Good morning, everyone, and thank you for joining us at the 23rd Needham & Company Healthcare Conference. My name is Gil Blum, and I am a Senior Biotech Analyst here at Needham & Company, and I cover the immuno-o ncology and gene therapy subsectors. It is my pleasure to have with me this morning Recursion's Chief Financial Officer, Mike Secora. So as a reminder to our viewers, any who are watching through the live feed can ask questions through the Ask a Question box below the video feed window. And with that, Mike, so a bit of introduction for our viewers. For those who are unfamiliar with this story, can you kind of walk us through what it is you guys at Recursion are trying to solve?
Excellent. Well, Gil, great to see you again, and always a pleasure to chat. And thank you very much for having me and Recursion be part of this great conference. To answer your question about, you know, what Recursion does, why do we do it, how do we do it, I think it's first, you know, we need to kind of begin by addressing the fundamental problem. And I think that the fundamental problem is that in the space of drug discovery and development, there is arguably great inefficiency in that it takes over $2 billion, 10 years, with a likelihood of success of less than 10% to bring a new drug through clinical development to the market. And I believe that that efficiency has continued to wane.
Yet we do find ourselves at a unique time and place where there has been this confluence of technologies across things like automation, robotics, tools to control biology, like CRISPR, tools to help design novel molecules across digital chemistry, and even tools across AI, whether that is to permute different kind of chemical scaffolds or do unique analysis to find novel insights and advance those novel insights. And I think that we have seen such tools affect, you know, a vast number of industries, and I think we are starting to see such tools now permeate the life science space. Now, Recursion, a decade ago, over a decade ago, was actually founded on this premise, that one could start to unify some of these tools, apply these tools to drug discovery and development, and bring about greater efficiencies.
Now, over that time, we have been able to demonstrate different statistics, highlighting how not only are we finding, you know, advancing programs more quickly compared to industry average, but also, more cheaply compared to the industry average. As well as in, I think, most recently in our 10-K that we had disclosed over a month ago, we started to highlight ways in which we are able to quantify Recursion's unique, data arbitrage, scientific arbitrage, looking at Recursion's massive proprietary data set. Where are we finding targets, that are not known or well known in the corpus of scientific literature? That itself starts to highlight, the greenfield space, the data arbitrage aspect that Recursion can start to leverage.
So therefore, we have started over that, over that decade time, to systematically go through and construct standardized data sets across phenomics, transcriptomics, in-vivomics, which is in the, you know, in vivo data, where we are knocking out every gene in the genome and looking at how one gene can relate to another gene, you know, a compound to another compound, a compound to a gene. And this starts to... As you design these experiments, and you design this data, to have certain properties in place, properties like relatability, meaning an experiment conducted today can relate to one you had done last week, last month, last year, and can relate to experiments that you do into the future.
It starts to give rise to a set of data that is connected, and this has given rise to over 5 trillion different searchable relationships that we are able to start to organize and rank by aspects of novelty, aspects of strength of signal, statistical significance, clinical tractability, chemical tractability. And all of that starts to really accelerate as you unite these different multimodal data sets across phenomics, transcriptomics, proteomics, et cetera, with, as we had a partnership with Tempus that we announced last November, actual patient-centric data across DNA sequence, RNA sequence, clinical record.
You start to get at these causal models of outcome. And we've already, you know, and some of the consequences of this approach is that we have formed some partnerships with the likes of companies like Roche and Genentech in the space of neuroscience, Bayer in the space of drug oncology. Roche, for example, has already optioned their first program last October to make great progress in the construction of certain maps in the space of neuroscience with them.
We advance our own internal pipeline, a vast pipeline, where we have multiple programs in phase II, with readouts happening, multiple phase II, readouts happening this year. And now that we're in Q2, I'm happy to say that we'll have our first phase II, readout next quarter. It's fun to say next quarter now, that's the first time. You know, then we also have exciting new programs coming behind those, like our program RBM39, where we have found a target implicated in CDK12 biology. We're finding novel target, building novel chemistry, doing novel biomarker development, as well as other programs like Target Epsilon.
