IonQ, Inc. Q1 FY2022 Earnings Call

Greeting, and welcome to the IonQ First Quarter 2022 Earnings Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. As a reminder, this conference is being recorded. I would now like to hand the call over to your host, Jordan Shapiro, please go ahead.

Good afternoon, everyone and welcome to IonQ's first quarter 2022 earnings call. My name is Jordan Shapiro, and I am the Vice President of Financial Planning and Analysis and Head of Investor Relations here at IonQ. I am pleased to be joined on today's call by Peter Chapman, IonQ's President and Chief Executive Officer; and Thomas Kramer, our Chief Financial Officer. By now, everyone should have access to the company's first quarter 2022 earnings press release issued this afternoon, which is available on the Investor Relations section of our website. Please note that on today's call, management will refer to adjusted EBITDA, which is a non-GAAP financial measure. While the company believes this non-GAAP financial measure provides useful information for investors, the presentation of this information is not intended to be considered in isolation, or as a substitute for the financial information presented in accordance with GAAP. This call discusses certain non-GAAP measures, and you are directed to our press release for reconciliation of such measures to GAAP. Before we begin, please note that some of our remarks in this call will be forward-looking. Therefore, please refer to the cautionary statement in today's press release for additional details about these remarks. Please note that forward-looking statements made during this conference call speak only as of today. Now, I will turn it over to Peter Chapman, President and CEO of IonQ. Peter?

Thanks, Jordan, and thank you all for joining us today. Roughly a year before we went public, we defined a multi-year plan to drive revenue growth and technical progress over six years. Today, we stand at the 18-month mark, or about a quarter of the way into that plan. I thought, I would use this earnings call to give a quick report card on progress to date and provide my thoughts on the roadmap ahead. Originally, we expected to achieve $114 million TCV contract bookings cumulatively in the first four years of the plan. On the last earnings call, we discussed that we are seeing an increased interest in customers purchasing dedicated IonQ quantum computers. As a result, we announced our expectation that TCV contract bookings could reach nine figures for the three-year period from 2021 to 2023, roughly a full year ahead of our original financial projections. Our original financial plan was based in part on the cumulative numbers of systems IonQ expected to bring online over that period. We projected one in 2022, two in 2023, seven in 2024, 17 by 2025, and 33 by 2026. We currently have two IonQ harmony systems that feature 11 physical qubits available via our cloud partners. We recently announced our latest generation system called IonQ Aria, which features up to 32 physical qubits and 20 algorithmic qubits. In addition, a second Aria system is currently under construction to meet potential rising customer demand. Today, we are excited to announce the next generation in our series of ytterbium systems called IonQ Forte. Our early internal results indicate that this machine has an even higher average two-qubit gate fidelity than Aria. In coming days, we will be issuing a press release on preliminary fidelity of the system and a blog post that will provide additional insight into how Forte operates. We anticipate that Forte will launch for beta access in the second half of this year. We also have our first system based on barium qubits, controlled by longer wavelength visible light; barium systems are expected to make the optical design much easier while conferring higher performance. To reach our 2023 goals and beyond, this year we're working to establish and optimize our manufacturing process to build systems at scale. We are focused on features such as system footprint, manufacturability, deployability, industrial design, and reliability, as part of this development process. With regard to the number of systems we expected this year, we are clearly ahead of plan. On the applications front, I am happy to report that early results with customers demonstrate that we are making serious inroads with quantum machine learning. Recently, we were able to show that we can effectively train quantum image classification models to achieve performance that rivals corresponding classical machine learning models while using only a small fraction of the number of training parameters. These learnings build upon the results we previously reported on the speed of the training of the machine learning models, and the results themselves were more expressive and better captured the signal of the original data. We are learning how this can lead to superior performance as we scale up to higher numbers of algorithmic qubits. In summary, IonQ is firing on all cylinders, making progress on or ahead of schedule when compared to our roadmap. But that's just what we've accomplished today. What's more important is where we are in our race to make quantum computers that broadly beat classical systems. The first thing we need to do is to push our systems to their limits on native gate fidelity, i.e., the fidelity that qubits can achieve without error correction. IonQ Forte is another step in that direction. But as you will see, with upcoming results from our barium system, adding this new qubit species to our arsenal pushes us even further along that path. Initial testing suggests that the fundamental error limit for two-qubit gates using barium qubits will be significantly lower than ytterbium qubits, with a fidelity limit of roughly 99.98%. We've already demonstrated in released data showing that our first barium system has extremely low state preparation and measurement or SPAM errors, another important type of qubit error. To go beyond the fundamental limits of native gate fidelity, one must implement error correction, which requires a higher number of physical qubits. Last year, together with our university partners, IonQ became the first quantum computing company to actually implement an error-corrected qubit. And we did it with a 13 to 1 qubit overhead. An important next milestone for the industry will be to run gates between multiple error-corrected qubits that result in lower two-qubit error rates. Even with the low overhead, we need more physical qubits to implement error correction. We've already proven our multi-core chips can fit more qubits into a single system. Our investment in manufacturing and engineering this year will enable us to continue scaling up our systems so that we can achieve error correction in the future. At the same time, we're also working to connect multiple systems together to combine their power using a technology called Photonic Interconnects. Our partner labs at the University of Maryland and Duke University pioneered this idea many years ago, with demonstrations appearing in peer-reviewed journals, such as Science and Nature. This technology has been reproduced by many other labs throughout the world. IonQ is working hard to translate these ideas from the laboratory into a commercial setting. The era of broadly applicable quantum computers is coming into clearer focus every day. To summarize, I'm pleased with the progress IonQ is making both commercially and technologically. We entered the public market as the first pure-play quantum computing hardware company, with the belief that we can develop world-changing technology. We have demonstrated the value of our systems numerous times with customers who are working to understand how to apply quantum computing to the everyday business problems in industries ranging from energy and mobility to machine learning and financial services. Financially, we are very well capitalized with cash, cash equivalents, and investments of approximately $586 million, which on our current plan we believe to be sufficient to get the company to profitability without the need to raise any additional capital. Last quarter, we were honored in the Time 100 Most Influential Companies list for 2022, alongside some of the most innovative companies of our era, and we intend to continue pushing the limits on Quantum Technology. None of this would be possible without the relentless efforts of the entire IonQ team, our respect for fundamental science, and our longstanding dedication to developing and commercializing cutting-edge technology. While there is still much work to be done, we are proud of our accomplishments so far. I encourage you to read through our 2022 letter to shareholders that outlines our vision for the future of IonQ and the quantum computing industry. We believe that the future is bright and are happy to have you alongside us for the journey. Next, Thomas will provide an update on the financials.

