Executive readout · one minute
Webcast research workspace
Read the call alongside every captured source. Transcript, audio stay in one workspace.
Conference · 2026-11-11
You came from your AI assistant. Add Equibles to it and ask about this data directly, free.
Executive readout · one minute
Read the call alongside every captured source. Transcript, audio stay in one workspace.
Research coverage
2 live sources
Switch sources without leaving this page or losing your listening position.
Open the source you need; every reader stays inside this workspace.
All right. Good morning. I'm Doug Schenkel. It's my pleasure to welcome Jurgi, the founder and executive chairman of the board at Sophia Genetics. We also have Kellen Sanger, the VP and VP head of strategy. I can promote you if I mess it up. I don't want to demote, but yeah, it's good to have you both here. Thank you. In terms of roadmap, we're going to discuss Sophia's strategy, commercial adoption, and business model. We're going to explore your work with hospitals, labs, and biopharma companies, along with all the other opportunities that lie ahead. So there's a lot for us to unpack. Maybe the best place to start is with the big picture. And Jurgi, you and I have known each other for a long time. I know a bit about your journey, but I'm really curious, as you've moved from the CEO role into the executive chairman role, can you take us? Back to the original idea behind SOFIA and how that founding vision has evolved and actually how you're going to be involved in the next chapter. Yeah, absolutely.
Pleasure. And great seeing you again. So when we funded SOFIA, it was quite clear, actually, that in precision medicine, data would become critical, actually more and more important. Initially, people were focused on just data production, but for us, it was clear that the value would be on the data themselves. It was very clear as well back then that genomics would become the key data modality in precision medicine, even including drugs, right? At the time, maybe it was not so clear, but for us, coming from research and systems biology, this was very clear as well. And by the way, because of these two, it was clear that the data would be very complex, hence why we thought there would be a need for a platform, AI native, to compute the data and make it at scale, right? And this is what we've been doing since the beginning. We put our platform in the market, it's called Sophia DDM, in 2015, and right now we're supporting over 1,000 hospitals in over 70 countries to compute data that they produce themselves. The one thing that at the time maybe we haven't purely, I would say, captured and matured over time is that the genomic data wouldn't be sufficient by themselves and that one would have to combine genomic data with other data modalities to follow patients longitudinally. This is something that I think started to get clear over the last five, six years rather than 15 years when we started at SOFIA. And as executive chairman now, I actually have the privilege to be back to real innovation technology science. So I'm not leading operations as he was. And as well, Kellen will be able to give you much sharper answers on the traction on the business-wise. But I'm really dedicated on things where we think the future will be, definitively AI, technology, new technologies to produce data, but as well new technologies to compute the data.
I'll stick to our topic list, but as I hear you describe this, I mean, this is really getting you back to, I mean, you're passionate about all things, Sophia, but this is what's always made you most excited, what you're working on now.
Absolutely, and it gives me the opportunity as well to spend more time with people who are defining the new category of companies and systems that will be set up.
Now, when I talk to people about Sophia, sometimes what I hear is a very interesting approach, but all of a sudden, you know, it doesn't seem like there's a shortage of bioinformatics tools. There's all sorts of different sequencing platforms. There's more data repositories. What do you think Sophia is solving for different constituents in a way that nobody else is?
I think in the end, it's the scalability of our platform, the accuracy of our algorithms, plus the ability of delivering new and subtile algorithms very quickly into the market, right? And it's not by chance. This is because since the beginning, we've accepted to be a tech player, to compute data that would be produced by hospitals. Data that would not be harmonious, would have to be computed with sophisticated methods to basically find the signal, and this has become our competitive advantage, a bit like Swiss watchmakers very much working, you know, on the mechanics of the Swiss watches. Now we're in capacity to deliver data compute at scale, accurate, and to make it adjust according to the new needs that are emerging in the market.
One thing you mentioned. In response to my earlier question was the scale of the company today. You're in 70 countries. Does the diversity of the data that you have in your... Data Lake, does that put you in a position where you think you're uniquely positioned to partner with biopharma in a way that others aren't?
