Greetings. Welcome to the IonQ fourth quarter 2021 earnings call. At this time, all participants are in a listen-only mode. A question and answer session will follow the formal presentation. If anyone should require operator assistance during the conference, please press star zero on your telephone keypad. Please note this conference is being recorded. I will now turn the conference over to your host, Jordan Shapiro, VP of Financial Planning and Analysis and Head of Investor Relations. Thank you. You may begin.
Good afternoon, everyone, and welcome to IonQ's fourth quarter and full year 2021 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 fourth quarter and full year 2021 earnings press release issued this afternoon, which is available on the investor relations section of our website at investors.ionq.com. 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. You are directed to our press release for reconciliation of such measures to GAAP.
Now, before we begin, please note that some of our remarks on 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 these 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. I first want to acknowledge the deeply troubling crisis still unfolding in Ukraine. It's a time of great concern for all of us, and it is simply devastating to see the impact it's having on the lives of so many people, including some of our IonQ staff members who have family in harm's way.
Before I get started, I thought I would share a little history of IonQ. When I joined IonQ three years ago, the board told me to ignore sales, ignore the competition, and ignore the press. Just focus on building the world's best quantum computer and bring it to market as quickly as you can. In some sense, a field of dreams approach. If you build it, they will come. That's just what we believe we have done.
We believe we have built the world's best quantum computer, as shown in our latest generation IonQ Aria machine. I tell you this story because it's important to understand where IonQ is in its commercialization efforts and the upside to come. At the beginning of 2021, we had no dedicated sales executives and were selling two 11-qubit machines.
Now having finished 2021, I am happy to report that even after tripling our original 2021 contract bookings forecast in September from $5 million- $15 million, we beat that number again to end up at $16.7 million for the full year. This resulted in recognized revenue of $2.1 million for the 2022 fiscal year, which was 31% above the $1.6 million we forecasted on the Q3 call.
This year, we are starting to stand up our production engineering and manufacturing departments. This group will be responsible for building all commercial quantum computers going forward, including for the first time, selling systems outright to customers, which we anticipate could start shipping as early as 2023. In our outlook today, we have not included any upside related to scaling the sales of these machines.
We believe that over the next two years, one or two system sales could push our combined TCV contract bookings over nine figures for the three-year period from 2021 to 2023. That said, the quantum industry is nascent, making it hard to predict which quarter particular sales contracts will land. As a result, we expect our contract bookings in the near term to continue to be very lumpy.
In 2022, we expect to more than double our sales staff, including our first salespeople in Europe. We are now selling our recently announced Aria machines that are more than 1000 times more powerful than last year's model. Our forecast for fiscal 2022 expects our 2022 revenue to be over $10 million at the midpoint, which will be approximately 5x our 2021 revenue.
We estimate full year 2022 bookings at $22 million at the midpoint. This represents a $7 million or 47% increase over our original forecast, which we set out before going public, and a 32% increase from our outstanding 2021 results. Thomas Kramer, our Chief Financial Officer, will go over the rest of our financials in detail shortly.
On the technology side, in February, we took a leap forward, not just for IonQ, but for the quantum computing industry as a whole. We announced the performance results for our latest generation Aria machine. This system achieved a record 20 algorithmic qubits with the independent industry benchmarks from the QED-C, showing that Aria clearly outpaces every commercially available quantum computing solution today.
We are confident that IonQ systems are the best in the world, and that we are only furthering our lead. Last week, we were excited to announce that Microsoft customers can leverage the power of IonQ's Aria series quantum computer on Microsoft Azure. We expect this will unlock the creativity of Microsoft application developers worldwide. IonQ Aria is only one part of the story.
As part of our plan to scale IonQ quantum computers beyond Aria, we announced in December that we plan to build new systems using barium qubits. Barium qubits have the potential to enable lower error rates and higher gate fidelities, which in turn could lead to more accurate and powerful quantum computers.
They can also be controlled by visible light lasers as opposed to ultraviolet lasers we use in our ytterbium systems. We believe visible light lasers could result in cheaper, more readily accessible components and a promise of faster gate speeds. Another crucial advantage of barium qubits is that they're easier to network together over long distances. Networking quantum systems together is key to making them more powerful and eventually to building a quantum internet.
In summary, while Aria is leading the pack with its ytterbium qubits, we believe the addition of barium qubits to power our systems could further our lead in system performance. We believe our investment in barium qubits is already paying off. Earlier this month, we published data from the forthcoming barium-based system, showing a 13-fold improvement in state preparation and measurement error, lovingly known as SPAM error in the industry, one of the key limitations to scaling quantum computers.
