Welcome back to our afternoon session of 2025's Global Healthcare Conference. I'm still Mani Foroohar, Senior Analyst, Genetic Medicines. I'm very fortunate to be hosting essentially the entire Schrödinger Management Team. I know you guys have a few slides you want to run through. Do you want to kick those off, and then we'll sort of run with that?
Yeah, yeah. Let me just kick us off a little bit, just give a real quick background on the company that might help with the discussion. Let me start by talking about our platform and why we developed this platform and what we're actually trying to do. The key to drug discovery is predicting the properties of molecules as accurately as possible. There are basically two approaches to doing that. One is you heard a lot about machine learning, AI, right? You've heard all about that. That approach can be useful. It's very fast, but it requires massive training sets. You keep hearing about that. Another approach that does not require training sets is sort of a first principles method using physics, actually simulating the property that you're trying to compute.
That has some very significant advantages, which is that it's very accurate, and it can be used on any molecule. It doesn't matter if it's a completely novel molecule, if it's in a new chemical space, a new target, it works just fine. It is slow. It's computationally expensive. By combining the two methods, now you have something really powerful. You can generate massive training sets for machine learning using simulation, using first principles methods. That's the platform that we've developed. It's allowing us to predict the properties of molecules with very high accuracy, but doing it on a massive scale. You get the advantage of scale and speed of machine learning with the accuracy of physics-based methods. That platform, we leverage in a number of ways.
One is we license the software that we've developed, this platform, this physics and machine learning platform, to life science companies and material science companies. That's one of the advantages of physics-based methods. It's sort of agnostic to the system. Physics is physics, right? It works on materials, works on biologics, on small molecules, on any kind of protein, any therapeutic area. It's physics. We license the software. We also establish collaborations with drug discovery, obviously drug discovery collaborations and materials collaborations as well. You can see there we have a number of collaborations. More recently, although it's not so recent anymore, we have proprietary programs. You can see there how many active programs we have. That's how we're leveraging the platform. There's a little sort of hint here by overlapping these different businesses.
There are a lot of synergies, which you can probably imagine there are. Maybe we'll get into some of those between these different businesses. Just some quick highlights. This will be the second to last slide. We had a great fourth quarter. You can see there the software revenue, the software part of our business grew 16% in Q4. Total revenue was $88.3 million in Q4. The full year was also a really excellent year. Our guidance was 10%-13%. We were very happy with that software revenue guidance result of 13.3%, $180 million in the full year. We guided to 10%-15% growth in the software business for 2025. You can see that over on the right.
You can also see that we guided to 45% to 50 million in revenue from the drug discovery business, which is a pretty big increase from the drug discovery revenue in 2024. I think this is a really important point, the collaborations have been going very well. That is why we have so many of the collaborations. Even existing collaborations, the one we had with Lilly, the one with Otsuka, those were existing collaborations. They were expanded last year. I think that is clearly a reflection of how well those collaborations are going. We are very excited to be presenting data for the first time on our own proprietary programs, clinical data, readouts from the three clinical programs that we have. I am sure we will talk about those. The first one being in the second, I was about to say in this quarter.
Second quarter.
Yeah, second quarter, yeah. Q2. I forgot what month we were in. In the second quarter of this year. The other two programs that we'll be presenting on later in the year. We're also very excited about the new technology that we've been developing. We put a lot of effort into this. We have a very large software development team at the company. We continue to advance the platform. We're pretty excited about this predictive toxicology project. This is a lot of failure of programs and drug discovery programs preclinically and clinically are the result of binding to off-targets. We've developed a physics-based method for predicting toxicity associated with binding to off-targets. We've enabled quite a few targets there. Maybe we'll talk more about that later too.
We've also released this year, actually released it last year, late last year, a platform for biologics. We continue to advance what's called a force field, which is sort of what underlies all the physics. This is allowing us to do, that stands for machine learned force field. That's allowing us to model, for example, the chemistry that occurs in batteries. That turns out to be really complicated. It requires sort of enhanced force fields. You can see that we're doing that. I think I'll just leave it on this slide. I think we should get to the questions. This just gives you a sense of the sort of advancements in the platform that span really the whole drug discovery process from after-target selection all the way to preclinical development. You can see there that we've been pretty busy advancing our platform.
Maybe we'll get into some of these new technologies as well. Hopefully that was helpful.
Awesome.
Get us started.
