All right, great. Welcome, everybody, to the JP Morgan Healthcare Conference. I'm Casey Woodring from the Life Science Tools and Diagnostics team, pleased to be joined by the management team at BillionToOne. We're going to do the standard 40-minute slot, corporate presentation, Q&A afterwards. Not much more to be said. Oguzhan, kick it off.
Thank you, Casey. At BillionToOne, we are redefining what is possible with molecular diagnostics, with a technology that, for the first time, achieves single molecule sensitivity and precision. We'd like to talk about our company through four pillars of differentiation that we believe make our company a different type of molecular diagnostics company. We believe that these four pillars are going to make us a category-defining company and the first one in our space to enter S&P 500. The first pillar, everything that we do, starts with a revolutionary single molecule next-generation sequencing platform that enables us to build category-defining products, products that are not just simply slightly better, but 10x better. We have a patented QCT technology that enables single molecule-level sensitivity, and we have been able to build differentiated products that move the field, both in prenatal and oncology cell-free DNA testing.
With these products, we have been able to grow rapidly. We have grown from $0 million- $334 million annualized revenue run rate as of the end of Q3 in just five years. This positions us extremely well, as this particular market exceeds $100 billion total addressable market just in the United States. We have been able to do this while also achieving superior gross margin profile. We have achieved already 70% gross margins, but this is despite the fact that many of our tests are still at subscaled ASPs, and we are using only 1/4 of our lab capacity, so we have significant potential for improvement for COGS per test reductions. Last but not least, we have been able to achieve all of this growth, this extremely rapid 100% year-over-year growth, along with GAAP profitability.
This is partially due to our culture of fiscal discipline and efficient operations that incorporate AI across all of our operations. Again, as of Q3, we have been able to show GAAP profitability and positive cash flows with only 10% of the accumulated deficits of many of our public competitors. Let's delve into each of these pillars and see how we have been able to achieve something that is pretty close to unprecedented in molecular diagnostics. The first pillar is having a revolutionary technology platform and building differentiated category-defining products.
We believe that BillionToOne is ushering in the next paradigm in molecular diagnostics, similar to how the 1990s were really defined by PCR for primarily infectious disease testing and genetic testing, which was then supplanted in the 2000s by the Human Genome Project and the first attempts in sequencing that resulted in NGS focused on specific mutations, which was then changed in the 2010s. As the sequencing cost decreased, the companies and scientists have been able to do deep sequencing for the first-generation prenatal and oncology cell-free DNA test. We believe that now and the future belong to ours because of our technology. Cell-free DNA is truly a remarkable biomarker, It is like a microcosm of the entire body. The potential is truly limitless, but it is also an extremely difficult biomarker. Sometimes you might have to detect a condition just from a single molecule that is present in blood.
In that particular problem, noise is so much higher than signal unless we solve some of the fundamental problems. We believe that our technology, single molecule next-generation sequencing, is designed to unlock this full potential of cell-free DNA. Really, the technology that we developed here is really simple and elegant. Many companies focused in this space on analyzing the data differently. What we realized here is that at the time that the data are being analyzed, a tremendous amount of noise has already been included, incorporated into the data due to the amplification biases and sequencing errors. So the differences can only be incremental at that stage. We see this across the board.
Even though one example is early cancer detection, even though companies use many different methodologies, from fragmentomics to methylations to machine learning approaches to really anything that you can imagine, and the performance across all of these methodologies seems very similar, and that is really because at the time that you are looking at the data, the noise has already been incorporated, so the differences can only be incremental. What we do is that we design and add artificial synthetic DNA molecules called QCTs, quantitative counting templates, to the sample before any amplification and sequencing happens. These QCTs are designed to mimic the original sample, but they have a single molecule embedded identifier. This allows us to know exactly how much amplification happened across different parts of the genome, what are the errors that have been incorporated into the data, so that we can remove that noise from that data.
