Rigetti Computing - Earnings Call - Q2 2025

Executive Summary

• Q2 2025 revenue was $1.80M, down sharply year over year (Q2 2024: $3.09M) and up sequentially (Q1 2025: $1.47M), with gross margin at 31% vs. 64% in Q2 2024; diluted EPS was $(0.13), and net loss was $(39.7)M, including $(22.8)M non-cash fair value losses on warrants and earn-outs.
• Announced general availability of Cepheus-1-36Q, the industry’s largest multi‑chip quantum computer, achieving 99.5% median two‑qubit gate fidelity and a 2x reduction in error rates vs. Ankaa‑3; 100+ qubit chiplet system at 99.5% median fidelity targeted before year‑end 2025.
• Balance sheet strengthened via $350M gross proceeds through the ATM program; quarter‑end cash, cash equivalents and AFS investments were ~$571.6M with no debt, positioning Rigetti to fund R&D milestones and scaleup.
• Demand headwinds tied to U.S. National Quantum Initiative (NQI) expiration and pending reauthorization; mix with UK NQCC development contracts weighed on gross margins; OpEx up on R&D and personnel costs.
• Key near‑term catalysts: continued tech execution (gate fidelity, qubit scaling), NQI reauthorization progress, and DARPA QBI phase selections; management reiterated 3‑4 year roadmap to quantum advantage under superconducting modality.

What Went Well and What Went Wrong

What Went Well

• Achieved general availability of Cepheus‑1‑36Q (four chiplets), halving two‑qubit gate error rates vs. Ankaa‑3 and reaching 99.5% median two‑qubit fidelity: “Just 6 months after our record performance with Ankaa‑3, we’ve once again halved our error rates with Cepheus‑1‑36Q”.
• Strong liquidity and no debt: completed $350M ATM equity raise; cash, cash equivalents and AFS investments of ~$571.6M at quarter‑end support R&D and commercialization.
• Clear scaling roadmap and confidence in year‑end milestone: “We expect to release a 100+ qubit chiplet‑based system at 99.5% median two‑qubit gate fidelity before the end of 2025”.

What Went Wrong

• Revenue declined y/y and margins compressed: revenue $1.80M vs. $3.09M y/y; gross margin 31% vs. 64% y/y due to mix and contract pricing variability (e.g., UK NQCC contracts).
• Larger net loss driven by non‑cash liabilities: net loss $(39.7)M includes $(22.8)M unfavorable fair value changes in warrant and earn‑out liabilities; prior year quarter had $3.4M favorable impact.
• Macro/government funding uncertainty: CFO cited NQI expiration and pending reauthorization as a revenue headwind; timing of DOE/DOD programs remains a swing factor.

Transcript

Speaker 5

Day and thank you for standing by. Welcome to the Rigetti Computing Second Quarter 2025 Financial Results Conference Call. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question and answer session. To ask a question during the session, you will need to press *11 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press *11 again. Please be advised that today's conference is being recorded. I would now like to hand the conference over to our speaker today, Dr. Subodh Kulkarni, Chief Executive Officer. Please go ahead.

Speaker 2

Good afternoon, and thank you for participating in Rigetti Computing's earnings conference call covering the second quarter ended June 30, 2025. Joining me today is Jeff Bertelsen, our CFO, who will review our results in some detail following my overview. Our CTO, David Rivas, is also here to participate in the Q&A session. We will be pleased to answer your questions at the conclusion of our remarks. We would like to point out that this call and Rigetti Computing's second quarter ended June 30, 2025 press release contain forward-looking statements regarding 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 described and are discussed in more detail in our Form 10-K for the year ended December 31, 2024, our Form 10-Q for the three and six months ended June 30, 2025, and other documents filed by the company from time to time with the Securities and Exchange Commission. These filings identify and address important risks and uncertainties that could cause actual events and results to differ materially from those contained in the forward-looking statements. We urge you to review these discussions of risk factors. Today, I am pleased to report that we continue to achieve our ambitious roadmap goals and maintain our momentum on the technology front, most recently by demonstrating the industry's largest multi-chip quantum computer with impressive performance.

Our multi-chip quantum computer, CFIUS-1 36Q, the industry's largest multi-chip quantum computer, is released for general availability and deployed on the Rigetti Quantum Cloud Services platform, QCS, and will be available on Microsoft Azure thereafter. Just six months after our record performance with ANKA-3, we have once again hugged our error rates. With a median 2-qubit gate of 99.5%, CFIUS-1 36Q has achieved a 2x reduction in 2-qubit gate error rate from our previous ANKA-3 system. CFIUS-1 36Q is the first multi-chip quantum computer in the industry to achieve this level of performance. With four chips, CFIUS-1 36Q contains the largest number of chiplets in a quantum computer and further validates our approach to scaling Rigetti Computing's quantum computers.