Thank you, Peter, it has been quite the quarter. Before diving into our financials, I would first like to take a few minutes to address some common questions we have received over the last quarter about the increasing likelihood that IonQ will sell entire quantum computers, in addition to selling access over the cloud. Let's start with why prospective buyers might want to own entire systems. There are two main reasons why customers might prefer to own the system. The first and most common is the desire to control when they compute. As of now, our cloud offering operates with a queue system, and there are certain times where a customer may be in line behind a large number of jobs waiting to be processed, such as on a Monday morning at 9:30 Eastern, when the U.S. stock exchange is open, and financial customers start trading. If you need quantum compute time to process something time-sensitive, for instance, a hedge fund financial model or an update to world conditions, this probably would not work well for you. So having a dedicated system and being able to define the queue is likely to be important to many users. The second reason is security and a need to keep certain sensitive data on in-house systems. Although we are already very diligent about keeping our customers' data secure, there are certain entities that, due to internal controls and procedures, cannot run specific processes or expose sensitive data to any machine besides their own. As we mentioned on our last quarter call, we are already seeing substantial interest in system sales stemming from private companies and governments alike. To that end, we are doubling down on making IonQ quantum computers government-ready. Our government sales efforts are already bearing fruit. In April, we announced that IonQ was part of a group of private companies and universities that received a Research and Development award from DARPA to fund a quantum algorithms benchmarking project. While contracting is still being finalized, we are the only quantum hardware makers selected for this multi-year award and are excited to contribute to an initiative that will estimate the resources necessary to run important quantum applications. This is something IonQ has long been a public advocate of, as we believe absolute transparency will benefit the quantum computing industry as a whole and give customers a standard with which to compare systems across different methods. Moving on to our financial results from the first quarter, we're happy to announce we achieved total revenue of $2 million and bookings of $4.2 million. Our investments in growing our sales team are starting to bear fruit. Moving down to the statement of operations for Q1 2022, our total operating costs and expenses for the first quarter were $20.2 million, up 171% from $7.5 million in the comparable prior year period. Breaking this down further, our research and development costs for the first quarter were $7.3 million, up 101% from the $3.7 million in the prior year period. Recall that as Peter noted, we are investing heavily in R&D given the anticipated demand for building more systems than previously projected this year. Our sales and marketing costs in the first quarter were $1.9 million, up 724% from $227,000 in the prior year period. This increase was due to the growth in our sales team, as IonQ has started to invest heavily in our commercialization efforts. Our general and administrative costs in the first quarter were $9.2 million, up 211% from $3 million in the comparable prior year period. This increase is largely attributable to the growth in stock-based compensation expenses, which were $4.8 million for the first quarter, compared to $1 million in the comparable prior year period. All of this resulted in a net loss of $4.2 million in the first quarter, compared to $7.3 million in the prior year period. It is important to note that these results include a non-cash gain of $13.4 million for the first quarter related to the fair value of warrant liability. We saw an adjusted EBITDA loss for the first quarter of $10.3 million, compared to the $5.5 million loss in the prior year period. Turning now to our balance sheet. Cash, cash equivalents, and investments as of March 31st, 2022 were $586 million. We believe this gives us sufficient cash reserves to get to profitability without needing to raise additional funds given our current roadmap. With this substantial cash position, the company's operations should be well-positioned to deal with the ups and downs of market fluctuations, energy prices, interest rates, and so on. Now turning to our second quarter outlook. Given that we are still at the beginning of our commercialization phase, I want to reiterate my comment from our last earnings call that we expect bookings to continue to be lumpy for quite some time. Our projection for the second quarter financial results reflects this lumpiness. In the second quarter we expect revenue to be between $2.3 million and $2.5 million. We anticipate second quarter bookings of between $3 million and $5 million. We are excited to share that we are increasing our expectations of full-year 2022 bookings from a range of $20 million to $24 million on our last earnings call to a new range of $23 million to $27 million. With that, operator, we would now like to open the line for questions.

