Rigetti (RGTI.US) Q3 Earnings Call: Revenue Misses Expectations, CEO Stresses Technology Milestones Delivered on Schedule, Sales Still in R&D Phase

After the U.S. market close on November 10, quantum computing company Rigetti Computing (RGTI.US) released its fiscal 2026 third-quarter earnings report and subsequently held an earnings call. The most closely watched aspect of the call was the contrast between the company delivering its technology milestones on schedule and revenue falling well short of market expectations. Addressing this discrepancy, CEO Subodh Kulkarni emphasized that the Cepheus-1-108Q system had achieved its 99.5% median two-qubit gate fidelity target, further validating the scalability of the chiplet architecture; on sales, he said that "everything is still in the R&D phase" and that sales volume is "relatively small in the grand scheme of things." The company reiterated its full-year revenue target of approximately US$20 million, contingent on the India C-DAC customer completing acceptance in the fourth quarter. The earnings report showed third-quarter revenue of US$1.411 million, down 27.5% year-over-year and down 72.5% quarter-over-quarter, roughly 60% below the market consensus of approximately US$3.52 million (Benzinga/FactSet basis). On a U.S. GAAP basis, diluted net loss per share was US$0.08 and net loss was US$3.137 million, including a US$25.291 million gain from the change in fair value of warrants; on an adjusted (non-GAAP) basis, diluted net loss per share was US$0.06, slightly worse than the consensus of approximately US$0.05, and adjusted net loss was US$20.468 million, widening year-over-year. The company reiterated full-year 2026 revenue of approximately US$20 million, more conservative than the FactSet consensus of approximately US$22.6 million. Following the release, the stock fell as much as 9% in after-hours trading; it closed the regular session at US$12.12, down 6.63%. I. Key Focus and Response The core tension of the call was that technology milestones were delivered on schedule, but commercial revenue was far from keeping pace, with the full-year target highly dependent on acceptance of a single order. Management was repeatedly pressed during the Q&A on the certainty of the sales pipeline and the full-year target. When asked whether sales could ramp over the next few quarters, Kulkarni said the company continues to engage with national laboratories, universities, and a small number of commercial customers engaged in quantum research, and that sales are expected to grow given the overall enthusiasm around quantum computing, but "they are relatively small in the grand scheme of things, because everything is still in the R&D phase." CFO Jeff Bertelsen added that interest from government laboratories and commercial customers such as TangleLab is growing, but the market "remains research-driven" and will not ramp significantly until quantum advantage is achieved. On the full-year target of approximately US$20 million being highly dependent on acceptance of a single US$8.4 million C-DAC order, management did not shy away, reiterating that recognition of the related revenue is contingent on completing installation and performance acceptance in the fourth quarter. Technology Milestones and Roadmap In his opening remarks, Kulkarni summarized the quarter's progress in three points: first, Cepheus-1-108Q achieved 99.5% median two-qubit gate fidelity, continuing to improve performance at the 100-plus qubit level and further validating the chiplet architecture; second, the company signed a US$100 million definitive agreement with the U.S. Department of Commerce; third, the company is advancing its next-generation technology roadmap. The company's superconducting systems maintain gate speeds of 5070 nanoseconds, roughly 10,000 times that of trapped-ion systems and 100 times that of neutral-atom systems. The next major system goal is to demonstrate, by mid-2028, a system with more than 300 qubits, approximately 99.7% median two-qubit gate fidelity, and gate speeds below 60 nanoseconds. Quantum Error Correction and Commercialization Pace In response to an analyst question about which metrics to track over the next 18 months, Kulkarni said the company will continue to improve its three core metricsfidelity, qubit count, and gate speedand will work with Riverlane to advance error correction and mitigation, disclosing progress when milestones are reached. He emphasized that error correction and mitigation only become meaningful at a minimum of 100 physical qubits and approximately 99.7% or even 99.8% fidelity, and that it is premature to talk about "logical qubits" before then. Government Orders and Sales Pipeline Kulkarni said nearly all developed countries have introduced quantum