Home / Transcripts / Xanadu Quantum Technologies Limited (XNDU) · May 14, 2026

Xanadu Quantum Technologies Limited (XNDU) Earnings Call Transcript

May 14, 2026

Toronto Stock Exchange CA Information Technology Software earnings 45 min

Earnings Call Speaker Segments

Operator operator
#1

Just a quick note that this call is being captured for later playback. At this point I'll hand things over to Brett Harris, our Vice President of Investor Relations. Brett, go ahead.

Brett Harris executive
#2

Thank you, and good afternoon to all of you. It's a pleasure to welcome you to Xanadu's first-ever quarterly earnings conference call. With me today are Dr. Christian Weedbrook, our Founder and Chief Executive Officer, and Michael Trzupek, our Chief Financial Officer. This call is also being webcast live, and a recording will be available on Xanadu's website for a period of time. The information we're sharing today is time-sensitive and accurate only as of the date of this live webcast, May 14, 2026. On this call, we may make forward-looking statements that reflect current expectations. Such statements rest on management's present beliefs and expectations as of today, and they carry a number of significant risks and uncertainties, so our actual results could differ materially. Anyone thinking about investing in the company's securities should refer to the discussion of these risks and uncertainties in the company's periodic reports filed with the SEC. Xanadu undertakes no obligation to update any of these statements. At this point, I'd like to hand the call over to Dr. Christian Weedbrook, Founder and CEO of Xanadu. Christian.

Christian Weedbrook executive
#3

Thanks, Brett. Good afternoon, all. Christian Weedbrook here, Xanadu's founder and CEO. Back in 2016 I launched this company on the strength of a white paper and the belief that photonics offered the correct route to a scalable quantum computer. Ten years of research and peer-reviewed breakthroughs later—and now with a public listing on Nasdaq and the Toronto Stock Exchange—that conviction has only grown stronger. For anyone new to our story: Xanadu is a Canadian quantum computing company driven by one mission—to build quantum computers that are useful and available to people everywhere. As a full-stack quantum computing company, we develop both the photonic hardware and the software required to run it. Our approach, which relies on photons, or light particles, we believe provides a distinct and durable path toward a large-scale quantum computer and a quantum data center capable of solving some of the world's most complex and pressing challenges. I want to pause briefly to acknowledge what this call represents—not merely a financial reporting milestone, but a public commitment to transparency, to accountability, and to delivering on our technological roadmap for building scalable, fault-tolerant quantum computers. I'll take a few minutes on Q1 highlights and the state of the business, then hand things to Michael for a deeper look at the financials.

Christian Weedbrook executive
#4

On March 27, 2026, Xanadu started trading as a public company, listed on both Nasdaq and the Toronto Stock Exchange, where it goes by the ticker symbol XNDU. That came after we finished our business combination agreement with Crane Harbor Acquisition Corp. Going public meant more to us than just a capital event. It sent a signal to the scientific investment communities that photonic quantum computing has moved past the experimental R&D stage and into a phase of scaling, where progress depends less on theoretical physics and more on practical engineering. The go-public transaction brought in $302 million in gross proceeds. Combined with roughly CAD 285 million of anticipated Canadian and Ontario government funding that is currently under negotiation, that accelerates Xanadu's path toward building a quantum data center by 2029, 2030. We hosted our inaugural Analyst Day on March 4 in New York City, and there we showcased technical progress, roadmap updates, and a deeper view of our commercialization strategy. To those who joined us, thank you. If you weren't able to attend, the materials are available on our investor relations website. Before I go deeper into our first quarter highlights, I'd like to give a broader overview of our technology for those new to Xanadu, along with its differentiators and our progress to date.

Christian Weedbrook executive
#5

Everything we do is guided by a single objective: to create quantum computers that are genuinely useful and accessible to everyone, everywhere. Let me take a step back and lay out our strategy. Here at Xanadu, we hold the view that photonics represents among the most credible routes for scaling quantum computing. There are five key reasons for this. First, the architecture of our system is modular and networked, which lets us scale systems via interconnects, in much the same way that modern data centers operate now. Second, we take advantage of the semiconductor and photonic supply chains that already exist, instead of constructing a bespoke manufacturing ecosystem from scratch. Third is computation at room temperature. Any approach requires some kind of cooling, be it cryogenics or laser cooling. At Xanadu, cooling is needed only to initialize, or turn on, our quantum computer, and beyond that point, the qubits, gates, and measurements all happen at room temperature. Fourth, photonics makes real-time and flexible error correction possible. This affords Xanadu more freedom to implement sophisticated error correction schemes that are hard to achieve in other architectures. Specifically, our error correction enables low-overhead approaches that cut down the resources needed for fault-tolerant quantum computing. Lastly, photonics delivers a fundamental edge in gate speed, with operations taking place at frequencies of 100 megahertz.