And then behind all of that is also our data strategy, where we are potentially looking to maybe partner with other large pharma companies or large tech companies, where we may think about novel ways to provide access to some of the tools we have developed, like LOWE, which is our Large Language Model Orchestrated Workflow Engine , and some of our large foundation models in phenomics and chemistry, so on and so forth. I believe that, you know, if we can continue to drive value across our pipeline, our partnerships, our data, and bring about novel efficiencies, I believe that we can have tremendous impact and arguably work on one of the great problems of our time that has, I think, scientific and social consequence.
Okay. So maybe kind of the focus this point a little further and make it clear for, for our audience. It's not just about the size of the data set, it is about repeatability and industrializing data. So in some ways, you're trying to address, you know, the inherent noisiness of individual investigators working across a very, you know, broad and diverse industry. And that's maybe, maybe a better way to kind of view the data sets. It's not just about how much you have of it, but how well it interrelates.
Well, I think that's quite fair. I think that, you know, we should probably just, you know, make an acknowledgement about some of the properties of the data that are pervasive in the life science industry, is that in many cases, analog is a standard. In many cases, data is often siloed. And in having siloed data, things can be forgotten. And, you know, if you are some company who maybe you had some endeavor to work on some large therapeutic area, but for whatever reason, maybe you have, you know, shut down that work in that space because you've reprioritized. What happens to that? Is that forgotten? And if maybe a half a decade or a decade hence, when you want to restart those efforts, do you bring back the data or do you recreate it?
There is this kind of perpetual, almost reinventing the wheel. It doesn't have to be. You know, I think that, you know, we're all quite reproducibility crisis that I would say is pervasive in this industry. Let's first begin with certain philosophical design choices or how to, how to even think about data around reproducibility, relatability, scalability, orthogonality, where, you know, can you not just have data that is relatable, you know, you know, from every experiment to every experiment, but can you also have data that is scalable? Meaning, the more that you accumulate, can you drive greater signal to noise? Can you quantify that, right? Can you also then build models that also highlight better, better properties around recall of canonical aspects of biology?
And can you also augment with different kind of orthogonal data sets for which you are able to confirm or find or refute a given hypothesis in driving that greater refinement. So, overall, Gil, yes, it's not necessarily how much data one has, but rather how much connected, relatable, good data one has. And if one can drive, you know, large amounts of that good connected data, then one can start to find novel insights. And I believe it's data of that sort is the differentiator when one is trying to bring about clarity in understanding a complex system. And biology, human biology is indeed quite a complex system.
Okay, so I want to start with a key focus for a lot of investors, and that's, you know, the news flow from your collaborations. And so what can you say regarding expected milestone payments from your pharma partners in 2024? Can we anticipate larger payments for the kind of foundational work that you're doing, for example, with your collaboration at Roche? Can you parse this? What guidance can you provide?
Sure. I'm happy to talk a little bit about that. Well, I think, I think that it's been fantastic to work with, work with partners like Roche and Genentech. And I think being working with leaders at Roche and Genentech, like Aviv Regev, who has been, I think, an outstanding scientist and leader. And, and I think, I think it's also remarkable, you know, her as a computational biologist leading early research, I think highlights notable change with how one needs to be thinking about digital-first approaches to drug discovery and development.
I would even just acknowledge that, you know, recently, you know, she gave a plenary talk at the AACR conference and, you know, she had a very large audience at that oncology conference, and talked quite favorably about how the. And again, that was just a recent announcement, but, you know, in the last, you know, week or so. But I also even just would call out that some of the work we've done with them, Roche had optioned their first program last October, I believe, finding a novel target that, you know, advancing.
And I think that was remarkable because it meant that Recursion was able to conduct a level of rigorous science, provide that to their committees, have that reviewed. I think, you know, even perhaps some of the folks that even within Roche may have not been fully committed to what we're doing, it's great to even kind of continue to win hearts and minds from folks, you know, in this space. I would also frame as we've given guidance there, Gil, around, you know, near-term potential milestones. There is a potential for option exercises for partnership programs, both with respect to, you know, Roche as well as Bayer, or options for programs that have advanced. Also call out that there is the potential for these large map-building initiatives.
