Prepared remarks
Good day, and thank you for standing by. Welcome to the Rigetti Computing Second Quarter 2026 Financial Results Conference Call. Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Subodh Kulkarni, Chief Executive Officer of Rigetti. Please go ahead.
Good afternoon, and thank you for joining us for Rigetti's Second Quarter 2026 Earnings Conference Call. I'm pleased to be joined today by our Chief Financial Officer, Jeffrey Bertelsen, who will walk you through our financial results in more detail following my overview. We appreciate your continued interest in Rigetti and look forward to answering your questions at the conclusion of our prepared remarks. Before we begin, I would like to remind everyone that today's call, along with our second quarter 2026 press release, contains forward-looking statements. These statements reflect our current expectations, objectives and underlying assumptions regarding our outlook and future operating results and are subject to risks and uncertainties that could cause actual results to differ materially from those anticipated. These risks and uncertainties are described in more detail in our filings with the Securities and Exchange Commission, including our Form 10-K for the year ended December 31, 2025, our Form 10-Q filing for the quarter ended June 30, 2026, and other periodic reports filed from time to time with the SEC.
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, please refer to today's earnings release on our Investor Relations website at www.rigetti.com and to the 8-K furnished with the SEC. Before I dive into the quarter, I want to frame our discussion around three key takeaways. First, we continue to demonstrate technical leadership with our Cepheus-class systems, including ongoing progress to improve the performance of our 108Q, Cepheus-1-108Q platform. Second, the letter of intent we signed with the U.S. Department of Commerce for up to $100 million in funding over three years further validates our superconducting chiplet-based approach and strengthens our ability to invest against our roadmap to Quantum Advantage.
Third, we recently announced an expanded collaboration with HPE and the Pittsburgh Supercomputing Center to develop a hybrid quantum-classical supercomputer, reflecting growing demand for our approach and positioning Rigetti to deliver differentiated quantum-enhanced HPC solutions. Q2 was another important proof point in our strategy to combine technical progress with real-world access and long-term strategic partnerships. Quantum computing remains a long-cycle opportunity, but we are increasingly seeing the ecosystem coalesce around platforms that can scale in a practical way and that are available to users where they already run their workloads. We believe Rigetti is well positioned in the environment for three reasons. We are focused on superconducting gate-based quantum computing, which offers a combination of scalability and speed that is difficult to replicate with other modalities.
We are pursuing an open modular architecture that allows us to integrate innovative solutions from partners to advance our technology faster, such as quantum error correction (QEC) technology from Riverlane and transduction technology from QphoX. We pioneered and continue to lead in chiplet-based architectures, which we believe provide a more practical path to scaling to 1,000 qubits and beyond than monolithic approaches. Let me start with our technology and product progress. Cepheus-1-108Q remains one of the highest qubit count, generally available gate-based quantum computers in the world and the largest modular quantum computing system on the market today. The system is built from 12 interconnected 9-qubit chiplets and is accessible to customers via Rigetti Quantum Cloud Services and through Amazon Braket, Microsoft Azure Quantum and qBraid. We are one of only three companies, including IBM and Google, who have delivered gate-based systems with over 100 qubits, and we believe that our modular approach gives us a strong advantage on the path towards higher qubit count systems.
Today, Cepheus-1-108Q continues to operate at a median single-qubit gate fidelity of approximately 99.9% and a median two-qubit gate fidelity of roughly 99.1% with gate speeds around 60 nanoseconds. This performance at the 100-plus qubit scale is meaningful, and our teams remain focused on improving fidelity throughout 2026 as we refine chip design, materials and fabrication processes and incorporate learnings from our R&D platforms. From a systems engineering perspective, our work this quarter extended beyond adding qubits. We continue to mitigate coupling interactions between tunable couplers that become more pronounced beyond the 100-qubit scale and to address coherence time limitations that are now the primary constraint on fidelity. As we discussed with investors in May, our current coherence times are in the 25 to 30 microsecond range, and we are executing on a set of chip design and material initiatives intended to roughly double or triple those times over the next several years.
