Good afternoon. My name is Quinn Bolton. I am the quantum computing and semiconductor analyst for Needham and Company. Thank you for joining us at Needham's 21st Annual Technology, Media, and Consumer Conference. It's my pleasure to host this fireside chat with IonQ from the company. On stage with me will be Inder Singh, Chief Operating Officer and CFO. We also have Hanley D'Onofrio, Head of Investor Relations in the audience. Inder, Hanley, thank you for joining us. I'm going to have Inder come up. He's going to go through a couple of opening slides, and then we'll start the fireside chat.

Thanks for having us. Pleasure to be with all of you. Thank you for joining. I thought I'd share a couple of slides to set the landscape for what we're going to talk about today. Not everyone knows quantum. Not everyone knows quantum equally. It's important to get the same set of facts first, I thought. Forward-looking statements. You read those really quickly as I went through those, I'm sure. But basically, we have been a quantum innovation company since the company went public in late 2021. And the company has scaled using ion traps as its modality. We are now selling our fifth generation computing device around the world and growing, as you will see in a moment. But we picked ion traps actually 30 years ago, way before I got involved with the company, of course. and our founder found that the modality of ion traps starts with some advantages. Lowest errors, highest coherence, things that matter if you want to drive computing power over time and scale over time. So whether lucky or smart, we ended up being the first company to reach 4.9's fidelity, which is an important metric to achieve on the path to getting computers that actually are self-sustaining in a way of being fault-tolerant. We also switched to using semiconductors going forward from using lasers in the past. So a year ago, this year even, four years ago, we were using lasers to hold ions in place. That's fine, but when you increase the number of qubits and the machines get more powerful, you have more ions, you have more lasers, it gets bigger, needs more energy, needs more maintenance. And also scaling becomes a challenge. So we decided to acquire a company called Oxford Ionics, which brought us a semiconductor roadmap for the next five years, which we've publicly laid out. And we are now able to use an existing supply chain, as well as manufacturing capability, that the semi-industry has used for decades. So we benefit from scaling faster and getting to more qubits faster. We also switched, as I said, from lasers to using electronic qubit control. Again, fewer lasers, fewer complexity, lower cost, and also easier to scale. And so electronic controls allows us to do the same things we were doing using lasers, but make machines actually smaller, cheaper, and more fieldable in more applications. And then also recently we announced our architecture for how we're going to get to basically having fault-tolerant computing. It's very cool. It's called Walking Cat Architecture. It's a 100-page read. Or you can ask ChatGPT or your engine of choice to summarize it for you. But basically, it's our engineering roadmap for how we're going to get to 10,000 qubits and beyond and create computing in the quantum world that is tolerant of any issues like faults and errors. As a company, we have virtually every metric that you see on here as industry leading, including the revenues that we reported and the guidance we gave for this year has a midpoint of $265 million for 2026. In our Q1, we reported record revenues, third quarter in a row of record revenue growth, fourth quarter in a row of record revenue dollars, and we continue. We're also innovating very heavily. We're spending more than $300 million in R&D to keep driving the innovation engine forward. Cash position enables us to look into the three- to five-year time horizon and not just the 90-day clock. So we're very focused on ensuring that we invest with longevity for success for our customers and be able to build the five-year roadmap that we have. So $3 billion of cash helps do that. And we now have 350 customers around the world that are using either our computing devices or our quantum security devices or our quantum networking devices, et cetera. So we've gone from one product company two years ago to making many, many products and having a quantum platform in place. Patents, of course, and then customers. you can see logos sprinkled around the bottom of the slide. The way we're going to scale from where we have been to 256 qubits, which we announced some breakthroughs on, and I'll talk about that more in a minute, to 10,000 qubits, to 100,000 qubits, to millions, and then after. So this is a five-year roadmap. We're all on the roadmap already. And this is all leveraging an existing semiconductor foundry process. and the foundry process that exists knows how to scale using multiplexing and chiplets and that's all that's needed for us to go from 256 to the 10k to the 100k to 2 millions and beyond and that's really important because that's how you get to faster solutions more powerful algorithms the ability to do more with computing and what's called quantum advantage more quickly so we have high confidence in our ability to execute on that roadmap. And then beyond that, we have the ability to interconnect computers together to almost act as one. So you can start having distributed computing in the quantum space, something that was unimaginable probably even a few years ago. So on the 256, we announced on the most recent earnings call that we've now finished the prototyping of that chip, which was an important milestone for us. Now that we've achieved that, we know we can use electronic controls on a semiconductor die in a mature node foundry called Skywater, and we can now build a system around it. So we're already building the system that surrounds the chip, and we'll be prepared to deliver that to customers starting next year. And we've started, importantly, to turn our attention to the 10K already. So not only are we building the 256 system, we're now figuring out how to do the 10K probably far earlier than we thought even three or four months ago. And happy to be making the progress that we're seeing here. And then, you know, 10K and beyond obviously becomes a question of simply scaling. But the main engineering milestones we've achieved with the 256, demonstrating that you can actually do iron traps on a semiconductor die. And then the scaling part of it, as I said, is an existing process that Skywater, who's our foundry partner, has. And plenty of knowledge around advanced packaging that they will help us with. Our compute devices are already deployed in the field, solving real-world problems today. This is just a sampling of the end markets we're in, life sciences, energy, logistics, you can see, finance, and engineering. And then customer logos who are actually making it happen. Each one is using it for a different use case. And so we are building the compute platform and the apps that go with it so that our customers can benefit. Sometimes it's in partnership with other ecosystem investment partners that we're also engaged with. All of the things I just mentioned are in the first tile here, quantum computing. In the last year, we transformed the company by building a networking capability as well, so we can connect quantum to quantum. A quantum cybersecurity as well, because we recognize when you get enough qubits, you can actually break encryption, and that's not very far away. So secure customers now. Quantum sensing to provide quantum sensing in denied areas where there's an adversary that locks in and distorts GPS signals for PNT. And then being able to deploy some of those sensors in space with our own constellation of satellites as well. So everything that we were as a company was in the bottom left there, and most other quantum companies are in that box as well. We've been building the platform over the last year plus, and we're happy where we are. We're happy with our momentum so far and happy with you and happy to take Q&A.