When the Roche and Genentech partnership was announced late 2021, there is up to or exceeding $500 million tied to these map-building efforts. Of course, not all at once, but still very significant. And I think should there be a potential map-building option in the near term, I think it highlights a number of aspects of Recursion. It highlights our ability to do large-scale cell manufacturing. I mean, in order to service the Roche Genentech partnership, we since 2022 have produced over 1 trillion iPSC-derived neurons, and we believe that would put us at one of, if not the largest, iPSC-derived neural cell producer on Earth in order to service this mapping effort.
It also, I think, highlights and validates Recursion's approach. It also, I think, highlights and validates Recursion's business model, and I think continues to highlight and credentialize how we can be an enterprise-scale service provider, an enterprise-scale solution provider, to large pharma companies, as well as large tech companies who want to kind of wade into this space going forward. So these are all things that we've given guidance to potentially watch for on the near term, which would have, you know, revenue consequences.
And as I think we've serviced our partners well, and I think worked well with them, I think there is the potential, as we've given guidance, for additional partnerships in large therapeutic areas, like perhaps CV metabolism or infectious disease or Immunology & Inflammation . These are other kind of large, complex areas that I think would fit nicely, highlighting the plasticity of the Recursion platform, while still allowing us to kind of focus on more precision oncology, rare disease efforts internally.
All right. I do want to focus on one specific venue as to kind of potential opt-ins. This is something I like to pay attention to. Are some of these programs reaching a stage of maturity where we could see a clinical opt-in?
Well, I think, you know, we continue to advance the programs, and we always have to be mindful and respectful to our partners as to what we can disclose, when we can disclose it. But certainly, when we hit a milestone, we will certainly disclose that stage. I think it's important to frame, Gil, that if we look at Target Epsilon, which was a target that we had worked on with Bayer initially in the space of fibrosis, as Bayer was going through a management change, they very much enjoyed working with us. They liked working with us. I think they understood deeply our platform, and I think respected how plastic we could be.
There was a desire to maintain that partnership, and as we were evolving that partnership to focus on undruggable oncology, which is their focus as a company, corporate focus, that was announced in November. Also in line with when we were announcing the Tempus partnership and how we were accessing this patient-centric data, particularly in oncology. At the start of January, I believe the first week of January, we were at JP Morgan, we talked about how we not just updated the partnership with Bayer, which now doubled the per-program milestones, but also we were in-licensing the most advanced program from that partnership.
This is Target Epsilon. It's a novel target implicated in fibrotic disease, and we have already brought that in internally and have already advanced that to the preclinical stage, where it is entering IND-enabling studies. So I hope that frames a little bit of where some of the work is with some of our partners, but of course, we always have to be respectful to them, and when we are able to, when we've reached a milestone.
Maybe you can walk us a bit through the decision-making process as it relates to opting in from some of these partnership programs for Recursion, kind of the way that you decided on the fibrotic program.
Yes. Well, I think in the case of Target Epsilon, I think that, you know, we had, I think we had been doing great work. I think we had a lot of confidence in the program that we had been able to advance. We acknowledged that, you know, Bayer was going through a change for their company and for their focus, but, you know, good science should prevail, and we were happy to option that at, I would argue, quite favorable economics, and I think we've disclosed that in some of our materials, like our 10-K.
And so as we, you know, potentially continue to advance that program, and we talked about that already in IND-enabling studies and maybe watching that to continue to advance. You know, I think it gives us optionality, whether we take that to the clinic ourselves, or whether we potentially look to eventually out-license that, or potentially, maybe repartner in the area of fibrosis with another partner.
Maybe kind of stepping a little back here and looking at partnerships more broadly, how should we view cash flows from partners as they evolve over time?
Sure. Well, I think if you look at the framing of some of the partnerships that we've had so far, there has been a meaningful upfront payment. If we look at, say, Roche Genentech, there are research-tied milestones, not tied to any programs, but rather in the construction of these maps. I think that's a novel evolution in partnership deal structures. Again, $500 million there. Then, you know, there is up to or exceeding $12 billion tied to up to or exceeding $300 million for each of the 40 potential programs, and then royalties on net sales. I think as we have different programs.
