Xanadu Quantum Q3 2026 Earnings Call Transcript

Xanadu Quantum (NASDAQ:XNDU) reported third-quarter financial results on Wednesday. The transcript from the company’s third-quarter earnings call has been provided below.

This transcript is brought to you by Benzinga APIs. For real-time access to our entire catalog, please visit https://www.benzinga.com/apis/ for a consultation.

View the webcast at https://edge.media-server.com/mmc/p/za2wjr8c/

Summary

Xanadu Quantum reported Q3 2026 revenue of $1.3 million, a slight decrease from Q2, with a net loss of $32.2 million, primarily driven by increased R&D expenses.

The company achieved significant technological milestones, including a world record for photon loss in an optical switch and strengthened its manufacturing capabilities through partnerships with GlobalFoundries, ASML, and Bluefors.

Xanadu Quantum is pursuing a full-stack strategy integrating photonic quantum hardware with its PennyLane software platform, targeting a fault-tolerant quantum computer by 2029-2030.

Partnerships with AMD and the launch of Backline software aim to enhance real-time processing for fault-tolerant quantum computing, with growing interest from national laboratories and other quantum companies.

The company is expanding its R&D capabilities and manufacturing foundation with the new Inception facility, supported by Canadian federal funding and DARPA, to accelerate the development of a quantum data center.

Future outlook includes increasing R&D and capital expenditures in Q4, with continued headcount growth and strategic investments in technology and partnerships.

Full Transcript

Jordan (Conference Operator)

Thank you for standing by. My name is Jordan and I’ll be your conference operator today. At this time, I’d like to welcome everyone to the Xanadu Quantum Q3 2026 earnings conference call. All lines have been placed on mute to prevent any background noise. After the speaker’s remarks, there will be a question and answer session. If you’d like to ask a question during this time, simply press star followed by the number one on your telephone keypad.

All participants are limited to two questions only. Thank you. I would now like to turn the call over to Brett Harris, VP Investor Relations. Please go ahead.

Brett Harris (VP Investor Relations)

Thank you. Good afternoon everyone. Welcome to Xanadu Quantum’s third quarter earnings conference call. Joining me today are Dr. Christian Wiedbrook, Founder and Chief Executive Officer, Michael Trupek, Chief Financial Officer, and Rafal Yanik, our Chief Operating Officer. This call is being webcast live and will be made available on Xanadu Quantum’s investor relations website. This call contains time-sensitive information accurate only as of the date of this live webcast.

On this call, November 11, 2026. This call may include forward-looking statements and information which are based on current expectations. These expectations and information are subject to significant risks and uncertainties and actual results may differ materially from the conclusions, forecasts, or projections in the forward-looking statements and information. Certain material factors and assumptions were applied in drawing the conclusions or making the forecasts and projections reflected in the forward-looking statements and information.

These risks, factors, and assumptions are discussed in our periodic filings with the SEC and applicable Canadian securities regulators, including our most recent quarterly and annual reports, and should be reviewed by anyone considering an investment in our securities. Management disclaims any obligation to update these forward-looking statements and information except as required by law. Now I’d like to turn the call over to Dr. Christian Wiedbrook, Founder and Chief Executive Officer of Xanadu Quantum, who will walk you through our hardware progress, our technical roadmap, and our partnerships.

Rafal will then cover software and Michael will take you through the financials. So with that, Christian,

Christian Weedbrook (Chief Executive Officer)

Great. Thank you, Brett, and good afternoon, everyone. We always like to begin with our mission, which is to build quantum computers that are useful and available to people everywhere. Xanadu Quantum was founded in 2016 to build fault-tolerant quantum computers using photonics, and today we have three generations of hardware released, over 160 PhDs, more than 150 patents and patents pending. We’re proud to work alongside leading industry partners and end users.

At the heart of our strategy is the conviction that photonics is the right foundation for a scalable quantum computer, and a decade of peer‑reviewed research has reinforced our chips compute without cooling, which can be made with silicon manufacturing processes and a compatible telecom infrastructure. Photonics also gives us faster clock speeds, flexibility in how we do error correction, and a computer that is modular and networkable by design. We believe photonics offers advantages over competing modalities across key metrics including clock speed, logical overhead, fidelity, connectivity, and room‑temperature operation. Collectively, these advantages position our approach to achieve fault tolerance and deliver commercially viable quantum computers.

We believe quantum computing represents a significant long‑term opportunity with a potential to create value across a broad range of industries. But realizing that opportunity requires more than developing the technology. We are working with partners and customers across multiple industries to develop real‑world applications and demonstrate where quantum computing can deliver meaningful value. Xanadu Quantum is building the full quantum computing stack as an integrated platform spanning photonic chips, manufacturing, networking and error‑correction software, and end‑user applications.