Christian Weedbrook executive
#6

That's somewhere between 10 and 1,000 times quicker than competing modalities. In the end, two main obstacles sit between this industry and a quantum data center: raising the performance of quantum computing modules so they achieve fault tolerance, and linking those modules together via scalable network systems. Fortunately, the networking hurdle is already behind us. Our Aurora system — shown in Nature — is the first modular, scalable, networked photonic quantum computer in the world, with 12 logical qubits spread over 35 photonic chips and 13 km of fiber optics, running at room temperature with real-time error correction decoding. Now that networking is handled, our main focus is boosting our photonic chips' performance by cutting loss via engineering enhancements. Progress remains genuine and ongoing. During 2025, we cut optical loss by 60% year-over-year. That represented a step function improvement, and we keep pushing forward. Our targets: a qubit factory in 2026 to 2027, fault-tolerant operations in 2028, a quantum data center in 2029 to 2030, and scaling to as many as 500 logical qubits from 2029 to 2030, while upgrading our manufacturing capabilities to keep pace. Hardware isn't where we stop. Companies such as NVIDIA have shown us that hardware by itself doesn't create transformational businesses.

Christian Weedbrook executive
#7

On top of that sits a software ecosystem that brings out the hardware's full usefulness and draws customers back. That is the reason we built PennyLane — an open source quantum software platform intended to serve as the second major pillar of our company strategy and a long-term competitive edge. Through PennyLane, developers, researchers, and enterprises gain access to quantum computing and explore it today while getting ready for the fault-tolerant hardware of tomorrow. The uptake is such that we now have 150 university partners around the world, among them University of Toronto, Johns Hopkins University, and the University of Maryland. Our three-year partnership via UMD's National Quantum Lab centers on workforce development and commercialization throughout the Mid-Atlantic. We are preparing the next generation of quantum scientists and engineers to program our platform. Those developers learning quantum computing on PennyLane now will turn into tomorrow's customers, partners, and colleagues. PennyLane ranks among the most widely used quantum software platforms. The adoption growth we see — over 35,000 active users and 200,000 monthly downloads as of the Analyst Day — tells us the quantum developer community is expanding quickly and Xanadu sits at its center. Although it is free today, enterprise engagements such as our research initiative with Lockheed Martin to advance the foundations of quantum machine learning create a path toward monetization down the road.

Christian Weedbrook executive
#8

Over in Canada, during March, we announced that negotiations are underway for as much as CAD 390 million in funding from the governments of Canada and Ontario, aimed at launching Project Optism and building out advanced semiconductor and photonic manufacturing infrastructure that would support Canada's quantum supply chain. Those talks are still in progress, and we're eager to keep investors informed as things move forward. During the quarter, we unveiled major new alliances with AMD, Lockheed Martin, TELUS, and Fidelity Center for Applied Technology, pushing commercialization forward in defense, finance, and telecommunications. These come on top of the 2025 partnerships we formed with Mitsubishi Chemical, Rolls-Royce, Riverlane, Corning, and Applied Materials. This ecosystem is deliberate. Quantum cannot thrive on its own. Working closely with end users is precisely how algorithms become real-world applications, and how we make sure the systems we build are practical and commercially valuable. Take AMD, for instance: together we're pushing hybrid quantum-classical computing for aerospace and engineering simulations, employing GPUs to run 20 qubit, 35 million gate quantum computational fluid dynamics simulations and speeding up workflows by 25 times compared with CPUs. Our client and research partnerships matter enormously for driving useful applications and demand, yet our large-scale hardware manufacturing partnerships — like AMD, Corning, and EV Group — represent a key structural advantage. Such partners contribute deep hardware engineering know-how and make rapid iteration at scale possible by drawing on proven high-volume manufacturing processes and existing production lines.

Christian Weedbrook executive
#9

With that, we can speed up development cycles, cut execution risk, and put capital to work more efficiently — which ultimately lets us put more money into chip production while bringing down cost per iteration. Pairing photonics-based hardware innovation with an easy-to-use software platform forms the basis of our strategy: deliver near-term utility while keeping a credible path toward fault-tolerant quantum advantage. Building quantum computers that are useful and available to people everywhere is our mission, and every investment we make is guided by it. Now, I'll hand things over to Michael, who will go over our key financial highlights.