It's our view that superconducting qubits are the leading modality for quantum computers due to their ability to scale and their ability to achieve gate speeds more than 1,000 times faster than other modalities like ion traps and pure atoms. Our superconducting qubits leverage technologies like chiplets that have been maturing in the semiconductor industry for decades. Use of these well-established methods enables Rigetti to scale its quantum computers to higher levels of performance and qubit counts. This legacy of technology curve advancement continues with the CFIUS-1 36Q architecture and includes the following features that contribute to improved performance. Transitioning from a monolithic chip to chiplets enables greater control over chip uniformity, which in turn improves performance. Leveraging chiplets also reduces manufacturing complexity and improves fabrication yield. Optimized 2-qubit gates enable faster gate times while reducing coherent errors, which improves fidelity and is important for executing quantum error correction techniques.

These improvements enable a 2x reduction in error rates. Advances in multi-layer chip and tunable coupler design also enable higher performance. Our industry-leading proprietary chiplet approach to scaling makes us confident that we will hit our end-of-year technology goals. We believe quadrupling our chiplet count and significantly decreasing error rates is a clear path towards quantum advantage and fault tolerance. We intend to continue our momentum and expect to release a 100-plus qubit chiplet-based system at 99.5% median 2-qubit gate fidelity before the end of 2025. While we are pleased with our sequential growth in quarterly revenues, we believe the achievement of our technology milestones remains the key metric to achieving our long-term success. On the financing front, I'm pleased to report that Rigetti has significantly strengthened its balance sheet.

During the second quarter of 2025, Rigetti completed the sales of $350 million gross proceeds of its common stock, pursuant to our previously disclosed at-the-market equity offering program. We are well positioned to support commercial scale-up of our superconducting gate-based quantum computers. Thank you. Jeff will now make a few remarks regarding our recent financial performance.

Speaker 0

Thanks, Subodh. Revenues in the second quarter of 2025 were $1.8 million, compared to $3.1 million in the second quarter of 2024. On a year-over-year basis, our revenue for the quarter was impacted by expiration of the U.S. National Quantum Initiative and its pending reauthorization in the U.S. Congress. Renewal of the U.S. National Quantum Initiative, sales to U.S.-informed governments, and Novera are all important to future sales. Gross margins in the second quarter of 2025 came in at 31%, compared to 64% in the second quarter of 2024. The lower gross margins on a year-over-year basis were impacted by revenue mix and variability in the pricing in terms of our development contracts, including our contracts with the UK's National Quantum Computing Centre for quantum systems, which have lower gross margins than most of our other revenue.

On the expense side, total OpEx in the second quarter of 2025 was $20.4 million, compared to $18.1 million in the same period of the prior year. The increase in total OpEx was due to annual salary increases, new hires, and higher consulting costs, mainly in research and development. Higher costs for our annual shareholder meeting due to the increase in the number of beneficial owners of our stock also contributed to the increase. Stock compensation expense for the second quarter of 2025 was $3.6 million, compared with $3.3 million for the second quarter of 2024. Our operating loss for the second quarter of 2025 came in at $19.9 million, compared to $16.1 million in the prior year period. We recorded a $39.7 million net loss for the second quarter of 2025, compared to a net loss of $12.4 million for the second quarter of 2024.

Our net loss for the second quarter of 2025 includes non-cash charges for the change in the fair value of our derivative warrant and earnout liabilities, which had a $22.8 million unfavorable impact on our net loss for the quarter. Derivative warrant and earnout liabilities had a $3.4 million favorable impact on our net loss for the second quarter of 2024. As of June 30, 2025, we had approximately $571.6 million of cash, cash equivalents, and available for sale investments and no debt. Thank you. We would now be happy to answer your questions.

Speaker 5

As a reminder, to ask a question, please press *11 on your telephone and wait for your name to be announced. To withdraw your question, please press *11 again. Please stand by while we compile the Q&A roster. Our first question comes from Troy Donavon Jensen with Cantor Fitzgerald. Your line is open.

Hey, gentlemen. First off, congrats on all the great traction here.

Thanks, Troy.