By developing each layer together, we can manage loss across the entire system, tightly integrate hardware and software, and optimize performance from the chip to the full computer. Since our last earnings call, the Xanadu Quantum team has delivered several exciting achievements that have furthered this mission, and I’d like to highlight five of them. First, on the technology front, we set what we believe is a world record for photon loss in an optical switch, an incredible step forward on our loss roadmap. Second, we signed hardware partnerships with GlobalFoundries, ASML, and Bluefors, which will help us scale our manufacturing capabilities.

Third, on the software side, we launched Backline with AMD and hosted our latest QHack developer conference, which drew a record number of attendees. Fourth, we are deepening our work with industrial and institutional partners. With Mitsubishi Chemical, we demonstrated our quantum algorithms for modeling advanced semiconductor materials. RBC and Booz Allen are using our software for workforce training alongside our work with Lockheed Martin. Finally, we added resources to keep executing, including new support from both the Government of Canada and DARPA.

We are making strong progress in scaling our R&D capabilities and accelerating iteration. All of this puts us on the path to a commercially viable quantum computer by 2029–2030. Before I go through the details, let me reiterate our philosophy and approach. To achieve our mission, we are pursuing a full‑stack strategy that integrates our proprietary photonic quantum hardware with PennyLane, our modality‑agnostic software platform. That path defines our core focus: building a fault‑tolerant, large‑scale quantum computer. This is the defining opportunity for Xanadu Quantum and requires our best people, resources, and attention.

Pursuing revenue before meaningful quantum utility or selling systems prematurely would ultimately divert talent and resources from our most important priority: delivering the hardware roadmap. We measure progress through execution against that roadmap and our key technological milestones, followed by partnerships that accelerate adoption of future commercialization. At this stage, these are far more meaningful KPIs than quarterly revenue. Now, having said that, we believe there may be opportunities to generate revenue from products or licensing built on the components we are already developing or have already developed on our path to building a large‑scale quantum data center. We see this as a low‑distraction way to build a customer base, bring those customers into our ecosystem, and introduce them to our broader offerings, including PennyLane.

Coming back to the hardware, our roadmap puts numbers on that focus. We target up to 200 logical qubits in 2029 with an error rate as low as 10 to the minus 8. The following year we target up to 500 logical qubits with error rates as low as 10 to the minus 16. We believe error rates at this level could be industry‑leading and would open up more use cases than are currently available to other quantum companies. And by 2031 and beyond, we are targeting 1,000 or more logical qubits.

We have a proven track record of delivering on our technological milestones as our Aurora system, published in Nature, linked four server racks into one fully networked quantum computer using optical fiber. Our earlier systems, X8 and Borealis, were published in Nature as well. We believe that combination of room‑temperature operation, manufacturability, and networking gives photonics a clear path to a quantum data center. Reaching those targets comes down to loss, which remains our most important engineering challenge. If we reduce loss below a critical threshold and combine that with error correction, we expect logical error rates to fall dramatically at the networking we have already demonstrated with Aurora, and that should let us scale to logical qubits without a hard ceiling. This is primarily a challenge of engineering, not physics. At our Technology Roadmap event in August, we published our loss roadmap for the first time.

It tracks the aggregate of 17 key optical components, expressed as a ratio of our hardware loss to the loss required for fault tolerance. So a lower number means we are closer. But before I go further, I want to clear up any potential ambiguity. In the loss roadmap we published in August, the loss plot behind it was drawn so that each year marks the beginning of that year, not the end. That means we expect to reach 10.2 times by the beginning of 2027 rather than the end of 2027.

The plot itself has not changed. We’re simply making the timeline it implies clearer. The rest of our roadmap was built on that same loss plot and remains the same. We continue to expect up to 200 logical qubits in 2029, up to 500 the following year, and 1,000 or more by 2031 and beyond. I am pleased to say we have made significant progress against this goal so far and will look to provide an update in the first quarter. Because loss reduction is an iterative engineering process that does not progress in a straight line or necessarily show measurable improvement every quarter, we plan to report our aggregate loss metric annually.

Let me walk through some of the components of that progress, starting with our optical switch. At 7 millidB, we believe it is a world record—the lowest reported loss for an electro‑optic switch of its class—and down from roughly 20 millidB that we reported last year. To our knowledge it is the lowest reported loss for an electro‑optic switch of this class. We would also like to share loss updates on two other key optical components that make up our loss roadmap.