Well, thank you for that overview. And, you know, we've been covering this space now for four years. I think investors are, you know, sort of coming up to speed in terms of their knowledge of quantum. But I'm sure there are still folks that are, you know, confused by the technology. And so maybe what do you believe is most misunderstood about quantum computing as an industry, about ion trap as a modality, and perhaps about the company that you might, you know, help clarify for us?

Yeah, I mean, look, I think quantum mechanics is, you know, something that PhDs build their whole life around. And you can spend your whole life studying it and be in good company. if you still don't understand it, because even Einstein said it was spooky science. You have to accept that it exists. You have to accept that it allows things to be done in an exponentially more powerful way. And if you accept those two tenants, then it's a question of how do we harness that? How do we actually build a company and a capability of applications that can be run on it? Things that require simultaneous calculations in fractions of a nanosecond that it would take years or longer for classical computing to do. And so not everything works best on quantum. Not everything works best on GPUs. You need both. You need QPUs and you need GPUs. And we're seeing more and more of that happen. As to the rest of the industry, there are different modalities in quantum. There's superconducting, there's neutral atoms, there's photonic, any number of them. And, you know, I think that because it's early stage for the industry at large, Every company is trying its best to get its modality to work over time. This company has been delivering four generations of computers already. Fifth generation is being rolled out to the market right now. Sixth generation is the 256 I mentioned. Seventh generation is our 10K. We're working on that already as well. So on computing, we have, we think, a very strong proposition for customers because it's all about compute power, of course. At the same time, it's about energy. And our form of ion traps uses less energy than superconducting and some others. So the total cost of ownership is not even a comparison, to be honest. So you basically buy the quantum computer. It plugs into a wall. It does not require a grid. It does not require refrigeration down to zero degrees Kelvin, et cetera, et cetera. No form factor is wrong. I'm not saying any other modality is not right. It's just you begin with a natural advantage, and then you build on that advantage. And that's kind of where we are. Excellent.

Is it great that you left this slide up? My next question is, you guys have transformed the company from sort of just the first bucket of quantum computing out of quantum platform company. And for investors who may not be familiar, what are some of the products you offer within networking, cybersecurity, quantum sensing? And how do they all tie together? What are the benefits of having a complete platform?