Expect more exciting news to come from our barium systems later this year. Given the state of the global supply chain, a natural follow-up question might be, where do we get enough barium qubits to serve our needs? In February, we announced the results of our public-private partnership with the U.S. Department of Energy's Pacific Northwest National Laboratory to secure a constant supply of barium atoms for our computers.
We could not be more thrilled to be working with the esteemed scientists and engineers over at PNNL to make barium-based quantum computers a reality. Next, let's talk about our research and development progress since our last conference call. Last month, IonQ announced that together with researchers from Duke University, we outlined a novel method of running a new set of N-qubit quantum gates that operate on many qubits in a single step, including the well-known Toffoli gate.
We expect N-qubit gates to lead to substantial breakthroughs in quantum algorithms, making them more efficient and powerful with fewer operations. What's more, this new family of gates can only be run on the proprietary architecture exclusive to our trapped ion computers. Our platform has the most algorithmic qubits in the industry, and now we're enabling those qubits to do even more.
We expect these gates to be essential to work across applications such as quantum chemistry, optimization, and machine learning. Last, let me discuss our progress on real-world quantum applications with customers. In January, we were thrilled to announce a new partnership with Hyundai to develop quantum algorithms to simulate battery chemistry for electric vehicles. This work has the potential to greatly enhance charging, discharging, durability, capacity, and safety of batteries.
With electric vehicle adoption increasing all over the world, battery development is a critical component in the transition to a more sustainable mobility. The Hyundai partnership follows a series of projects on quantum chemistry at IonQ. We have previously announced the work with Dow to demonstrate an end-to-end pipeline for simulating molecules, such as those present in fertilizer creation. Our work with Dow was an extension of previous IonQ efforts to simulate water molecules with our quantum computers.
The Hyundai project shows the natural progression of tackling increasingly complex molecules with each project. We believe the experience we gain from manipulating molecules in the new Hyundai partnership will be broadly applicable in related fields such as drug discovery and materials development. In summary, we have confidence in our 2022 contract bookings forecast of $22 million.
As we are getting ready to sell full systems over the next couple of years, we believe there is potential for a significant upside for the next 24-month period. Our team is filling out nicely with top talent joining both at the executive and individual contributor level. We're focusing special attention on scaling our sales force. We continue to lead the industry with our technical milestones and are already seeing our endeavor into barium-based systems paying off.
We are the leader in the cloud quantum computing market as the only company available on all three major platforms and compatible with every major quantum developer language. Although we did not discuss it in detail today, our patent portfolio continues to grow at a pace, which we believe will continue to be an asset for decades to come.
Between our financial results, technical progress, and significant cash reserves, we believe we are well-prepared to maintain our leadership position and drive the adoption of our quantum computing solutions. For more detailed information about Aria, algorithmic qubits, our barium system, SPAM error, and the Microsoft announcement, I point you to our website. With that, I'll turn it over to Thomas for an update on our financials. Thomas?
Thank you, Peter. Good afternoon, everyone, and thank you for joining us. I would like to start off by going over our results for the quarter. As Peter mentioned, we outperformed our outlook this quarter with $1.6 million in revenue and $1.5 million in contract bookings for the quarter. We ended the full year with $2.1 million in revenue, which was $500,000 or 31% higher than what we estimated on the third quarter call. Our TCV contract bookings ended up at $16.7 million for the year, which exceeded our most recent midpoint estimate of $15.8 million given on the Q3 call, which was already up 3x from our beginning of the year forecast.
Given that we are still at the beginning of our commercialization phase and that in addition to transactional-based cloud revenue, we sell large contracts where customers pay for reserved compute access, we should expect bookings to continue to be lumpy for quite some time.
We have been approached by several entities about potential system sales outright, which would further add to this trend if and when these sales should happen. Moving down the income statement, our total operating costs and expenses for the fourth quarter were at $12.5 million, up 153% from the $4.9 million in the comparable prior period. For the full year 2021, this totaled $40.8 million, up 159% from $15.7 million in 2020.
To break this down further, our research and development costs for the fourth quarter were $4.9 million, up 96% from $2.5 million in the year prior period. For the full year, R&D totaled $20.2 million, up 99% from $10.2 million in 2020. We expect these costs to continue to increase sequentially as we invest in building our next generation of computers. Our sales and marketing costs in the fourth quarter were $849,000, up 281% from $223,000 in the year prior period. For the full year, sales and marketing costs were $3.2 million, up 565% from $486,000 in 2020.