A lot of places to bounce around. I'll start with the more sort of, also the dry financial question, which I think I get a lot of as people try and parse out what guidance means. When we talk about a $45-$50 million of drug discovery revenue, what assumptions flow into that? What are kind of the upside, downside drivers in terms of events and partnerships? Is it driven by striking new partnerships? What are the variables that could move that number throughout this year?
Yeah. Our philosophy is that we try to guide to the most likely outcome or range of outcomes. We're not guiding to all possible range of outcomes. The drug discovery revenue consists primarily of the amortization of upfront payments that we recognize over time as we complete the work. We've received those payments. We've got the cash, and we're just recognizing the revenue as we do the work. It's subject to the variance of the project plan and does things go faster or slower, but that's not a huge amount of variance. There are small pieces in the guide, as there can be large pieces in our actual results that come from interim milestones. We get to a certain point, and we've delivered a DC or a candidate of some kind, and that can trigger bigger milestones that happen in 2023.
We probability adjust those outcomes because they're not in our control or not fully in our control at least. That goes into the 45-50. The last piece is, do we account for new deals, business development, new collaborations, transactions on proprietary assets? The answer is there's a tiny, heavily probability-adjusted contribution actually in that range. Achieving the range is not dependent upon that kind of uncertain outcome. We try to capture kind of all the opportunities, but it is mostly driven by those upfront payments and the amortization of those previously received payments.
As we look forward to the continued growth of the business, one of the things that we talked about on the last call and a couple of times over the course of the last year on our Polar Express bus tour, et cetera, was where we are in terms of the adoption curve on the part of pharma, large cap biotech, drug developer counterparts, broadly speaking. Where are we inning-wise in terms of that adoption in the R&D budget? Separately, where are we in terms of the partnering BD budget that looks more like what pops up in terms of collaboration, drug discovery, et cetera? Are those at the same stage of development, those conversations?
Yeah. Maybe I can take the first one. As broadly used as the platform is now by essentially every pharma company, every pharma company is using the software, we're still actually pretty early days. The reason is, first of all, a lot of these physics-based calculations I was telling you about and the incorporation with machine learning is actually relatively new. It required a lot of things to happen. First of all, computers had to get fast enough. We had to figure out how to run these things on graphics cards, on GPUs, which are much, much faster than CPUs. I mean, many orders of magnitude faster. That had to happen thanks to NVIDIA. We had to have a better understanding of protein structure.
That's relatively new, sort of structural biology revolution, which was fueled initially through X-ray crystallography, then cryo-EM, and then even computational methods like AlphaFold. We had to advance the technology in the way we have. It really has just come online relatively recently. We've only relatively recently really learned how to use this technology at the massive scale that we're using it internally. That is important because here's where we are really. We're using the technology, obviously, on the scale that is not limited in any way by the cost of license internally, right? The cost of licenses or availability of computers because of our relationship with Google. We're just using the software at the level that it needs to be used to achieve these kinds of successes, not only our internal programs, but the collaborations that we have, which are well-known.
That scale of usage is still many orders of magnitude higher than the usage across the industry. Now, the good news is there are some companies, and it's a non-trivial number now, which is good. It's not just an N of 1, that are getting pretty close to the scale of usage that we're using internally. That's encouraging. There's still a lot of room for growth from those companies. You have the companies that are behind those companies, right? I hope that gives you a sense of that it's encouraging where we are, but there's a lot more room to grow. I think we will continue to see broader and broader adoption at sort of massive scales across the industry over the coming years.
Karen, do you want to talk about the question of the budgets and between BD, collaborations versus software?
Yeah. I mean, I think there's usually a couple of different reasons for collaborations. It's either companies who are adopting, as Ramy's explained, where they are trying to understand how do we run this at scale. You've seen us do a lot of collaborations like that with Lilly and most recently with Novartis, where there's a blend of things going on. They're watching how we use the technology so they can get the most out of it. There's also the opportunity for us to actually partner early programs with those companies. There's a lot of discussion about technology and targets. That drives business development discussions. To your question about how much more of this can we do, obviously we have an active pipeline of proprietary and collaborative programs. We're always talking with pharma either about targets or technology.
I foresee that we'll be doing more of these collaborations in the future, depending on what the pushes and pulls are around targets, therapeutic areas, and technology.
Something that I hear from investors from a more bearish perspective is if you have a level of utilization of your platform internally, because the internal cost of capital, as it were, of using your own platform is quite low. It's at cost by definition internally. On a marginal.
That's right.
For a marginal unit, it's very low.
Yeah.
