes back to normal operations. Again, quantum is now.
We're delivering solutions, quantum key distribution solutions. You can see one of our solutions, the 1U rack. This is our fourth-generation system. It has a high technology readiness. It's manufactured in batches in our ISO 9001 certified facilities. Very important, in security, it's always important to have independent testing and evaluation. We were able, with this solution, to obtain the world's first security certification from the National Security Research Institute in Korea at the beginning of this year. We're working with several other regions or countries to extend this certification. This solution is used more and more with the increasing awareness of the quantum threat. Adoption, sorry, is accelerating. The solution is used by tier-one banks, telecom operators, and government customers worldwide as an example to secure data center interconnects.
It is also in use in several of the more than 40 quantum networks that are in operation worldwide. For example, in Singapore, Singtel is using it in its national quantum safe network to provide quantum security as a service. In Korea, SK Telecom has used it to secure its 5G backbone against advanced attacks. It's also important to understand that this solution is easy to integrate. IT engineers don't need a PhD in quantum physics to be able to install this kind of solutions. It's typically deployed as an overlay to a conventional network, and it integrates seamlessly with classical telecom equipment vendors such as Nokia, Ciena, or Cisco. I will stop here. I'll pass the word to Jordan Shapiro now, but I'd like to remind you one more time to come see the demo after the show. Thank you very much for your attention.
Thank you, Grégoire. As you've heard from Grégoire and Chris and others today, when IonQ was founded, we had the vision of scaling our compute with modular architecture based on photonic interconnects. What we didn't know at the time but soon realized is that customers came to us wanting to purchase not just quantum computers, but also quantum networks. What we thought was at first an anomaly quickly became a trend, which encouraged us to expand our business to quantum networking about two years ago, and earlier this year, to formally launch the Quantum Networking Division. So what are we doing at the Quantum Networking Division at IonQ? We deliver commercially ready quantum networks, securing the world's critical infrastructure and setting the groundwork for a vast interoperability between quantum devices.
We specifically build production-ready networking hardware and software, which allows IonQ to be uniquely situated to transmit quantum information from point A to point B. We're also compatible with existing telecom fiber, so we can embed ourselves into current data centers for AI and otherwise, and to lay the groundwork for the quantum data centers of the future. So let's talk a little bit about the devices that we build in the Quantum Networking Division and produce today. We have the leading technical, manufacturing, and commercial capabilities in the industry here at IonQ, which has allowed us to become the one-stop shop for enterprise-ready quantum networks worldwide. Here's a quick overview of those products. First, we create photon sources. We're creating the core connection that allows one quantum device to interact with another.
We also build quantum frequency converters, allowing different types of devices, any type of quantum device, to interact with the same network using commercial fiber. We build quantum detectors to allow us to know when entanglement is happening and to use that to make decisions for applications ranging from computing to networking and communications. With the addition of the new LightSync technology, the quantum memories, we're able to build quantum repeaters. This is relevant for networks that are on the ground because you can now extend that network further than you would otherwise be able to do. We also provide quantum switches so that our customers can choose which devices they want to interact with on their quantum network, and some of those devices might include quantum computers.
So we envision a world in which quantum computers can interact with each other, enabling new use cases everywhere from clustered to distributed quantum computers. Moreover, we build and offer timing synchronization devices, which are key for building quantum networks. So we make time controllers and time taggers, allowing for precise coordination of different devices that you're connecting. Last but not least, we offer quantum security endpoints. Today, we offer two types of endpoints: prepare and measure and entanglement-based. Prepare and measure is in production today with leading telco and financial institutions, as Grégoire mentioned. Entanglement distribution is also in production today, offering, in some ways, an even higher level of trusted security because you are able to take the same entangled information and distribute that across multiple cores in your network.
Both of these approaches leverage the core principles of quantum mechanics and physics to make data information theoretically secure, meaning that it can't be eavesdropped on without you knowing about it and being able to fix it. So in whole, we are building the full suite of quantum networking products, which is why we're the one-stop shop. That brings all of these devices together to supercharge data centers today and lay the groundwork for quantum data centers of the future. IonQ is building the infrastructure for how all quantum devices will connect. But our ambitions lie beyond that. While we're pioneering quantum computing on the ground, we have customers coming to us asking to take quantum networking to the next level: space. Our acquisition of Capella gives us components to advance our efforts to take quantum computing to aerial and space-based networks that are truly global.
We can leverage technology like the IDQ Centauris platform to build quantum-secured networks at ground stations. Cerberis also allows us to combine quantum key distribution, or QKD, with post-quantum cryptography, or PQC, both of which we offer here at IonQ. We like to say, "PQC everywhere, QKD where it matters the most." Soon, our quantum offering will be available for communications across coasts, borders, and oceans, thanks to our space technology. Beyond the pure communication use case, we are looking at distributing entanglement globally and will also launch space products there. We envision entanglement being generated and distributed across data centers across the globe to connect a variety of quantum devices, be those computers, quantum communication endpoints, sensors, timing devices, you name it, all globally. Importantly, the Quantum Networking Division is breaking boundaries every single day with our customers.
Today, I have the honor of sharing two customer announcements that we have been able to produce with the Air Force Research Lab in the U.S. First is that today, we've been able to show that in a single quantum computer, we can generate two zones for qubits. One that is a computing zone where you're running complex gates. And the second is a networking zone. And we can shuttle qubits from one zone to the other. That computing zone allows us, like I said, to run complex algorithms. The networking zone allows us to take that information and transmit it to another quantum computer, core to what we're trying to do in the Networking Division. The second announcement with the Air Force Research Lab is that we've achieved a major milestone to connect our quantum telecom wavelengths to classical telecom fiber.
