Q4 2025 Earnings Call Transcript

Good day, and thank you for standing by. Welcome to the Rigetti Computing Fourth Quarter and Full Year 2025 Financial Results Conference Call. [Operator Instructions]. Please be advised that today's conference is being recorded. [Operator Instructions]. I would now like to hand the conference over to your speaker today, CEO, Dr. Subodh Kulkarni.

Good afternoon, everyone, and thank you for joining us for Rigetti's Fourth Quarter and Full Year 2025 Earnings Conference Call. I'm pleased to be joined today by our Chief Financial Officer, Jeff Bertelsen, who will walk you through our financial results in more detail following my overview. Also with us is our Chief Technology Officer, David Rivas, who will be available to participate in the Q&A session following our prepared remarks. We appreciate your continued interest in Rigetti, and we look forward to answering your questions at the conclusion of our remarks.

Before we begin, I would like to remind everyone that today's call along with our fourth quarter and full year 2025 press release contains forward-looking statements. These statements reflect our current expectations, objectives and underlying assumptions regarding our outlook and future operating results. These forward-looking statements are subject to a number of risks and uncertainties that could cause actual results to differ materially from those anticipated. Such risks and uncertainties are described and discussed in greater detail in our filings with the Securities and Exchange Commission, including our Form 10-K for the year ended December 31, 2025, and other periodic reports filed by the company from time to time with the SEC.

We encourage you to review these filings for a comprehensive discussion of these risks and uncertainties, that could cause actual events and results to differ materially from those contained in the forward-looking statements. Rigetti undertakes no obligation to update any forward-looking statements made during this call, except as required by law. During today's call, we will refer to certain non-GAAP financial measures. For details on these measures and reconciliations to comparable GAAP measures and for further information regarding the factors that may affect Rigetti's future operating results.

Please refer to today's earnings release on Rigetti's website at investors.rigetti.com. Also the 8-k furnished with the SEC today after the close. Now turning to the business. 2025 was a year of technical validation and disciplined execution for Rigetti. We advanced materially across Fidelity, Scale and Architecture, while remaining realistic about time lines and commercialization.

Our focus remains reaching true commercially meaningful quantum advantage, not headline milestones. I want to begin by grounding today's discussion in how we think about quantum computing at Rigetti, because that perspective is centered to how we operate, how we invest and how we measure progress. Quantum Computing is not about replacing classical computing. It is about enhancing it.

CPUs will continue to handle sequential workloads, and GPUs will continue to handle parallel workloads, where quantum computing becomes powerful is in simultaneous computation, problems where thousands of variables interact at once and classical systems struggle to converge. That is the problem space we are building for. Our strategy has consistently focused on superconducting, gate-based quantum computing because it offers two fundamental advantages that matter at scale, speed and scalability. We are working with electrons, not atoms or ions, which gives us gate speeds measured in tens of nanoseconds.

And because this technology is grounded in semiconductor fabrication, we believe it offers the most realistic path to building large-scale systems over time. Over the past year, we made great progress towards what we define as true quantum advantage. I'm excited to share that Rigetti recently achieved a 2-qubit gate fidelity as high as 99.9% at 28 nanosecond gate speed on a ProDrive platform using our new proprietary Adiabatic CZ scheme. We are still maintaining 99.9%, 1 qubit gate fidelity, and we have also reported median 2 qubit gate fidelities of 99.7% on our 9-qubit system, 99.6% of our 36-qubit system and 99% on our 108-qubit system, what we call Cepheus-1-108Q.

Together, these milestones are a testament to our ongoing progress in materials, fabrication and system level design. They're helping us further narrow the fidelity gap between superconducting systems and other quantum modalities, while delivering speeds that are about 1,000x faster than some approaches like trapped ion or pure atoms. We successfully deployed multiple systems to the cloud, including an 84-qubit monolithic chip system and a 36-qubit chiplet base system. More importantly, we demonstrated that chiplet timing works in practice.

That matters because scaling to thousands of qubits on a single die is not realistic. Chiplets are how we believe quantum systems will scale in the real world. As we pushed beyond 100-qubits, we gained important insights. On our 108-qubit system, we identified tunable coupler interactions that emerge at higher scale.

We made a deliberate decision to delay general availability and address the issue. The executed architectural refinements that successfully improved system stability and control. That decision reflects our discipline and increases our confidence in our 108-qubit chiplet-based system, as we move towards customer readiness. That experience underscores why we have our own foundry.

Rigetti operates Fab 1, the industry's first dedicated and integrated quantum device manufacturing facility, which allows us to tightly couple design, fabrication and testing under one roof. This enables faster innovation cycles as we scale beyond 100-qubits and drives proprietary advancements rather than incremental work arounds. We see Fab 1 as a durable competitive advantage that accelerates our road map and create a meaningful barrier to entry as quantum systems grow in scale and complexity. That combination of scale, control and execution is also what our customers and partners are responding to.

