I think we are live. Welcome, everybody. I'm Joe Moore, Morgan Stanley Semiconductor Research. Very happy to have Inder Singh, CFO of IonQ. Thank you for coming.
Of course.
being here.
Thanks for having me.
I think we haven't had you since the Arm days, so -
Yes
... it's good to see you. You've been in the CFO seat, I guess a little less than a year, but you were on the board before that. What drew you to the IonQ opportunity?
Well, I mean, like I said, I was on the board. I actually joined the board when the company first went public.
Okay.
You know, quantum computing four years ago, like, no one was really talking about it seemed like it had promise, but you never knew for sure.
Mm-hmm.
Over the four years that I was on the board, I got to see the company actually scale the ability to make us a more powerful computing device year after year after year. Peter Chapman, who was the prior CEO, took it from basically, you know, almost zero revenue and created a product that was more powerful each time in the market, and other companies had gone public around the same time also. I. You know, being at Arm, I was like-
Yeah
Okay, this could be the future of computing, and wouldn't it be great to have, like, a front row seat, as part of it?" I was still CFO. We had sold Arm to NVIDIA. It was in that holding period, waiting for approval. This became a interesting company the more I watched it scale.
Mm-hmm.
It had the right engineering talent. It didn't need to go to market much because, frankly, you know, sales were still pretty small, and mostly it was research laboratories in the first few years or so. To see it come out of the research laboratories, and then importantly, when Peter Chapman announced his intent to retire and move on to other things, we faced a CEO succession, and I was lead director at that time. I was on the audit committee and we were looking at internal and external candidates, and it was hard to find somebody with 10 years of public company, quantum company experience.
Yeah.
Like, there were none. We looked at all of the other candidates and sure enough, you know, among our choices, we had a really good CEO, who I thought, you know, physics major at Cambridge, multidimensional thinker. Early last year, we appointed Niccolò de Masi as CEO of the company, and today he's Chairman and CEO. He very rapidly lived up to my best expectations as someone who's not just gonna be, like, a linear executor...
Yeah
... but looks at the market more broadly and says, like, "What are all the different things quantum or quantum-related that we need to have?" Kind of like Cisco, if you think about Cisco, the way it became the company, let's try to be that and look at the networking of quantum computers, the security that you need after a quantum computer breaks encryption, the space-based assets you might need to quantum secure, et cetera, et cetera. In September, I moved from the board to the company as COO and CFO, and it's been pretty amazing. It's been like, I don't know, five, six months. It feels longer than that, for sure.
Watching the company not only build a machine that each year was faster or had more qubits, but also a company that is building the applications that go with it. Because what good is the iPhone without the App Store, et cetera? You have a portfolio of products. A year ago, we had one product, and now we have a dozen or more. So many vectors of growth, and yeah, you know, as board members, you know, you get to see sort of, like, snapshots of the movie, but never the movie.
Yeah.
I wanted to be part of it.
Maybe as a relative newcomer's perspective to the company, can you talk about why investors should care about quantum? You know, we're not at the point yet where we're doing things that can't be done classically, but it's coming. What types of applications do you think will happen there? Should we think about, you know, the market cap of AI as being the territory you're going after? Just why should we care about quantum? Obviously, code breaking and things like that we get, but the sort of broader use cases.
Yeah. Those are great questions. I mean, I think the promise of quantum is around doing things that a classical computer can't do, right? There's this thing called quantum advantage, and at some point, you reach that, where classical computing just cannot do that many number of calculations or that many number of simulations of things before giving you the best answer. And forever that's been, like, a bit of a moving target, frankly.
Mm-hmm.
NVIDIA isn't standing still. I mean, they keep innovating, so classical computing moves as well. At some point, you cannot simulate a calculation on a classical computer that a quantum computer could do. We're closer to that than we were three, four years ago, for sure.
Mm-hmm.
It's, you know, humble opinion, maybe a year or two away before you get to that point. The power of quantum, as most people may know or not know, is the fact that it's, like, two to the N power versus two times N. The number of qubits that you add have an exponentially more powerful impact potentially. Quantum computers are right now still of different modalities, right? There's superconducting, and there's photonic, and there's annealing, and yes, there's ion trap, which is what we have. Some start with more errors, some error correct, some start with less errors. The point of the industry, though, is that there'll be one or two modalities probably that succeed in the longer term, to be decided which ones. Ion traps begin with fewer...