We've been able to demonstrate frequency conversion from 493 nanometers, which is the wavelength for trapped ion, into telecom. This is a major accomplishment. It allows us to put quantum computers onto the network. And it allows IonQ to be the first to do so. Impressively, we're also taking all this technology and using it in production today with customers. Customers can go work with EPB in Chattanooga, the Electric Power Board there, and use today an IonQ-powered commercial quantum entanglement distribution network. So today, customers can get onto this platform and use our technology live for entanglement distribution. World first. Meanwhile, in South Korea, SK Telecom, as Grégoire mentioned, is securing the nation's data through quantum key distribution. So this is sensitive AI and customer data securing their 5G traffic with IonQ. The important takeaways from today on the Quantum Networking Division are as follows.
Quantum networks are here. They're already here. They're securing the world's most sensitive data, and IonQ is building the foundation for the world's connected data in the future. Thank you very much.
Now, we're going to take a moment for some Q&A. So just to repeat, everyone has a QR code in front of you where you can pose any questions for our presenters. So I'm going to start with a question for you, Dean, which is, you presented today that we'll get to 2 million qubits in 2030. And IBM hasn't even announced getting to 2 million. What gives you the confidence that we can outperform IBM?
It's a great question. And so if you look at our roadmap, right, 256 next year, 10,000 the year after, 20,000 the year after, 200- 2 million, right? The basis for those numbers is not just thin air, right? If you think about the integration work that we have in front of us, this is in collaboration with Mihir and the LightSync team in terms of what does it take to start integrating the photon collection, the quantum memories into our systems. And you look at the technical kind of wickets we need to go through. Likewise, if you look at the technology for the Oxford Ionics, you saw the unit cells. You saw that we need to, the scaling strategy relies on replicating that unit cell. And it goes from our current 1D architecture to a 2D architecture.
The size of the unit cells, the ability to grow the unit cell, the ability to tile those relies on a progressive set of technology maturation, increases in die side length. At the end of the day, we've built our current technical roadmap based off of very sound technical reasoning and scheduling that is honestly completely doable and based in kind of sound engineering judgment and planning, right? We don't need miracles to happen. As Chris pointed out, we are relying on the power of the semiconductor manufacturing and processing industry to reuse the existing technology that they're already doing for that kind of unit cell replication. From an overall perspective, that 2 million at the end of the decade is exactly on target and exactly just a natural scaling of current capabilities that exist in the market.
And then maybe for you, Chris, what will Oxford Ionics leverage from IonQ's facilities to build the 256- qubit chip?
Yeah, I got a dinner actually up in Boston.
I think Chris Monroe said this earlier on, right?
Yeah, for the.
We use these barium qubits. We at Oxford Ionics use exactly the same qubits down to the number of electrons as IonQ does. Actually, when you look at the bill of materials, when you look at the physical system, when you look at the assembly requirements for this, they're 95% the same. This means we can essentially directly put these two things together. The teams are already now starting to plan this out. We just get direct speed up by doing that. We don't have to retool significantly on this. We essentially simply put these electronic integrated qubit control chips into the same kind of systems that we already have. Then it's just an integration challenge.
I was going to join you for dinner last night.
How about the fact that since we're increasing qubit capability so quickly, at what level will volume production actually be targeted?
Well.
Thank you. Sound good?
Sorry about that.
That's a great question. Ultimately, what we don't want to be doing is having to drive our own volume demand. What we can do is rely on things like the standard semiconductor supply chain, where we already have volume drivers that aren't us. Other people have already paid to solve the pain, and we can just use their learnings, and I can't say enough just how much the semiconductor supply chain is like magic. There's trillions of dollars of investment gone in to solve miserably hard problems, which we don't have to solve since other people already have, so for example, the semiconductor processes we use have been developed and matured for decades. We're not having to do anything new there. We're having to do designs, but the fabrication from those is already massive economies of scale.
Awesome. Maybe we'll turn to the networking side a little bit. We have a question that asks, could you please discuss how your quantum internet product secures information aside from eavesdropper detection?
Yes. So the quantum internet is a network that allows to distribute remotely entanglement across a network. And if you distribute entanglement, you have strong correlations between two photons. You can measure these photons. You locally get a random number, but it'll be the same random number at the other end of the network. And random numbers are a very important resource in cryptography because they are used to form keys. And so by using the quantum internet, you can establish keys in two locations, a sequence of random bits, and then use these keys to encrypt valuable information. There's really this two-step process: quantum communication first, and then encryption using keys of valuable data.
Awesome. Another question for you, Dean, to take it back. To what extent is the software vision more middleware versus an application layer-centric idea? And how will we monetize it by system sales or discretely?
Say the question one more time.
To what extent is the software vision more middleware versus application layer-centric, and how will we monetize it, system sales or discretely?
We've had to invest in the software stack, literally the full stack. It's not just the middleware that we've been investing in to date. We've talked about the need for qubit control at a very low level. That has been a software investment since day one, right? The capabilities that we have in terms of our qubit control differentiate us definitely in the market. As you move up the stack and start getting to our OS and the capabilities that it has, if you go even beyond the OS in terms of what we need to do in terms of job management, scheduling, all of those are differentiators for us in terms of what we can do with our systems and computers. And then as you go even further up the stack, you'll hear from Ariel Braunstein a little bit later today.
We have investments in our application work, right? We're building out libraries for chemistry. We're building out other capabilities. In the overall, I think the Fireside Chat session, you'll hear Martin Roetteler, who runs our solutions team. The team is working with customers, understanding their needs, taking that learning and reinvesting that learning into capabilities that then can be deployed to other customers. From an overall software perspective, we are trying to build an ecosystem. We're using customers now. They're getting familiar with our development environments. They're getting familiar with our capabilities. We just expect overall customer lock-in over time, the ability to take customers, get them familiar with being able to start running small applications, run larger applications.