We are also seeing increased demand for on-premises quantum systems, particularly for national governments and research institutions seeking direct access to hardware for hybrid computing and systems-level R&D. In January of this year, we announced an $8.4 million order from India's Center for Development of Advanced Computing or C-DAC, for a 108-qubit on-premises quantum computer scheduled for deployment in the second half of 2026. This system will be integrated into C-DAC's supercomputing environment and is based on our chiplet architecture, which is central to our scaling strategy. That order builds on the memorandum of understanding we signed with C-DAC to explore the co-development of hybrid classical quantum systems.

Taken together, these efforts reflect how customers are engaging with us not just as a hardware vendor, but as a long-term technology partner in hybrid computing environments. At the smaller end of the spectrum, late last year, we also announced purchase orders totaling approximately $5.7 million for 2, 9-qubit Novera on-premises systems. These systems are being used as test beds for quantum hardware research, error correction and internal capability development. Importantly, they are upgradeable, which allows customers to grow with the platform as their needs evolve.

Our Novera QPU also continues to be an ideal solution for customers who want to integrate our technology with their existing cryogenics and controls. We are pleased to announce that we have secured a purchase order for our Novera QPU from a Japanese research organization, which is scheduled to be delivered in April 2026. This will be Rigetti's first QPU to be located in Japan, and we are excited to be expanding into this new geographic region. A core differentiator for Rigetti is our open modular architecture. We do not believe the future of quantum computing will be built by any single company attempting to own the entire stack.

Instead, we have designed our platform to integrate best-in-class partners where they can move faster or deeper, than we can alone. We are partnering with Riverlane to advance real-time quantum error correction capabilities, as it is foundational to achieving fault-tolerant quantum computing. Riverlane has demonstrated capabilities that we believe will meaningfully advance that goal. We are also working closely with NVIDIA to support NVQLink, an open platform designed to integrate quantum systems with AI supercomputing.

This collaboration reflects our shared view that quantum computers will coexist with CPUs and GPUs in data centers as part of future hybrid computing environments. Another example is our collaboration with QphoX and the U.K.'s National Quantum Computing Center on optical readout of superconducting qubits. This work addresses a fundamental scaling bottleneck by reducing cryogenic heat load and wiring complexity. While this remains early-stage research, it illustrates how our architecture allows us to incorporate novel technologies that could materially improve scalability over time.

This ecosystem approach gives us flexibility, accelerates innovation and reduces execution risk as the industry evolves. The quantum computing market today remains research driver. Most systems are deployed to government labs, national research centers, universities and early commercial researchers. That is not a limitation.

It is a reflection of where the technology is in its life cycle. I want to be very clear about how we define quantum advantage because this frames our road map and our timelines. For Rigetti, quantum advantage meets outperforming classical systems on practical workloads in real computing environment for commercial applicability. We believe achieving quantum advantage requires several things to come together, scale, fidelity, speed and error mitigation.

Specifically, systems on the order of 1,000-qubits, 2-qubit gate fidelity approaching 99.9%, gate-speeds below 15 nanoseconds and integrated error mitigation. Based on what we know today, we believe we are roughly 3 years from reaching that point. That may sound conservative, but in a technology as complex as quantum computing, prescription and credibility matter more than bold claims. Looking ahead, 2026 is about execution and scaling.

Our near-term priority is completing deployment of the 108-qubit system at 99.5% median 2-qubit gate fidelity, which we expect around the end of March. Beyond that, our focus is to deploy a system with more than 150-qubits with an anticipated 99.7% median 2-qubit gate fidelity around end of December 2026. As far as we know, no one has demonstrated systems at that scale and fidelity using chiplet based architecture. In parallel, we'll continue advancing our chiplet architecture as the foundation for scaling toward a system of more than 1,000 qubits with an anticipated 99.8% median 2-qubit gate fidelity by or around the end of 2027. Chiplets are central to our strategy and represent the most practical path to large-scale systems.

We also will continue working to integrate error correction into the stack. Our work with Riverlane demonstrates ongoing progress in this area. From a market perspective, we expect 2026 to remain focused on delivering on-premise systems across government, national labs and academic institutions with select commercial customers engaged in quantum research. Finally, we strengthened our balance sheet.

We exited the year with approximately $590 million in cash, providing us with the flexibility and runway to execute our road map through the quantum advantage time frame. Our investment focus remains organic. We will consider M&A only if it meaningfully accelerates our road map, but we do not need acquisitions to execute our core strategy. To close, quantum computing is a long cycle opportunity.

It requires patience, technical rigor and capital discipline. We are not building for next quarter or next year. We are building for a meaningful durable impact over the next 5 to 10 years. Rigetti's strategy is deliberate.

We focus on speed, scalability and fidelity. We leverage a strong ecosystem. We define success rigorously, and we invest with a long-term view. Thank you for your continued support.

I'll now turn the call over to our CFO, Jeff Bertelsen, for a review of our financial results. Jeff?

