Schrödinger, Inc. (SDGR) Earnings Call Transcript & Summary

Welcome, everybody. My name is Chris Shibutani, member of the Goldman Sachs research team. We are extremely pleased to have Schrödinger with us here. This is an amazing cast of characters. I've known many of you for many years, looking forward to this conversation. I always think of these as unique opportunities to kind of just engage in a discussion, and the people really matter. And especially right now with the kind of multi-variegate business model that you have. Ramy, you've really put together quite a team. And Karen, you and I have known each other for a number of years. By the way, there's a fantastic Longform Podcast that talks about her journey as well.

But I want to start, as I always do, with having people just sort of provide us something of -- your personal, professional journey in particular, just so we understand, as the words come forth, where they're coming from and the lens that they're coming through. So, Karen, you first.

Yes. Thank you. Very happy to be here. Karen Akinsanya, I'm the President of R&D for Therapeutics at Schrödinger. I've been at the company 5 years now. Prior to that, 25 years in big pharma, most of that at Merck in clinical pharmacology and running therapeutic area teams as well as a short stint in business development.

Perfect. So, Mr. Porges.

Thank you, Chris. Great to be here. Geoff Porges, CFO at Schrödinger, for almost a year. Previously at SVB, or the bank formally known as SVB, at least, for about 7 years running therapeutics research, and before that, at AllianceBernstein and originally at Merck also back in the distant past.

Ramy?

My name is Ramy Farid. I'm the CEO of Schrödinger, I've been at Schrödinger for 21, 22 years, something like that. I don't know, I'm starting to look to 21 years. Before joining Schrödinger, I was on the faculty, the chemistry department at Rutgers. So I came from academia.

And we're headquartered in New York, and it really is kind of the melting pot. So investors should really try and bother Jaren, who's sitting in the front row here. There are other Investor Relations persons to actually sit down and talk to folks. I think that this particular trio has got a wealth of insight and knowledge that is really valuable because one of the things that's always really striking about you guys and others sort of broadly in your space, artificial intelligence, R&D, right? Let's create that as just sort of like a Venn diagram-type circle is that there's different dimensions to the business, different business models, different approaches to that. It's very important. I think that leadership has some degree of ability to sort of see across some of those while being also focused at the mission at hand. And so I think one of the things that's been interesting, particularly with the recent stock performance, that you are relevant and tangentially related to an [ EOD ] glow of the current investor activation around investing in the world where all could change as shaped by information management artificial intelligence, et cetera. And it's kind of interesting to see these things happen. Ramy, you've been around forever with this, right? And at the same time, you're sort of -- we've had these smart tools to do things better, engineer things, et cetera. And it's really captured the imagination of folks. And you also did a recent financing. I think there may have been a little bit of marketing around that. So how are you positioning yourself? And how will you position yourself to this audience of health care investors and people who are listening when I bring up the word AI. It's actually 2 letters, but you know what I mean.

Yes. So AI has been around for a very long time. And what it actually is, is machine learning, right? What -- it's machine learning. And what that means is that it's all about identifying a training set, a bunch of information and then trying to learn on that. Now that has some really exciting applications in quite a large number of fields, including in drug discovery, but it has a really severe limitation, which is what I just said. It requires a training set. No matter what you call it, you can call it generative AI, you can call it AI, you can call it whatever you want, deep learning. It requires a training set. And here's the thing. There are approximately 10 to the 60, 1 with 60 0s, ways of combining organic elements into a drug-like molecule. So just call that infinite. It's another way to describe 10 to 60. It's infinite. The size of chemical space is infinite. And what that means is that it's impossible to generate a training set that will capture an understanding of all of chemical space. It can't be done. This would be equivalent, by the way -- the analogy is the amount of experimental data we have now for molecules that in the whole industry is equivalent to one drop of water in the ocean. So this would be equivalent to saying, I now understand everything about the ocean by analyzing one drop of water. It can't -- that obviously is silly. So while there's a lot of excitement around AI and a lot of hype, and it's going to have a really big impact, there's this limitation that's really, really important to not lose sight of. So what are we going to do? As an industry, we cannot use machine learning alone to design molecules. We can't use machine learning alone to design molecules because it's too diverse, and we can't build a training set. So what we've been focused on for the last 30 years, actually in most of my professional career, right, the time I've been at Schrödinger, is how are we going to overcome this severe limitation. It's not going to be by saying, well, let's make a whole bunch of molecules. Because that would be equivalent of saying, okay, now we have 2 drops of water in the ocean, now we understand the ocean. That's silly. Or 3 drops or 4 drops, it's not going to work. There has to be another completely totally different way of doing it. And the way of doing that is, say, forget about machine learning, let's figure out a way to predict the properties of molecules using first principles, using physics, not machine learning, but actual physics. Let's simulate what happens when a molecule binds to a protein or when a molecule dissolves in water, and so on. And if we can simulate that at an atomistic level of detail using physics, using first principles, now we have something really powerful, which is that we can explore now or predict the property of any one of those 10 of the 60 molecules. That's really, really powerful. And that's not machine learning. But here's the thing. Those methods are computationally really expensive. They take a lot of compute, there's a lot of compute time. So now we have a dilemma. We've got a method that's really accurate that can predict the property of any molecule, any arbitrary molecule. That's really powerful, but it's slow. And then we have this other method that's super-fast, but has the severe limitation. You need a training set. So what we've done, and this is how -- this is really exciting, we've combined these 2. We've said, okay, now we have a way of generating massive training sets using the physics and we can use machine learning to try and understand something about this collection of molecules. And we're talking about on the scale of hundreds of billions of molecules that we can now analyze and decide which ones to make, for example, in a drug discovery project. So that's a really long -- I hope that was clear. I mean I think it's really, really important for people to understand this because there's a lot of hype and there's a lot of misinformation and there's a lot of exaggeration around what -- any one of these methods can do. And I think it's really important to be kind of precise about it and understand what we've done. What really differentiates us from all the other companies in this space that are just talking about AI is that we actually have a method that can really drive really -- in a real world application, can drive drug discovery projects. And I think we've demonstrated that over and over and over again. I think that's something that's really important to pay attention to. Now does this actually work? Well, we have all these collaborations we've been doing for a long time, and it actually works. We keep putting molecules in the clinic. There's even some in the market. We keep advancing our own program. And it's by understanding the underlying physics of these complex molecular interactions, which happens to be combining physics and machine learning. So is that helpful? I mean...