First, our on‑chip homodyne detector, which extracts the result of a quantum algorithm, has shown a two‑times loss reduction since this time last year and now sits at 77 millidB. Second, our fiber delay line, which serves as a quantum memory in our computer, has shown more than three‑times loss reduction since the start of the year. Behind these results is a simple strategy: running more wafers. Every wafer run and tape‑out—meaning each time we send a chip design to be fabricated—is another cycle of learning. More cycles gives us more data, lets us validate performance across a broader range of devices, and helps us turn design improvements into manufacturing results faster. We use two major materials for our wafer runs: silicon nitride and thin‑film lithium niobate.

Year‑to‑date we have doubled our silicon nitride wafer runs and increased our thin‑film lithium niobate wafer runs by 15% sequentially. That creates a positive feedback loop on loss reduction. More wafer runs speed up our learning, and faster learning drives progress. Our partnerships build on that momentum, and this quarter we highlight three. Each addresses a significant part of our roadmap—manufacturing scale, waveguide loss, and cooling. First, GlobalFoundries: this multi‑year partnership will scale our ultra‑low‑loss silicon nitride chips and our superconducting nanowire single‑photon detectors on GlobalFoundries’ 300‑millimeter manufacturing line in Malta, New York. These faster detectors raise our clock rate and reduce the physical footprint and resource requirements of our qubit factory, which is part of our system that generates our GKP qubits.

A high clock rate also lets us use less optical fiber, which reduces total path length and brings us closer to fault‑tolerant threshold. Second, ASML, the world’s leading supplier of semiconductor lithography equipment. To our knowledge, Xanadu Quantum is the first quantum computing company to establish a direct technical collaboration with ASML. We expect ASML’s expertise in advanced lithography to help us reduce waveguide loss, which is the number one cause of optical loss on our integrated photonic circuits.

Finally, Bluefors, to develop compact, modular cryogenics that bring detector cooling into a data‑center footprint. It’s important to note that our architecture requires minimal cryogenics compared with most other approaches to quantum computing, and this work reduces the size and cost even further. This design is expected to eliminate the need for large cryo plants and make quantum computers easier to scale. As a reminder, our qubits, gates, and measurements all operate at room temperature.

Cooling is required primarily for qubit generation, which is essentially the initialization stage of our computer. Partnerships are one way we are building capacity externally; the other is in‑house. This quarter we took a major step with Inception, our new 158,000‑square‑foot advanced photonics research and manufacturing facility in Toronto. Supported by Canadian federal funding under Project Optimism, this facility will feature advanced clean rooms, heterogeneous integration, and packaging capabilities, which are critical technologies for both AI and quantum computing.

Our existing packaging facility has already reduced packaging turnaround times from as long as three months to approximately eight hours, and Inception is designed to bring the same order‑of‑magnitude efficiencies to testing, measurement, and assembly. We expect to begin using Inception early next year. Taken together, these steps give us a manufacturing foundation for a commercially viable quantum data center. Now let me turn the call over to Raf to walk you through our software and partnership progress.

Rafal Janik

Thank you, Christian. Our software strategy is closely aligned with our hardware roadmap Christian just outlined. We are building the software ecosystem, developer community, and enterprise relationships that will drive adoption of our fault‑tolerant quantum computers as they become available. Today, our priority is to grow adoption, establish our platform as the foundation for quantum application development, and build commercial relationships early.

We believe this approach positions us to capture a meaningful share of the quantum computing market as the technology matures. PennyLane remains at the center of that strategy. It is one of the most widely used open‑source platforms of quantum computing. According to the Unitary Fund, approximately 30% of all quantum developers used PennyLane in the past year. Across the PennyLane family, including Lightning and Catalyst, we are seeing approximately 1.1 million package installs per month. PennyLane has also been cited in more than 2,500 academic papers. Our developer community continues to grow.

This year’s QHack, our seventh global quantum coding challenge, attracted record participation with more than three and a half thousand participants. These numbers show that developers are already using software to learn quantum computing, develop algorithms, and explore applications. As quantum computing advances towards commercial utility, we believe this established developer ecosystem will be a significant competitive advantage. As Christian mentioned, this quarter we launched Backline in collaboration with AMD. This extends our software infrastructure for fault‑tolerant quantum computing. Backline is an open‑platform solution that integrates classical CPUs, GPUs, and FPGAs with quantum hardware to support real‑time processing, decoding, and control required for fault‑tolerant quantum computing. Quantum error correction is one of the fundamental challenges in scaling quantum computers.

It requires processing large volumes of measurement data, identifying errors, and providing corrective feedback at extremely low latency. With Backline, we demonstrated end‑to‑end processing round trips of under 3 microseconds using commercially available AMD hardware. This demonstrates the ability to perform real‑time error correction using standard classical computing infrastructure without the dependence on proprietary hardware. Backline is designed to be extensible across different quantum computing architectures, creating opportunities beyond our own hardware platform.