Yeah, I mean, some of these are rather obvious if you think about the evolution of quantum computing, that having a quantum computer is great, but it needs to connect to something. How does it connect? It has to be networked. How does it get networked? You have to build a quantum network. So this company started developing quantum networking capability, and we're now deploying that in market. So you can connect a quantum computer that we make, frankly, to someone else's also. So we announced a DARPA project called Hark where we can live with different modalities even in the future. And we're the one company that can connect all of this together. So absolutely, quantum networking is something we have in market now is being sold. Some customers actually buy the network first and then the compute device and then other things. And having this platform allows us to have the customer start their journey in one product and branch into the other. And so quantum cybersecurity just acknowledges the fact that if quantum computing will break encryption, and I think everyone has now accepted that it will at some point, there's some debate about how fast that is. I think it's like almost here and the algorithms are getting simpler and the machines are getting more powerful that we're building that you need to be responsible and lock down networks in anticipation of RSA and ECC encryption of all forms being broken in the not too distant future. So as responsible stewards, we made sure we have a quantum cybersecurity business as well. Some people will start with that. Financial institutions might think about that one first, for example. And then quantum sensing, and quantum in space, round out the portfolio, because the other things that national security types of deployments require is not just the first three tiles. They require computing, networking, and cybersecurity, but they also require it to work in space. They require sensing in space that cannot be jammed, like GPS can be jammed. So we built this portfolio over the course of the last year or so. Our new CEO, as soon as he came in, he got to work. I was on the board at the time. And I was thrilled to see him go into action and basically expand the company into a platform company.

Excellent. You guys reported very strong first quarter results last week and raised guidance for fiscal 2026. Maybe talk a little bit about what drove the upside both to the quarter and the full year. Do you think this is more IonQ, company-specific momentum, or do you think the industry is starting to hit an inflection point in terms of commercialization and just growing acceptance of the technology?

Yeah, look, we watched everyone, of course, but I think, you know, we pay attention to what we're doing. And what we're doing is building the fastest machines connected to each other, secured in a way that I think is required and connected and basically meeting the customer where they are. And so what we're seeing in our business is the flywheel effect, as I said, of some customers saying, you know what, we're going to start with cybersecurity and then we'll talk to you about other things. So I'm saying we want to start with your compute device and try that out. and then go to other things. So we started to report this quarter, for example, that over a third of our sales, 35% of our revenues now are multi-product customers, so more than one kind of product, which I think is a growing trend and speaks to the power of the platform and the strategy that our CEO put in place. The ability to have one-stop shop with everything just works, you turn it on, flip the switch, and it works, is the engineering durability that we're trying to ensure is in everything. This allows us to be a very global company at the same time. So we're now in 130 countries with more than 350 customers, and the opportunity to cross-sell. So we've got every methodical pursuit and capture approach to go to market to ensure that no matter where the customer starts, we lock them in with us for our five-year roadmap that I mentioned, as well as the other products that we can bring to them.

You mentioned Niccolo came in, I think, just over a year ago. He's been very active on the M&A front. One of the questions we get from investors is, you know, you've done a number of acquisitions. How have you been able to integrate and bring all these different companies together? Obviously, first quarter results would say that you've done it pretty well. It certainly doesn't seem like it's holding you back, but just talk about the integration process of all these acquisitions you've made.

Yeah, so we put together an acquisition integration team last fall basically to say, now that we've acquired these assets, let's figure out the things that should be common among all of them. So, as you would imagine, payroll systems, IT, HR, finance, the traditional things that you would want to make sure are aligned for a public company like us that is scaling rapidly, we rapidly integrate it. IT always takes a bit to do, so you do it in phases and steps. But adding cybersecurity, increasing the level of our awareness of our own IP now becoming more valuable to potential adversaries, et cetera, those are all investments we made very quickly. We are now focused on continuing the innovation engine, to your point, ensuring that our core business, which has been computing, continues to grow. On that earnings call, we did guide to a doubling of that business again this year. So that continues to be strong, and the whole company is doubling at the same time. So the other things are starting to contribute to the growth just as much as computing is right now. We feel confident because we also increased our RPOs, which is a measure of future revenue backlog. And it is now at a record and, you know, well over a year of visibility into our revenue gives us more confidence that we are doing the right thing. Customers appreciate it. And the fact that we're getting customers saying, we actually want to buy your current generation, but we want to be first in line for your next generation. Well, that's nice to hear, too.