Whereas a pharma counterpart, that marginal cost is not zero. It's whatever they're paying you. Also, they have a universe of other software suites and partners, et cetera, that they're using. When you think about, when you apply that lens, where are we in terms of share of budget capture?
Yeah.
Which I think is maybe a more reasonable metric than what does it mean if we take all the pharma and they act like Schrödinger internally? I don't think that's like that's kind of an asymptotic number that's not reachable.
Yeah.
How do we think about share of budget capture?
Yeah. That's something we've looked at pretty carefully, Jeffrey.
We took the top 30 pharma companies by R&D spend. Then we compared that to their spend on our technology. Right now, the largest companies in that top 30 are spending less than 0.15% of their R&D spend on our technology. As they move to the right, as in increase their spend on our technology, we're still below that threshold. When a lot of people are asking us, well, what's going on in pharma? How do they feel about H cuts, FDA cuts, and everything? We are so small still that we are noise. We're actually value creating for these companies in the sense that we're enabling to discover molecules. We're less headcount, less molecules being synthesized. We aren't getting sort of negative pressure on us. There's a lot of upward room further above.
The other thing that I would say is we've seen biotech companies emerge that are computation first. They're sort of approaching the problem much the same way we approach the problem with people that we know. They have been phenomenally successful advancing drugs to the clinic, in many cases being acquired. They are also using the technology at the scale we are using it, which is millions of dollars worth per program. If you compare them to the big pharma companies, they are much more intensive in their use of the technology than the big pharma peers. It's just that it would be hard for a private biotech company, for example, to be spending $10 million worth of software or something like that. They're actually using it in the way we are.
They're leading the way for the industry, including for the pharma companies that acquire them.
As we think about the growth and penetration in the larger companies, I think a conversation that I have a lot of investors, in part because they do not participate in these sort of behind closed doors negotiations like this, is to what extent is Schrödinger obligated to show they are displacing some other less efficient cost versus to what extent are you guys obligated to the extent of capturing more revenue? Is it dependent upon showing, hey, this is faster or more efficient or look, we avoided this toxicology issue with X % of frequency versus your lowers of technology using before? Which is the key driver of novel partnership capture?
Yeah, Jeffrey, do you have any thoughts on that?
Yeah. I mean, I think we are.
It's an infinitely vast question I'm asking. I get it.
We aren't being judged on our ability to actually reduce cost in drug discovery. We don't have to go in and sort of run a project saying, OK, we can come up with a molecule for this versus that. I mean, the way I would like it, having at least seen it for a couple of years now, is it's almost like a light bulb goes off. The light bulb is, OK, as an industry, we're not going to design, I don't know, aircraft without using CAD/CAM. We're not going to make a cartoon without using a computer. We're not going to, I hate to use a close-to-home analogy, we're not going to run a hedge fund without having a data stream. We're not going to discover drugs without using computers. I mean, it's that simple. Progressively, the industry is realizing that.
When they realize it, it's like a light bulb goes off, as I said at the beginning. They say, wow, we've got to do this a totally different way. This is no longer a craft. This is an industrial activity. Then they adopt it at scale.
There's another, I think, metric that perhaps is worth considering, which is the quality of the molecules. We recently came across a statistic that over the last three years, about $40 billion of M&A transactions have been done in the I&I space. What was amazing when we looked at that was about $9 billion of that were associated with molecules that Schrödinger designed. So Nimbus is TYK2, Morphic's alpha-4 beta-7, which was a switch from a biologic to a small molecule. Quality of molecule, but also cracking really hard drug discovery problems. Currently, our collaborations with Novartis and Lilly are around those types of things where working with our platform means you can win on those target product profiles. That, I think, is a huge driver of it's an important metric, I'll say.
Sorry. We hear this pushback quite a bit about the scale that we're using it at. One of the things that I realized, I think, last year is that in our collaborations, which have been going successfully, the scale that we deploy the technology in the project, in one project against a target that's identified by a company, and that's a big software contract. The scale we're using it is several fold larger than their entire software contract. Part of the reason that Otsuka or Lilly or Novartis or any other company might come to us is to say, what can you do by bursting this technology to massive scale? That is what we're doing. I would say the pushback to that kind of resistance is it actually makes a difference when you ramp the scale way up.
That is why we're having all the success with these collaborations.
