Conference Transcript - IONQ 2026-06-09

Vijay Rakesh, Analyst — Mizuho

Thank you for joining us again for the Mizzou Global Tech Conference. I'm Vijay Rakesh, senior semesters and autosanalyst for Mizzou, and joining me is Inder Singh, CFO for AntQ. Thank you for coming to the Mizzou 2026 Global Tech Conference. We really appreciate it. AntQ obviously is one of the leaders in quantum computing. They are the biggest quantum computing company in the U.S. by revenues, pure play. and Q has a front row seat to one of the biggest transformations I guess in the computing industry from AI to quantum so and you know you guys have grown massively I think 2024 you are doing what 40 million in revenues and I think this year you're probably in the guide is 265 million or so so that's a that's a forex growth in two years impressive massive market cap of 20 billion but again thank you for coming to the tech conference and please join me in welcoming CFO Anderson thanks thank you so might be able to start off the with the big picture obviously quantum computing all in this year might be a billion in revenues for all the players combined you know we think as you look out to 2030, 15 billion, not big, but it's a 15x growth. Look out maybe 10 years, by 2035, we think it's a 300 billion market, which might be, you know, 10, 15% of the total compute time of 3 trillion. If you take Jensen's numbers, which, what would I know? He's been very good and very correct, so we'll take that. So I think the opportunity set is massive. Maybe you can walk us through what you are seeing, how the space evolved, what are the opportunities there, and then we can maybe dig through the roadmap for you.

Inder Singh, CFO

Sure, sure. And thank you for your comments. I think when I joined the board of this company originally back in 2021, it had zero revenue. I had a pretty impressive CEO.

Vijay Rakesh, Analyst — Mizuho

You won't go back that far.

Inder Singh, CFO

I had a pretty impressive CEO, you know, 30 people in a dream. And, you know, I was at Arm, had sold Arm to NVIDIA. I was CFO of Arm and looked at this as kind of the future potentially, right? No one knew. And, you know, to your point, five years later, here we are. And this company has grown not from one product, which we had until 2024, for to a platform company now with 10 or more products. So many, many TAMs, not just one TAM anymore, right? The TAM you mentioned, plus the TAM for security, plus the TAM for networking, plus, plus, plus. And so I think the opportunity set is bigger. It also does mean that you have to now, just as when I was at Cisco Systems, become the Cisco of this space, and we have that opportunity. We have to keep, obviously, innovating. uh yes when you know a really good metric of success is revenues because usually that means a customer is willing to buy your machine or whatever you're building and in our case we have not only a machine but our fifth generation machine that we are deploying today last week i was in switzerland meeting one of our largest customers and they have embraced three generations of our machines so i'm starting to see that stickiness flywheel effect that i've seen at other large companies happen and to your point our roadmap going forward and we of course have to execute but that roadmap is unmatched and so we have I think the right team the right opportunity we have to execute for sure we have to be able to have manufacturing capacity and all the things that you need to actually maintain that leadership and then not only in computing but also connecting computers together ours and somebody else's and then securing in in a quantum security way the networks that ultimately quantum computing will actually break right and that's getting much closer and yes i respect jensen and having been on the other side of the table with him when i was at softbank slash arm he has a vision and he executes for sure but i think even he's acknowledging that next to every gpu cluster there might be a cqvu one day right which is a big change

Vijay Rakesh, Analyst — Mizuho

You mentioned execution and obviously the roadmap. I think one of the things that you guys have noted is you have a 256 physical qubit system coming out. And just to set the table, the whole quantum computing performance is measured by the number of qubits that you bring to the table, whether it's physical qubits or finally there are corrected logical qubits, I guess. And you mentioned a 256 physical qubit system that can drive 12 logical qubits. How do you, is that on track? What is the interest level that you're seeing from customers? How does that raise the bar versus, or how does that raise the performance versus what you delivered on the prior generation? Yeah, sure.