We've seen this with our work with Hyundai and others in terms of their ability to take early work, leverage that, pass it along to other technical teams within the company, and then continue to drive and move forward.
Awesome. I'm going to take it back over to your side of the room, Jordan and Grégoire. We have a question about why use the quantum internet to connect to quantum computers across the world? What value does that provide beyond building two racks right next to each other? And on a related note, how does IonQ quantum networking fit into the satellite internet?
It's a great question, and thanks for the question, Hanley. There are many reasons that you could conceivably want to connect quantum computers at a broad distance, and today, the first and foremost reason is just availability. Quantum computers are new. They're coming out at new capability levels and new products, and so being able to reach across the border and in a trustworthy manner use someone else's quantum computer could be a clear use case, but in the future, we envision that there might be local reasons that you would want to have a quantum computer, for example, next to a local quantum sensor, and so the lower latency of that sensor and computer being next to each other, let's say in the U.K., allows them to interact more quickly, allows them to process data, and then send it over to a quantum computer in the U.S.
So there are many different reasons. The other thing I should mention is an application called Blind Quantum Computing, or BQC. This is an application where you can take a quantum algorithm and run it in an abstracted, hidden way on someone else's system so that you don't necessarily even need to trust the other person's system. It's a super secure methodology. It's one that's of interest to governments and enterprises worldwide that we're already working on. And it's another good reason why you might want to use someone's system across the world and run a secure algorithm, but know that no one's going to see what the work is that you were doing.
Thank you for that. Could you also expand a little bit on entanglement distribution that we have deployed and the work we're doing there?
Yes. So I mentioned one example of a commercial entanglement distribution network that we have at EPB in Chattanooga, Tennessee, and so today, customers can go online on that platform and start working with entanglement from one local data center to another in Chattanooga and demonstrate secure communications, demonstrate different use cases like QKD, but also experiment with new use cases on the enterprise side, so everything from instantaneous transmission of data to parallel processing across multiple systems. We've also announced that we are delivering a quantum computer to EPB in Chattanooga, so that will be the first site to have a commercial quantum computer and entanglement distribution quantum network co-located. You can envision that there are many opportunities there at the intersection of computing and networking once you have an entanglement distribution online.
Thank you. With that, I think we will take it back to the main agenda for the day.
Hello everyone. My mic is working. My name is Ariel Braunstein. I lead product and application for the computing side of IonQ. And I need a clicker. Here we go. And I would like to start today's presentation by acknowledging how confusing the quantum computing industry can be for people from the outside. Companies like ourselves are throwing a lot of acronyms at you and technical specs. And companies are talking about accomplishments and breakthroughs. And it's just confusing because nobody is talking about how all of these things add up eventually to actual value to customers. And value in quantum computing outside of research comes from applications. That is the expression of value for customers. And today, I want to try and fix that. So the interesting thing about applications in quantum computing is that every little choice affects application.
Every component, every architectural choice, every layer in the software that my colleague Dean was talking about in his presentation, they all affect eventually the outcome of these applications, what makes them viable or not viable, and all of these outcomes come down to four groups of things that we can categorize. One of them is what we call time to solution. How long does it take for an application to get to a solution? The next one is the accuracy of the solution. Then there is the cost to solution or the energy consumed in getting to a solution, and usually, it's a mix of all four, right? To help you understand how IonQ's roadmap impacts these outcomes, we selected four algorithms. These are not random algorithms. These are milestones in quantum computing. These are main algorithms in escalating sort of level of complexity and demand.
What we are going to do is talk to you about how these algorithms will be enabled by IonQ's technology roadmap and how does it compare to the known roadmaps of our competitors to give you a scale of actual outcomes of our choices and the choices of our competitors. So I will start this slide actually with a conclusion. And the conclusion is that it doesn't matter how you look at it. With every algorithm that you test, we end up having IonQ technology enabling a more accurate result that consumes less energy, gets to the result faster, and costs less. But let's look at some specifics. So the first algorithm that I would like to take you through here in this case is at the top. In AI, it's called multimodal reasoning.
So in this algorithm, again, IonQ is performing faster and consuming less energy end-to-end. And in this case, it's 40 times faster than the closest competitor that we were able to check. Next one is image change detection. It's used for satellite imaging. So think about the intelligence or insurance industry. In this case, IonQ technology enables 20 times less energy. And again, still costs less, energy consumption is lower, more accurate, and so on. Third, portfolio rebalancing in financial services. Again, IonQ technology, 170 times less energy consumed. And last in this example, Shor's algorithm, which you might have heard the name. It's sort of the measuring stick in quantum computing. It's like the ultimate evidence that quantum computing delivers on the vision that it was always conceived.
In this case, we predict in the next few years, RSA-2048, and that is used for secure web transactions, and ECC-256 used for cryptocurrency like Bitcoin and Ethereum, they will be compromised, and our technology for this type of work will already, so again, it will be faster, it will be cheaper, and two and a half times lower cost, right? It is worth noting that if you see the X's, the red X's, not every system out there will even be able to run these applications simply based on their stated roadmaps. These are their public roadmaps. But we still went through the exercise of evaluating to give them the benefit of the doubt. What if they could get to that scale? How will it perform?
What you can see here, it's simply even if they get there, the disadvantage because of their starting point is very significant because we simply built a better foundation, which is everything you heard here and from Niccolo in the beginning. Let's talk about this foundation. Let's talk about today. How do you compare across these foundations? The industry (and IonQ started in the beginning, the earliest days of the industry) spent the last decade building this foundation that we're talking about today. How do you compare between these foundations across companies? How do you know that our projections for the future are actually going to be there? The best way is to look at some benchmarks. Let's look at some benchmarks. The first benchmarks (and this is actually running today against the top line of our competitors).