I'm going to extrapolate from that to sort of the natural extrapolation, which is if going to massive scale in a partnership improves asset quality, the maximum improvement in asset quality should be internal, where you remove all constraints implicitly. Sort of the eating one's own home cooking kind of thing. What are we going to be looking at in terms of, in your mind, what are sort of the key metrics? We'll start with 1505, which I think is in 2Q. What are we going to be looking from that data set to say this asset quality is meaningfully differentiated and this is what you can get? Because it will be looked at not just as a data set for what's the DCF value of this molecule. That'll be the smallest part of that component.
For investors, it's how is this evidence that more light bulbs would go off in more offices in Switzerland, New Jersey, it's Japan, et cetera?
Yeah. First of all, maybe we could just say a word or two about 1505 for people who don't know what it is.
All right.
This is our MALT1 inhibitor. It is a dose escalation study. If people don't know what MALT1 is, it's in the BTK pathway. I think we're all very familiar with BTK. It's a very successful franchise in the pharma industry. We designed a molecule. It's an allosteric inhibitor with the objective of being extremely potent and shutting down MALT1 activity, which thereby shuts down NF-kappa B activity and can be additive to a BTK inhibitor. We think this is a super exciting area for B-cell malignancies and potentially actually other areas as well. What we did was we designed an allosteric inhibitor. First generations were orthosteric, not very well behaved, not very drug-like. This allosteric inhibitor that we've designed, the goal was to design one where you can literally shut down signaling of NF-kappa B. How are we measuring that?
We've shown in our phase I healthy volunteer study that we have a very well behaved molecule. It's behaving well from a safety point of view, PK and PD. What we're looking forward to this year is sharing that data in patients. In patients, we've been dose escalating in an all comers trial in B-cell malignancy patients, a heterogeneous population. What we're looking for is how well did it behave in terms of its profile, PK, PD, but also we're looking for clinical activity. This is a signal seeking study. It's a dose escalation trial. We want to see that just like in the healthy volunteer trial, that this behaves very well in patients and leads to clinical activity.
While it's an early set of clinical activity, not powered for efficacy or anything like that, that's what we hope to share in second quarter.
I just want to point out the TYK2 program, alpha4beta7 program, quite a number of other programs that have come from our collaborations have all leveraged the technology at scale, at the full scale. I think this is not the first time we get to see the impact of that. We've seen it over and over again. I think that has been pretty compelling. It is probably what continues to lead to more and more of these types of collaborations and the broader adoption of the technology. Yeah.
I think one of the disconnects that I've perceived has been that there is a unit that investors look at a molecule and a data set and they say, OK, how do I compare this versus this pivotal, this phase two J&J prior combination with this other? They tend to heavily weight, well, public investors tend to heavily weight late stage data sets. Whereas I'm going to stop right there and not turn this into a 17-part money question. They tend to favor late stage data sets.
Given that most of the light bulb moments that you guys have had have been partnerships that have been driven by data sets that were not necessarily even in humans yet, to what do you need to show, or I guess what metrics are followed most closely in these data sets in 2Q, but also we want, et cetera, later on in the year, that influence decision making at pharma partners, even if they may not compare directly to someone else's pivotal randomized phase three?
I mean, I would say the biggest litmus test perhaps is, is this a developable molecule for an exciting target? That is the litmus test for can this show enough activity to continue investing and to continue collaborating to demonstrate the full power of these assets? That does not sound very platformy. At the end of the day, when you're in phase one, that's what people are looking for is, again, does it hit the target? Has it got potential for combinations? What are the indications that we can possibly go after with this asset? Now, I'll just leave MALT1 for a minute and talk about CDC7. There have been previous CDC7 inhibitors. In fact, I think one of the earliest ones actually fell out of phase one within the first couple of months because of poor PK, hard to balance the PK, and also the selectivity.
With that molecule, I think the question is, how well behaved is this compound? Is it well tolerated? Does it have good PK? Can it continue in the clinic? In a dose escalation study, in a first in human, that's what you're looking for. While we're very excited about the potential of these assets in the long run, as you said, in large phase two studies, this year it's all about, are these good molecules? Can they continue in development?
One of the--I'm going to step back for a moment, if that's OK. When you think about the combo with BTK, how do you think about which BTK?
That's something that requires a lot of discussion. I will just give you my sense of this is that that landscape has continued to evolve. This is a, going on 20-year maybe franchise. You've got multiple generations of inhibitors. Now you have what's coming forward are degraders. I would say that there's been a lot of change in the BTK landscape over the last few years. You've got your covalent, your non-covalent. As I said, the degraders are coming. We are well aware that there are next generation BTK inhibitors that have pretty good safety profile and are expanding as they've landed. We're still figuring out which one we think is best to combine with. I ultimately think that the degraders, both on the BTK side—I talked about BTK—but MALT1 and CDC7 actually are both potentially also combinable with BCL2.