Inder Singh, CFO

So I think like every other quantum computing company two years ago that was trying to scale the number of qubits and so on, thinking that that's the measure of success and it's one measure you know we were as well we were using lasers control to control the ions in our case like lasers so the more ions you have the more lasers you need the more lasers you need the more complicated the machine becomes a bigger it becomes the more expensive it becomes and we realized that that's not a path to scaling rapidly so we acquired Oxford ionics which was already re-incubating the ability to take ion traps and scale them using CMOS, which is semiconductor technology, foundries that have existed for three decades, and know how to scale things by taking ion traps and scaling them. So 256 to me is an important milestone. Really what's most important in my mind is getting to what's called fault-tolerant computing. So that's what classical computing had to do to become useful. It's nice to say I have so many qubits, great. Yeah, but how do I actually use it? How does the machine actually find the errors itself, correct the errors itself, keep going? And I think that's the ultimate test. So like people asking about qubits and superconducting and all of that, that's great. We're happy to have that talk. But until you have a company that gets to fall tolerant in quantum, it's not useful for many of the things that you need to have. And we've published now our roadmap for getting not just the qubits, but exactly how to get the fall-tolerant.

Vijay Rakesh, Analyst — Mizuho

Got it. And I think to that, one of the things that you have released is that walking cat architecture, I guess. Pretty catchy name, but just wondering what that is, if you can kind of explain in kind of a layman terms of how that helps you get to kind of a fall-tolerant architecture.

Inder Singh, CFO

Yeah, I mean, look, I think walking cat refers to the Schrodinger's cat, right? which exists in two states so until you open the box those of you have read about Schrodinger decades ago he mentioned that you know you don't know if the cat in the box is alive or dead until you open the lid and then you disturb the state so without opening the lid it's actually potentially both right mind-blowing thing that that you know not even Einstein could quite gather so that's what the walking cat refers to walking is because we can actually take that cat state ion and move it across our entire grid and look for errors and and heal those errors so that's what that is now there are two companies that I know of that have published a wall-tolerant computing roadmap awesome yeah and IBM and you know I haven't seen many others do it yet, but I think those are the two that I've seen.

Vijay Rakesh, Analyst — Mizuho

And since you brought up IBM, as you look at your roadmap, and obviously both of you are in superconducting quantum, you are on the trap-down side. They are most on the gate side. How have you differentiated, or how do you think your roadmap comparatively seems to be better as you move towards four-tall-run systems? I guess.

Inder Singh, CFO

Yeah, and I think I just want to clarify one point. They are superconducting, right? So they're using the ability to take things down to zero degrees Kelvin. Those who studied about zero degrees Kelvin, it's when all matter stops moving. And so you have to be just above that, right? We're not that. We use ions that exist in nature at room temperature. So two different things. And I'm not saying one is better, the other is not. However, to get superconducting ions, you have to actually get the temperature down, create the matter, create the atoms from scratch. So you introduce levels of error that don't normally exist. Because creating matter involves errors, and then controlling it involves errors. So I'm not knocking the technology. However, it's expensive very hard to replicate you know it's IBM so they'll pull it off and the timelines that they describe I hope we're focused on machines that can operate near room temperature yeah that have a fraction of the cost require a fraction of the energy and for a customer and I speak with many of them they care about total cost of ownership yeah like you know do I have to like pay a big bill up front and then do I have to keep paying to keep running this thing And so I think there are some advantages that ion traps offer, which are less errors, they exist in nature, so you have a natural advantage with fewer errors, and then you can get to logical qubits faster.

Vijay Rakesh, Analyst — Mizuho

God. Before I go down the road map rabbit hole, when you look at, you mentioned cost of ownership. Maybe if you can help us kind of bracket that, how does the cost of ownership compare for you versus some of the other quantum computing modalities, whether, I mean, superconducting you mentioned, but maybe we can start.