Michael Trzupek executive
#10

Thanks, Christian. Before I take you through how we performed this quarter, I'd like to point out that I came to this company to help build out a rigorous financial and operational framework — one that underpins the long-term roadmap we're pursuing through 2030 and beyond. Our top priority in the coming few years is making headway on the roadmap Christian described and reaching as many as 500 logical qubits in 2029 and 2030, but at the same time we're establishing the groundwork for opportunities that generate revenue via a structure with several pillars: subscription-based cloud access, or quantum computing as a service, QCaaS; a SaaS-style revenue model for enterprise-grade PennyLane; sales of on-premise systems; IP licensing of our photonic subsystems; and possibly monetizing strategic partnerships. Now to our financial position: we closed the quarter holding CAD 272 million in cash. Our go public transaction brought in $302 million in gross proceeds, and together with roughly CAD 285 million in expected Canadian government funding, that substantially speeds up our path to constructing a quantum data center by 2029, 2030. With this capital base, we have the flexibility to execute on both our technical and our commercial priorities. One more thing worth noting: the financing support from the Canadian government won't show up as cash on our balance sheet at the outset.

Michael Trzupek executive
#11

What happens instead is that we anticipate receiving those funds gradually, in line with qualifying investments made as R&D work takes place. Our plan is to establish a CAD 300 million synthetic ATM facility, giving us room to raise capital as opportunities arise over time. Having assessed a variety of financing options, we think this structure offers what we anticipate will be the most cost-effective and least dilutive route for funding the business across the long term. To be clear on this point, the synthetic ATM shares will constitute a primary issuance, and every dollar of proceeds will go straight onto the Xanadu balance sheet, building up our cash reserves. Our strategy will be disciplined and opportunistic, and any capital raised would go toward supporting continued development and scaling of our quantum computing platform and technology roadmap. Importantly, having access to capital lets us innovate at a faster pace. We are speeding up commercial chip runs with our foundry partners, which raises iteration speed and enhances device performance within a shorter cycle. At the same time, we are growing our engineering organization through a deliberate shift toward bringing on more systems, software, and hardware engineers alongside our world-class physics team, in support of the move from research into product implementation. Let's turn now to our financials.

Michael Trzupek executive
#12

Revenue rose to CAD 2.8 million, up from CAD 700,000 in the same quarter last year, mainly because of revenue we recognized from our DARPA QBI Stage B participation. On top of that, we recorded roughly CAD 5 million in grant income tied to the Canadian Quantum Champions Program, along with other operating income. In our view, backing from both DARPA and the Canadian government strongly validates our research, our technology roadmap, and the commercial and operational promise of our platform. R&D expense grew to about CAD 17 million, versus CAD 10 million in the prior year quarter, which reflects ongoing investment in engineering talent, more wafer iterations produced together with our foundry partners, and broader process development work. G&A expense came in at approximately CAD 10 million for the quarter, and that figure includes about CAD 5.5 million of non-recurring costs linked to our recent SPAC transaction and our public listing on the Nasdaq and Toronto Stock Exchange. Adjusted EBITDA loss stood at CAD 13.9 million, compared with a loss of CAD 10.6 million in the prior year's quarter. Looking ahead, we anticipate that spending will rise in the coming quarters as we keep adding engineering talent and speed up our R&D investment through wafer and chip production. Let me briefly clarify our share count as of March 31st, 2026.

Michael Trzupek executive
#13

At the close of the quarter, we had 298 million shares outstanding. That breaks down as 255 million Class A shares and 43 million Class B shares. Among the issued shares, more than 250 million are held by legacy Xanadu shareholders, and those remain bound by the 180-day lock-up provisions put in place as part of our SPAC transaction. Those shares won't be freely tradable until September of this year. Given a strong capital position, expanding backing from both government and commercial partners, and a disciplined approach to scaling the business, we believe we're well-positioned to execute against our technical roadmap and long-term commercialization strategy. We remain focused on accelerating innovation, expanding our engineering capabilities, and advancing toward large-scale fault-tolerant quantum computing, while thoughtfully managing capital deployment and preserving financial flexibility. We're excited about the progress we're making and the opportunities ahead. With that, operator, we'd now like to open the call up for questions.

Operator operator
#14

Thanks. At this point we'll open things up for questions and answers. Our first question is from Nehal Chokshi at Northland Capital Markets. Go ahead, please.

Nehal Chokshi analyst
#15

Thanks, and congratulations on your debut earnings call—that's exciting. The Beatles tune "PennyLane" keeps playing in my head right now. With 35,000 monthly users, how does that stack up against a year ago? Could you clarify whom you see as PennyLane's nearest competitor? And where do they sit when it comes to monthly users too?