Opt in, there could be a revenue payment to Recursion as there is new maps that have reached a certain status, there could be a revenue payment to Recursion. And as we have maybe new partnerships being announced, that would potentially give a new, you know, new cash flow into the company. And I think all of this continues to drive Recursion increasingly to sustainable operation as we advance, as we mature, and expand our pipeline, as we continue to service our partnerships with enterprise-scale solutions, and again, continue to drive to greater sustainable operation.
Before we move to the clinical section, do you think that there is some, you know, maybe pent-up demand and some companies waiting to see, quote-unquote, "a clinical-level proof of concept"?
Well, I think that the what gives people or potential partners confidence, you know, varies a good bit. I think that it all sits within a spectrum. I think, you know, we can highlight a number of different successes that Recursion has, whether it be the phase I, trials that we've executed, the phase II, trials that we are currently executing and have phase II, readouts coming, whether it's that we have demonstrated finding novel targets. Now, in this case, multiple for multiple partners that they have had a lot of confidence in or have optioned already or in the data constructions that we have had.
Depending on the person, depending on the partner, you know, each of these give rise to a different level of confidence, and I think there is somewhere phase II, data might be it for them, just as it would be for different investors as well. But I do think that Recursion, over my time here and over the last decade, has shown itself, has validated itself across aspects of its pipeline, its partnerships, its data, and continues to do that. And I think what with that continues to define the space and bring leadership.
So shifting gears to the clinical readout that you mentioned is coming up in the third quarter from your CCM program. Maybe a good place to start, I'm sure not all people online right now are familiar with this indication. Can you give us kind of like the top background of this? What is it, and what is the epidemiology? And mostly, what is the current standard of care?
Sure, happy to. So CCM is a cerebral cavernous malformation. It is a rare disease affecting approximately 360,000 patients in the U.S. and EU5. It affects over 1 million patients worldwide. It is characterized by these. It's a disease of the vascular, characterized by multiple lesions in the brain that can be identified via MRI. For as large a disease as it is, there is no approved therapy, and most, you know, patients receive no treatment or some, only some kind of symptomatic therapy. Surgery may be attempted where it's possible. However, given that there are multiple lesions and they can be embedded within the eloquent structures, in many cases surgery is not tractable.
How this disease results is it can be a loss-of-function mutation from the CCM 1, 2, 3 genes, and, you know, ultimately is characterized by vascular instability, vascular leak within the brain. The symptoms that arise from here are neurological deficits, hemorrhage, and other potential symptoms. And I would say that, you know, this is... We're very much looking forward to this phase II, readout next quarter, where we also have Orphan Drug Designation from the FDA. And I know that this is a disease that is not particularly well known.
This is perhaps the case with, with many rare diseases, but it's, it's also quite fascinating how, and I believe that this disease could probably be quite underdiagnosed, because it's a slowly evolving disease. It's often identified, when someone has undergone some kind of, brain trauma, for example, head trauma, from a concussion or a car accident. They go into the ER, they get, you know, MRI, and sure enough, there are these, lesions that show up in the brain. However, given that this is a, you know, there is a familial form of this disease, it, it can be screened for. You can, know if you are a carrier of this, and, and this, it might, you know, kind of present itself.
So, kind of maybe to double down on that specific topic, I mean, pretty large, rare, and has pretty high prevalence, but not a lot of people in the investment communities are familiar with it. And I don't remember seeing any other development program. So why do you think that is?
I think, Gil, that I think there is a bit of a vicious cycle with some rare diseases, that if there is no therapy, then it can be underdiagnosed, and if it is underdiagnosed, then there is no knowledge of it, and it just kind of continues to, unfortunately, self-reinforce that lack of awareness. This, I think, also had been a disease that had been perceived as perhaps, you know, undruggable. And I know that in—when Chris, our CEO, co-founder, and Dean Li, his advisor in grad school, who's a, you know, vascular physician by background, they. You know, when they first started to kind of wade into the disease, they were starting to work with some of the gold standard models in this space. And ultimately.