I looked carefully at, as much as I knew. I mean, a couple of engineering degrees, you know, but enough to make me dangerous to look at the different modalities and realize that there'll be some applications that work better on one and some better on the other. For investors, I think right now there are a lot of companies that are private that want to go public, that are private, that wanna stay private, that are public already and available on the market. For me, it's a question of, is there a company that is building not just the iPhone, if you will, but the iPhone and the App Store at the same time? The ability to do the application that runs on that computer each successive generation. What we've tried to do with IonQ is to be that.
Build enough applications ourselves to make it relevant either in life sciences or the acceleration of drug discovery or material science, not everything. Also be able to keep that going for each successive generation with a roadmap that you can invest behind. If we were only in quantum computing as we were a year ago, that would be interesting itself. What Niccolò has done quite rapidly is add the networking piece, as I said, the quantum security piece, as I said, the sensing piece, and the other elements that you may need for different kinds of solutions for different kinds of customers.
Okay. Thank you. Those time frames of, you know, maybe being a year or two away, I feel like a lot of your quantum competitors seem to point more to several years away. Would you make the case that your technology is just gonna be ready sooner, or is there just a different definitional things about?
Yeah
readiness?
Yeah
how do I think about that?
It's a great question. I think there are a couple of sort of like leading indicators I always look at, right? When something moves from the lab to a commercial deployment, that's a good sign. When something is a repeat purchase by a customer that says, "I'll buy your last generation, this generation, and the next generation," potentially, that's a validation. When you start demonstrating, for example, as we did with AstraZeneca, the ability... By the way, that was in partnership with NVIDIA, right? You can argue which was the accelerator to the other, and I think it was the QPU probably that was the accelerator to the GPU. There are things you can do on the QPU that you cannot do as efficiently on-
Mm-hmm
... GPUs and vice versa. That kind of hybrid computing beginning to be deployed more and more is what I'm seeing. To me, that's another litmus test for has something reached the ability for it to be deployed in more places. I'm seeing more AI factory data center operators say, "We wanna keep you in there too." So those kinds of patterns, along with the use cases that I mentioned around AstraZeneca, protein folding with Kipu, applications that we will build for an energy grid operator, for example, that is one of our customers who bought our network first and then the computer and wanted us to help them build an application to optimize the electric grid.
Data center operators come on, and maybe you saw recently, data centers are falling off the grid just as fast...
Mm-hmm
...because they sense the electrical current being in imbalance of some sort, and they go to a backup power. Grids are having to deal with sudden drop-off loads as much as well. If you're a grid operator, you've got to keep it up no matter what. So those are the kinds of things that I think are real applications today. For us, like every successive generation, which is one a year, the roadmap that we've built now with a semiconductor- based going forward versus lasers in the past, allows us to maintain the cost advantage, I think, of the promise of quantum computing and also introduce more things in a more predictable way, using a foundry system that's been built over years and years and years now.
Okay. Then the kind of world-changing applications that you talked about on the earnings call, would that be the AstraZeneca? Like, what are those?
I think those are good examples.
Yeah
... of them, right? Well, world-changing is a very big thing, right?
Yeah. Yeah.
Like, looking for the causes of chronic diseases, trying to find how protein folding either affects disease of the brain or memory or not. These are real-life things that have already begun to be tried. Some of them get more capable the more qubits you have, as you know, and then others become enabled by more qubits down the road. Our roadmap that we've published is a five-year roadmap. It goes from 100 qubits, which is the next generation machine we're rolling out now. It's our 5th generation machine. The 6th generation moves to semiconductors, and that's where we think that you now start to get economies of scale, lower cost potentially.
The, you know, energy consumption, as you all know, and you know this, from all the work that you do on semis, GPU clusters, AI factories use a lot of energy.
Yeah.
Quantum computers of the kind like ion traps that we build.
Yeah
... you know, don't.
Yeah. Maybe talk about the modalities and why ion trap. I mean, it's always seemed to me aesthetically appealing to have something that can just sit in a box on my desk and do stuff rather than build low-temperature superconductors. Obviously there's success on all modalities at this point. Like, For technology that's getting relatively far along, there still doesn't seem a clear-cut consensus on which modality is best.
Yeah. I think. That's fair. I think that there are some who are choosing superconducting-.
Yeah
... and that's great. I think that, it requires.
Yeah
not just cooling, but like down to 0.2 degrees Kelvin. I've been told that it sometimes takes a month to turn a machine like that on and get the temperature down. Again, it could be something that is a long-term solution. There are others that are like still require scientific breakthroughs. Photonics tends to be one of the hardest ones, for example.