Since launch, we have seen significant interest from national laboratories, classical computing manufacturers, and other quantum computing companies. We believe Backline positions Xanadu Quantum at an important intersection of quantum and classical computing and creates opportunities for broader industry adoption of our software beyond our open‑source ecosystem. We are beginning to commercialize these software capabilities and build direct relationships with enterprise customers. Many of these capabilities originated from tools developed for our own research and hardware programs. With one of the industry’s largest software and algorithms teams, we have built significant expertise spanning photonics, quantum computing, algorithms, and high‑performance computing. This quarter marked an important milestone as we began generating revenue from our commercial software offerings.

With a small but growing base of enterprise customers, we are seeing early demand across several areas. First, enterprise training and workforce development. Organizations increasingly recognize the need to build internal quantum expertise rather than rely entirely on a limited pool of specialized talent. RBC, Canada’s largest bank, has selected Xanadu Quantum to deliver hands‑on quantum computing training for its team. Booz Allen is similarly training an initial group of its engineers using our software platform. Together with our workforce development activities at Lockheed Martin, these engagements demonstrate that major enterprises are beginning to invest in building quantum capabilities using PennyLane as the foundation.

Second, quantum application development and hardware readiness. We are helping enterprises understand where fault‑tolerant quantum computing could deliver commercial value and how specific applications would perform on our future photonic architecture. Through our simulation and resource‑estimation capabilities, our customers evaluate quantum algorithms, estimate hardware requirements and execution times, and assess potential performance relative to classical computing approaches.

This quarter, JSR, a leading semiconductor materials company, signed a statement of work for the feasibility study of these capabilities. These engagements are particularly valuable because they help customers establish a business case for quantum computing while giving us insight into the applications and performance requirements that will matter most as our hardware scales. Third, we continue to expand our commercial software portfolio by bringing capabilities developed internally—including AI‑assisted research tools and high‑performance simulations—to a broader set of customers. Some of these products are already in pilot and early commercial deployment. Our approach is disciplined. We bring products to market early, use customer feedback to validate demand, and focus investment on capabilities that demonstrate meaningful commercial potential.

Alongside these products, we continue to see strong interest from enterprises seeking to identify and develop commercially valuable applications for quantum computing. One such example is our strategic partnership with Mitsubishi Chemical, supported by the Canadian and Japanese national innovation programs. Our collaboration demonstrates the potential for quantum algorithms to accurately model light‑sensitive materials using extreme‑ultraviolet (EUV) photolithography, an important process in advanced semiconductor manufacturing.

We are now advancing that work towards a software pipeline that is ready for fault‑tolerant quantum computing to support the development of next‑generation semiconductor materials. Stepping back, our software business is progressing along three complementary paths. We are expanding one of the industry’s leading open‑source quantum developer ecosystems; we are developing critical software infrastructure for fault‑tolerant quantum computing, including Backline; and we are establishing commercial relationships with enterprises that are beginning to invest in quantum capabilities and applications. Software revenue remains at an early stage, but our progress this quarter demonstrates the growing commercial potential of our capabilities.

Our objective is not to simply build software products for today’s market; it is to establish the ecosystem, technology, and customer relationships that position Xanadu Quantum for long‑term growth as quantum computing becomes commercially viable. With that, I will hand it over to Michael.

Michael Trupek (Chief Financial Officer)

Thanks, Raph. I’ll cover our operating results, our balance sheet, and the funding that supports our roadmap. All figures I cite are in US dollars. Revenue was $1.3 million in the quarter compared to $1.5 million in Q2 2026 and $1 million in the prior year quarter. Our revenue will fluctuate from quarter to quarter based upon the timing of work performed on our various contracts and partnership-related revenue. As Christian mentioned, revenue is not expected to be a material driver of our operating activities.

In the near term, our focus remains on executing against our hardware roadmap. R&D expense was approximately $25.6 million, an increase of $5.9 million from Q2. This increase was primarily driven by strategic investments in our wafer spend to support upcoming tape-outs, process development activities tied to our new foundry, collaborations with GlobalFoundries, and to expand our engineering talent, which also contributed to higher compensation-related costs. As we said last quarter, we continue to expect R&D spend to increase over time as we add engineering talent and increase chip development, packaging, and improve all of our subsystems toward fault tolerance.