It's a bit for you maybe to start off with just use of proceeds. You've got a ton of money on the balance sheet now. I mean, is the intention to accelerate R&D, do a little M&A, or just kind of a cushion on the balance sheet to fund the operating losses?

Speaker 2

Our focus, Troy, continues to be on R&D development. We will obviously look at every opportunity to accelerate our timeline. Right now, we believe we are funding R&D adequately to hit the milestones that we have laid out. As you saw, we demonstrated a 4x9 qubit multi-chip system. We are deploying it as we speak. Our plan for the end of the year is to deliver a multi-chip 100-plus qubit system with 99.5% 2-qubit gate fidelity. From there on, to continue to increase the fidelity as well as qubit count using chiplet approach. Every opportunity we get to accelerate that timeline, we continue to look at it, and we will do so.

At this point, we believe we are still about three to four years away from getting to the 1,000-plus qubit, 99.9% fidelity with error correction and gate speeds of less than 50 nanoseconds, which is when we achieve quantum advantage. If we can accelerate that timeline using our strengthened balance sheet, as you correctly pointed out, we will obviously look at that. I believe right now we are still looking at roughly about four years to get to that quantum advantage point. Hope that answers your question.

Yeah, it does, very much so. Just with respect to OpEx, we'd assume just kind of sequential growth going forward, but no big, no big stepping, no big leaps in spending.

Speaker 0

Yeah, Troy, I think that's a good summary for right now anyway. As Dr. Subodh Kulkarni said, we're adequately funded in R&D, but we'll look for opportunities. Right now, I don't think we anticipate any significant uplift.

Okay, perfect. Maybe just one follow-up. Could you just give us an update on Quanta Computer, you know, what they're doing, what we can't see, you know, to kind of satisfy their commitment here with your investment or their investment in Rigetti?

Speaker 2

Sure. As we have disclosed in the past, Quanta Computer is a very strategic partner for us on the hardware side outside the QPU area. We continue to stay focused on the QPU side. Quanta Computer is investing right now on the non-QPU portion of the hardware stack. That primarily means control system and the rest of the hardware stack. Right now, their focus is to essentially come up to speed in control systems. Our goal is to get them up and running with control systems that work with our QPUs fairly soon, here in the next few quarters. Once they're up to speed in quantum computing and control systems, they will obviously accelerate development of that, allowing us more focus on the QPU side. They continue to be a very good and very strategic partner for us. The partnership is going really well.

We are excited to co-develop the quantum systems, our quantum systems with them. Hopefully, that answers your question.

Yeah, no, that's great. Guys, keep up the great work.

Thank you.

Speaker 5

Thank you. Our next question comes from David Williams with The Benchmark Company. Your line is open.

Hey, good afternoon, gentlemen. Congrats on meeting the targets on the CFIUS-1 36Q chip. That's impressive. I guess maybe, Dr. Subodh, the last time we spoke, you had confidence that you could get to this 99.5%, but said that you had a little bit of work to do, and you've clearly hit that here. How confident are you in being able to parlay that on the 100-qubit chip? Are there any major steps or challenges ahead of you in order to get that 100-qubit at the same fidelity?

Speaker 2

Thanks, David, for the question. Certainly, getting to four chiplets with 9 qubits to the 36-qubit level was a significant accomplishment this last quarter, and we are really happy to get that. Regarding your question about 100-plus qubit, we are confident we'll get there with 99.5% 2-qubit gate fidelity before the end of this year. The beauty of the chiplet approach is once the fundamental architecture is defined and the performance is there, scaling up becomes a lot easier by definition. That is the whole reason for the chiplet approach. You intrinsically are using the same 9-qubit chip multiple times, and that gets you better uniformity on your wafers. You get better yields, and it really allows us to get a perfect 9-qubit chip and then replicate it multiple times, which is why the semiconductor industry uses chiplets with the CMOS technology right now for all your advanced applications.

All the reasons that help semiconductor CMOS industry with chiplets are the same reasons why we chose the chiplet approach. Now that we have proven that it works at this high fidelity, our confidence is fairly high that we will get to 100-plus qubit and beyond, frankly. We really need to get to 1,000 qubit and multi-thousand qubits here soon to get to that quantum advantage point and then fault tolerant quantum computing beyond that. Our confidence is fairly high, but obviously, this is still technology development, and challenges will always be there. We are not taking it for granted by any means, and we'll continue to work hard to get it there. Hopefully, that answers your question.