Thank you, Subodh, and good afternoon, everyone. I'll spend a few minutes walking through our fourth quarter financial results, our balance sheet and how we are thinking about capital deployment as we continue to execute the road map you've just heard about. For the fourth quarter of 2025, revenue was $1.9 million, compared to $2.3 million in the fourth quarter of 2024. As investors have seen over time, our quarterly revenue profile continues to be influenced by the timing of system deliveries and government contract activity. That dynamic remained true in the fourth quarter, while we saw contributions from our contracts with NQCC and AFSOR (sic) [ AFOSR ], revenue variability at this stage of the market is expected and does not change how we manage the business or allocate capital.

Gross margins for the fourth quarter were 35% compared to 44% in Q4 of last year. Margin performance continues to be driven primarily by contract mix. Certain strategic contracts particularly with government and national lab customers carry lower margin profiles, but play an important role in advancing system validation, ecosystem integration and long-term positioning. Total operating expenses for the fourth quarter were $23.2 million compared to $19.5 million in the same period last year.

Spending remains concentrated in research and development, including engineering headcount, fabrication and system integration. Stock-based compensation was $5.6 million for the quarter compared to $3.4 million a year ago. Operating loss for the fourth quarter was $22.6 million compared to $18.5 million in Q4 2024. Our GAAP net loss for the fourth quarter of 2025 was lower than the GAAP loss for the fourth quarter of 2024, primarily due to the noncash change in the fair value of our derivative warrant and earn-out liabilities. On a non-GAAP basis, net loss was $11.3 million or $0.03 per share compared to a net loss of $14 million or $0.06 per share in the prior year quarter.

I want to briefly address the time line for revenue recognition with respect to the $5.7 million of Novera sales we announced late last year and the $8.4 million C-DAC order we announced in January. Regarding the 2 Novera sales for $5.7 million, we expect a little less than half of that revenue to be recognized in the first quarter with the balance recognized in the second quarter of 2026. Both Novera sales included lower margin dilution refrigeration systems. Therefore, we anticipate significant first quarter year-over-year revenue growth driven by a portion of the $5.7 million Novera on-premises system purchase orders expected to ship in Q1. Importantly, while individual quarters can move around, these contracts support a growing base of recurring and multiperiod activity.

Regarding the C-DAC order, we expect to recognize the revenue from that sale in the second half of 2026, following testing to validate that the system meets its specifications. The C-DAC order announced in January 2026 did not include ongoing maintenance or support. We expect to receive an additional PO for those services later in the year. Turning to the balance sheet.

We ended the year with approximately $590 million in cash, cash equivalents and available for sale investments compared with approximately $217 million at the end of 2024. We continue to operate with no debt. At our current operating profile, we believe our capital position provides sufficient runway to execute against the milestones Subodh outlined, including continued progress on scale, fidelity and system integration. Our approach to capital allocation remains disciplined and deliberate. The majority of our spending is directed towards core R&D activities that directly advance our technology platform.

We are not managing the business around short-term revenue optimization. We are managing it around credible long-term progress towards quantum advantage. We continue to evaluate longer-term fab and R&D capital needs, including the need for dilution refrigeration as qubit counts scale. Any future investment decisions will be driven by capability requirements and evaluated carefully against alternatives, including partnerships or shared infrastructure.

Our currently disclosed road map does not depend on near-term changes to our fab footprint. Our execution path remains primarily organic. We believe we have the available technical depth and internal capabilities required to deliver on our road map. At the same time, we maintain flexibility to evaluate selective opportunities that could accelerate progress in targeted areas, discipline and alignment with our strategy remain the filter.

To close, our financial strategy is straightforward. We are focused on maintaining flexibility, funding innovation responsibly and aligning capital deployment with long-term value creation. While quarterly results will continue to reflect the early stage nature of the market, our balance sheet position us to execute with patience and control. With that, I'll turn it back to the operator who will open the call for your questions.

[Operator Instructions] Our first question comes from Kevin Garrigan with Jefferies.

Thanks for letting me ask a few questions. So I guess just first, on the 108-qubit system, you made some significant progress on the QPU, but what are the key remaining gating items to deliver that as a customer-ready system?

Yes. Thanks for your question, Kevin. So as we said in our press release, we are on track to deploy the 108-qubit system around the end of March, with about 99.5% 2-qubit gate fidelity and 99.9% 1-qubit gate fidelity. As we mentioned in our prior press release, we intentionally delayed because of some interactions between tunable-couplers that happened at that scale, and that's what we are addressing.

We have done that. We feel pretty good about deploying the system here soon. Hopefully, that answers your question.

Yes, it does. And then as a follow-up, as the quantum supply chain kind of -- or just as the quantum industry scales, manufacturing capacity could become a pretty big constraint. And given Fab-1 is a key differentiator for you guys. Would you ever consider offering foundry or manufacturing capacity to others in the quantum computing industry?