G&A expense was approximately $9.2 million compared to $11.1 million in Q2. The sequential decrease in G&A expenses was primarily driven by lower professional fees incurred during the quarter, mostly related to transaction and regulatory fees. The quarter included approximately $1.1 million in nonrecurring costs compared to $2.3 million in Q2. Net loss was $32.2 million compared to $42.1 million in Q2 and $20.3 million in Q3 2025. The lower net loss compared to Q2 primarily reflects the absence of the $10.6 million fair value loss in earn-out share liabilities and the $1.9 million loss on derivative instruments related to our synthetic ATM, both recorded in Q2. The earn-out was settled in Q2. Adjusted EBITDA loss was $26.3 million compared to $21.3 million in Q2 and $15.4 million in Q3 2025.

The increase compared to Q2 was driven mainly by higher R&D spending to scale chip fabrication, including wafer consumption and headcount, offset by lower G&A. Capital expenditures, which we define as the purchase of PPE, were approximately $8.4 million compared to $6.1 million in Q2, primarily reflecting investment in specialized equipment to test, refine, and scale our chip development and packaging capabilities. In addition to significant investments refining the subsystems that will enable our fault-tolerant quantum computer, we also began incurring initial costs related to the development of our new Inception facility.

We ended the quarter with 361 employees, up from 306 at the end of Q2 and 277 at the end of Q1. We ended the quarter with $286.5 million of cash, down $26.3 million from Q2. We did not draw on our synthetic ATM facility in the quarter. We raised $67.2 million in Q2, leaving approximately $232.8 million available for use today. We also strengthened our funding position with nondilutive capital. We signed a definitive agreement with the Government of Canada for approximately $140 million of federal funding which funds Inception.

As Christian described, these funds are expected to be received over time as qualifying R&D investments are made rather than upfront as cash, and we announced provisional approval for up to $10 million of additional project funding under DARPA’s QBI. This is a one-year contract extension that builds on our Stage B work, and we expect to allocate the majority of the funding to accelerate critical components of our R&D roadmap. Looking ahead, we expect costs and capital expenditures to grow in Q4, driven by additional spend with GlobalFoundries, continued hiring, and equipment purchases. We expect continued headcount growth across hardware, software, and G&A. And with that, I’m pleased to turn it back to Christian.

Christian Weedbrook (Chief Executive Officer)

Great. Thank you, Michael. During the quarter we made meaningful progress. I think. Operator, we’ll turn the call over to Q&A now. I’m sorry. Go ahead, Operator.

Jordan (Conference Operator)

As a reminder, if you’d like to ask a question during this time, simply press star followed by the number one on your telephone keypad. You are limited to only two questions. We’ll take a brief moment to compile the Q&A roster. Your first question comes from the line of Kingsley Crane from Canaccord Genuity. Your line is live.

Kingsley Crane (Analyst)

Hi, thanks for taking the question. I want to ask on what you’re doing in the AI data center. So with HyperLight you mentioned that the ultra-low-loss switching platform could provide some power efficiency advantages in an AI data center. We’ve also seen another quantum photonics provider kind of pursue maybe a similar approach. So we’d just love to hear more on that, and just what the near-term opportunity could be, just any relationships with the customers yet.

Christian Weedbrook (Chief Executive Officer)

Yeah, absolutely. This is something that’s not just limited to this specific piece of IP around the thin-film lithium niobate switch. But I’m going to say broadly, one of the unique opportunities that Xanadu Quantum has is that we’re sitting on a large amount of photonic IP developed over the last 10 years that has direct applicability in datacom/telecom on the classical side, and then also on the quantum side in quantum communications and sensing.

So we are definitely exploring each and every one of these opportunities as they come up, both for existing partner companies and then potential other customers as well. But it’s a fine balance, making sure that we do not distract from the overall technical roadmap for us. Ultimately we want to build that utility-scale quantum computer that can drive the huge amount of value and societal change. If the opportunity arises that one of these devices would be commercially viable without the distraction from our goal, we would definitely pursue it.

Kingsley Crane (Analyst)

Okay, helpful. Thank you. And then with the GlobalFoundries partnership, I believe you had mentioned that you had adopted nanowire single-photon detectors, and I believe that the GKP result in Nature had used transition-edge sensors. So is that a change? Is that an evolution? Are those like mutually exclusive or—just would love to hear more about the evolution of the platform. Thanks.

Christian Weedbrook (Chief Executive Officer)

Yeah, thanks. So that’s a really good pickup. So overall, both of these devices, whether you use TESs or whether you use SNSPDs, their role is as a PNR—photon-number-resolving—detector. That’s what we actually need for our computer. So we were using TESs, and then we wanted to upgrade them. So basically think of it as a next-generation advancement, and we’re now using SNSPDs. The reason why we did that is really a great upgrade to our previous blueprint.