I'd like to highlight here again, with cell-free DNA testing, it is not the signal, but the noise that is important, especially as we get into lower and lower and lower limits of detections. When you are looking into MRD or early cancer detection, you have to look at things that are 0.0001% sometimes, very, very low, part per million levels, and at those levels, by looking at more and more locations, it's not really going to solve the problem. Removing the noise will. In many ways, our technology transcends the precision versus scale trade-offs of other methodologies. If you look at the precision versus multiplexibility curve, on the one extreme, you have the digital droplet PCR, which has been considered the gold standard for confirmation assays.
It is actually the only other technology that can get to single molecule-level sensitivity and precision, and it is used to confirm a finding, but the problem is that it can really only be used for specific mutations in each tube of blood. Of course, that doesn't work for population-level assays, so we have next-generation sequencing at the very other end, which is almost infinitely multiplexible, but the precision of the measurement is so poor that you can barely distinguish one copy of a gene from two copies of a gene. That is why we have unique methodologies like MLPA for such a problem. But in cell-free DNA, you are trying to sometimes differentiate 2 copies from 2.01 copies. So there, next-generation sequencing has a significant limitation. By the way, this is exactly why there are so many tumor-informed MRDs.
If you think about why someone does a tumor-informed MRD, it is because if you just look at the cell-free DNA, you wouldn't know what is noise and what is signal. You wouldn't know a mutation that you find actually is not a true mutation, but a sequencing error or amplification error. So by sequencing the tumor first, we essentially solve that noise problem. And our technology solves that noise problem without having to sequence the tissue. Single molecule next-generation sequencing at our company really powers an engineering biology approach to molecular diagnostics. So everything that we do is really design-based. At the time that we develop an assay, we actually know what the performance is going to be because we are almost approaching it from a mathematical physical perspective.
Every part of the assay development relies on quantitative data, which allows us to have a much more directed approach in improving our assays compared to more trial-and-error approaches that are quite standard in biology. So this approach allows us to build differentiated products on patented foundational technologies. It allows us to have predictable accuracy. When we first built our single-gene NIPT products five, six years ago, we have predicted what the performance is going to be, and years later, after clinical studies, that was exactly what the performance was. But it also allows us to have these precise laboratory feedback loops. We can see at each step of the lab where there is even a 0.01% contamination or where there is a little bit of decrease in yield so that we can root cause them and we can continue to improve these processes to essentially optimal efficiency.
With this technology, we have approached prenatal and oncology diagnostics. In both of these areas, we believe that current molecular diagnostics are limited. Prenatal testing has been primarily limited to chromosomal changes, and oncology has been limited by lower sensitivity and measurement noise. We first went into the prenatal market, where we faced tremendous competition. When we went into this market, there were so many companies that were essentially spending 10 to 1, sometimes 100 to 1, compared to the resources that we have had. We have been able to grow here by essentially approaching the most difficult problems that others could not solve. This field really was focused on aneuploidy and microdeletion testing. If you look at even microdeletion testing, the published results show sensitivities that are at the level of 75% or 80%.
We have been able to move non-invasive prenatal tests from these million base pair changes all the way to single base pair changes that result in cystic fibrosis, sickle cell disease, and other conditions that are important to screen and detect during pregnancy. It is very important to detect these conditions. They are recommended for screening in every pregnancy by medical guidelines, but the current workflow is extremely problematic. This is actually arguably the most broken healthcare workflow in the United States healthcare system. Despite the fact that billions of dollars are spent in screening the mother to see whether the mother is a carrier for one of these conditions, we are missing more than half of the affected pregnancies due to the difficulties of testing the fathers for these conditions.
And even if you can test the father, you are still at one in four risk, which causes essentially for people to undergo amniocentesis. And you are still at only one in four risk here, so you would need to do four amniocentesis for each true positive result. UNITY is a single blood test that directly identifies the fetal risk just from a maternal blood sample, simplifying this workflow. We really believe that UNITY will grow to be the new standard in prenatal care. Compared to traditional carrier testing, it actually matches the hypothetical scenario where every father is screened and every father is actually the true biological father. We know that just due to misattributed paternity, that is not the case.