Right, right. You touched a little bit on the technology roadmap, but just a couple of questions there. First, on the 256 system, you've got the test chip back from the fab. I know in other modalities, especially superconducting, as the qubit count goes up, calibrating these systems can be a very time-intensive process. Is that something that, you know, does it take a long time to bring up the test chip and start to get results back? And then are there any other gating factors on sort of your timeline to be able to deliver that system next year?

Yeah, that's a great question. I mean, we don't operate at zero degrees Kelvin, right? So things are harder at zero degrees Kelvin. We operate close to room temperature. And so for us, the testing, the fab, we're past that already. So what we thought would take nine months has taken just a few. So we're happy with the proving the chip strategy, first of all, getting the prototype in, as you said, and now we're building the system around the prototype already because the journey begins with the chip. It's the smallest component in the machine, but you have to deliver the entire integrated system. So the rest of this year, we're focused on taking that 256, building multiple prototypes of the system because you have to do the same thing as well, and ensuring that, again, when we deliver it to a customer, it just works, and it turns on, and it does the things they need them to do. These aren't labs we're selling to now. These are commercial customers that expect it to work day one and work all the time, if not most of the time. And that's the engineering threshold that we're shooting for.

I also wanted to dive a little bit more into the electronic qubit control technology acquired from Oxford Ionics. I think you're already starting to see some of the benefits of scaling. You went from 100 qubits in the old ion trap architecture now to 256 you know, with a roadmap already to 10,000, but can you talk about some of the other benefits, you know, does it accelerate or increase the gate speed of the systems relative to the laser-based modality, or are there other advantages that the electronic cubic control brings? Yeah, that's a

great question. I mean, it's, so I joined the board back in 2020, end of 2021. I was on the board watching the company grow, and we were using lasers, as you said, and lasers seemed the way they go. And so what we learned, however, for that learning curve, is that lasers can take you so far to a certain number of qubits and not much farther. And so when we just switched to the semiconductor roadmap, it was because we needed to get to 10,000, 200,000, 2 million. And using lasers, that becomes exceptionally difficult, not only expensive, bulky, large, energy-consuming, power. So, like, all of the things that you're trying to fight against go the opposite direction. With semiconductor approach, we're using mature nodes that are fully depreciated, therefore lower cost. We can drive down cost over time as we increase compute power. We have greater certainty because these foundries know how to do scaling in the semiconductor landscape, and that's what we're leveraging. It also allows us to have modularity. So beginning with the 10K machine, for example, you don't need to forklift the machine to take our next generation. It becomes a modular upgrade. Now you're locked in with your chassis at your customer, and you're swapping out some parts and upgrading the machine. That's a great lock-in for the customer, for us, for them to embrace our roadmap. And we just have to execute on the qubits, and we have high confidence that we can do that already.

You've talked and announced, I think, two sales or pre-sales of your 256 qubit system. But talk about just general interest from your customer base for that system. What kind of feedback are you hearing or anything else you can share on the pipeline you're building for the 256 Qubit system?

So our fifth generation is what we're selling into market today. And a quarter ago, I said that demand was exceeding our ability to meet the demand. We've now caught up with that, and we can now satisfy that demand. So this year, we're deploying our fifth generation machine. And as you noted, we've already started taking free orders for the next generation machine. And people want to be first in line to receive that because they know it's going to work if we've engineered it to work. And so, you know, I think if they own the tempo, it's easier for them to do the upgrade. They're easier for them to get first in line and other things like that. And that way they sort of lock on to multiple generations of your product, which I'm starting to see also. And the reason they're asking for this is not because they just want to have quantum. is because there's a realization that quantum is going to sit next to data centers. So GPU clusters, wherever you have a data center, over time, you'll see a QPU next to those GPUs. And these hybrid workflows that are being looked at are now becoming a pattern. So we have a Salesforce now that is looking at, okay, where are those deployments around the world? Where does it make sense for a QPU to be based as well? And then using that to filter our pipeline. So taking something that was five years ago in a lab into the commercial world, working at scale, and then building it on a semiconductor roadmap, and having the customer now lock into multiple generations. One customer we announced already had our Forte system from a year ago, has bought our Tempo system from this year, and wants to be first in line for the 256. 256 demand is going to, I think, be very strong, much stronger even than Tempo. And then when we get to 10,000, I think it goes even higher from there.

Does the 256 system, qubit system, get you close or already to quantum advantage on certain applications? Or do you think that's more the capability of the 10K system?