Christian Weedbrook executive
#16

Yes. Thank you, Nehal. On that specific comparison, we can follow up with you later. Broadly speaking, though, we can say it has been rising — certainly when you average across a seven- to eight-year stretch. That dovetails nicely, I think, with the second part of your question. Based on surveys and on what we hear from partners and people in the field, the two leading ones really are PennyLane and Qiskit, which is IBM's software offering. What we like about PennyLane, though, is that it is, if not the most widely used, then one of the most widely used agnostic platforms anywhere in the world. As you know, it runs on many different hardware providers, not only the photonic-based approach — but those are the top two. I would also mention Nvidia; they came late to the game, but they are doing some great work as well. PennyLane and Qiskit are definitely up there.

Nehal Chokshi analyst
#17

Thanks. I appreciate that. I’d like some clarity on this point—would Qiskit also count as a platform that works independently of hardware?

Christian Weedbrook executive
#18

That's a fair question. I'd put it this way: it's mainly focused on superconducting qubits. As you're aware, it's IBM's, and the real aim is to encourage and bring in as many users as possible for superconducting qubits. A few other platforms do offer it. As far as we know, none is available as broadly or in as hardware-agnostic a way as PennyLane.

Nehal Chokshi analyst
#19

Understood. All right, regarding hardware, could you provide any insight into how linear the decline in optical loss will be, moving from what I believe is currently 23 DB down to a target of roughly 1 DB across, I think, a three to four year period?

Christian Weedbrook executive
#20

It does appear linear, yet achieving that is frequently something you have to labor extremely hard for over a long period. At present, what we're seeing are incremental gains. Since roughly 2021, 2022, you can observe a reduction of up to 200 times along a single path. From the point we're at today, a 200 times reduction isn't going to happen, because we're approaching that line quite closely — the one normalized to one. At that point you reach threshold, and you have to go somewhat beneath it so as to be, just as the term implies, below threshold. That is the point where you can begin addressing significant customer problems. A major factor behind our raising capital and becoming public is that we're able to conduct increasingly more chip runs. Those chip runs truly rank among the essential elements for cutting loss. Frequently, if you pull it out, it's somewhat quasi-linear. It appears that way; chip runs may require anywhere between two and five months. After those come back to us, we test and measure the wafers. We figure out which improvements are needed past that point, then we revise our chip design and dispatch them again to various foundries.

Christian Weedbrook executive
#21

A lot of this happens via chip runs. Fiber-to-chip coupling is another area where we'll see gains and have seen them previously as well. Cutting those losses—how you get light into and out of the chips—matters a great deal. We're highly confident we can reach that line for us, hit it, and move past it. It simply makes it all the more certain, given that we raised a substantial amount of money by going public.

Nehal Chokshi analyst
#22

Okay, that sounds good. I'll hand things over now.

Christian Weedbrook executive
#23

I appreciate that.

Operator operator
#24

Kingsley Crane from Canaccord has the next question — go right ahead.

Kingsley Crane analyst
#25

Hello, and thanks for taking my question. I'd also like to add my congratulations on a successful first earnings call. Regarding the NVIDIA Ising announcement, it certainly validates the quantum sector. It helped lift the quantum sector. You know, I think the underlying tools are more modality agnostic, but photonics is more of an architectural advantage, along with clock speed and a different error correction mechanism than, you know, superconducting or neutral atoms. I would just love to get your thoughts on that announcement in general, and then just from a fundamental standpoint, how you think photonics could or could not benefit from what the work NVIDIA is doing there.

Christian Weedbrook executive
#26

That's an excellent question. One commitment we made from the very beginning was to be as transparent as we possibly can, and hopefully this response conveys that spirit—giving you a genuine sense of which things are substantive and which aren't. It's encouraging to see Nvidia involved here. They've been putting in serious effort for 2 years across numerous initiatives. As they've explained to us, their ultimate aim is straightforward: move more GPUs, which is entirely logical. Quantum companies want those GPUs for simulating algorithms. For benchmarking purposes. They're also interested in using them for error correction and fault tolerance. Now, it appears that stocks—Xanadu among them—had people hunting for patterns and concluding that many equities climbed thanks to that Nvidia news. We certainly don't want to question good fortune; seeing that was wonderful. But frankly, the underlying substance, in my honest view, didn't truly justify those share prices rising to the degree they did. I believe the bigger factor was simply that Nvidia is genuinely expanding its efforts in this area, which is a positive development.

Christian Weedbrook executive
#27

From Xanadu's perspective, what they actually provided wasn't particularly compelling to us, even if it was promoted as agnostic. There's a solid reason for that: the tools they put out we've already been using in-house for our photonic system for five or six years now. That was one piece of it. The other piece is a genuinely critical point. When you're building a large-scale quantum computer, you'll have plenty of quantum components, but you'll also need a great deal of classical compute. The competing approaches, when set against ours and Xanadu's photonic-based method, are considerably slower. And since they're slower, they're actually able to rely on NVIDIA's GPUs for the error correction portion. For our error correction, NVIDIA's GPUs are off the table, because we're too fast — which is a truly remarkable thing. Our clock rates beat ion traps or neutral atoms by, I believe, as much as 10,000 times. To put that in perspective, consider a problem those guys can crack. They get it done in a year — fantastic, since it would otherwise have taken millions of years. We can then crack that same problem in about eight hours.