Yeah, I can start and then maybe Kellen can give you some examples of recent work we've been doing with pharma, but definitively our moat is the network, is our algorithms and is the data. And the fact that we are plug and play in so many systems. It makes that the value proposition we can offer to the hospital, excuse me, to the pharma, is distinct and complementary to what central labs can do, right? So basically, this is being materialized on sometimes CDX efforts with the intent to be able to touch more patients around the globe, not only in the U.S., but as well, sometimes it can be around... Real-world evidence data production capabilities where pharma may eventually need to understand how people are being tested and treated today, how they're responding to the existing treatments in the market, to be sharper on how they're going to develop the next generation of treatments, and how they're going to design their clinical trials to maximize the chance of having the drug approved versus the standard of care in the market.
Yeah, and everything that you just mentioned is built on just this breadth of data. So you mentioned 75-plus countries. We're currently doing somewhere north of 400,000 patients per year. So one of the largest diagnostic players in the space, if you were to compare us to some of the others. And the global nature of what's being produced by the second, we can look. You know, by the minute, at every patient that's being analyzed across the entire globe, this is an incredible network of data that you're able to do things, like you just mentioned, building CDXs and real-world evidence and other types of products off of both biopharma and clinical.
Helen, actually, could you, just building off of that, recognizing... Sophia has, even though there's always been the focus on data, the tools really have been successfully deployed in the relatively early days at hospitals and labs. Do you think what we're talking about now is driving a notable change in pharmaceutical companies, biotech companies working with you? And beyond just the breadth of data, which is really important, are there other capabilities that investors should be contemplating that you have that others don't have when it comes to working with biopharma?
Yeah, there's, of course, the data, but also the network that exists globally, right? And so, Jurg, you mentioned that we're now working with BioPharma on CDX projects. This started with a lot of sponsored deployment. So, if you just think pharma is interested in having their therapies and drugs reach as many patients as possible across the globe, who better than the company with the largest global network of hospitals and labs and patients across the world? So just activating and identifying patients in different geographies, moving that into more regulated products with CDX, so identifying patients in conjunction to a specific drug and quickly turning that on across all geographies. So instead of a biopharma having to... Have many different vendors in the US, in Europe, in Latin America, in Japan. You're able to come to one network on one platform and look at all data that's being generated globally. And then as you run more CDXs and collect more data, you're also able to build interesting algorithms or predictors and answer specific questions for pharma about specific drugs. Specifically, exclusion or non-exclusion criteria are building what we're doing more recently and what we announced last week around AI predictors for certain indications. We announced bladder cancer. We've done similar in lung cancer and some in prostate.
What you described in the AstraZeneca collaboration, how important were the attributes you described in moving into a partnership with Astra?
Yeah, they've been obviously a great partner for us on many different areas, and starting in a lot of those data deals that we had mentioned where we were looking at different cohorts and therapies and how do we get the best drugs for their trials, but then after that drug is on market, how do we help identify patients across the globe? And so... Throughout all those stages, we've been proving our expertise and our right to play there, and as we start doing that with AZ, all of the other pharmas are now very interested as well because, you know, we're helping improve our risk and improve out our product.
Is there an opportunity through these partnerships to actually boost your clinical offerings? Obviously, CDX labels would help, but do you get some access? Like, is there any flywheel effect where you can learn from the biopharma and actually iterate on your own assets? Yeah, this has been the thesis since day one.
Even as he was describing his original hypothesis for the company, he was a bit humble. I remember he was showing me in one of the Series A fundraising decks. This is back in 2014. He was using the word artificial intelligence and applying this to genomic data and multimodal data in order to generate insights for patients across the globe. So quite visionary on some of the AI ML techniques. But no, absolutely. I mean, the clinical biopharma flywheel has been kind of native to the business from day one. The idea of once you have a CDX, for example, on the market, this helps increase the clinical utility and you have the premier application that others want across the world. The data that's generated on the clinical side creates interesting products for pharma. As you win more with Pharma, you're collecting more data about the patient, so getting closer to the patient, getting closer to the oncologist and the point of decision that's being made so that we can build additional tools, such as clinical intelligence on working with the oncologist and how you look at patients, so it all is very connected.
Amazing. I've made it to, I think, 11 o'clock this morning or something like that without saying AI. So I'm glad you brought it up. But you guys have spoken about moving from data interpretation towards clinical intelligence, including decision support tools and digital twin concepts. Which AI applications, like, where do you create value first? And then over time, like, how do we think about? More value being created, but also being able to monetize this from a franchise perspective.