Basically now it allows you to have a far faster computer. That’s the important thing. SNSPDs are orders of magnitude faster than TESs, but that speed also allows you to reduce the footprint of the quantum computer as well. So yeah, good pickup. That’s exactly what we’re doing. Thank you, Operator.

Jordan (Conference Operator)

Your next question comes from the line of John McPeek from Rosenblatt Securities. Your line is live.

John McPeek (Analyst)

Thank you, guys. Congrats on another quarter behind you, Christian and team. Two questions. First one, on the switch that you are working on with HyperLight, how does that—I’m not an optical datacom/telecom expert—how does that compare to what’s in the market? You know, you mentioned 7 millidB. How much better is that than commercially available products? And I just have a follow-up.

Christian Weedbrook (Chief Executive Officer)

Yeah, it’s actually better on two different areas. One of them is definitely on the loss. So this is quite a bit lower loss. Usually the amount of loss that typical optical circuit switches are able to tolerate is quite a bit higher than that. But what it means that you’re able to have then is a much larger number of channels going into a switch. So for us this is important because of the way that we select the qubits that we want to within our quantum computer.

We want to be able to, of course, keep as much of that light, but we still need a certain number of channels going in and be able to select from them. The classical datacom guys can handle a lot more loss. But of course, they want to move away from just 16 or 128 channels to as many channels as possible to drive their workloads. The other part that might have been missed is that it’s not only that it’s a lot lower loss, it’s also much faster. So if you look at competing technologies like MEMS liquid crystal switches, or even some of the integrated silicon photonic switches that might have been recently announced, they’re operating on reconfiguration times of milliseconds. Our switches operate at times of microseconds or lower. So this is something that’s incredibly important for our quantum computer, where we’re operating at megahertz to hundreds of megahertz.

We need to be able to reconfigure rapidly to select the correct qubits. But of course, this has a big impact on classical workloads as well.

John McPeek (Analyst)

And what do you think the market opportunity is for this device?

Christian Weedbrook (Chief Executive Officer)

I mean, within computing, it’s definitely quite large. In quantum computing, this is a critical component for our device. I personally wouldn’t want to comment on the market size of optical switching overall, but of course, it’s something that the team is keeping a close eye on to understand if, again, there is an opportunity for us to commercialize various different products based on this technology. We do know it’s increasing.

Jordan (Conference Operator)

Your next question comes from the line of Nahal Chokshi from Northland Capital Markets. Your line is live.

Nahal Chokshi (Analyst)

Thank you. Hey, that was a really nice press release yesterday with HyperLight. You guys talked about it a lot already. The context that I’m looking for, though, is being able to frame up the impact to the roadmap. And if you could define how many optical switches are present in Xanadu Quantum’s quantum computer, that would be really helpful in framing up the impact of this pretty strong technical release that you guys did yesterday.

Christian Weedbrook (Chief Executive Officer)

Yeah, we don’t give that specific number out publicly. I can say that the broader picture is that we have a lot of them, and there’s, as you know, there’s approximately 17 key optical elements where loss propagates through, and we need to reduce the loss overall. As mentioned in our talk, we’ll be giving updates yearly. And I can say big achievements like this that drives down the loss and is something we believe is a world first, helps us achieve and stay on our roadmap.

So that’s kind of how, you know, we’d like to comment on that question.

Nahal Chokshi (Analyst)

Okay, understood. And then the non-GAAP OPEX run rate for the September quarter—it was actually substantially below what we thought you would ramp to, and it does sound like you are going to continue to ramp. I would just like to get a sense as far as, you know, what’s the OPEX run rate you expect to be at once you get to the R&D scale that you are planning for.

Michael Trupek (Chief Financial Officer)

Yeah, thanks, Nehal. Good question. You know we continue to not provide forward-looking guidance. I will just reiterate what I said during my opening comments, which we expect it to grow during Q4, and you will see those results next quarter.

Jordan (Conference Operator)

Your next question comes from the line of Paul Traber from RBC Capital Markets. Your line is live.

Paul Traber (Analyst)

Thanks very much and good afternoon. Just wanted to switch gears to the software side. I was hoping to get your thoughts on the competitive implications of Intel acquiring Classiq, and then related to that, how do you see the various segments of the quantum software stack evolving over time? Meaning, do you see greater or lesser value at certain areas of the stack versus other areas?

Rafal Janik

Yeah, fantastic question and I’m going to say the ClassIQ acquisition is a great proof point of what we’ve been saying all along, which is that the software portion does represent a very large part of the problem and we’re very happy to have a leading platform again with over 30% of all quantum developers developing on top of it. In PennyLane, there are three different distinct portions of that stack. The first one is that user space, the one where the developers actually work.