They were finding failure in some of the experiments that they were doing, and this gave rise to this kind of unbiased approach that Recursion adopted, which gave rise to this unbiased phenotypic screening, gave rise to finding novel insight in the way that one could perhaps approach this disease. And so I do think that there is a vicious cycle of no approved therapy, under diagnosis, lack of knowledge, as well as the fact that this disease had been something that perhaps had just not been well understood for quite some time.
But I think the, you know, the work that Chris and Dean have done, the preclinical work that gave us confidence, the phase I work that we had conducted and now the phase II that we have been able to fully enroll, where we have, most of the patients have now gone through a, you know, that have gone on to 12 months of treatment, have opted in to staying on to the long-term extension. I think we continue to gain confidence in how we potentially could bring some novel therapy to patients in need here.
Maybe that's actually an interesting point to focus on here. How did you enroll this study? It's fully enrolled, and people are staying on an AAP.
Absolutely. So, to talk a little bit about, perhaps the, the, the trial design, sort of the enrollment criteria. So this trial initiated, this phase II trial, initiated, Q1 2022, a 12-month, treatment, where there's follow-up visits, approximately, every quarter or so. Again, this is a placebo-controlled trial. The criteria we had was, MRI, MRI-confirmed CCM lesions, so seeing that those lesions are indeed present, looking at, familial or sporadic forms of the disease, and then seeing that there are symptoms directly tied, to CCM. You know, I think we've been working you know, closely with, different, sites as well as different PIs. And again, happy to have that trial fully enrolled, where the vast majority of our participants have gone on to that long-term extension study.
We're gonna have a 12-month, you know, 12-month readout here. Lots of several MRI-based assessments, including size of lesions, number of lesions, the rate of change, and some report, clinician-reported outcomes as well. What kind of data and what magnitude of delta would you need to see to consider it as a significant?
Great question. Well, I think, you know, because we are, you know, as you framed, we are perhaps the most advanced company, we are the most advanced company in the clinic with this disease. We've been good partners with the FDA to look at different outcome measures. And I think, of course, there's outcome measures related to safety and tolerability, which I think has been well demonstrated in our phase I. We had a quite clean molecule there, and I think the choice of patients to opt into the long-term extension helps to kind of justify aspects of that safety.
But then there's a number of efficacy measures across imaging of CCM lesions, as you articulate, around number of lesions, size of lesions, rate of change of lesions, as well as perhaps some more softer reported outcomes both from the patient as well as from the clinician. And there's approximately 10 or so, you know, efficacy measures that the FDA wanted to see to start to really define the guardrails for this disease, knowing that we're pioneers within this space. Now, to your question about, you know, what would we want to see, what could be helpful in kind of advancing this program?
Well, I think, you know, some kind of, you know, coupling of the imaging data, looking at how the lesions are changing, you know, across these approximately quarterly visits that they come in. As well as, you know, seeing some of the outcome measures, either from the patient or from the clinician, around how they feel, how this makes them feel. Are they feeling as if they're achieving some kind of disease stasis? Just the feel of kind of a some kind of return to normalcy. That's important. So I think a combination of these things could be quite helpful in advancing this program.
A related question: What do you think are the most likely patients who would want or need this intervention? I mean, based on your safety profile, there's no reason for it to be restricted, but is it, do you think it'll be everyone or maybe surgery-ineligible patient, others?
That's a great question. I think that the safety profile suggests that one could probably be a bit more expansive. And certainly, this is a large enough population where there is no approved therapy. I would be happy for us to be, you know, be able to service those patients in that way. I think also, to your point, surgery-ineligible patients make for perhaps the most obvious in need of it. But, you know, we had a patient actually on-site about a month or so ago. We often have patients come on-site and talk to the team about their patient journey.
Who is a CCM, you know, she is, she's a CCM patient, and she was talking about how the disease presented in herself, where, you know, one day she just goes through this kind of moment of where she blacks out, loses consciousness. She falls, hits her head, you know, has a concussion, comes to, children get her to the ER. You know, she goes on an MRI. They see these lesions. The doctors themselves think, "Well, can't be CCM," right? There's this kind of bias among the diagnosis. And it's also around this time that I believe she hears from her father that he actually is a carrier of CCM and then this is kind of added to the calculus, and yet the doctors still don't pay attention enough to this.