Inder Singh, CFO

Yeah, I mean, look, as I talk to customers, they care about both, energy is becoming a big thing, right? So, they care about energy, they care about cost of operating the thing, cost of ownership, cost of acquisition. So, the first hurdle is always like, how much do I have to spend, you know, does it have an M behind it or a B behind it, if I want something that's going to be useful, right? So the machines today, okay, they're fine, but the machines that are going to offer a really quantum advantage are going to be much more sophisticated and complicated, bigger, more expensive. In the case of superconducting, again, it's kind of where we were two years ago. The bigger it gets, the more expensive it gets. And so, like, there's a 20, 30 times cost advantage that we have to start. And then now that we have semiconductor, actually the cost goes down with time as the machine gets more powerful and not the other way.

Vijay Rakesh, Analyst — Mizuho

Got it.

Inder Singh, CFO

So that would put you at a pretty significant advantage if you're looking at 20 30 times uh better uh cost of ownership versus a super conducting against with due humility yes yeah you know i think that um you know if you talk to a customer and say you know here's here's what it's going to cost you day one yeah and by the way we are on our fifth generation if you stay with us we'll get you the sixth and the seventh and today's machines ours anybody else's are forklift upgrades got it so you replace an old machine entirely with a new machine when we get to 10 000 qubits which is right after 256 it becomes modular upgrades got it so you could put the machine in and now it's some modules you have to swap out you get stickiness um lower cost yeah and so i think there's a natural advantage which is why right now we've been super focused on grabbing market footprint just grab market share Because then you have sticky customer relationships you can build on.

Vijay Rakesh, Analyst — Mizuho

Got it. I want to go down the, go to the KOTIA roadmap. You are at 12 logical qubits now on a 256 physical qubit system. You have noted about 800 logical qubits next year, which is an order or more jump in terms of logical qubit. Part of it might be the 412 architecture that you have been talking about, but it might be you can help us bridge how you get there, how you feel how comfortable you feel with that yeah because it's it sets you apart from the rest of industry nobody's in the hundreds of logical qubits yeah by next year yeah so might be yeah absolutely so so you know we've said this publication goes back to your very first question which is like getting to 256 is our first chip based system right and we said on the earnings call two quarters ago that we were making good progress on it already.

Inder Singh, CFO

We're using a US foundry to do that. That chip was a prototype was achieved in a couple months versus nine months which we had thought.

Vijay Rakesh, Analyst — Mizuho

Yeah.

Inder Singh, CFO

So accelerated much earlier and we have now have the same team that developed the 256 chip working on the 10,000. Still early days mind you. So I'm not saying that's anywhere near done but otherwise we might have been doing that a year from now instead of this year and so you know building the 256 gives you confidence that you can put ion traps onto CMOS and allow Skywater who we were using to help and they know CMOS like inside out scale that for us right and 10k is using the same 256 chip replicated and scaled so it's from your standpoint is basically scaling the 256 physical qubit chip to a 10,000 physical qubit chip to deliver the 800 logical qubits, I guess.

Vijay Rakesh, Analyst — Mizuho

Is it that easy?

Inder Singh, CFO

Well, you know, so far what we've seen with the 256 at least, getting from 100 this year to 256, we've been able to do. And so like the early, I'm not yet ready to report progress on 10k just yet. We will on earnings calls coming up.

Vijay Rakesh, Analyst — Mizuho

But we feel pretty confident that we know how to do that and spoken with the CEO of Skywater and his team yeah who demonstrate the confidence that they've done it before and it's about using semiconductor scaling basically and what's what's the response being from from customized because it's like we were saying before it's the 800 logical qubit is a massive step up it starts to now you start to be able to resolve real-world problems, real-world quantum computing problems, not the quasi traditional quantum compute, some in the middle. What's been the response from customers as you show them this roadmap that, hey, this is feasible. This is what the manufacturing partners are saying.