So the first one is an algorithm called QAOA. It's a very powerful optimization algorithm that we ran with our partners like Airbus and Oak Ridge National Lab and Ansys and others. As you can see here in this graph, the top line is IonQ. IonQ Forte is outperforming IBM's latest system, 34.7% higher quality to solution. And what's interesting in this example is that the more qubits we give the system, the bigger that advantage becomes. Next benchmark. The next one is—it's called QFT. It's the workhorse of quantum computing. It's already one of the most useful tools in signal processing. And as you can see here, again, at the top line is the orange line is IonQ. IonQ Forte is able to produce here 181% higher quality to solution. Oh, I apologize. This is 73. I'm going to the next one inadvertently. 73% higher than IBM and superconducting.
The next one I already led with a conclusion. This is a very interesting, very powerful algorithm, search algorithm called QAA, and in this case, you can see that the random dots that you see there are samples of a search performing on random numbers. The horizontal lines are the average score of the benchmark, and you can see here, this is not even IonQ Forte. This is the older generation. We ran it with IonQ Aria, and Aria is outperforming all other or the top option of the competition with 181% higher quality of solution. All of these advantages eventually come down to opportunities. As Niccolo pointed out in his opening, we have been there first. We turn on our systems first. We've been running with customers first. We've been testing these algorithms. We established a foundation that is simply more robust.
Therefore, we get opportunities ahead of everybody else. And the opportunities are emerging everywhere. You can see here examples in pharmaceutical, health care, logistics, service industries, engineering, financial services, security, manufacturing. These opportunities are popping everywhere. And it's up to us to capture them and convert them into actual business, which we are actively pursuing right now. It's all eventually about finding key algorithms, running them on powerful systems that are tuned for these algorithms because of the choices that we're making every day and using them to find the right use cases that address real business needs, right? So let's look at one example that we recently did. So this is a partnership between AWS, IonQ first and foremost, AWS, NVIDIA, and AstraZeneca. In this example, we chose to address a step in the drug discovery process.
It's a way of scaling up the yield of drug production by using a special purpose catalyst, and the way we did that is use quantum processing to accelerate a classical simulation process. It's a quite elaborate process within pharma, but that's the level you need to get to to extract the value, work very closely with these partners, and the result of it was a 20 times improvement of end-to-end time to solution in this, and as Niccolo said in the beginning, if you actually follow the QR code and look at the publication, the claims are actually much greater. The improvement in the steps that we are actually accelerating is more than 600 times faster, but the end-to-end, we wanted to be conservative, so we're showing you the 20x improvement there, but please, I encourage you to look at the publication around that.
That's clear evidence of advantage right now. The next example is very interesting. It's a new approach to allow for improving the performance and the cost of large AI models, things like LLM, but not limited just to LLM. What it allows customers to do is take their existing large models that they invested sometimes $1 billion just to train a large model, add a quantum layer to it, and use the quantum layer to retrain the LLM, or in this case, the AI model, to perform highly specialized tasks with very little additional training and very little data. It excels at situations where the available data is very limited. You want to get high accuracy despite the fact that you did not have a lot of data and you did not have to train this model with a lot of money.
As you can see here in the graph at the bottom, our energy measurements demonstrate that IonQ consumes energy linearly as the AI model increases, while GPU increases exponentially. It's also a way to reduce the energy demand in AI. Going back to the same four ideas where we excel. We're making it cheaper. We consume less energy. We make it more accurate, right? And we make it less expensive. Let's watch a little video about this unique new approach before proceeding to customer updates by Rima and Margaret.
Quantum fine-tuning is IonQ's hybrid approach that makes large language models, or LLMs, smarter, more energy efficient, and less resource intensive. By adding a quantum layer to any foundational AI model like ChatGPT, Gemini, or DeepMind, we can unlock measurable accuracy gains in data-constrained environments using fewer training cycles and less energy.
Today's LLMs, like GPT, are powerful, but they're generalists. Out of the box, they work well on broad tasks but struggle with specialized high-stakes accuracy like detecting subtle fraud, classifying rare medical cases, or parsing nuanced human language. Prompt engineering alone won't bridge that gap. You wouldn't run your most critical workflows on the same tools a student uses to write an essay. So why rely on an untuned model when accuracy is everything? The fine-tuning process retrains models for domain specificity. But the classical approaches to fine-tuning are slow, expensive, and energy intensive. Moreover, they often plateau with noisy, limited, or complex data. This is where quantum classical fine-tuning comes in. Rather than just improving the base LLM, quantum essentially customizes or repurposes the model for a specific business-critical task.
By adding a quantum layer that captures richer, higher-dimensional patterns, we augment classical AI to unlock capabilities that were previously out of reach. The result? Higher accuracy with scarce data, fewer training cycles, lower energy consumption, less resource intensive. In sentiment analysis, simple cases are easy for both systems. But when there is nuanced human expression, like short or mixed-tone examples, classical models often misfire, while quantum gets it right. Across our tests, quantum delivered over a 70% win rate over classical in nuanced cases, exactly where accuracy matters most. Overall, we saw a greater than 3% accuracy lift, cutting total errors by 8%. At enterprise scale, reasoning AI models like DeepSeek R1 don't just run a single step, but tens of thousands of iterative inference steps before surfacing an answer. So every 1% of accuracy compounds.