But then she gets screened, and I think she goes to see some specialist, and indeed, it's CCM. She has CCM. So I think that this is a slow-presenting disease. I think this could be well coupled with better screening. And I think with that better screening, I think one can, you know, help address a more pervasive, you know, population. But I do think that, you know, surgery- in eligible patients would certainly have the foremost need, perhaps for having some kind of therapeutic.
Maybe I have a final one here. So just given that CCM is one of Recursion's older development programs. How much weight should investors give to clinical success as validating for your broader tech stack?
Sure. Well, I think that as we have, you know, built our operating system over the last decade or so, we are always integrating new technologies, scaling technologies, and automating technologies in pursuit of additional insights, more refinement in those insights, and better drugs to patients faster, cheaper, as we look to have impact. And that means across different aspects, biological experiments, data tools, chemical design and validation, translation, et cetera. The fact that, you know, CCM had been one of our earlier programs, I think we had been able to identify and uncover a novel insight. I think for that reason, we have also we have good confidence in what we've been able to identify, you know, through the an earlier form of the operating system. I think our preclinical data helps give us confidence in that.
I think our phase I data helps give us confidence in that. I think, you know, the phase II that we will have and the fact that patients are choosing to go on to the long-term extension study, I think gives us confidence in that. And I think that we're happy to be working with the FDA to really set guardrails for how to think about, you know, measured outcomes for this disease. Like I said, I think that some combination of this imaging data and these important outcomes will start to give confidence that we're having some effect in this disease, and it's set up for how we think about next steps.
Okay. I do want to spend some of the remaining time talking about your tech stack, as you've done a lot of upgrades recently. So maybe a good place to start, much has been made about your ongoing relationship with NVIDIA, including an equity stake from the large chip manufacturer. So should we consider the relationship from for Recursion a specific one, or is this more a broad-based investment in use of AI for drug development? And this is kind of in light of recent announcements.
Sure. Well, I think that, you know, I think that, you know, NVIDIA is, has been a great partner for Recursion, not just recently, but, you know, we have a partner... You know, we've had-- We've been working with them for about five years, give or take. They helped us, you know, design our first supercomputer in late 2020, early 2021. I think they have seen us grow our data set many times over. They've seen us use our supercomputer in, in novel ways to drive insight. They've seen us acquire, companies like Cyclica, Valence, which help bolster digital chemistry and generative AI capabilities. These are teams that, you know, these large tech companies, would be quite familiar with, particularly as they've competed in certain open ML AI competition.
So I think that there's been some great respect, mutual respect from both organizations. And even at NVIDIA's GTC conference, you know, Recursion certainly mentioned in a number of comments there. And I think he's been, you know, a remarkable leader, knowing that he's been a founder CEO for 30 years. Often draws a lot of parallels between how the semiconductor industry had evolved, say, in the 80s and 90s, from where you're in a wet lab of sorts, designing chips, testing chips, all to kind of give you confidence that when you push that button on production, it's gonna be good to go. And that's very much almost like drug discovery a little bit, in that, you know, you're designing it, you're discovering aspects of a potential drug.
You're designing it, you're developing it. But when you're ready to kind of push that button, to you better make sure that's gonna translate. Otherwise, you're gonna give rise to significant failure. I think that there is parallels between those two industries that he often draws on, which I think are quite, quite true. And he's also demonstrated, I think, a certain keenness to follow this space, follow Recursion. And I think that, you know, I think as one views their company as increasingly focusing on data centers, hardware provision, and service of data centers, life science is very much a greenfield opportunity for which I think great augmentation and disruption can come from these tools, for which I think Recursion has been a pioneer of such philosophies over the last decade. And I expect...
You know, it's my belief that large tech companies like NVIDIA and others will perhaps increasingly wade into the life science space, as we've already seen with, you know, if you go back in the last year or so, you know, Oracle and Cerner, OpenAI and Microsoft launching BioGPT, Amazon being involved in their first phase I trial, acquiring One Medical, acquiring PillPack. There's, again, large tech seems to be taking to wade into this space in novel ways.