Inder Singh, CFO

Yeah, I mean, it's an unmatched roadmap, right? So no other company that I've seen has published a roadmap that even comes close. forget the 800 then you get to 80 000 when we have 2 million and so that roadmap no one has been able to publish and of course like the customer i was with last week they bought three generations including the 256 before we built the machine yet by the way and so they brought access to it in the beginning and so you know over time they'll they'll convert over to the 10k and so on what we're doing is making sure that as we build more powerful machines we're also building the app store for the iPhone so the the algorithms that will run on each generation of machine and we won't build all of them but we'll build some selected ones and so that's an also a differentiation I think for this company which is one of the world's largest if not the world's largest team of algorithm developers got it unless you scale from 256 to 10k physical qubits or eight in a logical qubits is the platform scalable meaning is the software compatible does do the OEMs have to do a lot of heavy lifting when they go to the internal logical qubit system how does that how is that transition yeah that's a great question I mean so again over the four or five years that I've been involved with this company we build our own machines right in our own premises so we have a factory in College Park Maryland and one in on the West Coast in Seattle and we are now going to have four of those one in in Oxford and one in Colorado. So we integrate all of the systems ourselves. And to build a 256, which is a different modality in a way, still ion trap, but semiconductor based, the chip is different. That's the size of your fingernail. The system around it is similar. So we have five generations of S-curve learning on what works and doesn't work, on the compiler, the optimization, the control mechanisms, the ways that you control what the machine actually does.

Vijay Rakesh, Analyst — Mizuho

Got it. One of the things that we have heard is to get to solving real quantum scale problems is you need 100 logical qubits. So the first question is, do you have the fidelity coherence level for those 800 logical qubits? And second, if that statement is true, that once you get to 100 logical qubit, you can now open the floodgates open for real quantum scale problems. Are you seeing that? Are you seeing a significant expanded interest level from customers to start engaging on much bigger projects, whether it is in molecular synthesis or pharma or biotech or cybersecurity, I guess?

Inder Singh, CFO

Yeah, where we're seeing the proof already with last year's machine, which was only 36 qubits, and this year's is 100, was early things that you can do potentially. And each generation, more qubits, more power. And as you know, like with quantum, it's not how many qubits you have. The power is actually two to the power of the number of qubits, two to the n. With classical computing, it's two times n, right? So this is the exponential power. So each generation, like four eclipses is the one before that. And so, yes, you can do many more things with 100 I'm not going to say that gives you a quantum advantage of course not I think quantum advantage happens when no classical machine can simulate what a quantum computer can do and I think that's closer to the 10,000 machine so whoever gets there first I think gets to that quantum advantage first 10,000 logical no total physical because I think 800 to your point you start unlocking incredible things yeah interesting so So you have grown 47%, I guess, or consensus has you growing 47% for 27, I'd be low-balling it, if you're seeing this massive explosion of interest.

Vijay Rakesh, Analyst — Mizuho

So we'll let you think about that.

Inder Singh, CFO

I've been in this business for too many decades to comment on that.

Vijay Rakesh, Analyst — Mizuho

Back to what you mentioned about cost of ownership. As you scale these systems to 10,000 logical cubists and 256, how much does the cost of the system go up? I know you mentioned it's a chip, so that's scalable, so it doesn't really scale. But how about this stuff around it? How does the system cost scale? Is it like 256 going to 10,000 is what, 20x? But does the system cost go up 20x?

Inder Singh, CFO

No, as I said, I think most of the system is largely the same. And as the number of qubits goes higher, it's a similar compiler. got it you know all of the infrastructure around it is about the same we said publicly our analyst day last September that the bill of materials for our machine is less than 30 million and so I think the cost actually goes down as I said right because the one thing that changes from machine to machine is the chip and the module that goes in there not the whole system yeah right so 10k and on it's the same machine literally yeah yeah so the performance per dollar actually for sure significant for sure if you measure it like cost per qubit there's

Vijay Rakesh, Analyst — Mizuho

no comparison and to that same TCO argument I guess how do you see the minute whether you're seeing customers start to now ask for more on-prem rather than using it on the cloud because you would assume like now that you're starting to get to some very, you know, attractive but, you know, very significant logical qubit numbers, you start to crack some of the quantum computing problems and you probably want to have some of these platforms in-house. What are you seeing that, are you seeing that mix start to shift more to on-prem versus quantum as a service?