strategies, and the company is closely engaged with the U.S. and U.K. governments and India (C-DAC), has deployed systems at the U.S. Department of Energy's Fermi National Accelerator Laboratory, the Department of Defense's Air Force Research Laboratory, and a national laboratory near Oxford in the U.K., and is building a new system at India's C-DAC. He said sales remain primarily to research-oriented customers, and the market will not ramp significantly until quantum advantage is achieved. U.S. Department of Commerce Agreement and Capital Bertelsen explained that the US$100 million in Department of Commerce funding will be drawn down gradually as monthly costs are incurred, all recorded in equity (additional paid-in capital), corresponding to approximately 7.7 million shares issued; the related costs are expensed normally as R&D expenses or capital expenditures, essentially equivalent to a US$100 million equity financing with restrictions on use of funds. As of the end of September, the company had not drawn down any of the funds. Cash, cash equivalents, and available-for-sale investments totaled US$497.1 million at period end, and the company has no debt. M&A and Full-Year Guidance Asked whether the company would step up M&A with approximately US$500 million in cash reserves, Kulkarni said it would only consider targets that could accelerate the roadmap or have key strategic value, that no action has been taken yet, that it remains focused on in-house development, and he reiterated confidence in the superconducting gate-based technology route. The company reiterated full-year 2026 revenue of approximately US$20 million, contingent on C-DAC acceptance in the fourth quarter, and management did not raise guidance. Earnings Call Transcript (translated with the assistance of AI tools): II. Executive Remarks Subodh Kulkarni, President and Chief Executive Officer Thank you for joining Rigetti's third-quarter earnings call. Chief Financial Officer Jeff Bertelsen will provide details on the financial results after my remarks. ... Before reviewing the quarter, I'd like to frame this discussion with three key points: First, we achieved 99.5% median two-qubit gate fidelity on Cepheus-1-108Q, an important milestone demonstrating sustained performance improvement at the system level above 100 qubits, not just on small research devices, and further validating our approach to scaling superconducting quantum computing with chiplets. Second, we strengthened our partnership with the U.S. government. During the quarter we signed a definitive agreement with the U.S. Department of Commerce for US$100 million in funding to accelerate three R&D areas critical to scaling superconducting quantum computing. Third, we are advancing our next-stage technology roadmap. Rigetti superconducting systems continue to achieve gate speeds of 50 to 70 nanoseconds, roughly 10,000 times that of trapped-ion systems and 100 times that of neutral-atom systems. The next major system goal is to demonstrate, by mid-2028, a system with more than 300 qubits, approximately 99.7% median two-qubit gate fidelity, and gate speeds below 60 nanoseconds; in roughly three years, the goal is to deploy a system with approximately 1,000 qubits, 99.9% median two-qubit gate fidelity, gate speeds below 50 nanoseconds, and integrated error correction and mitigation. III. Q&A Session TD Cowen analyst Krish Sankar: Sales Pipeline and the Interconnect Cost of Chiplet Scaling Q: How do you expect the sales pipeline or backlog to evolve over the next few quarters? A (Subodh Kulkarni): Thanks, Krish. We typically spend more time on technical metrics; on sales, we continue to engage with national laboratories, universities, and a small number of commercial customers engaged in quantum research. Given the overall attention on quantum computing, we do expect sales to continue to grow. But again, they are relatively small in the grand scheme of things, because everything is still in the R&D phase. Jeff, do you want to add anything on the pipeline or backlog? A (Jeff Bertelsen): I agree with what Subodh said. Quantum computing is still at an early stage, but certainly we are seeing growing interest from governments, government laboratories, and some commercial customers, such as the TangleLab opportunity you mentioned today. So directionally, interest is indeed growing, and as we get closer to quantum advantage and our systems become more capable, we expect to see continued increases in customer interest and growth opportunities over the coming years. Q: Subodh, you mentioned that the chiplet strategy has no limit at the largest scale. I wanted to ask whether, as scale increases, there is a need to address trade-offs in interconnect performance, fidelity, or speed? A (Subodh