In the case of ion traps, whether lucky or smart, the founders of this company, like 30 years ago, whenever they thought of this and wrote the paper on it, started with ions which exist naturally in nature, don't have to be created, have a charge to them, so you can actually say to an ion, "Do this, do that," so make it dance, if you will, entangle it, use it for computation. It naturally has fewer errors to correct to begin with, so you don't have the overhead of error correction as much. Then you have the ability to build applications that, as we're doing ourselves as a company, scale with the qubit count that you develop.
It's the logic, logical qubits or not the physical qubits that companies talk about often, I think. Joe, you know this, it takes standards sometimes for everyone to talk the same language. Those standards don't exist currently. I think that makes it easy for investors, companies, customers to get confused. Which one is better? This one says I have more physical qubits, this one says I have that. To me, in the end, it's all like, which application can you run first on which machine? That's kinda what we're trying to do both.
Mm-hmm.
Build the machine, build the application.
How does having that full suite of, you know, networking, sensing software that you've acquired, how does that help you to get there?
Yeah. It's I worked at Cisco and, you know, one of the things that I learned is obviously, like, be agnostic, like connect anything to everything.
Yeah.
Right? Number two, make sure that you have everything in one roof, and then you can start with a customer buying one aspect of your portfolio, moving to another. We already have examples of that. We have a customer that is an electric grid operator that had built a dark fiber network, high-quality infrastructure, and said, "I wanna actually put a quantum network on top of this and provide more connectivity to my growing list of enterprise customers in my region," and then said, "Well, I now wanna buy a quantum computer, so I can deliver more services and applications to them as an additional revenue source." We've had others that start with quantum computing and stay with quantum computing and don't need the networking.
We have in Chicago, for example, a computing device deployment leading to, okay, I need to connect to this other place as well. Our networking solution, which we're selling today globally, connects our computer to anyone else's computer, so it's agnostic. A lot of the work we did over the last four or five years was learning how you do that. How do you take a quantum computer, put it onto a telecom network, same signal being carried as DWDM and other things, and then translate it back to the frequency that you need to have for quantum computing? The other elements around sensing, you know, which is around making sure that GPS, which can be jammed, it has a replacement of some kind in certain environments, where, for example, the U.S. can depend on PNT, position navigation timing, and not face the same jamming.
That was an idea that came to us from a customer, and maybe more than one customer, who actually said, "Well, if you had that capability, we'd look at you, too." They may start in one place and end up in another place. The cross-selling opportunity, I think, those things when they start to flow through your P&L and your revenues start to grow, we just reported, as you know, $130 million, we guided for this year as well.
Mm-hmm.
You know, organic growth continues to be 80% or better.
Yeah.
I indicated 2016 could be even 100%, perhaps. That means your core business is growing and these other businesses are driving more opportunities for growth.
Maybe talk about that or 80% organic growth. I guess how much should we even be focused on revenue as kind of the real value of the company? I mean, obviously, this technology is kind of at a development stage. It seems most important to hit the proof points, but you have a lot of revenue as well, right? The fact that you broke out organic revenue growth is helpful to understand how much came from acquisition. What's driving the organic growth rates that you have, and how much should we care about it?
Well, look, as usual, product cycles drive revenue growth, right? The fact that we keep investing in R&D, which for us is like a mantra we have to stick to. From my Arm days, one of the learnings was like 80% of every $1 you spend should be on R&D. We're not there yet at IonQ. We're trying to get there.
Yeah.
The idea is keep investing in innovation. You know this better than anyone. That's how you maintain your advantage for customers. Ultimately, that's what we're trying to deliver. Hopefully, of course, for investors. The organic growth, you know, some of the metrics that I provided on the earnings call was actually from listening to investors.
Okay.
It was my second quarter reporting earnings. In the first quarter, I did sort of like a listening session around what else would you like to hear.
Mm-hmm.
The questions were around, like, is, you know, are you buying things because your organic revenue isn't growing? Or are you still selling mainly to labs? You know, are you still U.S.? Et cetera, et cetera. Each of the data points was designed to, like, provide transparency into the revenue sources. We talked about commercial, we talked about international growth, we talked about organic growth. I'll add more metrics as we go along, if it feels like it, you know, helps investors understand, like, how are you getting all of this growth?
I hear the both, the positives and the negatives, but you're kinda like, it's being too judgmental about revenue when we're sort of nobody's claiming that we're doing something yet that's a breakthrough. Like, we're kinda getting there, but that's when you get paid really.
Yeah.
Everything up until that point is gonna be development and is gonna be, you know, there's some related party and all that that goes into any kind of development relationships. I feel like it's really the technology proof points that I mostly would care about.