At this time we will be conducting a question-and-answer session. Our first question is from David Williams of Benchmark. Please go ahead.

Hey, good afternoon and congrats on the progress.

Thank you.

Yes, so I wanted to touch, I guess, real quick, maybe, Peter. But as you kind of think about the backdrop and quantum computing and the demand trends and you think about your customer base. Can you kind of help us parse out your customers in terms of those that are more commercial versus academic and research-driven? How much of your revenue, or how much maybe your bookings are driven by beyond maybe the academic world?

Let’s say, we have a wide range of different customers. It’s a good spread. Some of it, to be honest, comes from cloud companies and we don't know exactly how customers are using it. So there's kind of a wildcard there, if you will. Anecdotally, we see things out on the Internet, where students are using the quantum computers to do their quantum homework. Professors are using our quantum computers to teach classes and make YouTube videos, and then you have companies, large companies that are doing early work on these and producing scientific papers. So, you know, it is kind of a mixture of all of them. I'm not sure that I even would be able to generate an exact breakdown because we don't have full transparency into sometimes about the cloud companies and their users.

Okay, that's helpful and makes sense. And then maybe, there has been some discussion, I guess, about the ability of quantum machines to perform simple calculations, if you will. Can you talk a little bit about maybe the importance of algorithmic structure and why quantum machines are maybe not as useful in solving practical logical problems like your Casio calculator?

Yes, it's a great question. So quantum computing is likely to solve certain differential equations and at the same time is not designed for simple math. It’s hard to kind of wrap your head around. Adding one plus one isn't what quantum computers are designed for. If you want simple math, there is an app for that on your phone. And today, we're working on problems like general object recognition that look to rival the best classical technology today. I understand for everyone trying to understand this quantum stuff is hard. However, you can solve differential equations and at the same time, how can it be hard to do one plus one? We've never said that quantum computers are going to be good at everything; they're going to be very good, we believe, against a certain subset that happens to be commercially interesting going forward.

Very good, thank you for that. And then maybe just a little bit of color around Aria; it’s just coming online. Can you talk about the cubic capacity there of this system? How that compares in magnitude performance differential to the prior generations?

Yes, I'm going to switch this—since we have Jungsang Kim on the line here, I'll let him answer it since he's the expert.

Thank you for the questions. Yes, Aria is capable of being configured up to 32 qubits. We can adjust the number of qubits based on the optimal configuration and the application at hand. Currently, our performance metric of choice is algorithmic qubits. We believe it is currently operating with 21 physical qubits because that configuration delivers the best results for maximizing algorithmic qubits, allowing customers to run the most complex quantum circuits effectively.

Okay.

Did I answer your question?