The Software Development Kit for quantum applications. This is what we refer to as PennyLane. Within that there’s a whole host of tooling that allows you to develop algorithms faster and test them more rapidly and with greater accuracy. We’ve been developing this now for over eight years and really do have that leadership position. One that’s equally important to that is the middleware. Often we refer to it as compilation. If you’re following our stack, this is the project called Catalyst. We do again capture it under the broad term or brand of PennyLane. This is a really exciting area of research right now in that we’re finding order-of-magnitude improvements in terms of the efficiency that you get both in runtime and in the scale that you need of a quantum computer, by being able to optimize the compilation from what a developer writes to what would actually fundamentally run on the hardware.

And then the last part is really the part that we’ve entered most recently with Backline, which is the fundamental infrastructure that glues together all the subsystems needed to actually drive a quantum computer. A quantum computer is not just a QPU on its own, it’s a collection of classical and quantum compute resources working in concert. They require various latencies going down to ultra-low latency to support quantum error correction. And realistically, we’ve now built out the most performant platform to support that with Backline.

Backline is critical for us building our quantum computer. But of course, as quantum computers become more and more available, we will see more and more of that value push into the software layer. And we’re very happy to again have that leading framework to be able to broadly capture it across the entire ecosystem.

Michael Trupek (Chief Financial Officer)

You know, as we’ve said on previous calls, we always are looking at the market conditions and the valuation and we evaluate it almost on a day-by-day basis.

Jordan (Conference Operator)

Your next question comes from the line of Todd Coupland, CIBC. Your line is live.

Todd Coupland (Analyst)

Great. Good evening everyone. I wanted to step back from all the sort of the revenue crumbs that you put out during this discussion, whether it’s optical switches or software monetization, but then you sort of pull it back saying you don’t want it to be a distraction. So what should our takeaway be? There’s probably some, you know, revenue coming in from these areas before commercial quantum computer 29–30. Or is it just to highlight the significant IP in the company?

So I’m trying to process what you’re trying to say when you put this IP out there. But then you say we’re not going to be too distracted by it, so just give us a little more context on that.

Christian Weedbrook (Chief Executive Officer)

Yeah, I would say it’s still consistent in the sense that the things that we’re developing, we do patent, we do want to make sure we own that. We continue to develop PennyLane. And those two things are really good examples where regardless of whether we monetize them, we’ll still be doing them. And they’re very key for our ultimate vision of a large-scale quantum computer, what we’re calling a quantum data center. So the opportunities may or may not come up when it comes to things like monetizing PennyLane. We’ve mentioned in the past that we’re definitely open to that. Our first seven or eight years was really all about building the ecosystem around PennyLane, having really lower barriers or no barriers for people to adopt PennyLane. And we’ve been very successful with that. So now that we’ve built the ecosystem, it is an option for us to think about how to monetize it.

And it’s something we will think about. And the same with the licensing and the IP of any of our hardware. Our secret sauce and why we’ve been so successful is really the extreme loss reduction side of things, as we mentioned with some of our optical components this earnings call. So any of those, if we think about possibilities of productizing them, knowing that we need to do that anyways in terms of the low loss, or if someone comes to us and likes optical loss and how low it is, we’re open to those discussions.

What we aren’t in favor of, maybe another way to answer your question, Todd, is we’re not in favor of going into quantum security, quantum sensing, quantum internet or things of that nature. We think they’re great fields, but I think people sometimes may overestimate how hard it is just to focus on one thing and build something humanity’s never built before from a technology point of view, which is a large-scale quantum computer. So we have our hands full with that.

And I think that if a company like Xanadu Quantum is successful in that, that’ll be the biggest market share and the biggest market TAM overall.

Todd Coupland (Analyst)

Great. And then on the technical updates you know you had last quarter, you had some material ones recently with optical loss. Just give us an idea. What is the impact to DARPA C and a sense on a timing decision relative to Xanadu. Thanks a lot.

Christian Weedbrook (Chief Executive Officer)

Yeah, sure. You know, DARPA obviously likes to hear about these results and knows about them. They understand also loss is the biggest challenge. I would say that’s one part of DARPA. DARPA has other criteria that a company like Xanadu Quantum and anyone else needs to meet as well. As they’ve publicly stated they want to keep an eye on the technology and how it’s progressing. So we’re progressing very well. We’re in very close contact with them. I believe the year of the program is around November, December, but it could always be a little sooner as we saw four companies go into the next stage or it could be later as well.

And a lot of that is built on timing. It’s built on DARPA’s availability. And so we still have things that we need to report to DARPA. So no new update from that, but we’ll definitely keep you posted if we hear anything.