Yes, you're absolutely right. I think that the fact that we have now put most of the scientific breakthroughs behind us.
Mm-hmm
achieved four nines fidelity, for example.
There's a mouse over there. Sorry. Freaked me out.
Oh, we're San Francisco.
Yeah.
you know, others, other modalities still have some more work to do. I think what we now have to do is ensure that we can get from where we are today, which is our 5th-generation machine with 100 qubits, to 256 qubits, at the end of this year, 10,000 next year, 20,000 up to 2 million. I think that is a roadmap that we have been working hard to de-risk and execute, if not accelerate. Some of the things we've done are to ensure that we can achieve those sooner and move things to the left. The more you move stuff to the left, the earlier you can deploy some of the game-changing applications you're talking about.
Mm-hmm. Yeah. I don't mean any of that critically. I just at this point.
Of course.
... feel like, you know, really this is a technology that isn't really that economic yet. The development is gonna be what drives the revenue. I guess in that context, you're gonna be at GTC later this month at NVIDIA's developer conference. Obviously NVIDIA, I've sort of seen this evolution from sort of being dismissive of quantum to actually embracing it and potentially making a lot of money from it because he's got a lot of co-processors that sit in these quantum systems. How do you think about that relationship? How do you think about the advances that they're making, and is that something that will help you guys move forward as well?
Yeah. We've been working very, very closely with them, as you can imagine. I mean, our customers also want us to work very closely with them, right?
Mm-hmm.
They may start with a AI cluster, AI factory data center, and then some of the customers are saying, "I wanna put the GPU next to it." We work closely with them to make sure it inter operates with CUDA Q, of course. Like AstraZeneca was done jointly with them, right?
Mm-hmm.
It was an effort that AstraZeneca benefited from. It allows us to take that and replicate it with other pharma companies also. I see like a QPU, GPU environment where some things are done best forever on GPUs, perhaps.
Yeah.
Other things better done or best done on QPUs. I don't think one is replacing the other, for sure. I think one is gonna be augmenting the other. There may be quantum-only deployments. I can see those happening as well. Like more and more data centers around the world, we announced KISTI in Korea, for example, and they have a GPU supercomputer that they've deployed. They wanted the quantum computer to work with it, sit next to it, and that's what we're deploying there. Yes, we're gonna be at GTC. You'll hear more about what we're doing with them, but I see that as a long-term partnership.
Mm-hmm. Okay, helpful. Speaking of long-term partnerships, can you talk about the role of the U.S. government as obviously a big federal priority in quantum? How is that helping you guys?
Yeah, look, I think, you know, that many dimensions to that question for sure, so I'm gonna take it one at a time. I think one that you're referring to is what DARPA is doing right now.
Yeah
... in terms of its A, B, and C, sort of like, I call it good housekeeping seal of approval process to say, "This modality, this technology is ready to be deployed." There are customers that I think in the government that depend on that and look at that for validation of the technology. We work very closely with DARPA, of course, and we've gone through A and B, and C lies ahead. What we're starting to see is, you know, customer engagement already start to happen with government customers 'cause they've seen you go through A and B. They will continue to look at it, of course. Whenever C happens, you know, they'll either have waited for that decision that to happen or maybe made up their choices even earlier.
It's good to see that, it's good to be, frankly, among that process. The other is the QBI, and certainly that allows multiple agencies to also coordinate, makes it easier. I think that, you may see this as well in what you look at, virtually every country that we have met with, I have met with, is putting quantum up there among their top five, if not top 10 priorities as a nation.
Mm-hmm.
Yes, they want AI, yes, they want the ability to do genomics and other things, and yes, they want quantum also. It's not just the U.S., it's like virtually every government in the Middle East, in Asia. We've hired a new CRO, Scott Millard. He comes from Dell. He worked directly with Michael Dell and helped scale their AI business from like nothing to $20 billion-ish and more. You follow them even.
Mm-hmm
... even more closely, I'm sure. He's bringing this very disciplined scientific approach to say, "Pursuit and capture, target the right customer, right, you know, deploy." What I love about one thing he said was like, "Follow the AI money." Like, follow the AI money 'cause that's where you're gonna need a quantum computer either now or in the future.
Mm-hmm.
He was right. We're starting to see that play out.
Okay. There's a lot of AI money.
Evidently, yes.
Seems to be. okay. Any other countries to highlight? You said other countries outside the U.S. are critical.
Yeah, I mean, there isn't a country in the world Probably doesn't want to have something.
Okay.