Yeah, it can start with the value we're creating and how we're thinking about, you know, the evolution of that. So we started very much on harmonizing data. This was the way we would be able to enable hospitals produce high-quality data that they would trust locally. And then, at some point, we started to realize that this was good, but that eventually, beyond doing that for genomics data, bringing them capabilities on following longitudinally patients, including how patients were treated, how they responded to treatment could bring some additional value, right? So, I don't know if you read that, Doug, but... Today, it's estimated that we, as humans, spend about $250 to $280 billion in cancer drugs yearly, and the estimation is that by 2029, we, as humans, will spend almost half trillion, $500 billion in cancer treatments a year. So back to your question on AI utility and clinical intelligence, well, one will need to make best use of the data, because this won't be sustainable. As a society, we're not going to be able to continuously expand our expenses. So I think beyond harmonizing the data, at some point, it's going to be about bringing this clinical intelligence to this multimodal data, because of having computed massive amount of patient cases in decentralized setup. So, for example, as health care systems or insurance, think about who should be tested, who should be treated with a certain type of drug, when and where, and at what cost. So I think this is going to open a very, very large market. It's still very early. But on the data monetization, I think it's going to be about this kind of insight.
When you talk to biopharmaceutical companies, do they view it the same way you view it?
I think they understand, because for them as well, it's very important to make their own market sustainable. If, you know, we end up as humans not being capable of... Basically, purchasing or having the resources to purchase our own drugs is going to explode, right? And I think in that vein, you know, this part of their market access activities as well. And again, having technologies and companies like Sophia that can enable the system but enable the pharma understand, okay, these are the cases where it's not worth providing a treatment to a patient. While these are the cases which are very rare and very hard and where it's worth providing a new type of treatment or combination of treatment, despite it's expensive, because there is a clear benefit on this group of patients, it's something that is very, I think, pertinent, but is very much top of their mind right now.
I brought up the question of how biopharma thinks about this. There are new AI... Either AI companies that are moving into drug discovery and drug development or, you know, basically companies that have been set up as AI slash biopharma from the start. Are they partners or are they threats? You know, when I think of the differentiated foundational models you've built and I think about what they do, I could see where they would love to partner with you, but I could also see where You could start to overlap in terms of what's proprietary over time. What's the right way to think about that?
It's a good question. So we don't do drug discovery by ourselves, right? So back to the point of Kellen, our main advantage is the network. The scalability of our platform, the ability of computing data with very sophisticated algorithms. So when we are discussing with pharma companies in particular today, and maybe tomorrow with some companies who may be indeed AI-driven discovery companies, the main purpose for them to partner with us is the deployment capabilities. So it's not rather, you know, on the identification of a specific biomarker, that is from the identification of this biomarker, on the ability of repeatedly, replicatively, and with high accuracy, identify patients around the world with the same biomarker so that then they will receive a specific treatment. So I think we're rather complementary. We may be involved in some co-developments when it comes to new biomarkers, but we're happy if we're not the one owning the rights of the biomarker, but rather are being paid on deploying those biomarkers into the market.
Are you seeing interest from these types of companies?
Yes, we are.
I mean, if you just like all of the hypotheses in this space about AI, there's a central kind of agreement that the goal would be to help improve healthcare for patients and humans in general. And so I think when you compare and contrast the work that we're doing, we really understand the clinical deployment, so real-world use cases of AI in a clinical setting to determine how you might diagnose or treat or ultimately care for a patient. This is something that's missing with a lot of the theoretical models that exist where you're able to identify something that, sure, sounds interesting, but in our case, adoption and use is the true innovation.
I'm going to take a step back and talk about traction in the United States. And again, this is kind of ancient history, but for those who followed Sophia for a while, it's not where you started. In a way, you were playing catch up. You've done a really great job of doing that. What have you learned from bringing larger health systems onto the platform? And what determines how quickly those deployments actually, once you get in there, what determines success and how quickly you can ramp from a revenue perspective?
I will let... Kellan, give you the specifics, but as you said, we started probably the hardest way, and not because we chose it, just because we funded the company in Switzerland, we started in the clinical market, because we wanted to be this tech player owning or controlling that network. And then we moved to the U.S., and now we're moving to the biopharma, right? So, indeed, it has been a journey. And when we went public in 2021, there was questioning whether our model would be compatible with the U.S. model. And I think over the last two years, indeed, our performance has proven that even in the U.S. market, our model was very pertinent. Yeah.