A 97% accurate foundation model maintains signal across steps, while a 94% model erodes into unusable results. The difference isn't trivial. It defines whether an enterprise can deploy these systems in business-critical scenarios. IonQ's quantum fine-tuning model can enable high accuracy even at enterprise scale, and the smaller training footprint means faster cycles and lower energy costs. The future of fine-tuning isn't just smaller. It's smarter, and it's here today with IonQ.
Hello everyone. I'm Rima Alameddine, Chief Revenue Officer at IonQ. Thank you for being here today, so quantum is no longer a future promise. It is a present reality. Our customers realize that quantum computing and quantum networking will shape the future. So they are acting now. We have two types of buyers: mission-driven and ROI-driven customers. The mission-driven customers are investing because they are interested in fueling economic growth, attracting world talent, and driving innovation.
These are the universities, the countries, the states, the cities that are building quantum hubs and ecosystems to be leaders in the quantum economy. This category also includes governments. As we all know, governments are investing heavily because they know there's a quantum race out there, and the nation that masters quantum first will get to define the global order, then the second category are the ROI-driven customers. These customers are investing for a totally different reason. They are investing to solve their hardest problems, problems they're not able to solve today with current technologies. They care about accelerating growth, leapfrogging their competition, and defining industries. These are pharmaceutical companies creating new drugs, manufacturers building new materials, financial services customers building new cutting-edge models, and also enterprises in general that are looking to protect their data from intruders. That's where they use quantum security.
Our customers know that they need to invest now to be able to shape the future. And we're proud to help them lead the way. With that, I am going to turn it over to my colleague, Margaret Arakawa, who will share with you examples of our success and our momentum.
Good afternoon, everyone. Again, my name is Margaret Arakawa. I'm the Chief Marketing Officer for IonQ. Prior to joining IonQ, I had the privilege and joy of working at Microsoft for 20 years, in the old days, in the new days, and in the good days. I wanted to just step back just real quick to give you kind of a timeline on how and why we feel so confident. I picked the 1970s as a starting point for compute. And because I got to work at Microsoft, we cared about this one thing.
Remember the PC on every desktop? The PC on every desktop that became servers in every data center, but in the 1970s, it all started because PCs were being developed, and we were all of a sudden getting freed from the tyranny of a VAX machine, which I coded on, or a mainframe, so now you get a computer, as you all do, on your desktop or in your homes, so that started distributed computing and distributed computers. That then led to something else, and what that was, was the internet. I was overjoyed when the internet launched in the 1990s, but it actually launched in 1983, so to speak, when there was this thing, this protocol called TCP/IP that allowed people to communicate over the internet, or when the World Wide Web was actually delivered as a way to actually communicate.
The internet was built on the fact that we had computers on desktops. Because the internet now had a protocol, at the very beginning of the internet, as we all might know if you were around, it was neither fast nor secure. It got there. It got to the point where Netflix now can actually leverage that incredible speed, and I can have one gigabit downloads at home. What then did you get with the internet besides this distribution of these computers and ability to connect and transact? What followed was the cloud. I don't know really when the cloud was born, but I put 2002. The reason I put 2002 is why. What company launched a service that is the largest cloud service in the world? AWS. 2002 was AWS's Hello World. I decided that's when the cloud started because they leveraged what?
Amazon.com to build, and they are now running the world's largest cloud service. Could you have done that without the internet? And could you have done it without distributed computing? The next phase is AI. Guess why I picked that date? 2022. It's a really odd, terrible time to pick because actually, AI was officially discussed in the Dartmouth Conference in 1953. But I picked 2022 because that was a ChatGPT moment for people to understand how to be productive as a person. And then they understood now how do you gain productivity out of the cloud, out of the internet, and out of AI. So here's a bunch of companies that have been around: Microsoft, AWS, NVIDIA, Google, and Oracle, which had a great week earlier this week. The reason I put those down is all of them invested in this journey.
They all had to create the infrastructure, but all of them are actually cloud providers. They are working with NVIDIA. They're working with each other. Oracle's cloud actually runs on everybody else's cloud, so do we. Because the way I've learned, at least I learned, and on my continued journey, because I also interned at Intel, you have to invest in the infrastructure, and then you have to have a benefit for customers. The benefits for customers for IonQ today, what you can buy or what you can look at right there, we provide the basic things that customers want. They want to make sure that you are high performing, and Dean actually talked about, "We are on path. We will be delivering the world's most powerful quantum computers," and you require that in compute. You need to prove that you have high performance.
When you're networking, you need to actually care that you are doing it at lower cost, that you're doing it with scale. Chris Ballance talked about the lowest cost per qubit in the industry. Jordan talked about incredible scalability. Mihir talked about it from a photonic perspective, but when you think about scalability, think about all that infrastructure and all that investment you have to put in, and then Grégoire also talked about world-class security. I spent three years at Microsoft running a security group. I was told that I had to help Bill Gates prep for his last RSA Security Conference. I actually thought I was going to get fired because I was like, "I don't know if I can prep Bill Gates.
He's like brilliant." They said, "No, you have to tell him everything Microsoft does and sells at security and get him ready to get on stage." And the thing I learned is you got to work really, really, really hard. You have to keep at it. You have to have the IP. And you have people that want to make sure that their security is world-class. So with that, I will say that those companies did what we're doing. We have world-class sixth-generation hardware that we're selling. We sell it with an application layer that allows companies like NVIDIA to partner with us. When you see NVIDIA and you see quantum, they only can partner with people who have a software layer, hybrid services that allow a classical computer to talk to a quantum computer because it's a symbiotic relationship.