It did, it did. Thank you. I certainly appreciate that. And then maybe just one last one if I can, and maybe talk about a few, but I understand that the bookings are lumpy, but it looks like between this quarter and the guidance you're right around that midpoint of your bookings for the year. Just kind of wondering how we should think about maybe the second half in terms of bookings, what those expectations are? And I know previously you talked about maybe the fourth quarter being a little bit heavier, but just kind of how do we think about the bookings for the remainder of the year?

So as with old technology, we expect the bookings to be back-loaded. There is no good reason for it, it’s just how people spend toward their budgets. And we would, of course, if we have the opportunity to finish the sale earlier, we will, and all of the people listening here on this call will be the first to know, of course, that as we said on the call, we are able to raise our guidance based on the signals we're getting from the market.

Okay, fantastic. Thanks so much. Certainly appreciate the time and the best of luck on the quarter.

Thank you.

Thank you.

Our next question is from Ruben Roy of WestPark Capital. Please go ahead.

Hi, yes. Thank you. I guess, I wanted to start, Peter, if we could talk a little bit about the assumptions around the cumulative number of systems that you talked about getting up into the '30s by 2026? Maybe just spend a little bit of time on how you're thinking about those systems in terms of the ytterbium versus barium systems? Can we get the commercialization with the ytterbium, or to your point around error correction using barium qubits? Do we need to start to see some of those types of systems come to market before we can get to real commercialization of revenue? That would be helpful. And then also I guess around that same topic, if you could just talk a little bit; you mentioned that you're working on optimizing manufacturing and system footprint, etc. How are you envisioning the footprint of the systems once we get out to maybe 2025, 2026? Are these going to be smaller in form factor? I think at some point you had mentioned that we might see some rack-mountable systems out in the marketplace. So, an update on that would be helpful. Thank you.

Yes. Great question, Ruben. So I think it's true for any quantum computing company that the very first thing you do is look to optimize the native gate fidelity because that kind of determines how much overhead you will have to do for error correction. The more you can get that to be lower, the better in terms of the number of physical qubits you have to consume. So we are on one hand busily pushing in both ytterbium and barium to get the lowest number we possibly can. And so, we've talked about how basically at ytterbium we think that number is roughly 99.8 and at barium it's 99.98. So basically, kind of almost another 9, and then after that, we have to start thinking at least 9 in use case. We probably need to do error correction to bring in more fidelity, and that means we need more physical qubits to do that. So we're kind of pushing on the native at one time and now what you're hearing from us is we're starting to also push on this number of physical qubits. We have been doing that in the sense that, for instance, we announced a new chip where the number of qubits was much higher. We will continue to do that and increasing the number of qubits on the processors, the QPU itself. Now what you're hearing is that we are starting to think about building more quantum computers, so that we would have a pool of physical qubits to start to do error correction. The key kind of piece to that story is that as soon as you have multiple systems, you need to network them together. Because you need to do entanglement from one QPU to another, which means that they have to be networked. This is true, not just for IonQ, but I believe that’s true for any technology. So now the question is, we're starting to think about how can we manufacture more computers, more quantum computers, and now it's the natural time to sit down and say let's work on shrinking them. Let's start working more on reducing their costs in terms of manufacturing; those haven't been our primary focus really up until this point. And now going forward, we're standing up a separate team whose sole focus is kind of going forward. And if it turns out that we can’t make progress on, if you will, in terms of risk factor of one of the other qubits fees, it just means the error correction is coming sooner. That is what we believe. We're gearing up really for the end game in the race to the finish here. So that's the reason we're getting started. And as we talked about in the last earnings call, if we're starting to get interest from potential sales, well, we should start getting ready for that as well.

Very helpful. Thanks, Peter. Also, I wanted to ask about the Aria systems; you mentioned the second Aria system under construction. You also have the announcement around collaboration with Microsoft for the Aria. Are both of your Aria systems going to be available via the cloud, or is one going to be available via IonQ cloud? I would love some detail on how you're thinking about deploying those systems over the course of this year.

I'm not sure we've decided all the final details. But in general, what we tried to do is to build two systems in the same class. That way, if one is down for maintenance, you have another one that you can make available. And you're adding to extra capacity when they're both up and running.

Right. Okay, makes sense. And then just a quick follow-up question for Thomas; you mentioned the growth in the sales team. Can you give us an update of where you are in terms of how you were thinking about sales headcount for the year? How much further do you think you're going to go this year and maybe an update around any other sort of near-term headcount additions that you need to make around R&D or otherwise?