But if you actually look at missing paternal screening in this workflow, we are detecting three times as many affected pregnancies as the other current methodologies. This is such a large difference. If you think about Quad Marker Screen, which preceded NIPT for aneuploidies, that difference was only 10%. And today, NIPT for aneuploidies is already the standard of care with millions of tests that are being done. So that is why we believe that this is just a matter of time that this becomes the primary method to screen for these recessive conditions. Despite that, we didn't have the resources that many of our competitors had. We have grown to be the second largest prenatal lab in this space, and our ordering providers are our strongest advocates. So I have two quotes here that really, I think, highlight why providers choose us.
So in the first one, we have always known you have the better test, but we just love the competitor rep. We needed to make this change for our patients, and we thank you for not giving up on us. That essentially shows it does take time. There are incumbents here that we are taking market share from, and it does take time, but eventually, providers understand and believe that this is the best test for their patients. In the second one, it really highlights the unique value proposition. UNITY provided a high-risk call for cystic fibrosis, and it was confirmed. I am 100% confident I would have missed it using the traditional carrier method because my patients' partners rarely get tested.
In a real-world patient case, and this is actually cited in the news, so that is why I can share this, we have so many of these cases that are really changing even the prenatal treatment paradigms. This is the critical part. Great diagnostics changes even the treatment paradigms, where this was an unknown carrier status for a patient and patient's partner. UNITY showed high risk for cystic fibrosis for the baby. Diagnosis was confirmed via amniocentesis. In this case, the prenatal therapy with the CFTR modulators was initiated at 27 weeks, even before the baby was born. This baby not only did not go through NICU, actually passed newborn screening, which is incredible that a cystic fibrosis actually this baby is affected. We didn't change the fact that this baby has two copies of cystic fibrosis.
But because the treatment started early, the baby is doing so well that the phenotype-based newborn screening thought that essentially the baby didn't have cystic fibrosis. So with oncology with Northstar, we are redefining liquid biopsy for cancer care. We have two complementary products here that solve oncologists' two most important unmet needs when they get a late-stage cancer patient. The first question that they have is, what therapy will work best? And we have found that conventional liquid biopsies are missing actionable mutations. But once the therapy starts, the question is, is the therapy working? And the problem here is that the scan-based treatment response monitoring is imprecise and can be significantly lagging. So our solutions here on the therapy selection side is an ultrasensitive liquid biopsy that detects 50-plus% more actionable variants compared to other competitors, enabling better therapies. Not 5 or 10%, 50% more.
That means that even if you look at a cohort of five or 10 patients in a single physician's experience, there is at least one or two patients that actually get a better result, a better therapy, that do better because of the test that we provide. On the response monitoring side, we have a test that quantifies tumor at a single-molecule resolution so that we can find whether the patient's tumor burden is increasing or decreasing months earlier than scans. Again, without relying on sequencing the tumor to begin with. Everything here is tumor-naive, tissue-free. So 95% of our providers use Select and Response in tandem. Select to determine what therapy to use, response as the baseline measurement.
They follow up with more response tests in a typical scenario until they see that the patient is no longer responding to therapy, at which point they can use another Northstar Select test to determine what the next treatment may be. So we have done a head-to-head clinical study, and we did not bias this study in any way. We told the physicians that they can use whichever test that they want to use, whichever panel size that they want to use. We don't have the largest panel size. They can use it on whichever patient that they typically use it for. So just standard of care testing. We just asked on the same day that they are using one of these tests to send us another sample as well and we reported our results, and these competitors reported their results.
We show that in this prospective head-to-head clinical study that we detected 51% more SNVs and 109% more CNVs. On the Northstar Response side, we have shown that in an immunotherapy pan-cancer cohort that we are able to differentiate who is responding to therapy or not by more than two years. This is not an MRD study. This is a late-stage cancer study. In a late-stage cancer setting, having a two-year separation is pretty remarkable. This is a true real-world patient impact case. This is a patient in their 30s diagnosed with stage II rectal cancer. Provider ordered tissue next-generation sequencing. This provider actually did not believe in liquid biopsies at all. This was MSS negative, proceeded with chemotherapy. There was no targeted therapy or other therapy that they could find.