And I mean, more specifically, we've been on fire in the U.S. lately. It's been a huge growth driver for us. I think when you think at a macro level, what's happening in the U.S., and if you were to put yourself in the shoes of any of the large hospitals or health systems, you're just looking at a model where, first of all, clinical utility is being more proven. The reimbursement rates are increasing. The denial rates are decreasing. And so bringing testing in-house is becoming easier than ever and more profitable than ever. So you're constantly contemplating, should I let this patient leave my health network or my hospital, or should I do this testing in-house, be and reimburse the full services revenue on the patient, sometimes $2,000 or $3,000 per patient, pay Sophia a piece of that, and then profit ultimately upwards of $1,000 plus. which is a huge amount of economic influx for some of these hospitals and labs, and then on top of that, some of the other players in the space have proven that the genomic data of the patient is incredibly valuable to pharma, and so the monetization of that data on the back end, you know, we believe in not only driving those economics to the institution, but also helping them keep control of their data so that they can use this for research projects or for partnering with biopharma, and so it's just a mechanism for them to stay competitive, and I know we've mentioned that we've been growing 60-plus We have seen no signs of that stopping.
Maybe a weird question to ask, given the success you're having in the U.S., but at some of these sites, what are the biggest remaining bottlenecks? I guess part of it is, yeah. Yeah, that's true.
And so it takes, obviously, some time to complete the verification and validation when you're bringing a test for the first time. But in general, I, you know, there's some of the larger centers where we're coming in with one application. So we will land them on, for example, solid tumor or enhanced exome from the rare disease side. And then ultimately, we've done a very good job historically of expanding to additional applications. And so you're seeing some of the initial land thrust across large accounts in the U.S. So you mentioned that we've signed two of the largest health systems in the U.S. recently to our enhanced exome application. This is something where eventually you could expand within them, but it just takes some time.
Yeah, is that the right way to think about it? You get in the door with a beachhead. You prove that this can work in one application, and then the fan and the gear starts to go away.
Yeah, with the eventual goal of capturing the entire testing volume of the hospital. And some of this is us, you know, switching from homegrown solutions that they're doing themselves internally. Some of it is instead of sending to the central player, bringing that in-house. Some of it is changing from one provider to the other, and we're there to help them across all of those.
As you know, Doug, in the tech world, this is called the classic land and expand approach, right? It's almost like monopolies. Owning this street, and then you need to build your houses, and so every year, and Kelly knows that because he's very involved in the budget efforts, we need to arbitrate whether we want our salespeople to be more focused on land versus expense, and the two are important.
As you, from a partnership standpoint, how important is it to work more closely with the sequencing companies, essentially the technology platforms? Versus other players in the ecosystem in terms of just trying to remove any resistance to decentralizing with Sophia.
Yeah, look, so we are technology agnostic, right? And we accepted that basically hospitals and labs would need to have the freedom of basically producing the data with whatever technology they wanted. And this was not easy because, again, we had to harmonize the data, and hence why we build our algorithms. But definitively, for us, a way to go quicker to the market, and we're discussing with one investor just before our first chat, is the way to expedite the adoption. And I think all the system would benefit from that. I think the sequencing players would benefit, you know, from partnering with us because it means that, you know, there may be less, you know, Stressed by some negotiations with only the central labs that they can sell a bit more broadly to more people, that volume scales as well everywhere. Today, I don't know what are the exact numbers, but really when you get a solid tumor being diagnosed with CT scan, It's for sure less than 50% of the people who are being tested on their genomic profile, right? So I think working for us hand-on-hand with sequencing partners would be a nice way to, in general, increase the impact of the industry on society.
One more on the large health system, wins and lands and expands. How, like, how long does it take? And again, these are awesome wins. It's early, so maybe it's a little bit of an unfair question, but how, like, is it years and years? Is it quarters? Like, how do you guys think about how? How long is it going to take for this to actually show up in a meaningful way on the P&L?