That's why NVIDIA loves working with us and why we're able to work on incredible applications together. And finally, as you heard, this is the stack that we sell. We sell end-to-end, one-stop-shop quantum networking, a type of quantum that is applied to networking in a way that for the first time, you're going to have this belief and knowledge that your network is the strongest, most secure network capable in the world today. So these are the customers that have bought what we are selling. Incredible customers. World-class Fortune 500, Global 1000 customers like Hyundai, like T-Mobile, like Samsung, like Cisco. We also have, as Rima has talked about, we also have incredible ecosystem partners. And it matters. We have to scale our partnerships.
The last job I had at Microsoft, I was running Windows in the United States, a small little business that if I got a cold, Microsoft would be ill. So I made sure that we did well selling Windows. When I first got to Windows, I was running Windows 8, and all our customers were mad at us. And they should have been mad at us because we delivered a product to them that was a surprise. The UI had changed, and we didn't get feedback all along the way like we had traditionally done. By the time we launched Windows 10, we were excited. And what I focused my entire job at Windows and Microsoft was scaling via partners. Ecosystems are built with partners. If you don't have, we are the only ones that are actively working with Google, Amazon, and Microsoft, and all our quantum computers.
If you have a little credit card right now, you could go start today. You could learn how to code for quantum because we have to democratize it, and then finally, we have, obviously, our incredible government, national labs, universities. Silicon Valley started because of what? Because they had incredible smart people at all of these schools that were local to a lot of innovators and entrepreneurs, and that's how they became Silicon Valley. We are trying to create quantum innovation hubs all around the world, and we're partnering with universities and government labs and the government to make sure that people don't wait for quantum because quantum is now, so I'm going to take you for a really quick tour of the last year. The last year for IonQ was a tremendous year.
But I wanted to just highlight in one year alone who were these customers, partners, and universities, and what was the news that we actually delivered. The first one is the University of Maryland. We signed a $9 million partnership with them to help them continue to build their quantum capabilities. And again, we have to democratize quantum, and we have to make sure that students learn quantum. If they're working on a CS degree, they should also look at whether or not they should learn how to code for quantum. One of our largest deals, one of the largest in the United States of America last year in September of 2024, was the United States Air Force Research Lab. It was a $55.5 million contract. This is an incredible contract.
In fact, we've actually added to that contract and are now over $100 million in our government contracts. NVIDIA, like I said, we worked on a new chemistry application. We did it with CUDA-Q, which is their next platform for coworking, collaborating with the quantum companies. But you must look at quantum companies that invest in that because it's the only way we're going to scale. In December, we shipped a quantum computer to Europe. A lot of quantum companies have quantum computers in their labs. So that's something that most people don't realize. Amazon, Microsoft, and Google, they all have quantum computers. They are not commercially available. So they put our quantum computer on their clouds because we need to be commercially available because quantum is now. We also have landed as part of an anchor partner to the state of Maryland.
You'll notice, search state investment in the United States, country investment in quantum, you'll see that the state of Maryland put a big stake in the ground to say they are building a $1 billion capital of quantum. And they chose us as their anchor partner. We're showcasing quantum applications with Ansys. Who's Ansys? Ansys, which was just recently bought by Synopsys. They're one of the largest companies on Earth that work with high-performance computing companies that need high-performance computing. And they do computer-aided engineering or computer-aided design. So you need to work with the companies and the partners that have the most complex problems. AWS, as I said, great partners with them. They were excited when we announced global availability of our latest and greatest quantum computer, Forte Enterprise, at an a remarkable AQ 36. DARPA, I know that there's questions about what DARPA means and who they are.
Obviously, they were pretty much the inventors of the internet. They are investing because they know quantum is now, and they'd like to get the leaders in quantum to help them standardize and figure out the future of quantum. We were selected as one of the partners that are going to help DARPA and the government do that. In April, again, we landed a $22 million partnership with an electric power company, the EPB company that's in Chattanooga. They, who have often been incredibly world-class and leading, they actually have the first gig fiber to all of their citizens. They are now the first utility in the United States to have a quantum computer and a quantum network, both developed by IonQ. We, in April, were lucky to be selected as somebody that the National Institute of Advanced Industrial Science and Technology (AIST) , and it matters, right?
Because where is the most high-performance computing and compute in the world? It's in the U.S. I mean, there's other foreign entities that we shall not name. But the largest ones that we allied nations, it's the U.S., the U.K., Germany, France, Japan, Korea, all of those places. Japan selected us because they want to advance quantum and AI in Japan. Sweden, there's this incredible company, Einride. I don't know if you've heard of them, but they actually were creating autonomous vehicles that were electric. So they took two trends and put it together. And these interesting-looking trucks are actually autonomous and electric. And they were working with us because logistics and optimization is very difficult. We also partnered with AstraZeneca. That was something that Niccolo talked about. Why do we partner with a pharmaceutical company?
Because they have brilliant scientists that know how to discover and accelerate drugs using classical, and they partner with us because quantum will be able to discover and accelerate the next drugs that could cure cancer. That's what we all want. We want the cure for cancer. The longest development time is actually in testing and drug development. We have to accelerate that, and we're working with AstraZeneca to do so. In July, we partnered with KISTI, the Korean National Quantum Center of Excellence. Very excited. We are their primary partner for them to develop their quantum center in Korea. And then finally, an R&D breakthrough that we just announced. You've heard of De Beers, the diamond company? Turns out they also sell synthetic diamonds for materials that can be used to accelerate quantum computers.
Mihir talked about photonic interconnects and talked about all of the things he's doing to make sure that the speed between quantum computers and chips is actually fast. Diamond materials is the next generation that's going to help that. So I did want to leave you with quantum is now. All of those companies have invested in quantum with us. And we're partnering because it's the time. Like I said in that evolution, we didn't disappear. I've been in computing for 30 years. It's time now to invest in quantum and partner with those companies. So with that, I will let my next esteemed speaker come up. He is the Chief Operating Officer of IonQ, Inder Singh. Welcome.