One of the actual nurse providers in this case advocated for this patient, saying that they should also send a liquid biopsy. Even though the physician didn't believe in liquid biopsies, they agreed to send as a last resort a Northstar Select test. And we have found from blood, when tissue could not identify it, from blood, an MSI high status allowing the patient to start an anti-PD-L1 therapy, essentially an immunotherapy. Treatment changed from chemotherapy to anti-PD-1 therapy for this case and in clinical studies, this has been shown to be actually, in this case, 100% curative. So this patient might have died in a year or two and now has a huge chance for being completely cured and by the way, this physician now believes so much in liquid biopsy that sending a liquid biopsy on every single patient.
So with ordering providers being our biggest advocates, we have been able to grow rapidly. And our test volume grew 51% year- over- year. This is despite the fact that in many geographies, we don't have full coverage. So as we grow our sales team, we continue to expand our growth for our test volume. One of the other drivers of growth, and as you can see, it has been almost accelerating over time, is really due to the fact that we have been getting more and more in network as our test volume increases. It is very difficult, especially on the prenatal side early on, when you are out of network for physicians to use your test. So as our test volume increases, we become more in network, and we have now exceeded 250 million in contracted lives.
As our in-network status increases, we are able to convince more and more providers and more health systems to use us as their primary screening test. So there are multiple levers for further commercial growth. Across prenatal and oncology combined, we only have about 20% the sales force size of many of our competitors. So as we grow our sales team, we expect this to result in further growth. But even our existing sales reps are adding about 100 new clinics every quarter on average. So that is going to result in even further growth. And once we get one physician within one clinic, we see growth in those clinics over time. So 100% net test retention for the majority of our clinic cohorts. Our ASP is rapidly growing as well.
We have reached more than $500 per test ASP as we increase our covered contracted lives and as more and more Medicaids are loading and covering UNITY Carrier Panel PLA code. We believe that our ASP can continue to increase further with expansion of contracting and coverage. Really, the three pillars of ASP is coding, contracting, and coverage. And we actually have room for improvement in all three of these areas. So on the coding side, we have been able to achieve 70% gross margin despite the fact that we don't have ADLT pricing for any of our tests. So when and if we decide to go through FDA approval, that is going to result in a significant incremental increase for especially our Northstar Select test that can result in a significant improvement in our ASPs and gross margins.
On the contracting side, even though we are around 235-250 million in contracted lives, there is still room for growth. And every quarter, we are signing 10-15 different contracts, each of which incrementally increases our ASPs. But we also have a lot of opportunities for coverage. About 70% of our oncology tests are response, and we don't have Medicare coverage for our response tests. That is something that we are working on this year. So as we get more and more coverage for some components of our prenatal tests, as well as our core response tests, we expect our ASPs to continue to increase. So we are seeing with the growth of ASPs and growth of test volumes, we have seen exponential revenue growth, 0-334 million ARR in five years. And this is true across both prenatal and oncology.
We have been able to do this with superior gross margin profile. As I mentioned, we have achieved 70% blended gross margin with continuing expansion of our ASPs and continuing decrease in our COGS per test. Finally, we have been able to achieve GAAP profitability, something that eludes many of these companies, even though we are still early in our growth journey. We have achieved profitability with 10% of the accumulated deficits of many of our other public competitors, going from negative almost 500% in operating loss to positive operating income in year-to-date up to Q3 2025. We have guided the full year 2025 and full year 2026 to be GAAP operating margin positive as well. This is really due to a multi-prong approach. Every team is expected to increase their productivity every single year.
With reimbursement, as we increase our claim number by 50%, we are increasing our headcount only by 15%. Same with patient billing. In engineering, millions of lines of codes are AI-generated and accepted by senior engineers. More than 30% of the code in our code base is AI-generated. In COGS, there is no single silver bullet. We have completed more than 100 different COGS projects, each of which might decrease our COGS by $0.50 or $1. But in aggregate, it has made a tremendous impact in our COGS per test and we continue this 20 mi march in decreasing our COGS and increasing our ASPs and also, as I mentioned, we are only using about 1/4 of our lab capacity today.