No, it absolutely does. I guess the follow-up to that would be, do you think that your roadmap can be accelerated beyond? I know you've talked about three to four years, but it seems like you're making just such great progress on the scalability side that you might be able to accelerate that, even though maybe the error correction is lacking. Do you think you'll hit one of your targets maybe on the qubit side before you get to the others that you talked about that's three to four years out? Thanks.

We will certainly try to accelerate our timeline from that four-year to quantum advantage. Having said that, chiplets certainly help us quite a bit to achieving that milestone. At the same time, there are other important metrics as well. I mean, we talked about getting to 1,000 qubit for that quantum advantage or more, getting to 99.9% or better 2-qubit gate fidelity. Error correction, as you correctly pointed out, has to be there too. Also, improving the gate speeds to better than 50 nanoseconds, faster than that. There are other challenges in the dilution refrigerator. There's a lot of cables that we use right now. Right now, we are still using primarily coax cables. When you get to 1,000 qubit or higher, your density of cables and other components in the dilution refrigerator becomes quite intense.

You have to start looking at things like flex cable technology and other things that we will encounter. I don't want to make it sound simply that just because chiplets have been demonstrated, it's a relatively easy path, and we will be able to accelerate the timeline from what we have already told you. We'll certainly look at opportunities, but there are multiple dimensions we need to tackle, and that's where the number roughly four years comes from. Our view is that four years is probably the fastest any of us in the quantum computing space can get to quantum advantage. We've already quantified what our view is. To get to quantum advantage, you need a minimum of 1,000 qubits. You need a minimum of 99.9% 2-qubit gate fidelity. You need to be faster than 50 nanosecond gate speed, and you need error correction.

None of us doing gate-based quantum computing are there yet by any means, and that's where the four years comes from. Even though you may have heard about some companies talking about quantum advantage now and very soon, our view is that it's going to take time to hit those four things. Certainly, for some modalities like trapped ion and pure atoms, where they have fundamental science challenges to improve their gate speeds to get to these tens of nanoseconds. Right now, they are dealing with hundreds of microseconds, and they have some serious scientific fundamental inventions needed situation on hand to get to the gate speeds that you need to get to to demonstrate practical quantum advantage.

You continue to look at opportunities to accelerate, and we hope we find them, but a realistic timeline is what we are using for quantum advantage in the order of the four years with those four things that I mentioned earlier. Hopefully, that answers your question.

Yes, fantastic. Thanks so much for the color, and you've certainly done a good job hitting your milestones so far. We will certainly be looking for that acceleration. I appreciate it. Thank you.

Thanks, David.

Speaker 5

Thank you. Our next question comes from Krish Sankar with TD Cowen. Your line is open.

Hi, this is TD calling on behalf of Krish. Thanks so much for taking my questions. Subodh, if I could start first, I want to explore the M&A related question again. I guess, you know, just with the stronger balance sheet that you guys have now, I just want to get your view on kind of current valuations on quantum assets currently, and is M&A an important part of the growth story over the next year or two? Specifically, just asking related to more adjacent technologies, whether it's semiconductor manufacturing, advanced packaging, or software-related capabilities.

Speaker 2

Thanks, Yun. We will continue to look at opportunities where M&A could help us with our timeline. Our view is that we are very much in technology development right now. The timeline that we have laid out of four years to quantum advantage is primarily within our control right now. If we find opportunities in M&A where we can accelerate our timeline, we will certainly look at that. As of today, we don't see anything out there that can help us. We are in the leadership camp right now when it comes to overall quantum computing performance. There's probably a couple of tech giants that have one or two critical metrics that are ahead of us. Frankly, those tech giants are out of our league to consider M&A. Besides those kinds of opportunities, we really don't see anyone out there who's anywhere close to us.

We are quite a bit ahead of everyone when it comes to technology right now, except for a couple of tech giants in a couple of key metrics. We don't see any practical opportunity to use M&A to help us with our timeline acceleration. We'll continue to look at that, and if there are opportunities out there, we'll certainly not be shy to exercise those opportunities.

Got it. That's helpful. My second question is related to gate speeds. You kind of mentioned it at first that getting below 50 nanoseconds is important for ultimately reaching quantum advantage for the industry. I think previously you mentioned you guys are around 70 nanoseconds currently, and just kind of curious if you can provide some thoughts on the roadmap for getting to sub-50 nanoseconds. Also, what are the implications in terms of overall quantum system performance? Is it a benefit or a boost to coherence times, fidelity rates, or in terms of the overall productivity and performance of the systems? If you could help provide some color, that'd be helpful.