Yes, as we kind of talked about, we started off with a very small sales team last year. We had a set goal for the sales team this year, and then we believe we hit that in the first quarter or pretty darn close to it. In terms of the number of people who we wanted to get on board early, obviously, because there is no point in getting salespeople on at the end of the year; it's going to help your sales numbers. So we set an aggressive goal for kind of putting the team together so they are busily working on sales now, and I would expect that the sales effort will continue to grow. There is lots of greenfield in terms of opportunity, both domestically and also internationally.

Good update. Thanks, Peter.

Thanks, Ruben.

Our next question is from Richard Shannon of Craig-Hallum. Please go ahead.

Great. Thanks, Peter and Thomas for getting me in here. Just want to follow-up on a couple of questions here or topics today and brought up in the past year regarding the customers' interest in buying the system. I guess two angles to look at this: first of all, is there a real demand here for private companies buying a quantum computer in the near-term? And would those customers be using it just for themselves or potentially sharing it with a small group? Because it seems like it's kind of a purchase might be a little bit heavy for most organizations, when we finish the year, or is it something that it's mostly a government kind of sale here at least in the near-term?

You can put it on your credit card. So more seriously, what we're seeing is a lot of demand and interest there. Many times it's for an Aria, or a group, or an organization that's looking to get it for kind of a group of end users. Sometimes with, I’d just say it could be in a university setting or in the international or national setting or government. So, it's incentive in that context.

Okay, that's fair enough. Second question I had was on your new system, I guess. I wanted to—I think you described this as a ytterbium-based system, is that correct, Peter?

That's correct, yes.

Okay. So are there any advancements in the system here that are translatable or usable in the future barium systems? And do you expect to have and continue to develop next-generation versions of ytterbium?

I'm going to switch that to Jungsang, since we have the expert online.

All right, thank you for the question. Yes, these systems actually take a while to build, so we actually have been planning and developing the systems for quite some time. Obviously, I think we—at this point, we have this Forte, which actually includes a few new features. One of them is the way that we address this individual items to run logic gates; the beams are actually steerable. I think Peter mentioned that in one of the conferences at the end of last year, we have that technology. Those technologies and approaches actually scale and translate very easily into other atomic species or other qubit species like barium. So certainly many of these architectural features and underlying technologies are definitely aligned to all of our future systems, whether they are ytterbium or barium. At this point, we are very confident with our ytterbium technology; we can make—continue to make higher performing quantum computers with them, and we believe that trend will continue with Forte. But I think we are also going to make a lot of background understanding, so we can actually transition with barium. We are finding out some of our early measurements that barium systems do give us a lot more room in terms of higher performance systems, so that type of R&D is going to continue into the future.

Okay, perfect. Thank you, Jungsang. My last question is a follow-on from an earlier question today, which is understanding your customer set here. Peter, you mentioned that some of your customers have come through the cloud and you don't know identities, target, kind of, profile that. Is that suggesting that the cloud is actually a significant portion of that? I guess, my assumption was that most of those were coming from certain contracted customers and quantum is a small amount from the cloud. So can you clarify that for us, please?

We expect the cloud to be an important part of the quantum ecosystem going forward. It offers an easy way for customers to sign up and give it a try. Anyone on the call with a credit card could be up and running on one of our quantum computers pretty easily. So in that sense, it is particularly important. But the cloud is a commoditized market; it's intended to give you cheap access to hardware, and quantum isn't that; that place yet. We're at the early stages of the TV market, not the end, so it hasn't been commoditized. So one would expect that in a limited supply, that it would make more sense to have other business models that are not as favorable in terms of say the cloud. So, we think that the cloud stuff is very important; we're very happy that we've democratized access to these early quantum computers so everyone can get access to it and give it a try. However, we believe it will be a while before now the cloud market is the market for quantum.

Okay, fair enough. Appreciate the details, Peter. That was all for me. Thank you.

Our next question is from Scott Fessler of Morgan Stanley. Please go ahead.

Thank you. Would you guys announced or you named a couple of new large customers that were already trialing the beta for Forte launching later this year? Do you anticipate the new customers will headline these trials, or are they largely demoed by the same customers that are going on existing systems?

We are very much still in the early stages. We have usually when we release a new piece of hardware, the first people that get access to it are the internal applications team. As you can imagine, they are often working with customers, and so they might get early access to the system; there might be bugs. With that early access, it's always not perfectly smooth. But customers want the kind of latest and greatest, and so that's kind of part of the deal. They are likely to be the first ones to write papers or make announcements or something as a result.