This increase was due to further investment in our commercialization efforts and growing our sales team. Our general and administrative costs in the fourth quarter were $5.4 million, up 217% from $1.7 million in the year prior period. For the full year, G&A costs were $13.7 million, up 287% from the $3.5 million in 2020. This increase is largely attributable to a growing headcount and overhead associated with scaling, as well as incremental costs and fees associated with being a public company. We expect G&A costs as a whole to continue to increase sequentially as we hire more world-class talent to scale the business accordingly.
In our full-year results, you may also note the $4.3 million non-operating costs on our P&L, which is the portion of the transaction costs allocated to the liability-classified warrants from the original dMY III IPO. This is a one-time non-cash expense and will not be a recurring charge. We saw a net loss of $74.1 million in the fourth quarter compared to $4.9 million in the prior year period. For the full year, net loss was $106.2 million compared to a loss of $15.4 million in 2020. Included in the net losses for the fourth quarter and the year were non-cash charges related to the fair value of our warrant liabilities of $63.3 million.
All of this resulted in an Adjusted EBITDA loss for the fourth quarter of $7.9 million compared to a $4 million loss in the prior period. For the full year, our Adjusted EBITDA loss was $28.3 million compared to a $12.8 million loss in 2020. Turning now to our balance sheet. Cash, cash equivalents and investments as of December 31, 2021 were $603 million. Going forward, we expect our cash on hand to be sufficient to fund operations for the foreseeable future without the need to raise additional capital. Recall that our transaction in dMY III closed on September 30 last year and provided us with net proceeds of $573 million.
We continue to believe we are well capitalized and well-positioned to benefit from increased interest in quantum computing in both the public and private sectors. I also want to mention a new development on the fiscal infrastructure side. We will be opening a new facility in Seattle, which we believe will be important for both talent acquisition and scaling our production engineering function.
The Pacific Northwest has long been a hotbed of technology development, and quantum computing has been no exception. As we build new and lasting ties in the area with PNNL, the Northwest Quantum Nexus and other partners in the Greater Seattle area, we believe this will greatly benefit the company. We are looking forward to spending more time building relationships and quantum computers in and around the Puget Sound. Now turning to our first quarter and full year 2022 outlook.
This year we anticipate full year 2022 bookings of between $20 million and $24 million, with between $3 million and $4 million for the first quarter. As a result of bookings in previous periods and new 2022 activity, we expect revenue for the full year to be between $10.2 million and $10.7 million, with between $1.8 million and $2 million for the first quarter.
We also anticipate an Adjusted EBITDA loss of $55 million for the full year 2022 at the midpoint of our revenue guidance. Note that since we sell large contracts with terms stretching several years and with implementation typically lagging contract signing by a few months, we expect bookings to significantly exceed revenue recognition.
Until our business begins to scale in a meaningful way, it is possible that the timing of large deals may have an outsized impact on our revenue and bookings quarter to quarter. This means that we may not always see linear bookings growth, whereas recognized revenue will tend to follow a more traditional pattern barring any system sales.
Investors and analysts should also be aware that even if we see a spike in bookings in a period, the impact on recognized revenue may not be significant for several quarters. Our fourth quarter and full year results were encouraging, and we are seeing strong momentum both in external demand and internal technical advances. The world has been through a lot in the past few years when we have all learned the meaning of terms like pandemic and supply chain.
Inflation concerns and the unrest in Europe are unlikely to improve worldwide stability. 2022 brings with it its own set of challenges, but it's too early to identify the impact to the business environment. We feel comfortable saying we have no direct supply chain exposure to Russia, though we could still see adverse effects from the general slowdown around the globe.
That said, it is not clear that the current turmoil will produce headwinds that will outpace the demand from both corporations and governments as they seek solutions to previously unsolvable questions. In the absence of further unforeseen events, we believe we are well positioned to continue executing our plan and look forward to updating the entire investment community on our progress. Finally, it is my pleasure to welcome the newest member of our IonQ team, Laurie Babinski.
Laurie joins us as our General Counsel with deep legal experience from Intuit's Credit Karma. Laurie brings a wealth of IP experience in the technology sector and will be a wonderful addition to the team. We look forward to working with Laurie and welcome her contributions to IonQ and the quantum computing space as a whole. With that operator, I would like to open the line for questions.
Thank you. At this time, we will be conducting a question and answer session. If you'd like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star two if you'd like to remove your question from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. Our first question comes from the line of Ruben Roy with WestPark Capital. Please proceed with your question.