Good question, Stephen. I mean, the four things that we have mentioned to get to quantum advantage are the qubit count, which we believe has to be minimum 1,000, 2-qubit gate fidelity, which we believe has to be minimum 99.9%, error correction, and then gate speeds faster than 50 nanoseconds. Of the four things, we feel most confident that we will be able to get to gate speeds, faster gate speeds, relatively quickly. That's not the determining factor, if you will, to get to quantum advantage. As you correctly said, with ANKA-3, we are at about 70 nanoseconds. We are deploying CFIUS-1 right now, which is a little faster than ANKA-3, as my remarks pointed out. We are still quantifying it, but it will be in the 50 to 60 nanosecond type range. We certainly think we will be able to accelerate that faster.

Getting to 50 nanoseconds or faster is not that difficult, honestly. We believe gate speed is extremely important. Ultimately, you are building a quantum computer. Speed absolutely matters. Once you combine all the metrics, gate speeds are going to be critical when it comes to time for performing any operation and completion kind of tasks. Our view, as we have pointed out multiple times before, is that a quantum computer is not going to exist in a silo in some kind of a quantum network. It is going to sit in existing data centers with CPUs and GPUs in the form of a hybrid system. It will have to interface with existing networks.

I think our view is that you need to design a quantum computer that fits into the overall data centers, which means that your clock speeds and other metrics have to be commensurate with CPU, GPU clock speeds. For that, you do need the quantum computer to be faster than 50 nanoseconds. One would argue even that is on the slower side compared to CPUs and GPUs, but at least there is a chance to be able to use that gate speed to stay up with CPU and GPU clock speeds. Once you start talking the hundreds of microseconds that some other modalities like trapped ion or pure atoms use, you're really like 1,000 times, if not 10,000 times slower than superconducting quantum computers and certainly CPUs and GPUs. That makes it really hard to think about a quantum computer existing in current data center using current networks.

Our view of a hybrid system using existing networks really forces you to talk about tens of nanoseconds of gate speeds. Hopefully, that answers your question.

Yes, super helpful, Subodh. Just a quick follow-up or a housekeeping item for Jeff. Jeff, like post the equity raise, what share cash are we modeling for Q3?

Sure. I would say roughly $327 million-ish, roughly.

Perfect. Thank you so much.

Speaker 5

Thank you. Our next question comes from Nathaniel Quinn Bolton with Needham & Company. Your line is open.

Hey, guys. Congratulations on the nice results and the technical milestone for the mid-year. Wanted to start just with that sort of the roadmap on the size of the chiplet versus the number of chiplets in your tiled approach. You know, it sounds like you're going to stick with the 9-qubit QPU for the near term, but getting to 1,000 qubits, you know, if you stuck with a 9-qubit solution would require over 100 chiplets. I'm kind of wondering, you know, when do you start to see a trade-off between the number of chiplets versus the number of qubits on a given chiplet? Where's that sweet spot, do you think? Yeah, where do you think that sweet spot ultimately ends up?

Speaker 2

A good question, Quinn. Honestly, we don't know the answer right now as to when exactly would be the right time to transition from a 9-qubit chiplet to something higher. Clearly, we will do it before we get to 1,000 qubits, as you correctly said. Otherwise, we are talking about more than 100 chiplets. That will start putting unnecessary pressure on the packaging side, and there is no reason to push it that hard at this time. Certainly, there's more flexibility on the size of the chiplet itself. We'll stay with 9-qubit at least until we get to the 100-plus qubit milestone before the end of this year.

For next year's milestone, which would be higher qubit count and better fidelity than this year, we will look at options of staying with 9-qubit or trying to attempt something bigger, like a 16-qubit, some square number, so 16, 25, 36 kind of qubits for the next chiplet size. We are doing the work right now to decide which is the next optimal chiplet size. Certainly, once we go over a few hundred qubits, we will be using a higher qubit count chiplet to get to 1,000-plus qubit.

As you go to larger square qubit tiles, would that require any significant CapEx on the equipment in Fab 1, or do you think the existing equipment set should allow you to go to any reasonable square number of qubits on a single tile?

There will be some, there's always going to be some need for new capital or upgraded capital for our fab in Fremont, California. We continue to do the necessary investments there. We don't see anything. Our ANKA-3 chip, if you will, was 84 qubit at roughly 1 by 1.5 centimeter. The 9-qubit chip that we are dealing with, chiplet right now we are dealing with is 6 millimeter by 6 millimeter. Certainly, we have capability to handle a chip, if you will, up to 1.5 centimeters. We don't think we need something drastically different for fab to get a higher chiplet size. Packaging is certainly an area we are looking at right now as we start building more than 10 or 20 chiplets. Do we need better quality packaging equipment, or do we need something different? We are doing that work right now.