As we continue to increase our test volume, this is going to result in improvements due to fixed cost per test coming down, but it also creates labor efficiencies. So historically, we have been able to achieve 4x higher test volume with 2x labor cost. And we believe that we can continue this trend as well. And we are just getting started across our products, across our growth, across our GAAP profitability. So when we projected $65 million of revenue in 2023, that was up from $28 million in the previous year and it seemed very aggressive to really all of our investors. We have been able to deliver $72 million in that year. T hen we have projected $125 million for the following year as a private company and almost no one believed us, even the investors that actually invested in our Series D round.
We have been able to deliver $153 million in GAAP revenues in 2024. Last year at JP Morgan, we have projected for 2025, $220 million, and for 2026, $330 million for our revenues. For 2025, just a few weeks ago, we have guided to the $293 million-$299 million for our revenues. We expect to be easily in the upper end of that range. For 2026, we are initiating and expanding our previous projections to $415 million-$430 million. We have had a history of being able to significantly beat our projections, but I wanted to show $330 million here so that we are not expected to beat it by the similar 40% range here. We believe that our single molecule next-generation sequencing platform will drive our continued and rapid growth in $100 billion markets.
We are already capturing significant market share in the prenatal market, and we are the second largest prenatal lab here. In oncology, we launched therapy selection and therapy monitoring products that address a $20 billion U.S. market. We expect to expand to MRD this year, which is going to be a $30 billion potential opportunity. And finally, we believe that our technology solves some of the fundamental problems of early detection. In MRD, it is okay. It's possible to do a tumor-informed MRD, but in early detection, that is not possible and the fundamental problem of early detection is actually the same as tumor-naive MRD. And we believe that we have the really only technology that can address that fundamental problem and achieving very high sensitivity for early-stage cancer detection.
As a summary, we are transforming healthcare one molecule at a time, one patient at a time, with a unique technology that achieves single molecule-level sensitivity that has allowed us to build unique and paradigm-changing products. We have been able to grow exponentially even as we reach scale and we have been able to combine that with extremely efficient operations to have emerging GAAP profitability. As I said at the beginning of my presentation, our long-term goal, our five-year goal, is to build a category-defining company and become the first company in our space to enter the S&P 500. With that, I'm happy to answer any questions that you might have.
All right. Great. Yeah. Yep. Maybe we can kick it off just on the 2026 guidance you provided, specifically what sort of growth you're expecting between prenatal and oncology in 2026. And just curious to hear what you think is what are potential upside drivers for the 2026 numbers you've laid out today.
So we are not dividing the guidance for prenatal oncology just yet. We will be providing every quarter our prenatal oncology revenues. But of course, oncology being at a lower base, we expect oncology to grow much more rapidly than our Prenatal business from a percentage year-over-year perspective. And we are already seeing kind of some of the earlier indications of that.
Okay and then on the bottom line, the 2026 guide for GAAP operating income, you only noted that you're expecting that number to be positive. So maybe just how should we think about the drop-through of gross profit dollars to the operating line in 2026?
I'll let Ross handle that.
Yeah. I think in terms of the drop-through, we're not going to be real specific in terms of what level of operating income we expect in 2026. I think there's still various areas of the business we may choose to invest in, but we do want to operate with GAAP profitability both in the near term and the long term. I'd expect our gross margins will probably be at similar levels in 2026 to where we ended up in the latter part of 2025, high 60-ish% level. But it's going to depend on where we really invest in SG&A and R&D as to how much drops the bottom line. But we think we'll be solidly GAAP profitable on an operating income basis.
And how's the sales force build-out progressing? What was the rep headcount at the end of 2025? And does your guidance for 2026 really depend on hiring a specific number of sales reps?
Our guidance doesn't depend on hiring a specific number of sales reps because we believe that we have the sales reps that are necessary to be able to provide this level of revenue. That said, we will continue to grow our sales team both on prenatal oncology side to be able to set us up well for not just 2026, but 2027 and beyond.