Thank you. Thanks very much for the questions. Hi, Peter. Congrats on the continued progress and a great finish to fiscal 2021. I guess the first question I had is just around the commentary around system sales and being approached by entities around system sales. I think in the past when we've spoken about revenue rec over the next several years, system sales was a little bit further down the list in terms of how you're looking at, you know, your overall hardware. You have hardware in the cloud, you've got engineering services, et cetera. Has anything changed with your view on how hardware system sales will progress?
Are they potentially going to progress more quickly because of the barium qubits that you guys are working on or anything changed in the way you're thinking about rev rec over the next three to five years? Thank you.
Well, there's a couple of pieces. One is, we do recognize to get to scale that you have to be able to manufacture multiple quantum computers, and in addition that, you want those quantum computers to be standardized and
You know, easily manufacturable, which means you need to be able to support them and all the rest. What I would say is we're ahead of schedule in terms of the technology plan. We're starting now to work on that particular problem.
You'll start to see. You know, we've already said to the market that we are going to work on photonic interconnect to start to connect these things together. That's a way to get to much larger quantum computers. What is new is that we have seen a great deal of interest from several customers where they have expressed interest in getting you know, a full system. That is new from what we had said before.
Great.
And, and-
Congrats on that.
the other-
I'm sorry, go ahead.
I'll just add a little bit, just given, you know, the impact potentially on these, it obviously would drive kind of maybe a single sale would make a quarter be pretty lumpy. You should expect a certain amount of lumpiness going forward as we see these.
Sure. As a quick follow-up then, my question for Thomas, just around the improvement in bookings. Is that fairly broad-based, Thomas? Is it or is it more concentrated either from government or some commercial partners that are, you know, coming forth? Any additional detail around the bookings progress would be helpful. Thanks.
Absolutely. You're right to ask about the rev rec. It's obviously different for system sales versus more system as a service sale. However, we thought it would be prudent of us to just raise that. It's not a flag, it's more of like it's a good thing, but we want to raise it before it happens so that we're not all surprised. When we look at the lift in our bookings, we're seeing a continued interest from both corporate, academic, and government institutions.
Currently, we won't be breaking out individual sales, but we are very encouraged to see both how varied the interest is in quantum as well as how many jobs that are submitted every day from people we don't even know who are, because they're coming in through AWS and Azure and Google Cloud.
Sure. Okay. Thank you very much, guys. Congrats again.
Thank you, Rudy.
Thank you.
Our next question comes from the line of Scott Fessler with Morgan Stanley. Please proceed with your question.
Hi, guys. Congratulations on the great quarter. You know, following up on the system sales question, what impact should we expect that has on the operating model, you know, particularly around margins?
We are not foreseeing any immediate change to the operating model. What you will see is that once we implement the system sale, you would see a bump both in bookings revenue and margins because you're able to take down all of the benefit of the revenue to the current, whereas we will not stretch out all the expenses. It will be a bump to the margin, but it wouldn't be an immediate. Like, it wouldn't be outside of the model because we still would recognize all the same costs. It just happened in one period instead of several.
Got it. If I could get a follow-up in as well. Could you just share some early customer feedback you had from the Aria launch?
Generally the customer feedback is really very, very positive. People are impressed with what this new hardware can do, and are excited to see kind of being able to take their applications to the next level. So far the customer feedback has been tremendous.
Excellent. Thanks, guys.
Thank you. Our next question comes from the line of David Williams with Benchmark Company. Please proceed with your question.
Hey, good afternoon. Thanks for the time to let me ask the question. Clearly making some very nice headway here, some good progress. But I wanted to ask is, as you kind of think about the landscape and understanding of the technology and applications, how do you envision the competitive dynamics developing over time? It seems the size of the opportunity would provide room for I guess multiple technologies to coexist without restrictions to growth. Just kind of curious how you think about the competitive landscape and maybe your business just kind of given the strength of your systems.
You might hear from some competitors where they think that there'll be kind of certain niches that they'll be able to kind of be the leader in. It's interesting depending on the qubit technology when you have to design the chip, you know, before manufacture time.
That's where you sit down and make a business decision because of this early stage of their processors that maybe a particular chip might be designed in a way that might help a particular application. Obviously one manufacturer can't go and do all chips for all applications. That leads to that segmentation. Interestingly enough, at IonQ, we do the design of the chip, if you will, at runtime.