Well, actually, we do offer Fab-1 as a foundry to select customers, specifically the DOE, DoD and The U.K. National Government. So these are our customers, they use our systems, they have deployed our systems over there. And as part of the overall technology partnership package, we do allow them to run experiments where we become the foundry.

So we already do that, and we will continue to do those kinds of arrangements with, for select customers who have interest in developing their own chip architecture, chip designs and so on.

Our next question comes from Troy Jensen with Cantor Fitzgerald.

Congrats on all the progress and milestones achieved last year. But maybe just Subodh for you, I just want to make sure I get this correct, there's been a few different numbers kind of quoted in your press release about the single and dual gate fidelity. So when you launched this chip at the end of March, can you just clarify exactly what you think the fidelity levels will be for single and dual mode or dual gate?

So when we deploy this 108-qubit systems towards the end of March, our 1-qubit gate fidelity will continue to be at 99.9%, and our 2-qubit gate fidelity, the median number is expected to be about 99.5%. The reason we started clarifying 1-qubit gate fidelity, frankly, because there are many other quantum computing companies that are confusing everyone by reporting 1-qubit gate fidelity instead of 2-qubit gate fidelity. Historically, as you know, we have always focused on 2-qubit gate fidelity because that's really the most important metric when it comes to entanglement and so on, but some of the other quantum computing companies are routinely reporting 1-qubit gate fidelity and then comparing their 1-qubit gate fidelity our 2-qubit gate fidelity numbers.

So to avoid that confusion we have started reporting both numbers right now, so again, I'll reiterate our 1-qubit gate fidelity has consistently been at 99.9% or better for a few years now. Its the 2-qubit we monitor closely, and that's what will be about 99.5% median when we deployed the 108-qubit system by the end of March.

Awesome. If I could toss in two more quick ones. But 28 nano second gate speed, where were you guys at previously? And then also just an update on DARPA, where you guys stand with that?

So sure. So probably one of the most exciting parts about our press release this afternoon is the achievement of 99.9% 2-qubit gate fidelity, along with 99.9% 1-qubit gate fidelity and 28 nano second gate speed with our proprietary what we call Adiabatic CZ gate. CZ is a standard that many of us use in quantum computing for general purpose quantum computing. And this proprietary version of CZ gate us allows to get this incredible performance.

We really believe this is a great milestone to -- for us to follow because now we know it is possible to get 99.9% 2-qubit gate fidelity with our current designer architecture. So very important milestone that takes us with the confidence that we will be able to deliver a 1,000 plus qubit system in a couple of years with 99.8% or that kind of 2-qubit gate fidelity. So we're really proud of that accomplishment. Regarding DARPA, specifically, we continue to work with them.

We are confident that we'll get into Phase B. As we have discussed in the past, its an open-ended program. Once we reach certain milestones, they will get us into Phase B. They have given us a list of things that we have to address, mostly related to error corrections and a few other things.

And we are working on them as we speak. So we feel pretty good that we should be in Phase B by the end of this year or thereabouts.

Next question comes from Quinn Bolton with Needham & Company.

I guess just wanted to come back on the 108 qubit system. Obviously, the end of March is just a few weeks away. Are you guys already at the 99.5%, meeting 2-qubit gate fidelity in the lab and you're just sort of going through the process of getting the system online? Or is there still work to do on chips -- fabbing chips and tuning the process to get to that 99.5% 2-qubit gate fidelities?

Quinn, It's a little more complicated than that to give a simple answer like that, partly because when we bring up a new system, there's a lot of qubits obviously, 108 qubits is a lot of qubits and we bring different parts of the grid up, look at different edges and internal parts of the grid. So yes, there are multiple areas where we already are at 99.5% or better, but obviously, the whole grid is not at 99.5% median. Otherwise, we would have deployed it right away. So we are -- we did a chip redesign to address the coupling issues.

We are collecting data, verifying that all the data is consistent. So when we deploy, we will be confident that this is the right system to deploy. Just a little context. 108-qubit system at that level of fidelity and that gate speed, about 50, 60 nano-second is a really good system. I mean when we look at the overall industry right now, as far as we can tell, the only one who has anything in that league or better would be IBM at 120 qubits with their tunable coupler.

They used to have 156-qubit and 6 coupler, but they moved to tunable coupler and they're at 120. Everyone else is much more than that. And certainly, when you see announcements from companies like trapped ion or pure atom companies, as far as we know, nobody has even approached 100 qubits yet. There's a lot of press releases that go out, but when we go and see actual deployments, from any of these companies, no one is in that range. So we will be only the second company as far as I can tell, to reach 108 qubits deployed on a cloud.

Just wanted to put that perspective in place.

I appreciate that. The second question is for Jeff. Jeff, you gave us some sense that the gross margin on the 2 Novera sales in the $5.7 million purchase orders, we're going to be carrying lower gross margins because of dilution refrigerators. Can you give us any sense what level of gross margin would you expect on the $5.7 million?

And then I guess a similar question on the C-DAC 108-qubit system. What type of gross margin would you expect on that system? Is that also low because of the dilution refrigerator? Or is that expected to be a higher gross margin sale?