Jordan (Conference Operator)

Your next question comes from the line of Desmond Webster. Desmond’s line is live.

Desmond Webster (Analyst)

Hi, good evening everyone. Thank you for taking my question. So your algo team just figured out how to dramatically reduce the Toffoli count for QROM, you know, which is awesome. So I’m curious, has something your algo team discovered ever made your hardware engineers go back and change what they were building? Like, I mean, what’s the best example of that happening at Xanadu?

Rafal Janik

Yeah, it’s a fantastic question. So first of all, that QROM, there have been multiple discoveries along the path. One of them was definitely developed. The other one that I really want to highlight is that it was driven through an agentic AI discovery. So we’re able to do something in two weeks that would have kind of classically taken us more than six months, manual labor-wise, to do. And that QROM advancement was really a great proof point of how much this entire field is accelerating and how the tooling we’ve been building internally has been able to push that forward.

To your other part of the question of what is maybe the co-design between our hardware architecture and our algorithm development. It is a very synergistic process and one thing that might not be apparent from the outside is that in addition to all the fantastic IP we’ve built out on algos and hardware, one of the real strengths of this company is the digital twin that we’ve been able to build of our system. So when algorithms is now developing a new approach, they’re able to propagate that down to understand how that will impact not just the underlying architecture and the resources required to run that algo, but also the error-correcting algorithm that will be implemented. And similarly, as our error-correction experts come up with new approaches and new logical architectures, we are able to very quickly and clearly propagate that into resource estimates of our top algorithms.

So I’m going to say it’s a very synergistic activity between those groups and we’re continuously advancing the tooling to tie that closer and closer together to be able to more quickly iterate on that loop.

Desmond Webster (Analyst)

Excellent. Great. Thank you everyone.

Jordan (Conference Operator)

A final question comes from the line of Ella Tulchinski from Morgan Stanley. Your line is live.

Ella Tulchinski (Analyst)

Great, thank you. You mentioned packaging turnaround fell significantly. So I was wondering, as Inception comes online, what part of the process do you expect to speed up the most and what does it mean for how fast you can move on loss?

Christian Weedbrook (Chief Executive Officer)

So definitely packaging will continue to accelerate, but maybe more exciting than just the turnaround time is actually the volume that we’ll be able to do in order to hit our next technical milestones that were kind of addressed in our analyst day presentation. The volumes have to increase and the facility we have here at our existing headquarters is just going to be insufficient to be able to do that. The other part where we’re definitely increasing capacity by a huge amount is on test and measure.

The volume of chips that we’re getting through every quarter now is ever increasing. We are sitting on tens of terabytes of data: designed, simulated, manufactured and tested chips. This all is able to feed back into our agentic design and expert design systems to generate the next iteration of manufacturing runs. But all this is predicated on having that test and measure capacity. And finally the third major aspect of that Inception facility is a newer push which is into heterogeneous integration. And this is really a capacity that doesn’t exist at this level in quality anywhere in the world that is accessible to us. And this is the ability to take silicon nitride, thin-film lithium niobate and our III-V semiconductor-based photodiodes and actually integrate them into a single package that will be necessary for building out both our QPU and the switches required for a quantum computer.

Great. Thank you.

Jordan (Conference Operator)

That concludes our question and answer session. I’ll now turn it back over to Dr. Christian Weedbrook for a closing remark.

Christian Weedbrook (Chief Executive Officer)

Great. Thank you for that. Thank you for everyone for your attention today. We really appreciate it. Just a quick summary of today. During the quarter we made meaningful progress across our business. On the technology front, we set what we believe is a world record for photon loss in an optical switch, advancing ultimately our loss roadmap. We also established important hardware partnerships with GlobalFoundries, ASML and Glyphosate to support our manufacturing ambitions.

On the software side, we launched Backline with AMD and hosted a record-setting QHack developer conference. We continue to deepen our relationships with industrial and institutional partners including Mitsubishi Chemical, RBC, Booz Allen and Lockheed Martin, demonstrating the growing range of real-world applications for our technology. Finally, we strengthened our financial resources with additional support from the Government of Canada and DARPA providing further support for execution against our roadmap.

Thank you again, I appreciate it.

Jordan (Conference Operator)

That concludes today’s meeting. You may now disconnect.

Disclaimer: This transcript is provided for informational purposes only. While we strive for accuracy, there may be errors or omissions in this automated transcription. For official company statements and financial information, please refer to the company’s SEC filings and official press releases. Corporate participants’ and analysts’ statements reflect their views as of the date of this call and are subject to change without notice.