We don't think some significantly new equipment with very, very high price tab is needed at this time to get to the higher qubit count.

Got it. Okay, that's great. I wanted to move on just to, you mentioned one of the four requirements for quantum advantage would be the quantum error correction. I know this year's milestones are really around the tiled approach and hitting 100 qubits with 99.5% fidelity by year-end. As you get to that 100-plus qubit solution by year-end, when do you think you start trying to implement the low-density parity check error codes that I think you guys had submitted as part of your QBI DARPA program?

Correct. Quantum error correction is obviously a very important area long term, and it will become more and more of a discussion as we get into 2026 and beyond. This year, we are focusing much more on the fidelity side and increasing the qubit count to more than 100 qubit. We continue to work on quantum error correction on our own along with our partner, River Lane, Cambridge, UK. We have done some excellent work demonstrating real-time error correction to some level, low latency error correction, and we will take that work to get to real-time error correction soon. There is a lot of work to be done on that front. We believe we need several hundred qubits at 99.7% or 99.8% or something along those lines to truly demonstrate the value of error correction in a real-time sense.

We are not quite at that point on the hardware side to try the sophisticated error correction codes like the QLDPC code that you referred. We are a year or two away from starting to do that kind of work. Hopefully, that answered your question.

Yeah, it did. Lastly, just any updated chatter on when the DOE, National Quantum Act, or the reauthorization of NQI might make it through Congress? Does it feel like there's any momentum there? Is it going in front of committees? Are there hearings being held in Congress to try to advance that bill towards signage?

Yes, absolutely. It looks like there's bipartisan support. There has been bipartisan support for a while, and it continues to be the case. There are several versions of the NQI reauthorization bill that are in different committees, and a lot of hearings have happened in the last few months along those lines. The House has multiple versions. The Senate has multiple versions. Nothing has been consolidated down to a single version yet. We hope that happens in the next few weeks or months, and it becomes the NQI Reauthorization Act. Obviously, we are looking forward to getting that done and signed, but it hasn't happened yet. Certainly, support seems to be there, and all the hearings that we have participated in ourselves as well as following look like it's going to happen. It's just a question of when, not if.

Got it. Okay. Thank you.

Thanks, Quinn.

Speaker 5

Thank you. Our next question comes from Richard Shannon with Craig-Hallum Capital Group. Your line is open.

Hi, Subodh. This is Tyler Anderson on for Richard, and congrats on all the work this quarter. I was wondering, do you have any feedback or updates from QBI or NQCC to give?

Speaker 2

I mean, we certainly talked to both of the organizations, the DARPA organization as well as NQCC organization, on an ongoing basis. They are very much aware of our progress. With NQCC, as we have disclosed in the past, there are several active projects that are going on right now, one of them being upgrading their existing 24-qubit system to what we have right now in California, the 4x9 qubit chiplet type system. We are going to be working with them to upgrade their system, along with demonstrating some other fundamental technology blocks like optical interconnects and other things. Those projects are ongoing. We will continue to disclose appropriately as we hit some technology milestones or publish some papers. With DARPA, we clearly are in phase one right now. They will be narrowing the group down for phase two before the end of this year.

We certainly are optimistic, given our results and where we are, that we will make it to phase two, but it's ultimately DARPA's decision. Our key differentiation from everyone else is our open modular approach as well as the chiplet design. We clearly believe this is a leadership system that we have introduced with four chiplets. I'm sure that will play a huge role in DARPA's decision-making. We continue to stay optimistic on that front, but we will find out when they decide before the end of this year.

That's great. Do you have any timeline on when you plan to reach a 16-tile chiplet? Are there any learnings that you've had from transitioning back to the chiplet approach?

Oh, there's plenty of learnings that we have derived from going from monolithic chip to chiplets. A lot of it is part of our know-how, as well as many of the patents we have filed in this area. Certainly, you can take a look at our patent portfolio. Many, many of those patents have started issuing now, and you can take a look at what exactly we cover in those patents. It's obviously a very important, critical piece of IP for us to be the first and foremost in demonstrating chiplet and having a proprietary approach to scaling up. There's a lot of learning. You clearly have to design the qubits appropriately because you are using chiplets, not a single chip. You need to adjust the geometries and the wiring layouts and so on. Not very different than what we have learned in the CMOS industry when we use chiplets.