Yes. I guess the way I would answer that one is, I don't know that we want to get into quoting exactly what the gross margins are for, competitive reasons and whatnot. Our typical Novera systems without the dilution refrigeration have very high margins, definitely higher. With the dilution refrigeration, it's a resold item, you really can't mark that up.

So they are going to be lower than maybe what we would see with some of our other Novera sales. And regarding the C-DAC system, again, I think for competitive reasons, we won't comment on the gross margin profile specifically. I mean, it is a very important strategic account for us, and we're happy to have it. And it will definitely contribute to our sales growth next year, but don't want to get into the margin specifics.

Got it. And then just a final clarification on the C-DAC order. It sounds like, does the entire $8.4 million rev rec once validation testing has been complete? Or is there a possibility that the $8.4 million could be rev rec'd over a couple of quarters in the back half of the year?

No, it won't be spread over time. It will be rev rec'd all at once at a point in time once we've -- once it's been installed and we're able to demonstrate that it's meeting its specs. So it will be more like a traditional system hardware sale as opposed to rev rec over time.

Our next question comes from David Williams with StoneX.

I guess maybe first, if you think about what you're doing with NVIDIA on the NVLink there. Can you talk maybe a little bit about the progress and anything that's maybe developed over the last quarter in that regard?

Sure. Thanks for the question, David. So our view is that quantum computing is not going to exist in a silo. It will be part of a hybrid ecosystem.

So as we have said multiple times and even in this press release, we believe that CPUs will continue to be in data centers doing sequential computing, addition, subtraction, that kind of stuff. GPUs will continue to be in data centers, doing parallel computing. What quantum computing will take over is simultaneous computing part that is currently handled by GPUs. So effectively quantum computing becomes an accelerator to a GPU for select applications where you have simultaneous computing.

That view is consistent with NVIDIA and some other companies, that's where we are partnered with NVIDIA on the NVQLink. A critical part where we think it gives us huge advantage to be doing hybrid computing with superconducting gate-based quantum computing, which is what we do, is it gets to the speedy area. Because we are dealing with tens of nanoseconds in gate speeds, as you can see, our standard product is in the 50, 60 nanosecond and with this new gate that we announced, we are talking about sub-30-nanospeed gate speed. That commensurate with CPU and GPU gate speeds, and that really allows a practical hybrid quantum ecosystem to evolve.

When you compare that with some other modalities like trapped ion or pure atom, they're talking about hundreds of microseconds. In fact, if I recall this correctly, the most recent number from IonQ is 600 microseconds. So that's 30,000x slower than where we are. Just to repeat, I mean, we are talking tens of thousands of times lower speed with trapped ion type modalities.

And that creates a significant challenge for them to talk about a hybrid quantum ecosystem. So that's a huge advantage superconducting gate-based quantum systems have, where we have tremendous gate speeds commensurate with CPU and GPU allows us to do that kind of stuff. So that's why we have partnered with NVIDIA, we demonstrated at GTC in October last year, how a concept would look like. And we'll continue to do that with them.

They are not only a company from the HPC environment that shares this view. Other HPC builders are also sharing similar views. So you are going to see more and more companies talk about hybrid quantum computing environment with HPC and quantum computers, particularly superconducting gate-based quantum computers coexisting together. Hopefully, that answers your question.

Yes, it certainly does. And then maybe secondly, just kind of looking at the landscape for M&A or acquisitions, it seems like there's a fairly ripe environment of different enabling type technologies that are out there. So maybe just discuss the landscape, how you see it, and if there's areas of the stack where you could benefit maybe that could help accelerate your roadmap?

Yes. As we mentioned, we would certainly be open to M&A if it helps accelerate our roadmap. Our roadmap, again, to repeat, we are talking about more than 1,000 qubit system at sub-50 nanosecond gate speed and 99.8% median 2-qubit gate fidelity in a couple of years. As far as we can tell, that's a very impressive system that we may be one of the only ones, if not the only ones to be able to get a system to that level of performance.

So to get a system there and to try to find accelerating points where we can acquire someone to help us accelerate that roadmap, at least we haven't seen what exactly is out there that would help us accelerate that roadmap. Right now, clearly, our chiplet strategy is a critical component of us getting to 1,000 qubits. And there we are the pioneers. We have the IP.

We have the know-how. As far as we can tell, we are the only ones who are practicing chiplets in real life. So really, no one else can help us to accelerate that one. When it comes to the other components of the stack on the control system, as we have already disclosed, we are partnered closely with Quanta Computer in Taiwan, and they are a top player in CPU, GPU servers in the cloud, and they fully understand that control system part of the stack.

So we feel pretty good about who we have partnered with. I've already mentioned NVIDIA for NVQLink and the distribution layer software like CUDA Quantum. Other areas, we have Riverlane for error correction, QphoX for optical signaling, those kinds of partnerships. So we will continue to monitor the situation.