This includes joint IP with Fermilab where capping niobium superconducting contacts with tantalum has demonstrated meaningful coherence time improvements in fundamental experiments, and we are now incorporating those learnings into production chip designs. We are also refining deposition, oxidation and etch processes to smooth interfaces in the Josephson junction area, which we believe will further improve coherence and ultimately gate fidelities. Beyond Cepheus-1-108Q, we remain on track with the chiplet-based roadmap we laid out earlier this year and have begun investing in dilution refrigeration and infrastructure that can support 1,000-qubit systems. Our objective remains to reach approximately 1,000 qubits, two-qubit gate fidelities of 99.9% and gate speed below 40 nanoseconds in roughly three years. I would now like to spend a few minutes on the U.S. Department of Commerce letter of intent we announced on May 21.
Under this LOI, the department has selected Rigetti for a potential award of up to $100 million in funding over three years to accelerate superconducting quantum computing R&D that addresses key technical challenges in scaling and advancing our systems. The contemplated transaction structure includes the department receiving an equity stake in Rigetti, consistent with the total amount of funding, and the funding itself would be allocated under the CHIPS Research and Development Office broad agency announcement pursuant to the CHIPS Act. We believe that this prospective partnership reflects the view that quantum computing will have far-reaching impacts on national security, economic interest and overall prosperity and that superconducting qubit platforms are central to U.S. leadership in this field. It also reflects the administration's broader CHIPS R&D investments in quantum computing and related technologies designed to build domestic industry, create high-paying jobs and strengthen technological resilience.
For Rigetti, this LOI is strategically significant in several ways. It would allow us to tackle key scaling bottlenecks more rapidly associated with multichip architectures. It would enable us to accelerate multiple generations of superconducting quantum processors and associated control electronics at Fab-1, our dedicated quantum device manufacturing facility. It would deepen our collaboration with the U.S. government at a time when global investment in quantum computing is increasing and geopolitical competition in this domain is intensifying. It is important to emphasize that the LOI is not yet a definitive agreement. Any final structure will also include issuance of securities to the department and will be subject to customary approvals and conditions. We view this prospective partnership as aligned with our long-term strategy and our commitment to disciplined capital deployment. Jeff will provide additional context on how we are thinking about this funding opportunity within our broader capital framework.
Turning to customer momentum and market traction. Our strategy is to meet customers where they are across public cloud, hybrid infrastructure and dedicated quantum systems. On the cloud, Rigetti Quantum Cloud Services, Amazon Braket, Microsoft Azure Quantum and qBraid provide global access to our systems, including Cepheus-1-108Q. We continue to see engagement from researchers and enterprises seeking to experiment on one of the most capable generally available gate-based platforms where they benefit from ease of use and consistent uptime. In parallel, we are building a growing base of on-premises system deployments. Earlier this year, we outlined three systems targeted for delivery in 2026: Novera systems and a 108-qubit system for C-DAC in India. We remain on track and are seeing additional demand for Novera QPUs from national labs and universities with recent wins, including the University of Saskatchewan and a research arm of a large Japanese conglomerate.
These deployments deepen technical engagement, create multiyear usage pathways and generate high-value feedback that informs our roadmap. At this stage of the market, we are prioritizing customers committed to active experimentation and collaboration, and we are deliberate in how we build those relationships. As part of this customer momentum, we recently expanded our collaboration with HPE and the Pittsburgh Supercomputing Center. Rigetti will deliver a 9-qubit Novera quantum computing system into PSC's new TangleLab test bed, which is being funded by a National Science Foundation grant. Working with HPE, this deployment is designed to integrate the Novera system with an HPE-powered supercomputing environment so that researchers and industry users can develop and test hybrid quantum-classical workloads on real hardware. We view this as an important step in moving our HPE relationship from early experimentation toward commercially relevant quantum-enabled HPC solutions.
More broadly, global investment in quantum computing continues to expand, led by governments and national labs and increasingly complemented by commercial interest. Outside the United States, we are seeing structured programs such as the U.K. government's six-year ProQure initiative, where we intend to invest up to $100 million over time in systems, talent and infrastructure. In India, our 108-qubit C-DAC award reflects national-level interest in superconducting platforms. And across Europe and Asia, we are seeing a growing number of coordinated quantum initiatives. Commercial revenue remains early, but engagement is increasing across industries such as materials, logistics and financial services as they explore hybrid and quantum-inspired workloads. As system performance improves and the industry moves closer to quantum advantage, we expect commercial adoption to build on the government and research foundation in place today.