Inder Singh, CFO

Yeah, I think that what I'm seeing is, to your point um there there are data center operators who are putting a quantum computer next to their data center yeah because they're envisioning this world of hybrid computing already right and to get hybrid computing you have to have both machines yeah like the ai factory and the qpu so they're there i think it's a natural fit and i'm seeing like three data points now in a row that lead me to believe that's a trend there are others who may not need an entire quantum computer all the time and may want to just access it in a cloud environment and ours has been behind all three major clouds since inception um so you know we'll offer it both ways as i said on the earnings call we we want to meet the customer where they are right if they can buy the machine and pay for it fine if they just want access to it that's okay too and we can enable both and over Over time, there might be hybrid deployments where someone says, I want a machine on-prem, but if I need more compute power, I want to be able to reach that through the cloud. We can enable that, too.

Vijay Rakesh, Analyst — Mizuho

And to that point, you had NVIDIA on the AI side talk about where they see AI quantum computing as an adjacency where you can run CUDA and CUDA-Q on the software side, have a NVQ link to connect the AI platforms to a quantum computing platform, is there, are you seeing a real demand at the CSPs of the cloud for the ability to move workloads back and forth, or is it more of a defensive mechanism to, okay, I need that technology, I better look at it now?

Inder Singh, CFO

Yeah, I think that as I saw with classical computing in my career, CPUs and GPUs still coexist. you know like right right so cpus haven't gone away and i think qpus will come next to these two so it's going to be some things that only can be done on the quantum computer and you can't send it back to the ai factory uh and maybe vice versa like llm training probably gpus are are perfect for that yeah but when you're trying to simulate a hundred million scenarios in a nanosecond and get one answer which is the best then only qpus can be able to do that. So I think that workloads will naturally start to be sent to the right machine. Now, you know, are there workloads that can be split? Yes. Then you need a hypervisor layer to decide which one goes here, which one goes there. And I think the industry has to evolve to that.

Vijay Rakesh, Analyst — Mizuho

Got it. And so to that point, as you look at your customer pipeline, whether it's academic or government or enterprise, where are you seeing a real thirst for, you you know, okay, I need this eight-tonological qubit system. The logical qubits is great. Where are you seeing the highest, biggest interest, I guess? I mean, we step back and we think, yeah, molecular synthesis, all those sound great. But from your standpoint, like I see look, because you have a front row seat to that, but where are you seeing customers really saying, okay, I need this system right away?

Inder Singh, CFO

It's where you would expect, where the number of inputs is too large for classical computing to try to even handle. So the number of possibilities, just impossible. So life sciences, financial services, material science, battery chemistry. So those are the things where you have way more things to model that classical will never model. So early adopters will be industry like those. And so we've done enough proof points with customers that we've announced. AstraZeneca in terms of accelerating drug discovery, for example. Protein folding is around the corner. We've done the first instance of it. The second, third is down the road. Those are things that, you know, small molecule modeling of drugs, which has been impossible historically. Those are things where absolutely there's a relevance to the machine that we build that does quantum computing equally. the the as these machines get more powerful ours will and we're building the algorithms that optimize to run on them they will do things like break encryption in the not too distant future so as responsible citizens if we're creating the ability to crack encryption we need to have the ability to deliver a solution like the antidote to that poison right as i call it so we have a security business that we acquired that is a world leader in post-quantum security so that customers who want it now can get it now knowing that at some point whether you believe jensen or you know google right uh and others were saying it looks like we're almost on the cusp of that that we need to have that as well so like some begin with just saying i want the security now right right like financial services firms increasingly are saying we want we want the security now right and we don't want to wait till it's too late uh you know when when and this is a national security matter, like countries are racing to this. And whoever gets there first, you know, countries want to make sure they're there first. So our platform strategy is around having all of that, including like connecting computers to other computers on the ground, in space, perhaps in a computing way, putting sensors in space that can't be jammed, things like that.