Kulkarni): Good question, Krish. So far we have not seen any limits in connecting chiplets, because we use an interposer, connecting two chiplets through the interposer, and we have not seen performance degradation. That doesn't mean a larger grid won't bring negative effects at some point, but with the current 12 nine-qubit chiplets, we haven't seen any issues. So when we talk about 1,000 qubits, we're roughly at about 30 36-qubit-type chiplets, and we don't expect issues. That said, this is indeed a key area we continue to monitor, to see whether fundamental limits emerge as we scale. But for now, we don't see any limits. Needham & Company analyst Quinn Bolton: Government Program Progress and How Commerce Funding Is Booked Q: First, congratulations on the progress with Cepheus-1-108Q. Subodh, I wanted to start by understanding the interest you're seeing from government laboratories or government agenciescould you update us on the latest progress with the DARPA QBI program and the U.K. ProQure program? A (Subodh Kulkarni): Sure, Quinn. Overall, as you can see from the news cycle, nearly every developed country in the world now has some kind of quantum mission or quantum strategy, because everyone understands the importance of quantum computing to national security and its potential over the coming decades. We are particularly closely aligned with the U.S. government and the U.K. government, and increasingly with the Indian government through the C-DAC system. We continue to engage with other governments interested in quantum computing, but we are careful about how many governments we engage with, because each requires dedicated resources and personnel support. Obviously, we have installed quantum computers at a Department of Energy laboratory (Fermilab in Chicago), a Department of Defense laboratory (the Air Force Research Laboratory in upstate New York), have a system at a U.K. national laboratory near Oxford, and are now building a system at India's national laboratory C-DAC. We will continue to pursue these strategically meaningful collaborations that both advance technology and generate sales. We are deeply involved in the U.S. government's Department of Commerce initiative, the CRDO programthe US$100 million we mentionedwhich is extremely valuable to us, helping us develop technology and accelerate our timeline. We are also continuing discussions with DARPA on QBI and with the U.K. on ProQure, and will announce separately when there is disclosable progress. Q (to Bertelsen): You mentioned that costs under the Commerce milestones will be expensed or capitalized as incurredso how does the US$100 million flow through the balance sheet, cash flow statement, and income statement? It seems there is not yet a US$100 million receivable. A (Jeff Bertelsen): The funding will flow entirely through equity. The full US$100 million from the Department of Commerce will be allocated to the approximately 7.7 million shares we issued, recorded in equity. We draw down the funding monthly as costs are incurred, recording those funds in additional paid-in capital each month. So each quarter, as we draw down funding, you will see cash inflows in the cash flow statement, while costs incurred under the program are expensed normally as R&D expenses or capital expenditures. You can think of it as a US$100 million equity financing with restrictions on use of funds. Q: But that US$100 million in cash is not currently on the balance sheet, right? It will be drawn down gradually as costs are incurred? A (Jeff Bertelsen): Correct. As of September, we had not drawn down any of the US$100 million, because the program had just started and no related costs had yet been incurred. Simply put, drawing US$100 from the Department of Commerce means debiting cash and crediting equity. Q: So the income statement is unchanged, the US$100 million in costs still flows through R&D, and there will be US$100 million of inflows in the cash flow statement to offset those costs, right? A (Jeff Bertelsen): Correct. Alliance Global Partners analyst Brian Kinstlinger: Metrics to Track Over the Next 18 Months and the Timing of Error Correction Q: The next major data point on the roadmap is mid-2028. Over the next 18 months or so, should we expect to see incremental improvements in QPU scale, fidelity, or speed versus Cepheus? Which KPIs should we track during this period? A (Subodh Kulkarni): Good question, Brian. We will certainly continue to improve fidelity, qubit count, and speed, and will discuss error mitigation and correction in conjunction with our Riverlane collaboration as progress is made, disclosing when key milestones are reached. We currently achieve 99.5% median two-qubit gate fidelity on 108 qubits, higher on sub-lattices, with gate speeds