Let me briefly connect these developments back to our long-term roadmap. We remain focused on a clear sequence designed to position Rigetti to reach quantum advantage in roughly three years. Near term, that means driving Cepheus-1-108Q toward a median two-qubit gate fidelity of approximately 99.5% later this year while maintaining our gate speed advantages. Beyond that, we are working toward deploying higher qubit systems that leverage our chiplet-based architecture as the foundation for scaling beyond 1,000 qubits with fidelities and speed better mitigated and eventually fault-tolerant computation. In support of this roadmap, we continue to invest in Fab-1 and associated refrigeration infrastructure as well as partnerships with organizations such as Riverlane, Quanta and QphoX to integrate innovative solutions into the stack. The prospective CHIPS Act funding would further reinforce our ability to execute on multiple generations of processes while maintaining prudent capital discipline.
We intend to update our published technology roadmap later this year once we have additional operational data from Cepheus-1-108Q and clearer visibility into subsequent system deployments. Our objective is to provide investors with a transparent view of the milestones that matter most for Quantum Advantage, including specific targets for qubit count, fidelity, coherence time and gate speed. With that overview, I will now turn the call over to our Chief Financial Officer, Jeffrey Bertelsen, to discuss our financial results and capital deployment in more detail.
Thank you, Subodh, and good afternoon, everyone. I will spend a few minutes walking through our second quarter 2026 financial results, our balance sheet and liquidity and how we are thinking about capital deployment as we continue to execute on the roadmap Subodh described, including the potential implications of the Department of Commerce LOI. For the second quarter of 2026, revenue was approximately $5.1 million compared to $1.8 million in the second quarter of 2025. The year-over-year increase was driven by on-premises Novera QPU sales, reflecting recognition of previously announced Novera purchase orders. Gross margin for the quarter was approximately 43% compared to 31% in the second quarter of 2025, with variability driven by contract mix and pricing and the relative contribution from the Novera QPU sales. Total operating expenses for the second quarter were $30.3 million compared to $20.4 million in the same period last year, with spending concentrated in research and development, including engineering headcount, fabrication, chip design and control electronics development as well as investments in refrigeration and infrastructure to support higher qubit count systems.
Stock-based compensation for the quarter was approximately $7 million compared to $3.6 million in the second quarter of 2025. Operating loss was $28.1 million compared to $19.9 million in the prior year period. On a GAAP basis, net loss for the quarter was $52.6 million compared to a net loss of $39.7 million in the prior year period, with results again impacted by noncash fair value adjustments to derivative warrant and earn-out liabilities. As we have said previously, these fair value adjustments can introduce significant quarter-to-quarter volatility into our GAAP results and do not affect how we operate the business or allocate capital. On a non-GAAP basis, which excludes stock-based compensation and fair value adjustments to warrant and earn-out liabilities, net loss for the quarter was approximately $16 million or $0.05 per diluted share compared to a non-GAAP net loss of $13.3 million or $0.04 per diluted share in the second quarter of 2025.
Let me provide a bit more color on revenue drivers and how we are thinking about the remainder of the year. As we discussed on our prior call, we expected strong year-over-year revenue growth in the first half of 2026, driven by the previously announced Novera purchase orders we received late last year and earlier in 2026. In Q2, we continued to progress on the $8.4 million C-DAC order for an on-premises 108-qubit system in India, which we still expect to recognize in the fourth quarter of this year following installation and performance acceptance testing. More broadly, our revenue profile continues to be influenced by the timing of system deliveries and government-funded projects. We view this variability as inherent to the current stage of the quantum computing market and not as a driver of our long-term capital allocation or technology strategy. Turning to the balance sheet: we ended the second quarter of 2026 with approximately $541.3 million in cash, cash equivalents and available-for-sale investments compared with $569 million as of March 31, 2026, and $589.8 million as of December 31, 2025.