Vijay Rakesh, Analyst — Mizuho

I mean, obviously quantum computing, you know, I think this opens up a huge opportunity, it opens a huge window, I guess. And unless you open the window, you don't know what the opportunity size is. But that also begs the question, if you look at the AI side, you're spending going to 800 billion this year, a trillion next year. And people look at, okay, what does the cloud spend? You know, how many gigawatts is it? So how much do I need to spend on AI? How do you size the quantum computing opportunity? Like if, how do you look at what is the problem set and what is the quantum computing dollar that can be assigned to that? And to that point, how do you look at the different markets? Because that's the challenge, I guess, investors face. How do you size this?

Inder Singh, CFO

This is a fundamental change that, frankly, I'm not sure anyone could really model. Your $15 billion number earlier, I've seen 30 in other numbers as well, right? So I don't obsess on that, right? What I obsess on is if our revenues exceed the entire rest of the industry, their works revenues, okay, we are a pay setter, right? And so that's something that, not that I rest on that.

Vijay Rakesh, Analyst — Mizuho

I'm saying even the 15 billion, when you look at 2030, let's say, the AI CapEx number, the AI spend or AI market time is 2 trillion, 3 trillion, right? Some 2 to 3 trillion, let's say. 15 billion is like a rounding error.

Inder Singh, CFO

It is, and I think that's where when you think about rounding error, you think about a machine that can do way more than what the trillions might be able to do. When you think about ours plugs into the wall, doesn't require an electrical grid, doesn't require zero degrees temperature, et cetera, et cetera. At some point, that wins the argument. Right.

Vijay Rakesh, Analyst — Mizuho

And the flip question to that is, once you get to these 800 logical qubits and a thousand logical qubits, why can't that 15 billion be significantly higher, given exactly what you just mentioned, that it's scalable, it's lower cost, it opens up a whole new window, which, you know.

Inder Singh, CFO

I think the 15, the humble opinion, I think the 15 billion is way underestimated. Like, who knew ChatGPT before OpenAI announced it? Who knew that we'd be spending trillions on that, like three years ago, four years ago, right? And I'm saying this, you know, we may be four years behind AI, or three years behind AI. But at some point, I think there will be a realization. The good news is with our form factor, you don't need to spend a trillion. I'm the board of our data center company, and they're talking not in gigahertz of compute power, but in how much energy actually is being consumed as a measure of the power of the data center. That's not a conversation you have to have with our computing.

Vijay Rakesh, Analyst — Mizuho

Got it. And you guys have put out a 2030 target of 80,000 logical qubits, I guess. Is that primarily based on just scaling your sky water engagement? Yeah.

Inder Singh, CFO

And look, I don't want to make it sound easier than it is. It's always hard. Right. But yeah. But we are now leveraging mature nodes in semiconductors that are, you know, basically fully depreciated. Yeah. So low cost to begin with. And in our partner in Skywater, a technology company that knows how to scale.

Vijay Rakesh, Analyst — Mizuho

Got it.

Inder Singh, CFO

And, you know, we're unique in the semiconductor roadmap in that way.

Vijay Rakesh, Analyst — Mizuho

Yeah.

Inder Singh, CFO

Because, again, I think we learned two years ago what I think others will learn two years from now. Right. which is there's a limit to which you can scale without using something like a semi-knocker.

Vijay Rakesh, Analyst — Mizuho

Yeah, yeah. I want to quickly go to the manufacturing side. We've got like three minutes left here before we take some questions from the audience. You have your manufacturing partner, SkyYT. Can you talk to how, you know, do you have multiple fundrais? Obviously, you are in the process of an acquisition there. do you see IBM start to compete in that space in manufacturing because they just got a billion dollars in the U.S. government, I guess? Maybe you can address how you see that.