of about 60 nanoseconds. Whenever these numbers improve, we will disclose them and deploy as appropriateit doesn't make sense to deploy to the public cloud every time a single number improves; we will carefully judge whether it is worth end users switching to a new platform. You should continue to focus on our three key metrics: two-qubit gate fidelity (along with other fidelities such as parallel fidelity and readout fidelity), physical qubit count, and, critically, gate speed. We avoid loose use of the term "logical qubits"we note that some companies in the industry have begun to do so, but logical qubits should be nearly perfect, i.e., error rates at least better than 10 to the minus 10; talking about logical qubits at error rates of 10 to the minus 3 or minus 4 would confuse everyone. Q: When does error correction become a priority in your QPU development? Could you explain in terms of QPU scale and fidelity? A (Subodh Kulkarni): Sure. Our view is that error correction and even error mitigation only begin to take effect at a minimum of 100 physical qubits and approximately 99.7%, or even 99.8%, two-qubit gate fidelity. Before that, talking about error correction and mitigation is meaningless, which is why we have talked about it less previously. We work closely with Riverlane, and we also conduct our own error correction development and have published several papers. As we approach 99.7%, 99.8%, and with our sub-lattices already at 99.7%, we can begin to talk about error correction and mitigation at that level. Below 99.7% fidelity, whether using FPGAs, CPUs, or GPUs, the errors generated far exceed what can be corrected and would be overwhelming. Craig-Hallum Capital analyst Tyler Anderson: Co-Development Model for Cryostat Architecture Q: Regarding the cryostat architecture, does this require collaboration with existing dilution refrigerator manufacturers? Is this a dilution refrigerator with a different overall size, or an adjustment to the internal shape or configuration? I wanted to clarify which R&D you are taking on yourselves. A (Subodh Kulkarni): Good question, Tyler. Fundamentally, under this Department of Commerce initiative, we bear responsibility for delivering these milestones. The two items you mentionedminiaturization of readout electronics and a new cryostat architecturewe clearly won't do entirely ourselves; this is an extremely complex area with many potential partners who excel in specific domains, and we use a partnership model where appropriate. On readout electronics miniaturization, we are working with several partner companies to select appropriate mature technologies to deliver the Commerce milestones, while understanding their interaction with QPU technology. On the cryostat architecture, we are not a cryostat company ourselves; we currently source off-the-shelf dilution refrigerators from three companies depending on the application, and we are working with one of them to co-define this new architecture and how it will help us move toward 10,000 qubits and beyond. So on these specific milestones, you will see us use a partnership model. Cantor Fitzgerald analyst Kieran McCabe: M&A Opportunities and Capital Expenditure Decisions Q: With approximately US$500 million in cash and a market pullback, are you seeing inorganic (M&A) opportunities become more attractive? Or is the trade-off between in-house R&D and M&A roughly the same as earlier this year? I have a follow-up. A (Subodh Kulkarni): Good question. We certainly continue to monitor the market and look for potential opportunities, but will only consider them if they can help us accelerate the roadmap or have truly key strategic value. Obviously we haven't seen any targets yet, which is why we haven't acted. Fundamentally, we remain confident that the superconducting gate-based technology route we have adopted is the most likely to dominate the market, because we have advantages in speed and scale. Looking at the current landscape, our Cepheus-1-108Q, as I said earlier, is second only to IBM's 120 qubits among systems on the public cloud in terms of qubit countwe are right behind, then Google's 105 qubits, with everyone else below 100 and most companies far below 50. So M&A for the sake of M&A, or M&A to boost sales, is not the right strategy at this stage, in our view. We remain focused on developing key technical metrics; but if an opportunity with major strategic value that could accelerate the timeline arises, we would consider it. Q: My follow-up is related. The Commerce agreementI'm thinking more broadly, not just the US$100 million to you and funding for other quantum computing companies, but also investment in fabricationonce in place and finalized, has it changed your future capital expenditure decisions in any way?