Yeah, and it's starting to show up. So even this recent quarter is one of the first quarters in which they started to ramp, and they're using some of our technology and routine. But typically, it takes between six to nine months. So we talk a lot about bookings. Our typical sales cycle is we're there at the account. We're talking to them for several months, and then ultimately, they sign to adopt our application. They have a pretty good idea of the number of patients they see every year, and so they commit to a certain amount of volume. We call this a booking, and then it takes six to nine months for them to eventually ramp up completing verification and validation. Some are faster than others. We're very good at helping them along the way because we've done this so many times, and so in many cases, we offer services to help them expedite. But in general, the pros and cons of this is we have a very good insight into what our revenue will be. So we said on our earnings call that today we have 95% of visibility into 27 revenue, right, because we're signing business today that will ramp and grow throughout 27. And in addition, we have very little churn. Our annual revenue churn is less than 1%, which is probably world-class in terms of metrics. So the platform is very sticky, and once we're in, it's very easy for us to not get displaced but also expand with the customer.
The six to nine months is probably on the initial application, right? So then there's, you know, that's almost completing the land in a way, right?
And then the expand comes after that. Which is typically much quicker. Once they've already used our platform and they already are using it in routine, it's adding another application to the platform, which goes much quicker.
How do you think about menu in a, you know, sitting here today and there's going to be more and more applications over time, but again, probably using land and expand. too much, but as you expand, how broad can the menu get in these sites?
Yeah, so our menu today is a nice proxy for the entire testing market. We offer everything from hereditary cancer to hemont to solid tumor to liquid biopsy to others on the rare disorder side, and so we're seeing growth across all of those applications. We typically will land with one of them and then expand across the others. We have amazing new applications in liquid biopsy, Which is with our MSK access test, which we now have in 100 institutions globally, which is a pretty incredible number in such a short period of time with liquid biopsy going quite quickly. But the others are as well seeing growth. And there might be new ones in the future,
like MRS, transcriptomics, right? We end up, you know, as you know, we were probably best in class more than first in class, so we let others demonstrate the clinical utility of... A new application or a new data modality like transcriptomics, and then we help hospitals to adopt and scale that, rather than demonstrate the clinical utility for us.
Looking ahead, how do you expect the mix of core clinical... And even as I'm thinking about this, it's probably hard to parse it out, but clinical versus software versus biopharma, you're an important clinical partner for a lot of your customers, but there's a lot of tech attributes to the story. So how does the mix of revenue evolve over the next several years? Yeah, and all of it is software for us because we're all working on a platform, right?
So today, the biopharma business and working with pharma companies is less than 10% of our revenue. We think long term, this could be half of our revenue. This is just the growth we're seeing. We're signing really large accounts, really large deals. And in the medium term, it could be 50% of our overall revenue, as with the US, right? The US today is maybe 20%, and this could become as well 50% of our revenue in the long term.
And then from a profitability standpoint, does this mix? It doesn't. There's no negatives in this mix. It's more leverage on the existing infrastructure, correct?
Yes, it's a lot of leverage. I think as a company right now, we're approaching some very exciting milestones. We're reaching $100 million of revenue. We're growing 25-plus percent. We're going to be profitable from an adjusted EBITDA perspective in the fourth quarter of next year, which we've mentioned this quarter. We lost $6 million of adjusted EBITDA, and so we're dropping down year-over-year or each quarter around 60% of our incremental revenue growth. So you're seeing the leverage of a true software native company.
I wanted to close with, and you kind of touched on some of these already, Kellen, but if we think about two or three key milestones over the next 12 to 18 months. What would you want us to be looking for as we think about success of the business in the medium term? Call it 12 to 18 months.
Yeah. So financially, I just mentioned a few, and obviously we're committed to continue growing at our current levels or higher. But I think from a strategy perspective, we have four pillars going into the next three years. So we're going to continue to win in genomics. Our clinical genomics business is growing north of 20% on a volume and a revenue basis, which is great. On the CDX perspective, which is pillar number two, which is winning more CDX deals with pharma, we just won our first two in our company's history, which we're announced in Q2. So to continue making progress there, moving into more regulated products, collecting more data for the patient, which is Pillar 3, and building a real-world data product. We just this quarter announced an AI predictor with AstraZeneca for predicting lung cancer, so doing more of these and leveraging our data. And then the last is building into clinical intelligence, so getting closer to the patient, closer to the point of decision-making, working more with the oncologist to build tools for clinical decision support.
That's great. All right. Well, we'll leave it there. Jurgi, it's always great to see you. And Kellen, I'm really glad you could join us. This is fantastic. So thank you guys for taking the time. Thank you very much.