We can do at runtime. We can basically design, if you will, the chip. You know, the perfect chip for your application. We kind of see our devices unique in that sense, in that it can be morphed on the fly for any application, which makes it more, you know, a better machine than, say, some of the competitors.
You know, at some point we'll have so many qubits that you can throw away qubits and, you know, do a least common denominator approach to quantum, but that's not in the near term. IonQ basically doesn't have to make that decision because we basically wire up the chip at runtime, and so it can basically be anyone else's chip and all of them combined at the same time.
Thanks so much for the color. And I guess one of the other areas I know that investors and, as I think about your customers' difficulty understanding maybe some of the benchmarking and how you stack up with the other ones. I know that the algorithmic qubit is an important metric there. Could you kind of walk us through the value there and why the algorithmic qubit is more important to think about versus maybe just the standard qubit and how you stack up in
Yeah
some of the others.
First thing is the benchmarks themselves. These were selected by the QED-C, which is an industry consortium made up of all the quantum companies. First thing to know is it wasn't, if you will, an IonQ choice of algorithms, although we obviously had, you know, input to the selection with many other companies.
We think it's important because, you know, these benchmarks represent the same kinds of programs that we think customers are gonna want to run. In that sense, they're the benchmark is closer to being representative to what a customer will run. Other things are much more esoteric, and they might be looking at kind of one particular, you know, component and not all of them combined into one thing.
This is very similar, if you will, to the classical world. There's a number of classical benchmarks which everyone uses in HPC or even laptops, which are all application-oriented benchmarks. We think, you know, it's important that an independent organization, you know, like QED-C or maybe IEEE should be the one who's deciding what the benchmarks are. An improvement from what's happening today would be to have an independent third party to actually run them. Instead of the individual companies running them, we'd be in favor of all of that because we think we'll easily continue to lead.
Very helpful. Thanks so much.
Thank you. Our next question comes from the line of Richard Shannon with Craig-Hallum. Please proceed with your question.
Well, hi guys. Thanks for taking my questions. Maybe one for Thomas here, more of a tactical question on your guidance here, both on revenues and bookings. I guess probably more so pointed to revenues, but maybe you want to comment on the bookings part as well. The ranges here you have for this year are fairly tight, which would imply a sense of precision here. Wanted to get your sense of why so precise. Is this you have a lot of visibility into these numbers? How would you characterize the upside potential as we go through the year?
Excellent question, and thank you for raising it. Obviously, it is hard to make predictions, particularly about the future, but when it comes to revenue, it is actually not that hard because the way bookings tend to turn into recognized revenue is straight lined on the types of contracts we have sold up until now.
There is an implementation lag between when we sell it and when they start producing revenue of roughly three months. You can do that also knowing that most sales in a quarter unfortunately tend to happen at the end of the quarter. You can build your rev rec schedule that way. When it comes to bookings, we have taken the approach that we know what's in our funnel. We know how we score that funnel. Could we outperform? Yes.
However, obviously every company can also underperform. We think it will be hard to see us come in at less than 20, so we're looking at this as a minimum 20 bet. What we could have bluebirds, which is why we brought it up. However, we don't think that they would happen as quickly, but they could. That's why we're actually raising that, so that we will not be surprised. We would all like to have that surprise when it happens, though.
Okay. That's very helpful, Thomas. Thanks for that. Maybe a multi-part here on kind of the sales opportunity here and the structure and workforce you're supporting. I think you talked about doubling the sales force this year. I guess the first question to that is, how similar is this to kind of the assumptions built into your models in the SPAC process?
Is this above or below that, in terms of pace? Then I think you said specifically the sales force would be built in part to kind of ramp out Europe. Maybe if you wanna characterize the opportunity you see, outside of the U.S., Europe versus Asia and others, just to give us a sense of where you're focusing and why.
Just in terms of head count, it's about what we said in the original pipeline model or what we thought. You know, the point though, to be honest, is that, you know, last year we started with almost zero in terms of sales team. On the sales side, we're really just getting going, and that's really the point. And we're making good progress on that in terms of the number of people this year. In terms of locations, you know, we see customers overseas, which would be in Europe and, you know, Asia, which would be Japan and South Korea and Australia. There seems to be a great deal of interest there as well.
Okay. Great for that. One last question for me, I'll jump out of line. Peter, I wanted to get your sense of the overall environment here. You've seen, as the first one coming to the public market, a couple of spinouts, a couple companies, one not quite in, one all the way through the SPAC process here. A lot more interest and activity and funding in this space. You've got a fairly substantial cash position here that you seem to be able to fund all of your operations organically here.