We continue to operate with no debt. Given our current operating profile, we believe our capital position provides sufficient runway to execute against our technology and system deployment milestones, including continued progress on scale, fidelity, system integration and our planned investment in the United Kingdom. Capital expenditures in the quarter were primarily driven by investments in Fab-1 and additional dilution refrigeration capacity to support higher qubit count systems. As we noted previously, we expect 2026 CapEx to be elevated relative to prior years, largely due to investments in dilution refrigeration and fab equipment. Our approach to capital deployment remains disciplined and consistent with what we have discussed on prior calls. The majority of our spending is directed toward core R&D activities that directly advance our technology platform, including our chiplet-based architecture, control systems and cloud integration.
We are not managing the business around short-term revenue optimization. We are managing it around credible long-term progress toward Quantum Advantage and commercially relevant systems. Should we move from the current LOI to definitive agreements with the Department of Commerce, we would expect the CHIPS Act funding to be deployed in a way that is tightly aligned with this philosophy. That means focusing the capital on specific technical programs that address scaling bottlenecks and packaging and enhancements while being mindful of potential dilution associated with any equity issuance. To close, our financial strategy is unchanged from what we outlined last quarter. We are focused on maintaining flexibility, funding innovation responsibly and aligning capital deployment with the long-term value creation of our technology roadmap, including the CHIPS Act funding. Quarterly results will continue to reflect the early-stage nature of the quantum computing market and the timing of large system contracts, but we believe our balance sheet and capital discipline position us to execute with patience and control. With that, I will turn it back to Subodh for closing remarks before we open the call for questions.
Thank you, Jeff. We are encouraged by the progress we are making on our technology roadmap, the strengthening engagement we are seeing from customers across cloud and on-premises channels and the strategic support we are beginning to see from governments such as the United States and the United Kingdom. We remain focused on delivering against the milestones we have laid out, including fidelity improvements on Cepheus-1-108Q, higher qubit systems and disciplined execution of our capital plan. On behalf of the entire Rigetti team, thank you for your continued interest and support. Operator, we are now ready to open the call for questions.
Questions and answers
We will now open the call for questions. And our first question comes from Brian Kinstlinger of Alliance Global Partners.
On the last conference call, you highlighted coherence time as the primary limitation on your new architecture to fidelity and that you believe you could address that challenge. You gave some details around that today. When is a reasonable time frame where we can expect to see that impact fidelity?
Indeed, as we discussed on the call, coherence time is the limiting factor for two-qubit gate fidelity at this time, and that's why we are so focused on improving coherence time. Right now, it's in the 25 to 30 microsecond range. We want to obviously bump it up as much as possible, but we think a factor of two to three improvement is practical in the near term, and that's what we are working on. The kinds of things we are working on, which we disclosed during the call, include smoother interfaces at the Josephson junction level and capping niobium with tantalum, which seems to improve coherence time; we have published some papers along those lines. Those are the kinds of things we are doing. We definitely expect fidelity to increase at the 108-qubit level before the end of this year, and we'll update everyone when we are ready to deploy a new system. I want to put some of these developments in perspective.
Keep in mind that there are only three very capable more-than-100-qubit systems in the gate-based world right now: IBM at 120 qubits, Rigetti at 108 and Google at 105. Many other companies remain well below that. We get incredible feedback from end users who are using our systems every day that our system is one of the easiest to use and has consistent uptime. So we want to be careful that before we upgrade our system with higher coherence time and higher fidelity, we have checked off all the other metrics and do not inadvertently go backwards in technical progress. We are careful with how we are doing the experiments and how we will deploy changes in the actual system that is deployed for customers. I hope that answers your question.
Does. If I can just ask one follow-up. A few months back, there was a change in leadership in DARPA's QBI. Has this had any impact on Rigetti's chances of getting to Phase B? And do you think this is still achievable by year-end?
Certainly, there was a change in leadership at DARPA. Our plan stays the same. DARPA gave us a set of areas where they wanted to see improvements to get into Phase B. We continue to work with them. We have periodic calls and meetings with them, and we will continue to demonstrate improvements in the areas they identified. We still feel optimistic that we will get into DARPA Phase B relatively soon. It's impossible to predict exact timelines on these things, but we continue to make very good progress and remain optimistic about entering Phase B. Along with DARPA, there are other initiatives, as I mentioned earlier, such as the U.K. government's ProQure initiative and other programs worldwide. So we are engaged in multiple initiatives; it's not just DARPA alone. As you saw, the Department of Commerce in the U.S. created a separate initiative focused on accelerating roadmaps, and we are part of that as well. DARPA will continue to be an important player in the quantum ecosystem, but it's one of several programs we are participating in.