And if we believe that someone can help accelerate our roadmap faster, we are certainly open, but at this point, our roadmap clearly is dependent on just executing our plan. So that's what we have said that most of our plan is organic right now, not contingent on M&A.

Our next question comes from Krish Sankar with TD Cowen.

I have two of them. So first one, if I remember right, I think by end of next quarter or so is your time line for making the decision of maybe building another fab or outsourcing it. So I'm kind of curious where we are in that and have the recent acquisition by your competitors kind of changed that decision-making thought process?

Well, we already have our own fab that is existing in Fremont, that's what's producing our wafers every day right now. Regarding the need for a new fab, we have mentioned that there may be a potential possibility that we may have to invest in a new fab, but we have been very clear, Krish, that we do not believe we need a new fab to get to quantum advantage. So we are -- certainly, for the next 3 years, we think our existing fab is capable of getting us to quantum advantage, which is 1000 qubit 99.9% type fidelity. The fact that we just disclosed a 99.9% 2-qubit gate fidelity performance with our existing fab is a proof point that our existing fab is clearly capable of taking us there.

Regarding our competitors buying CMOS foundry for a quantum fab, we are not quite sure why they did that when they had already -- I believe you're referring to IonQ buying SkyWater. We are not quite sure why they did that because as far as we knew, IonQ had already invested in a separate fab in Washington State 3 years ago, and them purchasing SkyWater, we are not quite sure and SkyWater's primary business, most of their business is CMOS foundry obviously. So we are not quite sure what exactly the rationale was, you need to talk to them. But we certainly do not believe that we need to be using any other fab except our fab for the near term.

Longer term, obviously, we have said there is potentially a need for a fab to -- there are multiple initiatives being looked at right now, including there are foundry options out there, and we'll certainly take a look at foundry options, other initiatives that are being looked at and decide what is the next step. But again, to repeat, we feel pretty good about our existing fab and feel confident that it will take us there to quantum advantage, which is about 3 years from now.

Got it. Very helpful. And so just as a quick follow-up, kind of curious on the government funding thing. Last year, we had DOE announced new funding. Have you seen any activity from them? There's also any latest thoughts on the U.S. NQIRA? And any of the sovereign initiatives that you're seeing that could benefit Rigetti?

Well, certainly, there's a lot of initiatives being discussed at the U.S. government level to support quantum computing. But as we can all see, there is no bill that has been signed and appropriated yet. There seems to be bipartisan support for this NQI Reauthorization Act different, both the House and Senate, there seems to be in support, but it hasn't yet led to a bill that is signed and appropriate, which we are all eagerly awaiting for. It looks imminent.

Everything suggests that such a bill should be signed here soon, and that will significantly help companies like us, but along with us other companies that play in this ecosystem too. So we certainly are supportive of those kinds of initiatives and hopefully, they happen. DoD funding continues. As you can see, we clearly are already getting funded from places like Air Force Research Lab part of DoD, and we'll continue to look at other opportunities with DoD and other areas of the government.

We certainly are a critical part of the U.K. government's initiative in quantum computing and there are multiple new initiatives being discussed by the U.K. government right now, and we certainly will take a look at those kinds of opportunities. And of course, we announced that, we are the first company that the Indian government has really chosen to get their quantum computer installed. We are really proud of that accomplishment. So when the first quantum computer is procured by the government of India, we believe it will be ours before the end of this year.

So we feel pretty good about being closely affiliated with U.S., U.K., and now the Indian government, and we'll continue to monitor different initiatives and funding activities going on.

Our next question comes from Craig Ellis with B. Riley Securities.

Subodh, in your prepared comments and in the press release, there was a note of Novera QPU sale to a Japanese research entity. I'm wondering if you could just tell us a little bit more about that. And then use that to elaborate on what the pipeline is looking like as you engage more broadly with other international entities?

Yes, sure. Thanks, Craig. So as we disclosed, we did get an order for a 9-qubit Novera from a Japanese research organization. We always ask their permission if we can disclose their name.

In this case, they didn't want to for confidentiality reasons on their side. But it is a premium -- premier Japanese organization. And once they give us permission to disclose, we'll be happy to disclose. Overall, we feel pretty good about interest increasing to get on-premises quantum computing.

As we have already disclosed, we have 2 upgradable 9 qubit systems that we are going to deliver in the first half of this year and on 108-qubit system to the government of India in the second half of this year. We have disclosed this Novera order and there a few more Novera potential orders here in the pipeline, and we'll disclose them when we get them. And if they give us permission to disclose their name maybe we will obviously do that. We are certainly talking to different government entities within U.S., U.K., India and some other countries too.

And we believe the demand for on-premises system will continue to grow. As you can see, just by adding the numbers that we have disclosed, we want going to see significant year-over-year increase in sales this year by just delivering the systems that we have already received orders for, and we'll continue to see that in the future. We believe that as we get closer to quantum advantage, which is about 3 years from now, you are going to see significant spike in interest as we get closer to that milestone. And that makes sense because that's really when people start seeing practical benefits with quantum computing, and you will definitely see more and more commercial authorization starting to show interest in on-premises quantum computing.