As these machines become more powerful, and I mean, like, beyond 100 qubits, probably 256, maybe 10K. At some point, I think we start seeing similar export control questions that we need to deal with. We are starting from the get-go and saying, look, like, assume that regime is in place already. Let's ensure that we're selling a computer to a country where we would be able to, like, support that customer over a long period of time. There's some countries we'll just not do business in, as you can imagine. There are many countries, however, that are wanting it, we can do it, but even then we'll check with the U.S. to make sure that this is going to be something that we can support for multiple generations of machines over time.
Okay. You talked on the earnings call about demand exceeding supply. Can you talk about that a little bit and how you think about the supply chain to serve that demand?
like, there is no manufacturing foundry.
Yeah
for a classical computing machine. We had to build our own factory, right? All of the generations, the five generations I've talked about, including the current one, we build in Seattle.
Mm-hmm.
We design between College Park and Seattle and build and then deploy at customer sites. We're a manufacturing company as it is. The demand we're seeing for the 100 qubit machine that's rolling out this year is higher than we had anticipated based on prior demand projections for earlier generation machines.
Okay.
I think the applications and the machine may be hitting the spot for many and more customers.
Okay
Because they want to do that. I started on September 5th, I think. By September 10th, we had already started deciding what we need to de-bottleneck, either in our ability to manufacture or in our ability to deploy. Both are equally important. Again, the COO hat focuses a lot on that to make sure that we can meet the demand. Yes, absolutely, there's demand for more Tempo machines than arguably we can manufacture today.
Okay.
Thankfully, that facility in Seattle does have an extra floor that we also own.
Good
with plenty of real estate if we had to expand. However, as we move from laser-based machines, which are, you know, prominent, these days, to semiconductor and electronically controlled ion traps.
Mm-hmm
We can start moving into a classical semiconductor foundry with older nodes.
Okay.
Not our own factory. At some point, the manufacturing can be stopped in Seattle and moved over to an existing foundry.
That's helpful. I just have one more question, and then we can open it to the audience if there are any. Your cash balance, $3.3 billion. How do you feel about that? Do you wanna raise more? The company does continue to be acquisitive. You know, just how do you think about the cash balance and the uses of that cash?
Yeah. Look, I think having a strong balance sheet, having cash, gives you line of sight to being able to invest for the longer term.
Mm-hmm.
One of the pitfalls of all public companies is the 90-day cycle and the obsession with this year's financials and like, you know, what will the world think if I spend more on R&D, et cetera, et cetera. The ability to have enough cash to think, like, multiple years-
Yeah
Not just for acquisition, but even just organic investment in R&D, infrastructure, and so on, gives you the confidence, but also it gives your customers the confidence.
Mm-hmm
that if they purchase something for you're gonna be around for a while. In my experience, probably in yours, that's very, very important to have.
Well, you are competing with IBM, Honeywell, people like that have a lot of resources as well.
They do. Of course they do. I respect IBM in everything that they do.
Yeah.
Right? I want them to keep investing in this area. I hope they'll keep investing in this area because we need to have an industry.
Mm-hmm.
You can't just have an industry with one company. IBM, of course. Superconducting, their solution, of course. What's not to like? Again, like what we're focusing on is a bill of materials that is less than $30 million.
Mm-hmm.
As the machine gets more powerful, we start making it on semiconductor base, it should get cheaper, not more expensive.
Mm-hmm.
Ours plugs into the wall.
Mm-hmm.
Superconducting needs a little bit of a grid. They're probably-
That is cool to see, I have to say, just seeing the facility. Yeah. It looks like it just plugs in, just set it right next to my desk, yeah.
The point is that when you think about a customer who wants to invest for the longer term.
Mm-hmm
-and thinks they're getting, you know, the 100 now, the 256 then, and 2 million eventually, they're gonna be living with that investment choice for a long time.
Yeah.
With quantum or as with anything, it's not just about the initial cost of the machine that you buy, it's also the running cost once you have it. The cooling alone, keeping everything low temperature, Helium-3, et cetera, et cetera, et cetera. TCO compare, you know, for an ion trap type thing to an equivalent, let's call it generic superconducting, is not even a comparison.
Mm-hmm
in terms of TCO. Not everyone's thinking of the TCO always.
Mm-hmm.
I do think that over time, that'll become a compelling argument for why doing really, really high-end computing with lower energy consumption and lower cost, frankly, becomes an advantage.
Okay. Great. Let me see if we have questions from the audience. Okay. If not, we can wrap it up there. Inder, thank you so much.
Awesome. Thanks for having us. Really appreciate it.
Yeah. Thank you.