And our next question comes from Krish Sankar of TD Cowen.
The first question on the current fab facility in Fremont. With the current tooling, what's the annual QPU production capacity? And what kind of CapEx requirement should we expect for that?
So Krish, right now, the fab we have in Fremont is a 150-millimeter fab. Capacity is not a concern at all. We have plenty of capacity in that fab because each quantum chip is capable of significant computation. We make a number of 9-qubit chips right now, and with our chiplet architecture we are using 9-qubit chiplets to build 36-qubit and 108-qubit systems. In the future, we believe we will make larger chiplets, such as 36-qubit chiplets, that will take us to 1,000 qubits and beyond, and we will continue to increase chiplet size. Right now, the 9-qubit chiplet is roughly 6 millimeters by 6 millimeters, so we can build many 9-qubit chiplets on a single 150-millimeter wafer and run multiple wafers per day. Capacity is not an issue. The challenge is capability: there are some limitations with 150-millimeter equipment and because it's not an automated line. We believe there are limitations that come with a non-automated line and the 150-millimeter size. So we continue to look at alternative options for upgrading the fab, but it's mostly for capability, not capacity.
Got you. Very helpful, Subodh. And then a follow-up on the HPE 9-qubit Novera system: will you be providing just the QPU chips, or will it be the QPU plus dilution refrigeration, controls and other infrastructure? And what is the timing for delivery on that?
Yes, it's a great question. We're really excited to partner with HPE and the Pittsburgh Supercomputing Center. This is an important initiative to demonstrate how hybrid computing works. Today, there are few places where you can actually demonstrate hybrid computing on real hardware. Our contribution to the NSF-sponsored project is a full 9-qubit system: the QPU plus dilution refrigeration, the control system and all associated infrastructure. HPE's contribution is the HPC side and the interfaces, which we will be working on with PSC. It will be an exciting project to bring the best of HPC and quantum computing together and allow global access to hybrid computing experiments. Because we use superconducting gate technology, we have a significant speed advantage: our gate speeds are in the 50–60 nanosecond range, which is orders of magnitude faster than trapped-ion or atom-based modalities. That speed advantage will be apparent in hybrid computing, because modalities with very slow gate speeds have trouble keeping up with CPU and GPU latencies. With superconducting systems, you'll see how speed helps in overall computation and arriving at correct answers much faster than other modalities. We're excited about the partnership and demonstrating hybrid computing, particularly with superconducting technology.
Maybe just to follow up on that with Jeff's last question: timing of delivery and how to think about revenue recognition for this?
We believe the delivery will be sometime in 2027. We don't know the exact quarter yet—the program just kicked off recently. As we have more details, we will be transparent about the timelines. But at this point, it will be a 2027 delivery. We'll let you know once we know exactly when it will happen.
And our next question comes from Troy Jensen of Cantor Fitzgerald.
Maybe a quick question to start with, Subodh. Any thoughts on the executive order? Has there been any change in government activity post signing of that order? I'm wondering if it's mainly security more or just system level, but I'd love to get your thoughts and if anything has changed since then.
With the executive order passed by the President, there has been increased emphasis on quantum computing. A number of initiatives have started because of those executive orders. They are generally in early stages, but you're already seeing examples, such as the Department of Commerce initiative using the CHIPS Act. We definitely expect more funding to flow into the quantum ecosystem from the U.S. government because of those orders. There are various bills being debated now—whether it's NQI authorization funding for the Department of Energy and national labs, the NDAA which includes quantum line items, and other legislative activity. Overall, the executive orders highlighted the importance of quantum computing for national security and strategic reasons and are helping free up funding. You are already seeing that with the Department of Commerce, and we expect more activity from both the Department of Energy and other agencies.
Makes sense. Congrats on the HPE announcement—obviously that's huge. I assume there are many potential data center partnerships. Is HPE more focused on networking or servers? Any more details on that would be great.