Hopefully, that answers your question, Craig.

Yes, that's very helpful. And it's nice to see the revenue momentum on the Japanese research sales, Jeff, are you expecting those to rev rec this year? And if so, can you give us a sense for whether that would be in the first half of the year or the second half of the year?

Yes. The Novera to the Japanese organization, we expect that to ship in April and to rev rec in Q2.

Our next question comes from Richard Shannon with Craig Hallum Capital Group.

This is Tyler on for Richard. To the Japanese organization that purchased your system, did they already have multiple [ dilution ] fridges installed? And are they just testing different components in the stack or different combinations of components in the stack? And I have one more follow-up.

Well, certainly, we know they have 1 dilution refrigerator because they have ordered core Novera QPU and not a whole 9-qubit upgradeable system. We are not sure of what other modalities they have tested and within superconducting, if they have looked at any competitive solutions. As you probably know, Tyler, IBM really doesn't offer something like a 9-qubit system, Google doesn't offer a on-premises qubit quantum computing system. There is an organization -- a couple of organizations in Europe like IQM and QuantWare, they can offer smaller qubit count systems.

But frankly, our performance is significantly better than those kinds of competitors. So we believe the Japanese organization did their homework and decided superconducting quantum computing is an area they want to invest in. And within superconducting to get started, Novera is a perfect solution to get your researchers familiar with quantum computing and starting to learn about quantum computing ecosystem and work on algorithms and applications and stuff like that.

Got it. And you had mentioned you're doing work with -- sorry, I'm at the airport. You're doing work with Riverlane on the scalability of error correction. Could you just elaborate on what that means?

Sure. What we have disclosed is that we have partnered with Riverlane based in Cambridge, U.K. It's a company of probably our size, about 150 employees, excellent quality work they do in error correction software, and we have published some papers that anyone can take a look at. We showed some concepts of how real time error correction will work.

We have shown a path to how their error correction hardware will closely integrate with our hardware. So they will be a core part of our stack, if you will, and how that will scale up as we go up to 100 and then 1,000 qubits and beyond 10,000 qubits. So our roadmaps are well aligned. We work very closely with their team.

So effectively, we view the error correction as a key part of our stack going forward. Hopefully, that answers your question.

Yes. Is there just any update on like number of qubits for one of their systems? Is there any change in that?

Not in that sense. I mean number of qubits and the raw fidelity, obviously, that all those things come from us. Where they start coming in, is error mitigation and error correction area. Obviously, that's the layer that is critical when you start talking about quantum advantage.

So really, the benefit that our customers are going to see with Riverlane error correction is when we start approaching quantum advantage. Right now, they can test it. Riverlane will offer them their services. They have physical products to ship, if you will, and we can demonstrate that our systems are closely working well together.

But clearly, we are not at a point of 1,000 qubits at 99.9% type fidelity where error correction can really be demonstrated to show practical benefits. So we need to get closer to quantum advantage before end users will start seeing the real value of error mitigation and error correction.

Our next question comes from John McPeake with Rosenblatt Securities.

Thanks, that's on getting to the error rate that you need to get to by the end of the quarter. Question on the 150-plus qubit system that you guys had originally planned for by the end of this year at 99.7% 2-qubit gate fidelity. Is that still on cards?

Yes, absolutely. That is our roadmap. We will deploy this 108-qubit system soon and then plan is to get to 150-plus qubit around the end of this year. We want to be careful.

I mean, any time you say by the end of year, it always creates a challenge for December 31st and so on. I mean, these are extremely complex systems. These are extremely complicated technologies that we are developing. So what we have said is more than 150 qubit, about 99.7% median 2-qubit gate fidelity around the end of this year.

That's our next milestone. The bigger one that we are really excited about, and we are focusing very much on that is more than 1,000 qubits, closer to 99.8% median 2-qubit gate fidelity, less than 50 nanosecond gate speed in about a couple of years. So that's where a lot of our work already has started. So we'll certainly get 150 plus delivered around the end of this year, but most of the effort right now has already started on the 1,000 plus qubit in a couple of years.

And as I mentioned earlier, I mean, even 150 right now, there is no one our with 150-plus qubit system with the exception of IBM's old technology, where they had 156 qubit and certainly 1,000 is going to be a big milestone for the whole industry, certainly for us, but for the whole industry. We don't see how other modalities, even though they claim they will be there, you look around trapped ion or pure atoms or any other modalities, none of them, as far as I can see, have reached even 100 qubits or even anywhere close to that. So when they have the roadmaps talking about getting to 1,000 and tens of thousands, it's a roadmap.

In the case because we are using chiplet technology and because we have semiconductor fabrication, and we know how to stack up the chips, we feel pretty good about our executability of our roadmap. So yes, to answer your question, absolutely, 150 plus around the end of this year and more than 1,000 around the end of next year. That's the plan.