Christian Weedbrook executive
#28

I mean, eventually that’s going to be our real differentiator — one of the big advantages of photonics, you know. What it comes down to is that the GPUs aren’t fast enough, since we’re extremely fast, so we end up having to use FPGAs. You probably saw — I believe it was right before our IPO — one of the major announcements for us over the past few months, involving AMD, and so we’re working with them because FPGAs are much quicker for this specific job. That’s why, you know, that’s how what we’re doing ties into NVIDIA’s announcement.

Kingsley Crane analyst
#29

Thanks. That helps clear up a few of those misunderstandings. On the financial side — this is for both Christian and Michael — you've mentioned not going after non-recurring engineering revenue and services that might pull talent away. Over the last couple of quarters, you've done quite well with professional services. It would just help investors to know, as enterprise interest grows, where that line sits. How do you weigh taking on those kinds of engagements versus turning them down? And just a reminder on linearity — what should we expect for that line for the remainder of the year? Thanks.

Christian Weedbrook executive
#30

That's a great question, and Michael and I tend to put it this way: for the most part, what we're after is partnerships rather than customers. Of course, a number of those partnerships will inevitably turn into customers, and some already have. Our approach at Xanadu is to think long-term. When we take the long-term view, we're looking at 2029 and 2030. Imagine we've already arrived there. We've got our large-scale quantum computer, which we refer to as a quantum data center, with hundreds up to 500 logical qubits performing everything you'd want such a computer to do. Picture that: it's accessed over the cloud, and at that point we'll have a subset — a small, key group of customers — putting it to use in areas like drug discovery, pharmaceuticals and so forth. Then you work backwards and ask, "How do we reach that point?" We reach it by staying focused on hardware milestones. We reach it by being selective about our partnerships. Over the last few months, we've announced a few partnerships. AMD was mentioned. Lockheed Martin is another really great one. These are partners that, hopefully, will come along with us over the next few years until we hit 2029 and 2030, and potentially become some of the first customers of this.

Christian Weedbrook executive
#31

Our mindset is really about partners, not customers. That said, Lockheed Martin is a paid partnership, so in that case we truly get the best of both worlds. I honestly believe that when we take a long-term view on revenue, keeping our eye on the prize is essential, because the pie is bigger by far orders of magnitude once you have a large-scale quantum computer, instead of selling, you know, a quantum computer for a few million CAD, where GPUs can genuinely simulate the same answer. In the end, as you said, that takes our best talent off the prize. That's roughly how we see it.

Operator operator
#32

Todd Coupland from CIBC has the next question. You're up.

Todd Coupland analyst
#33

Good evening, everyone — thanks for joining. Christian, I'd like to raise the subject of DARPA Stage B. Would you be able to bring us up to date on where things stand with it, and let us know when we should expect decisions regarding Stage C?

Christian Weedbrook executive
#34

Thanks for asking, Todd. There really isn't anything new to report. As you're aware, Phase B lasts a full year. It might wrap up sooner, or it might take longer — that depends on DARPA's preferences and on how well any of us deliver. What I can tell you is that we'll have a clearer picture near the end of the year. Our year concludes around November. Like I said, it could come earlier or it could come later. They've had other companies visit the Xanadu site, and so far we've built a strong rapport and a collaborative relationship with DARPA. That's really all we're able to share, and frankly, it's all we know. The team still has several more months of work and deliverables ahead.

Todd Coupland analyst
#35

Yes. If that result comes to pass, how big a validation would it be?

Christian Weedbrook executive
#36

That is a really solid question, since Rigetti is doing impressive work, and as far as I know, they either weren’t included or weren’t included right away. That doesn’t mean they won’t be included at a later point. I’m not sure how much impact that had on them, so it was sort of a single data point. I would say, though, that it is difficult to determine. You sort of imagine maybe you could be penalized for not being included or perhaps you could be rewarded for being included, but it is very early to know. I would say it is excellent that we are in and that we reached Phase B, and Phase C, the incentive has always been CAD 300 million. Honestly, I don’t know, and I don’t believe many people know what the outcome will be. It is genuinely difficult to read the market for quantum companies, how things are responding. I would say, you know, on the basis of $300 million, it would be excellent to be included, and it is not that we don’t want to be included, and we are doing everything we can to be included. We’ll see. This is probably a quick, good segue to sort of remind everyone, perhaps on the call, that we are in final negotiations of, you know, separate from that $300 million, final negotiations of CAD 390 million.