We're excited to partner with HPE. They're the top player in HPC. The partnership will evolve as we get into demonstration. It's not exclusive; we will talk to other HPC players and HPE will also engage broadly. The exciting part is bringing state-of-the-art HPC together with a state-of-the-art quantum computer and examining how to partition workloads—CPUs for sequential, GPUs for parallel, and quantum for simultaneous computation. With superconducting speeds that match CPU and GPU latencies, we can better demonstrate the benefit of quantum computing. We'll be able to show reductions in time to results, improvements in accuracy, and address applications that are challenging for current HPC setups. We're looking forward to assembling the system between now and 2027 and making it available to the public then. We'll continue discussions with HPE and other HPC partners on how to leverage quantum computing for various workloads.
And our next question comes from John McPeake of Rosenblatt Securities.
Great. Subodh and Jeff, congrats on the execution here, steady progress. I have a couple for Jeff. You talked about CapEx. But deferred revenues ticked up more than in any quarter I've seen in the history. And you actually had your largest billings quarter if you look at the change in deferred and revenues. Could you talk a little bit about what that's about, if anything?
Sure. That mainly relates to C-DAC; with that order there were prepayments that added to deferred revenue. So that was the main driver.
Okay, that makes sense. And then one for Subodh. I've heard compliments about your QPUs from some optical interconnect companies. I'm curious if you've had any discussions with respect to your open architecture potentially doing exploratory work on optical interconnects for your Novera QPUs?
John, thanks for sharing that. We consistently hear from end users that our QPUs are reliable, consistent and easy to use. We talk to multiple optical players. Among publicly disclosed partnerships, we have mentioned QphoX, where we're working on transduction—converting microwave signals to optical signals and vice versa. Right now, our 9-qubit and 108-qubit systems rely on coax signaling. We plan to use flex-cable technology to scale to 1,000 qubits, but beyond 1,000 qubits, optical signaling becomes necessary. Optical interconnects are the path to tens of thousands and hundreds of thousands of qubits. Early research with companies like QphoX is promising, but it's still early; we need more work before that research converts into a definitive roadmap and timelines. It's encouraging, however, that optical signaling appears viable for very large-scale systems.
And our next question comes from Quinn Bolton of Needham & Company.
This is Shadi Mitwalli on for Quinn. Now that you have your 108-qubit system up and running in the cloud, can you talk about what you are seeing in regard to user engagement? I have a follow-up after that.
As I mentioned earlier, we are getting excellent feedback from end users who are using our 108-qubit system. We consistently hear it's one of the best systems out there. To remind everyone, only three more-than-100-qubit systems exist in the world: IBM at 120 qubits, Rigetti at 108 and Google at 105. Many others are well below 50 qubits. Thousands of customers are using our platform via AWS, Azure and other cloud platforms, and the feedback is that our system is easy to use, consistently available and reliable. Most of the work users are doing today is research: algorithm design, fundamental exploration of quantum computing and fitting quantum into their ecosystems. We acknowledge we are not yet at Quantum Advantage; we need to get closer to 1,000 physical qubits, approach 99.9% two-qubit fidelity, and incorporate error mitigation or correction before we can demonstrate broad tactical workload applications. For now, our systems are built primarily for research, and that is what users are focused on.
Got it. That's great to hear. On the $100 million LOI, it sounds like this will likely be milestone- or technical-based. Can you talk about what Rigetti would need to achieve to unlock the funding?
Good question. The agreement is not definitive yet, so we're still discussing specifics with the Commerce Department on what exactly is expected to receive the $100 million. Broadly, we know the areas of focus—cryogenics, miniaturization of electronics in the readout chain, and accelerating our roadmap. The overall goal is to accelerate our roadmap, and we're defining which milestones can be accelerated. It's still a work in progress; once it becomes a definitive agreement, we'll disclose more detail on the specific milestones and timing.
And our next question comes from Tyler Anderson of Craig-Hallum.
This is Tyler Anderson on for Richard Shannon. Quick question: for the HPE and PSC deployment, is the dilution refrigerator and control unit yours, or will that be provided by a third party?
It's our system. We design and build the full system, although we source components: we buy dilution refrigerators from vendors and partner with Quanta for the control system. So while third-party components are used, the overall system is designed and built by Rigetti.