There you go.
Thank you. Well done. It's hard to top that. So thank you for putting me right after you.
It's great to be here with all of you. I get the easier part because you get no forward-looking statements from me today. But I do want to make sure that you appreciate what you heard from my lens. I've joined IonQ formally. I was on the board of directors. I was lead independent director, head of audit committee. I became many of those things when I was CFO Arm. Some of you may have heard of Arm. It's a semiconductor design company in the U.K. I was helping it get ready for an IPO. They wanted me to then sell it to NVIDIA and then back to IPO again. Exciting journey. To me, it was sort of like the present and past of computing. And maybe, look, I'm not going to diminish it. Maybe the future as well. So classical computing is on a track.
You heard all of the Jensen Huang statements. You heard of the GPU successes, et cetera, et cetera. That marches on. But if you think about kind of how these things are coming together over time, what's happening in semiconductor space? You're actually going to smaller and smaller and smaller nodes. Do some of you follow semiconductors? I'm guessing some of you do. You don't have to admit if you don't want to. If you look at 5 nanometers going to 3 nanometers, going to 18 angstroms, going to 14 angstroms, they're heading down the path where quantum mechanics is, and quantum computing follows a different set of laws and logic, wave theory versus particle theory, and leverages immensely, more exponentially more power if you can start to manipulate things at that level.
I remember when the company approached me to join their board. First they wanted me to join as CFO when I had sold it to Arm. They were like, "I can't do that." When I did join on the board, I asked our CEO at that time, "I'm thinking of joining this board. I think quantum is the future of computing." I won't name the individual. He's not there now, so it's not Rene. When he said to me, "It'll never happen. It'll never happen. Quantum's never going to happen." He could have been right. He had been doing it for 27 years. Could have been right. At that time, most quantum companies that I was looking at to compare IonQ to had no revenue. There were R&D shops, pre-revenue companies wanting to go public. This was one of them.
And when I looked at what this one had, and I looked at the leadership team, and I looked at the capitalization of the company, I looked at the comfort that this one had the funding, over $500 million in cash, to invest, then, okay, you increase the chances of success. You have to have the right scientists and engineers and researchers and marketing people and everything else. And what they've delivered is, you'll see on the next graph, is kind of that growth trajectory. That's the revenue performance of the company. 2021, when the revenues were $1.6 million, that's when I joined. And when I looked at the roadmap that I was hearing from Chris Monroe, Niccolo was on the board. He had been involved heavily in the company. I got comfort that we have the money to invest. And now the question was, can we actually build?
These revenues are by selling machines. I haven't seen another company in quantum have this kind of a trajectory in selling machines. I'm not trying to brag on this. I'm just saying these are just facts. Also, they're not forward-looking comments, so I'm not saying past is prologue, but our goal is to make sure we do the things on the right-hand side of the page going forward, which is grow our customer adoption, and you've seen, I think, lots of names and examples of how that is happening. Why is that happening? Partly because of the reasons that Ariel talked about, the kinds of applications that might be coming, the things around drug discovery, just things around battery chemistry. How can I make an EV go longer? There are problems that are intractable problems that classical computing might solve one day, but takes years and years.
And a machine like this, which can, instead of doing serial processing of information, can do simultaneous processing of all the different possible scenarios and tell you what to look at. And I'm not saying it's the panacea. I'm also not saying you should think of this like if quantum wins, GPU loses. No. I think we end up in a hybrid world where both coexist. There are certain workloads, like large-scale matrix multiplication. When you're doing LLM training, that's what that is, a table of weights and a table of data. And you're smashing them together and multiplying them many times until you get the right answer and the training model. Maybe quantum's not exactly the best at that. But you saw an example of what you can use quantum for. Take that and make it better.
If you think about the intersection of a Venn diagram of AI and quantum, and that crossover point is where you can really turbocharge models. I'm not predicting that will happen. I'm not saying that that's kind of what we're investing for. That's what we're enabling. So the diverse business comment there is important for you to understand. I think you saw a number of speakers present. I've listened to them. So this is, I think, day five on the job for me as COO and CFO from the board into the company. But I did not want to miss this journey. I didn't have to do this. I was on a number of boards. Niccolo called up and said, "Hey, you've become a sounding board," et cetera. Et cetera. You want to join, want to be part of the excitement? I couldn't say no.
And so, how can you say no? How can you say no when this technology can enable the betterment of humanity? I've always worked for a mission-oriented company my entire life. I won't. Industry owned and operated. It was classified at the top secret level. But we couldn't have done it as a government without our industry partners. And we won that race. And we won that Cold War. We're in a race today. We're in this room. We're live. We're in public. We're a public company. We're sharing with you our plan to win the race. And I agree with Inder. I'm thrilled. There's an ecosystem beyond this room that's also involved in winning this race, many of the partners that we showed you here today.
But unlike the Olympics, where if you come in second in the race and they give you a silver medal, there's no silver for second here. We must win this race. And that we, yes, as IonQ, I'm here for a reason to help win that race. But we as a country, we as an economy, we as a way of life have to win this race. It's that critical. It's that consequential. And so a blessing for me. I've had a life of service. I feel this role here is just a continuation of that service to a mission above self. Yes, we're a company. Yes, we want to impress you with our earnings and our growth, which we are confident we will. But what really generates my passion here is that focus on winning that race for the country, for our way of life.
And so very, very proud to be part of the team, Niccolo. Thank you for pulling this team together. Thank you for joining us, Robert. And thank you all for being on stage with me. And I'm going to move on to close our session here. So I feel that it's important to be a disciplined business when you're a public company. And we're going to finish exactly on time today, which I feel is a real feat. I want to thank our team for organizing a fantastic morning. I'm going to tell you a little bit about why I'm so excited to be here and why I'm so excited about where we're going. As you just heard from Robert, this is not just about national security, but also national economic security. There is no aspect of applied science that we don't positively impact. This is an amazing company.