Christian Weedbrook executive
#37

That looks good. Nothing further to report on that front. We're still in the final stages of negotiation. And then, of course, there's the Canadian QBI, known as the Canadian Quantum Champions Program, which the government has indicated they intend to mirror what's happening with the QBI program of over $300 million. As far as we're aware — to the best of our knowledge — that is still moving forward. Regardless of what happens, we have plenty of opportunities in how we or others move ahead to Phase C.

Todd Coupland analyst
#38

On operating expenses, if I may, what's the run rate excluding those non-recurring items? It's roughly CAD 21 million. As we move through 2026, at what pace should that grow? Thanks a lot.

Michael Trzupek executive
#39

Right. As we've discussed, our plan is to give guidance later this summer on that as well as other engineering metrics. We aren't ready to do so today. Let me return to what I've told you before: clearly, when we raised this capital, the purpose was to speed up our engineering growth and our wafer starts. You should expect that to increase over the year, and later this summer we'll be in a position to share more specifics.

Todd Coupland analyst
#40

Thanks. Appreciate it, Michael.

Operator operator
#41

Our following question is from John McPeake at Rosenblatt Securities. Go ahead, please.

John McPeake analyst
#42

Thank you. Christian, Michael, and the team, congratulations on the listing, and welcome to these quarterly calls. Could you remind us how many silicon turns will be needed before you reach the loss rate you're targeting, and how long each of those turns takes? Will the EV Group partnership do anything to accelerate that? I noticed that news the other day.

Christian Weedbrook executive
#43

Thanks, John. On the last part of your question, probably not. As for the first part, to be completely candid, pinning down precisely how many chip runs there are is difficult, and a couple of reasons explain that. This process is extremely complicated. Our approach is to handle every photonic integrated chip design in-house here in Toronto, then dispatch them to various foundries across the globe. That is where things become somewhat challenging, since our computer is built from three or four distinct substrates — silicon nitride, lithium niobate and others. As a result, we partner with the finest foundries capable of working with each of those specific materials. Because of the different materials, we don't just have several chip runs — for each material we also have several foundries. On top of that, we actually operate at different production levels. For example, we collaborate with a foundry — a university foundry in Waterloo. That one is purely for prototyping, which lets us grasp our chip designs far more quickly and essentially feed them into the big foundries. We also work with medium-style foundries, dedicated ones such as NY CREATES in Albany.

Christian Weedbrook executive
#44

We also partner with big foundries such as Tower and others, and we've collaborated with GlobalFoundries too. Another player in large-scale manufacturing is UMC. Applied Materials is another exciting partner we're engaged with. Many of them are also deploying to 300 mil tools. Looking at the tools, we're running some on 200 mil and 300 mil. On the whole, I'd say that's likely a clearer way to frame it, instead of counting how many runs. The more work we complete and the quicker we complete it, the sooner we'll reach the loss reduction. That's the way we approach it, in any case.

John McPeake analyst
#45

All right. Regarding the qubit factory — could you expand a bit on what that actually means within the roadmap?

Christian Weedbrook executive
#46

Absolutely. I keep coming back to the name, because it's a good one — it tells you this is a kind of factory, in other words we're producing a large quantity of something. What we're producing in large quantity is qubits. Our qubits are photonic, to be specific they go by the name GKP qubits. That's simply a label taken from the authors behind the original paper. Basically, it's one variety of photonic qubit. We showed one off — by now it's getting close to, say, 10 or 11 months back — and it was published in Nature. That's very helpful for our approach, since it lets you put up with losses, which is the very thing we're working to cut down. If we can find ways to tolerate them at particular levels, that works just as well. Those are the kind of qubits involved, and this factory is how we generate as many of them as we like. Once you've got them more or less on demand, you can then put them to work doing computing. There's another reason we split it out that way — not just its importance, but the fact that it's the hardest piece, the last hard piece of our computer, which means that once we crack it, and we're well along the road to cracking it, everything else becomes roughly 2 orders of magnitude simpler to pull off.

Christian Weedbrook executive
#47

A great number of the components that we needed to get right for the qubit factory—picture that as a sort of nucleus or core, and once we crack it, everything else essentially flows from those outcomes. Yeah, that's basically what it is. Past that, what you might be focused on is fault tolerance. We've got these qubits. We've got a factory producing them right now, and we need to make them even better, which means reducing loss. Honestly, that theme of loss reduction carries across all the years. As we mentioned with our Aurora computer last year, everything is in place to genuinely change the world. The computer architecture exists. You even have a qubit factory, sort of an MVP version. That's the reason we need to keep on developing it. It truly is an exercise in loss reduction. Obviously you have to bring all the components together, and even if you achieve that loss reduction—say today, for example—you still have to build a computer, and that involves simultaneous component improvements as well as assembly taking place over the years too.