Great. So the issues you're resolving around the 108, are those critical to reaching these roadmap milestones? Is that -- is that the way I should think about it like once you resolve this tuneable coupler?

Yes, absolutely. I mean, the issue that we are resolving as we speak with the tuneable coupler and that's why we did the chip reiteration, it's critical to not only 108, but 150 and everything beyond that. So every time we advance to a certain level, we take advantage of all that when we go with the next systems. That's why, I mean, when we hear some companies talk about how they are going to get to 1 million qubits without demonstrating even 10 qubits, we are very skeptical of those kinds of claims.

And frankly, I mean, even if you look at some companies trapped ion companies like IonQ, who acquired Oxford Ionics technology at a couple of qubits. And that's where they are right now, at a couple of qubits and to certainly say we'll be at 1 million qubits next year. We remain skeptical of those kinds of claims. I mean, unless you can demonstrate 10 and then 100, we just don't see how you can go from 2 qubits or 5 qubits to certainly a couple of million qubits in a year or 2 years.

Okay. And then just last one, if I could. The 1,000 qubit machine by the end of 2027. Can you give us some kind of sense as to how many logical you might be able to squeeze out of that? I don't know if you're thinking about the Riverlane error correction solution or some other error correction? How should I think about that? Because it's a great physical number, certainly 99.8% on that one.

Yes. So really to start talking logical qubits, you need to get to closer to 99.9%, which is the quantum advantage point that I've talked about. In general, in the superconducting gate-based quantum computing technology space, we are in, typically, the number that we use is roughly 10 to 50 physical to logical qubit, depending on the exact fidelity and stuff like that. Maybe 10 to 100 in the worst case.

But that's the ratio. So you have to divide that number by 10 to 100, to get the number of logical qubits. Once you approach 99.9% level. Now, I know that there's a confusion going around right now in the industry because, again, some trapped ion and pure atom companies have started reporting physical to logical qubit ratio of 2:1 and in some aggressive cases, 1:1, I just want to caution you that they are using a very weird definition of logical qubit in that case.

They are not talking about the logical qubit as a perfect qubit. They're talking about logical qubit having a fidelity, and in some cases, their logical qubit fidelity is actually lower than the physical cubic gate fidelity. So it doesn't make any sense that they are using that language, but they are and that's unfortunately confusing a lot of people as to logical qubit and what does it mean and stuff like that. Hopefully, that answers your question, or maybe I confused you some more.

Our next question comes from Brian Kinstlinger with Alliance Global Partners.

Well, my question was kind of answered. But I guess I'm curious, Subodh, if you can touch on or elaborate on either in Japan or India your customers, what the evaluation process was like. I think you said in Japan, maybe you weren't sure of the other modalities. But what was the competitive landscape in time?

Well, certainly, almost every national lab, university or any potential -- even commercial customers, they are fully aware of different modalities, the pros and cons. And once they decide to choose superconducting gate-based modality, and usually, the reason is for the obvious things that we have been saying, which are scalability and gate speed. Those are the huge benefits with superconducting modality. Everyone understands that fidelity is the main challenge in superconducting modality.

So usually, that part, most customers that we talk to have done on their own. Usually when we talk to them, they have already gone through that process, but then they are starting to look at different competitors within superconducting gate modality. Obviously, IBM is always there in the fray. Sometimes we deal with companies like IQM from Finland or QuantWare from Holland or some other companies around the world.

I mean, the main thing that differentiates Rigetti is our open modular architecture. We have this innovative way of incorporating third-party solutions that allow us to come up with innovative solutions faster. Examples being Quanta Computer for control system or NVIDIA for NVQLink or distribution layer software like CUDA Quantum or Riverlane for error correction. Most customers like that kind of an open approach because usually they have some ideas on what other things they would like to try with quantum computing because they're also doing research at this point.

And the other part where we really outshine our competitors is chiplet. Everyone sees that chiplet is a very, very possible way to scale up long term. We allow our customers to upgrade as you can clearly see our current 2 orders that we are fulfilling are upgradable 9-qubit systems. So once they get 9-qubit up and running, we fully expect them to ask us to upgrade them to up to 108-qubit, sometime next year because of the same dilution refrigerator with some changes in cables and connectivity will be able to handle 108 qubits or even more in the future.

So main differentiator, the reason they choose us are our open modular approach and our chiplet approach and a few other things, but those are the customers that end up choosing Rigetti. Hopefully, that answers your question.

I would now like to turn the call back over to Dr. Subodh Kulkarni for any closing remarks.

Thank you for the thoughtful discussion today. We are excited about the momentum we are building across our technology roadmap, our partnerships and growing engagement from customers around the world. Our team is focused on executing with discipline and delivering systems that enable meaningful progress in quantum and hybrid computing. We appreciate your continued interest and look forward to sharing our progress in the quarters ahead.

Thank you.

Thank you. This concludes the conference. Thank you for your participation. You may now disconnect.