Even there, you have your first-gen inhibitors. Now you've got your degraders coming through. If you think about where we'll be with these types of novel agents, MALT1 and CDC7, in about three years' time, the landscape's going to have changed, I think, again. There's the opportunity not just to combine with the first-gen or second-gen BTK inhibitor, but also potentially with the degraders that are coming through. That's something we're obviously thinking about. Where will these types of assets be? Who are the best partners? Who are the best in terms of the growing franchises? Those all have to be taken into consideration. As we've said in the past, we're focused today on monotherapy showing activity. We ultimately think that there is a very important opportunity to combine with BCL2 and BTK.
I think that will evolve over the next couple of years.
The debate that I hear on this topic—I know I'm drilling down a pretty small part of the story, so forgive me—the debate that I hear is whether or not that decision, that combination decision, is really a biological and clinical one, or if it's largely like a corporate dollars and cents strategy question. Because on some level, you're going to show the drug characteristics of your own agent on top of some degree of inhibition by whatever mechanism.
Yeah.
Separately, you're going to want to eventually pursue a combination strategy that will have regulatory implications and maybe partnership. How do you think about balancing those two? Are they sufficiently independent that they're just separate time horizons? Two years, two, three years apart doesn't matter.
I think there's a time element to it. I'll also say that there is a biological piece to this. Patients with these types of diseases—let's just stick with lymphoma for a moment—they want to be on drugs that are well tolerated. They're living for a long time. They don't want overlapping tox or anything like that that's going to make their quality of life bad. We also have to think about the overlapping tox potential, which agents are going to be well tolerated in combinations. Even triple combinations are being talked about. There is that biological lens on it, I would say. There is also then the sort of commercial and regulatory strategy. There are many different types of lymphoma that these agents can work in.
I think while, for example, MALT1, I think, is a driver in certain subsets, the question is, where are you seeing the most active combinations that would lead to a regulatory and commercially successful opportunity? I think without picking my favorite BTK or BCL2, I think it's important that we understand the breadth of opportunity so that we do signal seeking in combination in the right setup. That may not be ultimately the right regulatory combination. Maybe a different outcome with a degrader, for example.
Does that imply rationality for like a basket study approach? Or is that just too complicated?
I think it's a little too early to say.
Yeah. No, I think that's right.
Yeah.
I'm going to hop over to an actual stock question, if you'll forgive me in the last couple of minutes or so. You have a hybrid shareholder base. And some were very focused on traditional tech-enabled business services type metrics. I don't want to use SaaS incorrectly, but SaaS-ish.
Yep.
SaaS-y. Oh, I've got to use that. Conversely, biotech investors who are more focused on traditional biotech metrics show me a late stage data set. This data set versus J&J is that data set versus AbbVie’s and so on and so forth. Clearly, you've committed to growth in the top line that exceeds growth in OpEx and monetizing the benefits of operating leverage. That plays well, most well with a group of investors who would much rather see the pipeline agents find a home somewhere else in exchange for some amount of capital. That’s not necessarily the right answer on an NPV basis. How do you balance being judged by two very different groups that have almost different philosophies of investing? How do you think about balancing those two?
If you say the software investors are the left hand and the biotech investors are the right hand, both of them do not want you to be raising lots of capital iteratively. We think that we have a self-sustaining sort of biotech model. We also have a very interesting, steadily growing, incredibly sticky software business. Some people would like the kind of financial model of that to be really tightly tuned so that we are keeping track of every dime that we spend and delivering a certain EBITDA margin and that sort of thing. Even that is not that attractive at the level that we are at. We think that we have a huge TAM. We are continuing to invest in the platform to develop all the capabilities that Ramy had on that slide to go after that TAM.
Even if we were just doing software, we would still be investing in that platform. We think that we have a model that prevents the dilution of a typical biotech model. We're not trying to do a secondary at this conference or something, which unfortunately a lot of people we know have to. We also are in a great position to deliver a growing base and provide a foundation to our stock. Look, it doesn't feel great for us. It doesn't feel great for so many people in the industry right now. We're in a really stable position and incredibly well capitalized. We consider ourselves fortunate.
We have run over. Thank you so much for joining us once again. I look forward to continuing the conversation.
Thank you.
Thanks.