John McPeake analyst
#48

That is useful, thanks. Best of luck to all of you.

Christian Weedbrook executive
#49

Much appreciated.

Operator operator
#50

Our following question is from Quinn Bolton at Needham & Company. Go right ahead.

Quinn Bolton analyst
#51

Thanks for giving me the chance to ask. Christian and Michael, congratulations on the going-public transaction. Christian, perhaps we could begin with you. I just wanted to ask — you brought up the rising costs associated with chip design and the effort to get additional wafers through the foundries. What share of the loss-reduction improvement stems from chip design as opposed to process node or process technology advances? At Analyst Day, I believe you referenced items such as the need to enhance surface roughness, which sounds like it relates to etch and depth rather than simply how one designs the component. Could you perhaps quantify how much of it comes from design versus process technology?

Christian Weedbrook executive
#52

Yes. To follow up on what you mentioned, it's partly one and partly the other. Putting an exact percentage on it is difficult. My sense is the larger share comes from the chip designs themselves. The explanation for that is we rely on photonics, and we're able to build on what the telecommunications sector has already achieved, which is why we can collaborate with UMC, Tower, and others, since they've likely been doing photonics work for decades at this point. So what that gives us is access to many of their existing tools, and the nodes are mature already. That said, we do have a group we refer to as our fabrication process engineering team, and they work on-site alongside these foundries, offering input on how to adjust the material structure somewhat, you know, but never anything requiring a shift to an entirely different material, for example. We rely on lithium niobate, which is thoroughly known and extensively used in the transceiver market, Datacom and so on. To bring the loss down further, some process development adjustments are needed, but it isn't a case where we must purchase, at least for now, an entirely new tool or swap lithium niobate out for another material.

Christian Weedbrook executive
#53

That point is really crucial. Another thing that matters a great deal, and that we're also observing — Applied Materials is a solid illustration here — is that the newest equipment and processes tend to be described by, well, saying you're working with 300 mil tools. An increasing share of what we do involves 300 mil instead of 200 mil. We continue to run some on 200 mil, but the encouraging part is that you're tapping the most sophisticated equipment that's already out there. We didn't need to purchase, you know, build them ourselves. Those cutting-edge tools will also assist us in cutting the loss, since they're essentially state-of-the-art.

Quinn Bolton analyst
#54

Understood. I wanted to circle back on PennyLane. I believe you said the monetization approach for it eventually would involve enterprise licenses of some kind. I'm curious whether PennyLane supports compilation right down to a particular hardware platform. And if that's the case, could there ever be a chance to move toward a more usage-based model for PennyLane — one where you're able to, say, monitor consumption of hardware resources and bill for it accordingly?

Christian Weedbrook executive
#55

Keep watching for news on that final piece. With luck, we might be able to share a positive update shortly. That ties into the broader notion that we held back on providing many specifics — well, just, you know. For anybody who has dealt with open source, contemplating ways to commercialize it is a familiar consideration. We are exploring that. So please, you know, keep watching for news on it. You know, returning to your initial question, the software stack sitting right at the top is extremely, extremely agnostic — take PennyLane. It works like this: you enter a command, say, I don't know, you want to run Shor's algorithm, whatever the case may be. In itself, you don't truly concern yourself with the hardware. As you move down the stack toward the far opposite end, you require software that drives the hardware. It's a gradual spectrum. Moving from top to bottom, and finally at the base, it truly must match the particular hardware. We possess in-house software and tools closely tied to our photonic chips. That likewise applies to our photonic chips. Other photonic firms are doing excellent work, yet it wouldn't really be transferable to what they have. It is highly specific. I believe that holds, unless everyone is building superconducting qubits or everyone is building ion traps — you truly must remain within, you know, your own modality.

Christian Weedbrook executive
#56

That said, when you move down from PennyLane, you do enter the compilation layer that you brought up. Within that, as one subset, we offer Catalyst. To the best of our knowledge, Catalyst ranks among the most widely used compilation software available — perhaps even the most widely used. It is agnostic, which means we can use it, and others can also use it on their own hardware. I think once you go below that level, you begin reaching territory that has to become even more specific. That is how we think about it. We have been promoting Catalyst heavily, and we intend to do more and more of that, together with the broader package, which is PennyLane.

Quinn Bolton analyst
#57

That’s great. I appreciate it.

Christian Weedbrook executive
#58

That's not a problem at all. Thank you.

Operator operator
#59

Our following question is from Antoine Gara at Wedbush Securities. You have the floor.