Maybe a couple on prenatal. Can you discuss your growth strategy going forward to maintain the strong growth profile you've seen the last few years, particularly regarding the opportunity to penetrate into health systems and your land and expand approach that you've talked about before?
We have been able to deliver this level of growth despite the fact that we have been really the only company in our space that didn't have Epic Aura integration. This is an area that we are now investing in, and we are an Epic Aura partner. So we believe that this is going to be a big driver of health system growth. We are continuing to see rapid growth in our Prenatal business. That is why we have been able to provide this guidance. The momentum in our both Prenatal and Oncology businesses are just incredibly strong. So we believe that this is a relatively conservative good guidance to start the year.
And then on the competitive front, as other players in the prenatal space start launching competing offerings to Fetal Risk Screen, can you just elaborate on BillionToOne's data moat, its importance to continuing to win share here for you guys, and then really the core technological differences that will help BillionToOne maintain its competitive edge?
I think it comes down to two primary things. One is essentially having a technology that actually does not need partner testing. We believe that a lot of the competing products, even though they are being framed in the same way, they are primarily, especially on the clinical study side, being used after partner testing is done. That is a very different modality, very different type of using the test. And those tests essentially do not pinpoint the actual causal variant.
So even apart from the data mode where we have really large clinical studies with, in some cases, more than 100,000 pregnancies that we have conducted here, the easier thing actually to explain to the provider that everyone else is doing a correlation. They are not really finding the causal variant, and we are finding the causal variant. That is a very simple thing to explain, and that shows essentially why our technology is superior to others.
Maybe moving on to oncology here the last few minutes. You've seen impressive growth in the competitive therapy selection market. Just can you discuss what's driving growth there and the importance of the head-to-head data that you have with Northstar Select just in terms of winning new accounts?
Certainly. I mean, 50% is such a large number that you can go to a provider who is using liquid biopsy and tell them that, "Here's our data." But everyone has their own marketing claim. Everyone claims something that is impressive in oncology. We can tell them, and we have told them, and this is a great commercial strategy for us, is to offer that opportunity for head-to-head in their clinic. They only need to test on 10 head-to-head samples to be able to find that there is one or two patients that have a much better result with our test. T hat has been one of the key strategies where many of these providers who used to be really big advocates for other liquid biopsies to switch to our tests.
Okay and then maybe just one on MRD. You've indicated in the past that your test will outperform other tissue-free tests and could be comparable to tissue-informed tests. So what's the pipeline for data generation for your MRD test? And which indications are you considering targeting first?
So we are generating data. We are going to be doing clinical studies across all major cancer types. But the first available data is going to be in colorectal just so that it would be much easier for providers to compare results head-to-head because that is one area where everyone else has a lot more data. It is not because a tumor-naive MRD is particularly beneficial for colorectal cancer. In fact, colorectal cancer is one of the few cancer types where there is plenty of tumor that can be sequenced. But we believe that that is going to make it clear that this is as good as, if not better, than the leading tumor-informed assays.
Okay. Maybe on a related note, just you've achieved impressive margin profitability at a much smaller scale compared to competitors. What key organizational differences have enabled BillionToOne's industry-leading profitability? And can you discuss any ongoing projects aimed at further reducing COGS, particularly in oncology as you expand into MRD, for example?
I mean, I think I would encourage everyone to read the founder's letter in our S-1 because it really highlights that we really care about every dollar that we spend. There is a really big focus on productivity. But part of it is also how we approach R&D and product building. While many other companies are spending hundreds of millions of dollars in clinical studies, we are paid for those clinical studies.
When one of the biggest institutions in the world looks at many different competitors and chooses us and decides that they would rather use us and pay us for those clinical study tests as opposed to us paying them millions of dollars for that opportunity is one of the biggest differences as well. When you have category-defining products, everyone wants to work with you. You can generate clinical study data and other important data in a way that is so much more efficient than many of our competitors.
Got it. Well, it looks like we're at time. Thank you to the BillionToOne management team for joining us today. Thank you for all of you for coming to the conference. Have a great rest of the week.