I think each generation is exponentially more powerful than the prior one. So if you're in quantum or want to be in quantum, it doesn't matter which country in the world, frankly. I mean, every country wants to be in quantum in their top three priorities. And that's the trend we're seeing. So they want to be in quantum with whatever you have available that works today, and algorithms can be run on today. The more powerful the machine, the more powerful and complex the algorithms that you can run. So if you want to, for example, we talk about breaking encryption. If you want to do something like that, you're not very far away from being able to do that. You know, it's probably a few thousand logical qubits that you need to have. The algorithms are, of course, getting simpler, so that might get closer. So we're probably, you know, two years away, plus or minus from that. And 256 is the first proving ground. As soon as we get 256 and we have that capability, upgrading to 10K becomes so much easier.

You touched on sort of quantum networking. Maybe share just some of the advances you're making in quantum computing, the ability to sort of convert sort of the ions to photons so that you can then network them potentially to quantum computers of different modalities.

Yeah, so this has been a multi-year journey as well. And our partnership with U.S. government in some areas has allowed us to create capabilities of connecting a quantum computer over a traditional telecom network to other things. So that we've had. And now with Hark, which is the announcement that we made, we can connect that to different modalities as well. So if someone says, look, I have a superconducting thing deployed over here, I have an iron trap over here, or something else somewhere else, we are indifferent as to which modality is being used, and we can connect all of that. And we're probably the only company right now that I can think of, frankly, that can do it today versus promise to do it three or four years from now.

And IRQ, I believe, is developing the technology that converts sort of the ion qubits to sort of photonic qubits or to at least entangle. Are you also developing those technologies to sort of interface with superconducting or other modalities, or do those companies have to develop sort of the conversion to the, I guess, transduction to a photonic-based interconnect that you then hook your networking into?

Yeah, this industry will need to standardize on certain APIs probably over time, I think. So what we're trying to do is make sure that we establish a de facto set of requirements, a default standard, if you will, a de facto standard for all things to work together. And because of the presence and the amount of deployments we have already, we're in good position to begin to do that. Not to dictate what anyone else to do, but just to get one thing to communicate with another. We also work with the NVIDIAs of the world and others to connect into data centers as well. So there are ample instances, not just of quantum to quantum, but frankly quantum to GPU, which is here and now. And that is something that we are uniquely positioned to do. But to your point, the team is, we acquired a team called LightSync about a year ago. They do optical interconnects that have to be put in over long distances to connect things over distances. Those things are available to us today, and we can begin to deploy them into the market. We've announced a number of quantum networking deployments in Europe, in the U.S. as well, and I think that will only grow. We've announced one in Chicago, for example, one in Tennessee, and in many countries in Europe.

You talked about the architectural blueprint that you published, I think it was last month, of the Walking Cat architecture. I haven't gone through all 100 pages yet, and from what I saw, it got pretty technical pretty quickly. But can you sort of summarize the key highlights of that paper? And why was it

important for IonQ to publish that roadmap? Well, I think quantum computing, and frankly, any computing, is about getting to fault-tolerant computing. And in order to get to fault-tolerant computing, von Neumann published a paper, actually back in 1945, I believe, that was the basis of computing in classical computing become fault-tolerant. This is the quantum parallel to that. So it has similar elements required for fault-tolerant computing. What we published is our engineering roadmap for getting to fault-tolerant. And once we get to the 10K, which I think a lot of magical things start to happen, including the ability to get to using the walking cat architecture that you described, which, yes, 100 pages is long. ChatGPT will summarize it for you if you want. But what you see in there is the building blocks of how you do that, the specific step-by-step of how you do that. One of the really cool things about quantum is that if there's a eavesdropper trying to look in on the network, the network basically a connection collapses. Well, how do you then look for errors and things like that across the wide area network? And that's part of the cat phases in this walking cat analogy. And so there's, I won't take the audience through all of that in this thing. But that is our architecture for delivering fault-tolerant computing in the next two years or so. And so we think that we have the right approach to it. We're the first to get to it, we believe. And so it was responsible for us to sort of lay that out for our customers to understand exactly how we're going to do that.