Antoine Gara analyst
#60

Oh, thank you for the question, and congratulations on the dual listing here. I just had a high-level question. Roughly a month ago, a recent paper came out in Nature Communications that showed quantum state teleportation between physically separate, independent photonic emitters is seemingly achievable at pretty significant fidelity levels over meaningful distances. I believe the demonstration was carried out over a 270-meter link. I'm just wondering whether you could speak to this breakthrough. What does it mean from a Xanadu perspective? What does it mean for the broader space, or for the photonics modality, you know? Thank you.

Christian Weedbrook executive
#61

It was a fantastic outcome, a fantastic paper. More and more work along those lines keeps appearing broadly speaking. What set that apart was really demonstrating it through a communication or an application angle, and here that meant teleportation. You know, our Aurora paper that came out in Nature more than a year back now is really much more — you know, their paper was solid, but ours is, frankly, considerably more advanced technologically and more intricate, and it accomplishes far more, since you're now linking quantum computers or server racks together. For example, we could have pursued an application such as teleportation and sort of demonstrated something comparable, but the real point was showing how to connect them, with the knowledge that we can run algorithms afterward. All in all, a strong academic outcome. That paper you brought up was a good result. You know, the Aurora paper represented a far more sophisticated networking system from a quantum computing standpoint. It involved something like 35 chips across 13 kilometers, 4 server racks. Extremely involved, and, you know, the team takes great pride in that accomplishment.

Antoine Gara analyst
#62

Thanks a lot. Appreciate it.

Christian Weedbrook executive
#63

Appreciate it.

Operator operator
#64

Our following question is from Tyler Anderson of Craig-Hallum Capital Group. Go right ahead. Mr. Anderson, your line is now open. Please proceed with your question.

Tyler Anderson analyst
#65

Apologies for that—I’m juggling 2 calls. Thanks, guys, for taking my questions. This is Tyler Anderson, sitting in for Richard Shannon. I was just wondering whether you could walk through the milestones you anticipate hitting within Optimism, describe what success for that program would look like, and share its timeline.

Christian Weedbrook executive
#66

I appreciate that. At this moment, we aren't free to discuss the milestones for Project OPTIMISM. The government prohibits us from doing so since we are in final negotiations. What I can perhaps mention is that they align extremely well with what we must accomplish for our roadmap—the high-level roadmap we referenced—moving through the qubit factory, fault tolerance, and ultimately a large-scale quantum computer. They correspond very closely to that, and that is basically what the Canadian government's funding will support. For further details on our roadmap, stay tuned, as Michael said, until toward the end of this summer, when we'll have a detailed roadmap. With luck—fingers crossed—OPTIMISM will be announced by that point, and we can provide you a better answer.

Tyler Anderson analyst
#67

All right, thanks. Would you be able to explain what has shifted in the way you talk with customers now that the company is publicly traded?

Christian Weedbrook executive
#68

That is a fair question. So, what I can say for sure is that since becoming a public company, our recruiting efforts have improved. I'd describe it as somewhat greater momentum, though the organizations looking to enter quantum—and these days that includes an expanding share of big enterprises—are already engaged with us in conversations. We have had relationships with them for a long time. New ones keep forming. On that front, little has really shifted. It is difficult to gauge. My guess is they might now be somewhat more at ease, given that going public makes us look like a company built for the long haul instead of a feisty young venture. We have not truly observed anything more substantial. Nothing much has actually changed just because they might say, "Oh, you're public now, so we'll start working with you." That is certainly not what is happening. Alongside other excellent firms, we rank among the leaders in this area, and they have consistently recognized that, especially regarding PennyLane and our photonic hardware. Perhaps we can give you an update on the next earnings call, after we have spent a bit more time as a public company, to see whether things have shifted, but our reputation was already strong. I doubt it has really moved much.

Tyler Anderson analyst
#69

Understood. Well, people matter above all else. In quantum there aren't many folks around, so hearing that is excellent.

Christian Weedbrook executive
#70

Precisely. Yeah, that's correct.

Tyler Anderson analyst
#71

I appreciate that.

Christian Weedbrook executive
#72

I appreciate that.

Operator operator
#73

That wraps up the Q&A portion. Dr. Christian Weedbrook, I'll hand the conference back to you for any closing comments.

Christian Weedbrook executive
#74

Well, I appreciate that. I’d like to thank everybody who joined the call today. We’re grateful for all of your time and for the questions, and we look forward to carrying the relationship forward after today. Thanks to all of you.

Operator operator
#75

That brings the conference to a close. We appreciate your participation in today's presentation. You're welcome to disconnect at this time.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Xanadu Quantum Technologies Limited transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Xanadu Quantum Technologies Limited earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.