How do you see the market and the need for M&A, either in terms of, you know, technical teams or even kind of organizations, either hardware or software, that might be of interest in kind of a plug into what IonQ has already done to date?
Certainly we are keeping an eye out and, you know, for those particular opportunities going forward. That's an active task at the current time. I do think just in general that, while on one hand there's more money going into many quantum companies, at the same time, I expect a great deal of consolidation to happen over the next several years.
I think you'll see some players who will decide that their technical approach to their hardware is too far behind and will throw in the towel. I expect too, that we'll see new players who have interest in quantum that haven't expressed it today, because I think a lot of tech companies are just starting to realize that they need to be a player in quantum.
There's a number of notable players yet that, you know, no one knows what their quantum strategy is, and so those are new opportunities for, you know, for partnership and the such going forward. That's kind of it. It's certainly an interesting time in quantum.
No doubt about that. No doubt about that, Peter. Thanks. I think that's all the questions from me. I appreciate it, guys. Thanks.
Thank you. As a reminder, if you would like to ask a question, please press star one on your telephone keypad. Our next question comes from the line of Quinn Bolton with Needham. Please proceed with your question.
Good afternoon. Thanks for letting me ask a question. I wanted to start just on the outright sale of the quantum computers. Maybe if you could, you sort of talk to us about sort of your thought process behind how do you go about pricing these systems? Is it based on gate fidelities? Is it based on number of algorithmic qubits? And a follow-up question is, what's your manufacturing capacity of number of systems per year to the extent you know, you start to get these orders for full-blown quantum systems?
It's certainly a great question. Let me just point the first thing, which is obvious, that you know you can't use the same sort of model for classical as you can for quantum in terms of price performance because as you saw just here, these new machines are 1000x more powerful than last year's model, so they can't be 1000x more expensive. In the future, it's just gonna get worse and worse, right? You'll get to a you know sometime in the future where the next model will be you know 100,000x more powerful than the previous model. You clearly need a more aggressive price performance mix up than what you get today.
Today, we expect in the classical world, kind of every year or 18 months, using Moore's Law, to double the performance and be able to buy the same laptop, if you will, for the same price in that same period. Another way is it's, you know, it's half the cost. That leads to, you know, an observation, which is, price per qubit needs to be going down in every generation.
As much as we're focused on the technology to improve the quantum computers themselves, we're also likewise focused on the manufacturing cost going forward and shrinking the costs in every generation. Which generally means, probably for everyone but us included, including us, which is they have to get smaller. 'Cause every time, you know, if every generation gets smaller, generally the things get cheaper.
We're just standing up a manufacturing group this year whose sole purpose it is to work on that particular problem set. We will not deliver, I don't believe, first computers off that line until early 2023.
Got it. Understood.
What's that?
Yep. Thank you. Thank you, Peter. A second question. I just had. You know, you went through a number of the benefits of the barium-based ion trap technology. Wondering if there's anything that you think still on the critical path that you haven't been able to overcome on those systems? Or do you think at this point you've sort of overcome all of the, you know, the challenges and it's really just, you know, sort of blocking and tackling to get the barium-based systems into production?
As we've kind of said all along, everything that we're doing, we have shown working in a lab once before. We do not need a breakthrough in manufacturing, in materials science, physics. No breakthroughs are required. We're largely an engineering organization where we're taking what the two co-founders have done at the university laboratories and productizing it. The short answer is, you know, I don't think there's anything. There's nothing that's standing in our way. There's no magic that we have to figure out going forward to getting to a much larger quantum computer.
Thank you. Just quickly for Thomas, you gave an Adjusted EBITDA forecast of $55 million loss in 2022. Can you give us any sense on what CapEx may be or what a good estimate for cash burn over the year would be?
Yeah, just one second while we pull that up. Obviously since we are manufacturing hardware, there is CapEx involved that will sometimes outpace our OpEx. However, we do think that that's going to level out at I'd say 3-4 years from now. I would expect that our CapEx this year should be around $23 million, more or less.
Okay. Perfect. Thank you very much.
Thank you.
Ladies and gentlemen, we have reached the end of the question and answer session. I will now turn the call over to management for closing remarks.
Well, thanks everyone for joining today. I'll just actually make a comment too that I'd like to thank all the staff members at IonQ who've worked long hours to make these results possible. Thank you so much. We are, you know, extremely excited by the results from last year and the continued growth we expect this year. Last year was an exciting year, and we expect this year to continue that pace. Thanks again for joining.