Inder Singh, CFO

Yeah, there's more than one foundry you can use. But that foundry needs to focus on quantum, not just classical semiconductor. So Infineon is another foundry, for example. And we were looking at Infineon in the very beginning when we acquired Oxford Ionics. And Infineon is on a learning curve. they're huge yeah like their quantum thing was a teeny tiny part of that equation so when we looked at how fast can they scale for us and when they realized how quickly we were getting to some of these milestones and the volumes we were talking about it wasn't going to come together and what we got with Skywater was a company that knew about quantum to begin with a little bit enough where their learning curve was not as steep and their willingness to invest in it higher. And they've demonstrated they can do in two months what would have taken nine or 12 maybe elsewhere. So there will always be more than one foundry. I'm glad. You need that, right? Because without that, you don't get an industry. And for sure, Skywater does work with other quantum companies. They will keep doing that. And to your point, there will be others that want to get into quantum. That's a great validation of the space. I actually love that.

Vijay Rakesh, Analyst — Mizuho

And with Skywater, is that part of how you are getting to that 10,000, 80,000 or a million physical qubits, I guess? That roadmap is something that you're already working with them two, three years out, I guess.

Inder Singh, CFO

I mean, one step at a time, right? Two of G6, we needed to really get comfortable that they could do, right? And they have done. 10k we're working on with them now and so everything is a step if we can get them to move our tech roadmap to the left and get to these milestones even faster which is part of the thesis i think yeah then absolutely they remain the one i think that we will be leaning on and depending on yeah um never say never if you need a second source you can always get a second source yeah you know we have a balance sheet that's among the strongest if not the stronger in the industry. And the ability to invest for multiple years and deliver in the shorter term as well.

Vijay Rakesh, Analyst — Mizuho

Yeah, fantastic. I mean, you guys have executed well. I think your logical and physical Qubit roadmap looks significantly well above where the peers are.

Inder Singh, CFO

I like it when I don't have to say that.

Vijay Rakesh, Analyst — Mizuho

Yeah, there you go. No. Maybe we'll take some quick questions from the audience. Got a minute here. Go ahead.

Inder Singh, CFO

All right. Quick question. I think...

Speaker 2

Most of the industry today talks about, you know, general purpose computing, you know, over the next few years, curious how you think about vertical specific, you know, applications, you know, within financial services. Do you guys think of, you know, buying, building, partnering, and what's kind of like the mindset behind that?

Inder Singh, CFO

Well, you know, there are, what we look for are algorithms that can solve multiple industries problems, right? So there's a thing called Hamiltonians, which is like a common denominator among algorithms. And if you can optimize Hamiltonian-based algorithms, you can serve multiple different industries, including financial services, life sciences, and others. And so we don't want to spend on everything. We want others to innovate on top of our platforms and develop their own. But for the ones that we can address, we'll address ourselves. And so we have drawn together algorithm developers from really AI companies, like the trillion-dollar companies that Vijay was talking about. And they've come together at INQ basically to deliver on something that is beyond AI at this point. So I think that we will end up partnering with others like NVIDIA, which we do, because our machines are being deployed next to their machines. And all the things that Vijay was saying earlier are true. like our things have to be interoperable with theirs. But at the same time, like the ability to deliver what AstraZeneca wanted from us required both NVIDIA and us to go in together to do that. So absolutely, partnerships where it makes sense. And as I said earlier, like in our networking strategy, we want our network to be able to connect any one's computer to anyone else's quantum computer. And that's how you become like a de facto leader over time, which I learned from Cisco and others, which is like, connect anything to anything. And, you know, there's room for a superconducting future. There's room for an ion trap future. There's room for maybe others as well. Don't try to pick those. Just make sure that you connect everything together. Build the best one you can, which we're doing. So our competitor is us. So like, you know, 256 now, 10,000 next year, 20,000 a year after that, 200,000 a year after that. that's what we're focused on uniquely and then don't obsess over where everyone else is help everyone in fact succeed and it keeps you honest and it keeps you innovating and keeps you a little paranoid great I think that brings us to the top of the hour so it's been an absolute pressure and thank you for joining us thank you