Our employees love getting up to work in the morning. Our whole team does because no matter what your thing is, whether it's pharmaceuticals through to logistics, through to cryptography, through to defense, through to material science, through to sustainability, you make a bigger impact here than you will or can anywhere else. We are absolutely in it to win it, as I think you've seen this morning. We have shown this year that we will absolutely spend money to make sure that happens and to move things to the left on our roadmap because time is everything. It's not time we're actually not that worried about competitors, as you heard. A five-year lead is an astonishing one. Steve Jobs, 2007, launched this thing called the iPhone. Some of you in the room might have it.
He said at the time, "I think I have a two-year lead over Nokia and RIM." That two-year lead converged and converted into a $4 trillion market cap difference. Right? I think we have a five-year lead. We're going to continue to press our advantage by investing appropriately. The reason the lead matters so much, as I said, is not about commercial competitors. It's about nation-states. Nation-states are, of course, ferocious competitors. IonQ is tightening our security protocols, if you will, because what we're doing impacts so very much. You've seen today also that the very best talent in the world wants to be here. Right? Rick Muller only applied to IonQ. That tells you a lot about what he's seen the last 26 years looking at the government's quantum programs and what they've invested in, as well as how they've interfaced with other companies.
You've seen that A-pluses want to work with A-pluses, and so the Chris Monroes and Chris Ballances and Mihir Bhaskars and many, many more that we have in our applications team, Martin, Marco, Rick, et cetera, and there's a whole nother layer or two below that who don't necessarily report to me directly, as they say, but are equally vital in continuing to prevail and advance our lead. Everyone wants to be here because everyone knows that one plus one equals 30. Right? We have all of the world record holders at this company for fidelity. Right? Chris Monroe used to hold that record until, ironically, the other Chris at our firm stole it away from him.
We also have all the world record holders for, and I never get the exact phrase right, but basically quantum networking transmission speed was held by Mihir Bhaskar until, guess what, Chris Monroe stole it away from him at his lab at Duke. How amazing is that? Right? We have all the world record holders at our firm. And that gives us a darn good chance of holding all the world records for the next five years and the next 10 years. Success begets success in every industry that has this amount of applied science meets fundamental science, meets great engineering, meets commercialization. And you've seen today how extensive our lead is, not only in quantum computing, but how extensive it is in quantum networking. Now, you heard from Inder and Jordan and Grégoire and Marco about the criticality of quantum key distribution for our nation and for our corporates.
Everything I'm saying today applies to the Five Eyes and to NATO and all allies. At the end of the day, the cybersecurity market is truly enormous. It's enormous today. Everybody needs QKD today. It may not be the only thing you need, but it's definitely one of the things that everybody needs. When I think about who wins computing revolutions, I said at the start of the day that it's all about miniaturization, unit economics, and mass market penetration. This is the only firm on the quantum computing, if you will, revolution that is actually leading on all of those vectors. The most logical qubits, the lowest cost for a fully fault-tolerant machine, unbelievably low cost per logical qubit.
If you look at some of the data we shared today and divide numbers by the numbers of logical qubits we have, you come to two orders of magnitude at least as a lead, and this matters, right? Because as we've seen in every revolution, the more people that use things, the more that gets built on top of it. Right? Collaboration matters. Chris Ballance and I will be in the U.K. together next week, just as we have been here together, because what the U.S. and U.K. do together matters a lot, and what the Five Eyes do together matters a lot, and setting IonQ with Oxford Ionics LightSync and our friends at Capella as the opportunity to drive advancements in collaboration and cooperation between nations and between agencies is a dramatic accelerant, as well as differentiation.
The talent we've assembled is not only about the hardware, the compiler, the cybersecurity, or the computing, but also the application layer. And I said at the start today that I think quantum machine learning, quantum AI is probably going to drive even more upside than we have demonstrated on our application roadmap. When I think about the next five years, IonQ is positioned to continue to win, continue to advance our lead, and accelerate it. I think it is very likely that we're going to look back on 2025 and 2035 the way we look back on 2015 from today's vantage point.
We're going to say, "How crazy was it that we thought that nuclear reactors that are powering data centers the size of football fields is the future?" The future is going to be power-saving, cost-saving, and strong AI models that are perfect for quantum computers, which, by the way, the regulators like too, right? Because quantum computing, despite being about superposition, actually is pretty deterministic. We know what the algorithms that we run actually do. So the EU is going to like that. U.S. rather is going to like that. I think we're going to move into a quantum world. Yes, I agree. CPUs and GPUs are still on our phone. GPUs, CPUs, and GPUs will grow the total compute market together. But there's no doubt that every time you move to more powerful machines, the majority of the value capture goes toward the leader in that sector.
This is an industry that will have the same network effects of everything else in the tech sector. Right? There will not be 45 companies in the CPU space, the GPU space, the QPU space that succeed. There'll be one big one. There'll be a second place. And there might be some people fighting over the last 5% of the market share. Right? We absolutely believe we're going to be the 100-pound gorilla of the business of quantum. That's quantum computing, quantum networking, quantum sensing, quantum and space, how this connects together. We're the only company that can do it all, can win it all, and stitch it all together. And with that, I'm going to thank you so much for your attention this morning.
We look forward to having lunch with everyone and allowing you to attempt to ask us the questions that would be a Reg FD violation if we answered them. Other than that, thank you so much for your time. I really appreciate it. I look forward to working with them. Thank you for your support, and thank you for your trust.