And I think that represents an opportunity for companies who are able to build and command some purpose-built datasets fit for machine learning and AI applications like Recursion has done. Going back to some of your first questions around the designs of different datasets, around relatability, repeatability, scalability, orthogonality, all of that, just tech-centric ways of thinking about data, which we have embraced from our onset.
So inclusive to your NVIDIA relationship, you're added to BioNeMo, which is effectively their marketplace for AI-powered drug discovery models. So what ultimately are you listing on this marketplace, and how are you gonna, how can you monetize this relationship? Should we think about this as a revenue opportunity standalone, or more of an awareness driver, or, you know, co-drive collaboration over time with additional funding? How is this best viewed?
Sure. Absolutely. So at the JPMorgan conference at the start of the year, NVIDIA had announced how Recursion would be hosting a model on BioNeMo. And since that time, we've started to host a model called Phenom- Beta. This is a foundation model predicated on some of our phenomics experiments. We launched this model on BioNeMo as a non-commercial model to give people more awareness of this model. But certainly, we were able to kind of.
We can see what that access looks like, and should different commercial designs want to be put in place, then this gives rise to how we might want to structure enterprise-scale solutions with some kind of potential partner. I think that the way we think about our data models, as well as our partnerships in general, is that we often look at that, you know, business line from enterprise-scale solution provision, whereby, you know, we want to be working on large opportunities, large scientific areas, that are going to be transformational for us, for our patients, for the partner, and also for our shareholders.
I think whether it be tools like our Phenomics foundation model, whether it be other foundation models that we might want to, you know, consider making available either in chemistry or other kind of areas, or tools like LOWE, which is our LLM Orchestrated Workflow Engine, it's potentially likely that we would want to perhaps pursue commercialization opportunities by giving access to those tools via some kind of large partnership, maybe a new partnership, where those tools are made available. There may be some kind of economics tied to that access, that partnership, knowing that, again, we want to be able to be working on transformational problems, transformational partnerships.
We are almost at time, but I want to address one more key aspect, and that's Recursion added a lot of technology in the last year.
Mm-hmm.
How is the integration process going for all of these different collaboration and tools, both internal and external? Are you just gonna keep expanding your database, or is there a point of diminishing returns?
Great question, Gil. I think that over the years that I have known Recursion and been working at Recursion, I would say that we add many novel capabilities almost every six months or a year. So, that hasn't changed in the time that I've been here. We have acquired multiple companies, multiple technologies, developed them internally, all in service of of again building datasets that, you know, highlight across biology, chemistry, patient outcomes. How are we finding the right kind of insight that can be driven to a program efficiently, cost-effectively, and at scale, knowing that, you know, many of these programs, there's many diseases that need to be pursued.
We talked in our 10-K of even this year, potentially, running, multiple hundreds of proto programs through our operating system, for which some of those we might advance internally, some of those we might look to out-license, some of those we might look to partner, so on and so forth. It's about how you continue to bring volume, increasing volume at scale with these capabilities, and I think in a new way, in a novel way for yourself, for your partners. So that's a long way to kind of frame that we will continue to add capabilities, but it's not just about diminishing return, it's about that orthogonal nature. You know, if we've gotten to some point of having a good, robust amount of phenomic data, now let's build the transcriptomic data that helps confirm, refine, or refute.
We've now, I think in the 10-K, we talked about how we've amassed over 700,000 different transcriptomic observations. We now are operating as perhaps one of the largest transcriptomic data producers on Earth. You can imagine that also scaling to things like proteomics, right? You can also think about how we're thinking about novel biomarker development, how we're thinking about even industrializing clinical development itself, all to kind of provide that full-stack solution from target identification to clinical trial execution, perhaps, you know, beyond.
Okay. Well, we are a little past time, Mike. Again, thank you very much for your time today and walking us through not the easiest story.
Always a pleasure to connect with you. Always happy to be chatting about this value proposition, which, you know, I view as really tackling one of the great problems of our time, and I believe we can do this. We can have tremendous impact in science and in society.