If you look at all the work that happened a decade or so ago in the CMOS industry to incorporate chiplets in advanced devices, some of the things are similar that we are learning, but some are because by definition we are coupling qubits across the chiplets through an interposer. There are some new things that we are uncovering as we go along, and that's where a lot of the know-how is getting developed and IP is getting developed.

Okay, are you planning on staying with a square layout for your tiles? Piggybacking off of Quinn's question, do you have any range that you could give for the logical qubit overhead for QLDPC code?

For the time being, we'll continue to stay with the square chiplets. Right now it's nine. As I mentioned to Quinn, we'll probably look at a higher number before we go to 1,000 qubits. We are pretty sure we'll look at a higher number before we go to 1,000 qubits. Up to 100, we'll stay with nine for sure, but beyond that, we'll look at a higher number than nine. Regarding this whole discussion of logical qubit, as you probably are well aware, there's no clear definition of logical qubits. A lot of it depends on the error correction and how you lay it out. We'll continue to stay with physical qubit and fidelity, which basically gives you effectively logical qubit. Once you are in that 99.9%, 1,000-plus qubit range, we believe we will get to an overhead of 10 to 1 or better, but it's all projections right now.

No one has demonstrated anywhere close to 100 logical qubits yet. Our projections say that once we are at that 1,000-plus qubits, at 99.9%, 2-qubit gate fidelity or better, we will be in that range, but we need to get there. A lot depends on the definition of logical qubits and error correction and so on. We'll continue to use physical qubits and 2-qubit gate fidelity and metrics like that, which are clear and not controversial. Otherwise, once you say logical qubit, you have to define what a logical qubit is, how you did your error correction, and then numbers are all over the place at that one, making it very difficult to compare. Hopefully, that answers your question.

That does. I got one more, speaking of controversial. Does having this chiplet, is that garnering any more attention towards people getting any more on-premise systems, whether it's from people who you've already sold to or someone new?

I mean, we certainly, right now we deal with the U.S. government, the UK government as our two primary customers, if you will, and we continue to talk to some other governments on a selective basis. All of them are very interested in the chiplet approach. Our belief, and many of those customers' belief as well, is this is truly the only scalable way to get to more than 1,000 qubits. None of us see how you can take a single monolithic chip and take it to more than 1,000 qubits and certainly to tens of thousands and hundreds of thousands of qubits, which you eventually have to get to a fault tolerant quantum computer. Everyone sees chiplet as a necessary component to get to fault tolerant quantum computing.

When we talk to the DOEs, DODs, the UK National Quantum Computing Centre, they all understand the strategic value of chiplets and demonstrating that to get to fault tolerant quantum computing. Hopefully, that answered your question.

It does, and I agree with the statement on the need for chiplets. Thank you. I appreciate your time.

Thank you, Tyler.

Speaker 5

Thank you. Our next question comes from Brian Kinstlinger with Alliance Global Partners. Your line is open.

Great, thanks for taking my question. Are we both can agree that the most important metrics today are based on progress and your roadmap? You've clearly stated that you still have four years left on that roadmap to achieve quantum advantage. Is there some combination of your four metrics that begin to drive revenue or larger scale orders in your opinion?

Speaker 2

Certainly, we agree that right now it's all about technology development and technology metrics. That's the most important thing. Obviously, sales, we monitor, we report. We are dealing with primarily government labs and academic institutes right now. Sales, as you call them, they are more like research contracts. They are one-off. We will continue to participate with DOE, DOD, UK government, and other governments as appropriate. We really don't believe that those numbers, one-off numbers, they can fluctuate very widely, are really representative of what's happening. At the same time, yeah, the government, national labs, universities are interested in getting on-premise quantum computers for research applications, not for production workflows, not in their data centers or anything like that, but for research applications. As we continue to get closer and closer to quantum advantage, you are going to see more and more of those orders.

If you look at the national quantum missions, if you will, of various countries, starting with the U.S., we are talking substantial numbers. I mean, the NQI reauthorization, the number we talked about is $2.5 billion over five years. That's roughly $500 million a year. The DOD DARPA initiative is already, they have disclosed it to be more than half a billion dollars for the current QBI initiative, and there will be more projects like that from the DOD side. You go to the UK, you are talking of hundreds of millions of dollars, and many other countries in the Western world, along with some other select countries in Asia, friendlier countries in Asia, they're all talking about hundreds of millions of dollars a year. Some of that will be used for on-premise quantum computers, and we will continue to look at those opportunities to participate on a selective basis.