Okay. Just a follow-up there then; I know it's a slightly different topic, so as far as for Thomas, for the growth in or upping the yearly guidance for total contract bookings. If you roughly split the customer set into like three broad categories between government, academic, and commercial, where do you see the greatest growth coming that gives you conviction and taking up the contract bookings estimate?

So we actually see a lot of inbound interest, and what we don't know is who is going to break first. What we're presenting is a weighted view of our sales funnel, and together they form our forecast. So I can't tell you which one it is, because at our early stage of commercialization, every sale is binary; it happens or doesn’t. I don't know, but I will also say, as we stated before, we do not intend to break up segments just yet. I think we're too young a company to do that.

Okay. All right, thank you.

I'd say you’d have to give us a little more time to have enough data to be able to draw an appropriate timeline.

All right, thanks.

Our last question is from Quinn Bolton of Needham & Company. Please go ahead.

Hi, sorry I was on mute. Just want to say congratulations on the progress to date. Had a few questions, first on Forte. Does Forte use the same quantum processor that Aria does? So it's a 32 physical qubit system. And a related question, do the steerable laser beams or the steerable laser system give you better gate fidelity? Does it give you faster gate speed, as a result, with the different system?

Yes, this is Jungsang. Yes, I think that's a great question. So the first question is, we are actually we've actually been on a trend to make sure that our quantum computers become more software configurable, meaning we don't like to eventually get to a point where the number of qubits is really not tied to any of the hardware, the underlying hardware components in the system. I think Forte takes a big step in that direction in the sense that in the previous systems we had 32 fixed beams, and that's what limits the number of maximum qubits to 32. But with this steering system, we actually have that very much under software control. At this point, Forte can actually—the control system we have in Forte connects to address more than 40 qubits, although there are other reasons why we are actually not exactly operating at those levels, yes. But these are the kind of technological pieces we're putting together so that in the future our quantum computers will become much more scalable and not necessarily tied—the qubit numbers are not necessarily tied to the hardware. And then, the second answer to your second question is, yes, it indeed turns out that using this beam steering approach, we can actually make the laser beam that addresses the items to be much more stable and actually be less susceptible to some fluctuations in the system. Also, indeed we are seeing that our gate fidelities are getting better. When I think we can also deliver more power to the system, so I think we have a lot of room to continue to improve the gate speed and the systems as well.

Great question.

Great. Thank you for that. Second question is just on now that you've announced Forte and its beta availability in the second half of this year with broader availability next year. I think Peter, in the past, you've sort of said that the public cloud systems may be sort of the generation minus one. So as Forte becomes public or becomes available, do you expect Aria to sort of come to the cloud, sort of, the second half of this year once Forte is available to private beta customers?

Yes, no, I think you're seeing what I said in action, if you will. Aria is just coming to the cloud, and at the same time we're announcing the current generation. And we haven't announced that that would go to the cloud yet.

Understood and then lastly—go ahead.

Yes, no, it's just kind of following that trend.

Got it. And then lastly for Thomas. I'm wondering if you might be willing to sort of give us on the revenue forecast of $10.2 million to $10.7 million this year that you're reiterating. Can you give us some sense of what you would think the EBITDA loss might be at that revenue level? Just trying to get a sense of what you expect OpEx might look like over the balance of the year?

So I think that we shouldn't expect huge variations on that front, because when we do—increased bookings towards the end of the year, or in the second half, there will be less impact on the revenue recognition line, which is what flows down to EBITDA.

Pardon me; I think you did $10.3 million of adjusted EBITDA loss in Q1. Do you think something in the $40 million to $50 million EBITDA loss is the right level to be thinking about based on the $10.2 million to $10.7 million revenue? I mean, I assume that OpEx probably has some upward bias over the course of the year?

So we’re holding from our previously stated EBITDA loss of $55 million. It should not increase to last because we have to build these things as well, of course.

Okay. Perfect, thank you.

Ladies and gentlemen, we have reached the end of the question-and-answer session. And I would like to turn the call back to Peter Chapman for closing remarks.

Well, thank you all for joining our Q1 2022 earnings call and for all your thoughtful questions. I'd like to also thank the entire IonQ team for all their hard work that has brought us to this point to new heights for this quarter. We look forward to seeing everyone in the Q2 earnings call next quarter. So thank you again for joining and have a great evening.

This concludes today's conference. Thank you for joining us. You may now disconnect your lines.