Got it. So it's sort of trying to validate your roadmap, show you've got the technical capability or the technical roadmap to get to fault tolerance was really the highlight of why it was important to publish? Yeah, I think that because one of the questions

at the very beginning you asked was around the different modalities. And all the different modalities are different places in their journey. Iron traps are further along. Our iron trap architecture is further along. And it was important for us to then say, okay, how do we go from here to the promised land of getting fault tolerant computing and quantum advantage? And we did that to demonstrate that we know exactly how to go from here to there. I wanted to ask, you mentioned

Skywater technology acquisition you announced, I think it was in January. Shareholders voted, Skywater shareholders voted in favor of that transaction last Friday. One, why was it important to bring that manufacturing capability, the semiconductor foundry in-house? And how is that progressing in terms of the tape outs and your ability to spin products?

Yeah. I mean, so far, so good. So that deal requires, of course, HSR approval before we can close. But as we look at it, we've been already prototyping the 256 chip that I was talking about in their foundry. It's months ahead of schedule from what we thought it might take at an alternate foundry. So they've been able to help us in ways that I think would have taken us longer or maybe cost more even. The 10K is where they're experts. So their ability to scale semiconductors on the kind of mature nodes we need is unique. And so we've seen them now early signs of like really good progress of getting the scaling done as well. They also have a very advanced packaging business capability, which is important for us on the other end of having the chip ready as well. And so those capabilities made it very interesting for us to actually own the excess capacity. They serve other quantum customers, of course, and so do we. We don't talk about it much, and the other quantum companies don't mention it much, but we sell components to the other quantum players to help them build their machines already, as IonQ, and so do Skywater. And so we will respect that arrangement in being the merchant supplier to everyone. We just want to accelerate our roadmap. And what they allow us to do is to use their excess capacity with our investment dollars to actually scale.

And once that transaction closes, I think you disclosed in the first quarter, you spent something like $12 million that showed up as an expense to IonQ, which probably was reported as revenue. At Skywater, once you close, that sort of, that vertical, that stacking goes away and you sort of get rid of that sort of expense or that extra margin. Yeah, there's a couple of things.

I mean, one is sort of the short-term costs of a commercial relationship, of course. The other is, like, the more you make them successful, the more likely they were to raise prices on us a year from now, two years from now, three years from now. Like, if you're creating the future TSMC of Quantum, like, do you really want to do that? Or do you want to enable it so that it has your capital assets available to scale to help others, of course, and us? And so for us, it's a little bit of making sure that they can serve the whole industry. That's one. Number two, the U.S. has a strong interest in maintaining quantum onshore. Unlike semiconductor, which kind of spread around the world, and I think we're trying to reverse that trend from three decades ago, there's a strong interest, at least among national security customers that we've spoken with, in maintaining onshore manufacturing capability just as much as design capability. And our balance sheet of $3 billion helps Skywater actually be successful at that model. And so it's a mutual win-win, I think. Again, it requires approval by antitrust, of course. We remain confident that will happen by the second or third quarter, as we said on our earnings call. And, you know, trying them before even the transaction completes and having success on 256 speaks volumes for why this makes sense.

Got it. I wanted to ask now about some of the applications, and I'm going to kind of merge two questions here. The first is just, can you highlight some of the applications where you're starting to see greater commercial interest? The second part of the question, I think there's been a perception that a lot of the quantum revenue to date has been government-funded or academia. Talk about your commercial revenue, because I think it was a number you disclosed in the first quarter that probably caught some people by surprise, because it's more than half of the business now.

Yeah, exactly. I think that, and these are the logos on the slide, they're actually in real time using our existing compute, not even the future compute that we're going to create. And you can see the sectors that are adopting it, life sciences, financial services, some obvious ones. Material science also is another obvious one, I think. And so these are customers that have seen benefits of quantum even without the 256, frankly, even before the 100 qubits that we're rolling out into the market today. And they see the promise of what you can do with more computing power. That's why I think you're starting to see these hybrid deployments happen. And GPUs are terrific. And, you know, I worked at ARM and CFO of ARM, so I understand CPUs and GPUs. We've always lived with hybrid computing. I think we will still be in hybrid computing with QPUs sitting next to these others. And that time is here now. And so what we're seeing flow through our P&L now is the success of the compute solution that we have becoming more relevant to data center operators. We're saying, I want that also. And, you know, then they'll start to talk about networking and other things as well. So the AstraZeneca's of the world, if you can do drug discovery 20 times faster, well, that's worth looking at. You know, the kippus of the world, which is not necessarily on the slide, that have looked in the past at protein folding are now looking at optimizing a financial portfolio. These are applications that either we will build ourselves or in partnership with someone else. And we can't do it all, but we have about five or six different verticals we'll target with our own apps teams to develop capabilities and algorithms as we build the compute platform underneath it as well.