Even though it's not our focus, we'll continue to look at those opportunities, and we are pretty confident we will get those opportunities. By no means do we want that to become the focus of the company while we are continuing to work on getting to quantum advantage as fast as we can. Hopefully, that answers your question.

Yeah, great. Thanks.

Speaker 5

Thank you. Our next question comes from Craig Ellis with B. Riley Securities. Your line is open.

Yeah, thanks for taking the question, and congratulations on the progress technically, and Subodh on the extended visibility that it gives you to get to a 100-qubit system and beyond that. I had a question related to how the technology advances and the way you collaborate with your partner, Quanta, on the way things progress. I would expect that at some point when you scale up to larger qubit sizes, that I'm not sure what the thresholds might be, but at some point there would be systems implications, and a system for a certain qubit count would have to evolve to one of a higher qubit count.

The question is, how do you ensure that Quanta is progressing with the system development issues so that as you scale up to 100, multi-hundred qubits, 1,000 qubits, that on the system side they're delivering on time and that the entire system is going to be one that works really well? Thank you.

Speaker 2

Good question, Craig. Certainly, anytime you do a strategic partnership with anyone, you have to worry about those exact questions that your partner is capable of up to speed. They are not the ones who are going to slow you down and so on. We will continue to work closely with Quanta. They are a very, very capable company, as you know. They are the leaders in CPU, GPU servers on the cloud right now. They have the number one market share for GPU servers. They have a significantly capable and large technical team, and they are putting some of their best people on the quantum computing program right now. We have seen no indications that they are going to drop the ball on their side. They're very actively involved. Right now, we continue to make our own control systems, but we continue to basically get them up to speed.

They are very capable on the CPU, GPU side. When it comes to the hybrid system side, they are going to be teaching us effectively on the CPU, GPU side. I believe the collaboration is working out great right now. It's still very early days. They are coming up to speed. In 2026, at least certainly before the end of 2026, I believe we will start using control systems from Quanta, and then they'll start getting into the rest of the hardware stack. Given the overall timeline for quantum advantage of about four years from now, and that's when the volumes will start picking up in two to three years, I believe Quanta is very well positioned to help us with the acceleration ramp.

That's where they clearly bring their strengths of high volume, low cost manufacturing of these GPU servers, and that's really where we will start getting the benefit of Quanta's capabilities a couple of years from now when we start talking higher volumes. Hopefully, that answered your question.

It does. Thanks, Subodh. The follow-up question relates to the flip side of the current state of the government funding resolution issue. No, it's not yet signed. I wouldn't think that that would preclude you from interacting with national labs or the DOE and talking about roadmap issues, technology progress. Can you just talk about the things that you're able to do with some of those entities to position the business best for when we do get those funding resolutions and the team can better realize the related revenue opportunities from them? Thank you.

Yeah, it's a good question. I mean, we continue to engage very actively with both DOE and DOD right now. Even though the NQI reauthorization has not been signed and appropriated yet, our relationships continue to be very strong. If you physically visit Fermi Lab and the SQMS center in Fermi Lab, you will see several dilution refrigerator systems over there incorporating our chips. One of them is the full-fledged 9-qubit system that we deployed last year. There are many other systems that are using our chips, and various experiments are being done. Along with it, we talk with other DOE labs as well. You obviously are familiar with our involvement with QBI and the DARPA initiative. Nothing has changed from an interaction standpoint. Everyone continues to be very interested in superconducting technology and our open modular approach, particularly the chiplet approach in superconducting computing technologies. All that is progressing.

Obviously, they are stranded without getting additional dollars from the NQI reauthorization. They need more money to continue their experiments. Overall, the technology is progressing well. We continue to do our roadmap. They are continuing to do their work. It's just that all of us would like government to fund these initiatives at a higher level than what the current situation is. As far as I can see, our internal roadmap has not been impacted that drastically because of the lack of NQI funding so far. We certainly want our government to step up and start funding these initiatives.

Got it. Thank you, Subodh.

Thank you, Craig.

Speaker 5

Thank you. This concludes the question and answer session. I would now like to turn it back to Dr. Subodh Kulkarni, Chief Executive Officer, for closing remarks.

Speaker 2

Thank you for your interest and excellent questions. We look forward to updating you with our progress in future quarters. Thanks again.

Speaker 5

This concludes today's conference call. Thank you for participating. You may now disconnect.