That makes sense. Regarding what is getting your coherence time up—was integrating these technologies required to reach the fidelities you were aiming for originally, or have you been deploying some improvements progressively as needed? I have a quick follow-up after that.
When doing advanced R&D on complex systems, we continuously experiment and iterate rather than holding items in reserve. We've been working with Fermilab on capping niobium with tantalum, and those results are promising. There are reasons we haven't deployed all changes in the system currently available to customers: the live cloud system is performing well, so before upgrading it we exhaustively test changes in separate systems to ensure we don't inadvertently degrade other metrics. We are deliberate: we make changes, test extensively and only deploy when we're confident we won't harm other aspects of performance.
That makes sense—doing changes one at a time. If you implement these improvements and see a big impact on coherence time, adiabatic gates have shown fidelity increases on the 36-qubit QPU. Could these factors together accelerate your roadmap if everything goes well?
Yes. Adiabatic gates are promising; we published a white paper last year. We're using a version of adiabatic gates today in our 108-qubit system, though not the full implementation disclosed in the white paper for reasons related to controlled deployment. Work is ongoing in that area, and we plan to include adiabatic CZ gates as part of our roadmap for future systems.
And our next question comes from Gary Mobley of StoneX Group.
In the past, you did work for the U.K.'s NQCC, and I believe that was at one point a substantial portion of your revenue. What's the prospect for renewal funding or reengagement there going forward?
That's a good segue into the U.K.'s broader quantum ecosystem. The U.K. government announced a multiyear initiative called ProQure. It's similar in concept to DARPA QBI, though the path and milestones differ. Both aim toward fault-tolerant quantum computing in roughly six to seven years. We're actively involved with the U.K. government and our system is at NQCC today—a 36-qubit system that has received very positive feedback for ease of use, consistent results and reliability. We are working with the U.K. on the ProQure initiative and hope to have announcements as their decisions are made. Our plan is to remain in the U.K. ecosystem and continue investing there.
When you examine your pipeline of potential deals, how does it compare today versus a year ago? Is it skewed toward commercial experimentation, government-funded projects, university research or something else?
We see a lot more interest today compared with a year or two ago. That interest is broadly across governments, universities and commercial companies. Governments are increasingly active—the Department of Commerce CHIPS initiative and U.K. ProQure are examples. India is active as well. Universities are getting more involved beyond historical hubs. We're seeing more commercial organizations interested in on-premise research systems: they understand these are research systems, but they want to experiment and learn. That commercial traction contributed to sales growth in the first half of 2026. So overall, interest has broadened across government, academia and commercial customers, with the commercial interest being a newer and encouraging development.
And our next question comes from Nehal Chokshi of Northland Capital Markets.
It looks like the 36-qubit median two-qubit gate fidelity improved from 99.5% in the first quarter to 99.6% in the second quarter. Could you say what drove that improvement?
We work on fidelity improvements continuously. Improvements at the 9-qubit and 36-qubit levels—and to a degree at 108 qubits—come from design optimizations, fab process adjustments and material changes such as tantalum usage. The 9-qubit level is around 99.8% two-qubit fidelity, and the 36-qubit median improved to about 99.6% due primarily to design optimizations: how lines are routed and how coupling is engineered. We are incorporating those learnings and will continue to apply them to 108-qubit systems and future larger systems.
So to be clear: the improvements implemented on the 36-qubit system have not yet been applied to the 108-qubit system, or are they in flight to see if they translate to the larger system?
You're correct. We are deliberate about making changes. We don't flip many variables on the live cloud-deployed 108-qubit system that customers use daily. We test and validate improvements on other lab systems and are in the process of incorporating those learnings into subsequent 108-qubit systems rather than changing the production system in place immediately. That way we avoid inadvertently degrading performance for users.
I show no further questions at this time. I'd like to turn it back to Subodh Kulkarni for closing remarks.
Thank you. In closing, we are encouraged by our progress in the second quarter and remain focused on executing against our strategic priorities as we advance Rigetti's position in the quantum computing ecosystem. In the coming months, we will be meeting with investors at a number of conferences and related events, and we hope to see you there. Thank you for your continued interest in Rigetti.
This concludes today's conference call. Thank you for participating, and you may now disconnect.