And just the, sorry, the commercial revenue percentages.

So, you know, two quarters in a row now, we've talked about commercial revenue basically being 60%-ish of our revenue stream. And what I wanted to point out with that metric is it's not in the lab anymore. It's not in the lab anymore. It's being deployed in enterprises as a compute platform, and those that want it over the cloud, we can do that as well. And yes, you're right. I mean, three years ago, two years ago even, the customers were government labs saying, will it work, will it not work? We're now able to take it and make it work actually out in the field. And so an electric grid operator has deployed it. They're actually getting their first deployment happen right now as we speak. And so these are non-lab customers. So the commercial metric speaks to the adoption of the platform.

Excellent. Just sort of shifting back to sort of the acquisitions the company's made over the past year plus, there had been, I think, some perception or misperception that a lot of your growth was inorganic, and you disclosed you grew 80% organically or organic growth was 80% in 2025. It will accelerate in 2026. is organic mostly a proxy for the computing business at a high level or, you know, can you just talk about the growth rate of the organic business? Because again, I think there was some misperception that, you know, perhaps the acquisitions were the primary source of revenue

growth over the past year. Yeah, thank you for the question. I think that's why we introduced the notion of organic growth, right? And organic is compute. The definition, my definition of organic is compute. And so it's what brought us here over a five-year journey of creating machine after machine after machine. And so that is growing, as you said, 100% this year up from 80% last year. So there's no slowdown. As we go into the 256 and the 10K, demand signals tell me that's going to have strong growth for a long time. We just have to execute on our roadmap. And we need the talent, and we have the talent now to do it, and the capability with Skywater to execute on it. You know, the guidance for this year says the company at the low end of our guidance will double year on year. And, you know, so within that, I said organic grows 100%. Well, the company's doubling and organic is doubling, then everything else is also doubling. And we're happy to see the engines on all of the cylinders start to fire. It's not just the acquisitions for sure. Yeah.

I almost feel silly asking a question, but I'll ask it anyway because the company is in investment mode. Your EBITDA losses are meaningful today. You've got $3.1 billion of cash on the balance sheet. When you close the Skywater acquisition, there will be consumption of cash there. So just generally, how are you feeling about the balance sheet, particularly post the Skywater close, your ability to fund the business, but then also to give yourself, you know, additional capacity to go out and do other M&A to the extent you find attractive assets in the market? Yeah, I mean,

look, I think the $3 billion of cash and obviously the investments that we're making in our business today, even before Skywater, are really important for us to have. They give us the financial firepower to do things that other companies would find more difficult to do. Not that we're necessarily doing it for that reason. We're also doing it because our customers need to know that we're publishing a five-year roadmap, and we have the dollars to make it happen. So it's important to have it for a number of reasons. We have access to capital. You know, there are times when we actually say we don't need the capital, and it's being offered to us as well. So we'll be selective about it, but we know that as long as we keep executing and we're doing it with success and putting up the numbers, which really matters, north of a quarter of a billion dollars of revenue this year, up from $130 million last year, and so on. Those numbers invite folks to say, okay, you're going to be responsible stewards. Of course, we will focus on EBITDA. Right now, it's about market expansion. It's about delivering the platform. It's the R&D dollars I talked about. Every metric is, in our view, world class. We intend to keep driving that. And so absolutely, there will be a time to focus on why do I have 10 factories? Do I need one or two? Why am I doing things in 14 places and I can do them in fewer? This is not that time. We need the innovation cycle to keep firing. Do the integration of the acquisitions where it makes sense, G&A, finance, et cetera, and then do the other integrations, I'll call it, to harvest cost over time. And then also, interestingly, that once you have scale, it becomes easier to drive margins higher anyway. So the semiconductor roadmap allows us an easier path to get to the margins. And even in the first quarter, frankly, our organic gross margins went up, So I was pleased to see that. Over time, we'll work on making the rest of the platform more profitable as well. And part of that is like manufacturing strategy.

We are at the end of time. Really appreciate you joining us for the Needham Conference. Thank you for your participation. Thanks, everybody.