I think it is because people...

AI is changing the world [ methods ] and...

Exactly. So I mean I sort of like held before you this large platter. And then sometimes people put some scary things, like physics space, and they're like, oh gosh, it's not my favorite topic when I was an undergrad, et cetera. So I mean concretely speaking, it's what's defining...

We don't have a choice.

Yes. But I think it's concisely descriptive of the capabilities. And the very fact that you've had this capability for decades, and it really takes that long to almost sort of formulate it into a business. I always think things like science, the business and then the investment. And so we're on is perpetual journey, which there's various laps and degrees of maturation and reshaping that's ongoing here. But you do have actually a very genuine business. And that's now actually starting to show that these are sort of tulip bulbs and seeds, some germinate over years, over months, et cetera. And some of these are starting to bloom. It's hard because we're sort of looking at the dirt and know that there's something underneath there. So that's one of the things that investors hate. Their ability is limited on these things. But some of the really compelling proof points are showing some beautiful bouquets of what could happen as manifestations of your capabilities. So -- but I think investors have a hard time, particularly since you went public, understanding sort of the rhythm of the business. And so at the core of it, there's a software business model, right? There's been a very professional packaging and the availability that is user-friendly so that you can interface with your customers who are also trying to develop drugs in a smarter, faster, better, more precise way. To have the end in mind. And thinking about that, Geoff, in particular, you're coming in here with the breadth of your background, you probably watched the stock before you made your decision to come. And there were often these moments of terror called earnings. And it was where there was a reckoning of recognition of the malalignment between expectations and measuring businesses like this every 90 days and stuff. Ramy and I have had many conversations over the years about how can we do this. Should we do rolling 3 quarters or whatever? Geoff, talk to us about the business and how you studied it to the point that you said, "Yes, I'd like to actually be one of the managers of this and a voice for this business to the investment community because that's been one of the [ middle ] somethings that people have struggled with.

Yes. Look, it's challenging because we are simultaneously an operating business and asset business. We're generating very attractive, fairly rapidly growing revenue from a global array of customers. We've said 1,750 customers over $1,000 a year, and we're seeing customers buying $5 million-plus a year in software, the largest customers in the world. And we've guided that, that's going to be a sort of mid-teens growth opportunity for us this year. And we think that, that's a very sustainable kind of long-term growth opportunity for the business. And generally, we've indicated that, that business is sort of operationally breakeven, although we carry out that we are absolutely committed to investing in the core technology, the platform that Ramy described, both the physics-based methods and the machine learning component to it. We're not sort of idle or complacent and waiting for competitors to catch up. In fact, I'd say we're gapping away from our competitors. So that does take an investment. So we have a very nice software business that my conclusion was very attractive, very durable, sticky and was essentially contributing cash that could be invested into the platform and the proprietary portfolio. But of course, the ultimate test of a drug discovery technology is whether it works, right, whether it comes up with drugs. And as you can imagine, I did quite a due diligence on that, and I think that runs pretty well on the board. You just have to look at Nimbus. And frankly, the Nimbus transaction occurred after I joined, and that was sort of a nice sealing of the deal, but...

He gave me all the credit, [ actually ].

Yes. Like Morphic and Structure and the BMS collaboration. We announced recently the SOS1 that we discovered in the BMS collaboration had transitioned into development in their portfolio. And there have been a number of other -- those sort of collaborations that have been successful. So compared to most other platforms, I think that's a very large number of runs on the board in terms of validation. The typical platforms, 1 or 2 programs that work out. And this, I think, is sort of every year, year after year after year. So with our other platform, software business growing nicely and then a commitment to deploying capital internally to grow the opportunities that Karen's team identified for advancing proprietary medicines. And ultimately, I think we think that's the way to create the most value from a breakthrough technology. And we're seeing those investments. We're very excited about the opportunities for them, and we're sort of seeing a fair amount of validation, those targets, so we can talk about those. So I think that the mix of businesses was very compelling in the current market. We're very well capitalized. We're in great shape to both support the platform and the proprietary medicines. So the business is in a really good position with multiple paths to creating value. So that was what I saw from the outside is what I see from the inside.

And I think one of the things that perhaps health care or therapeutics-focused investors tend to struggle with is not that it's a crutch, but it's a tool and we all know how to extrapolate prescription data from IQVIA, et cetera. And it's like who are the customers? And we love doing epidemiologic bottoms-up model. It's 5% penetration to third-line treatment, et cetera. Twelve months of duration of use, and you come up with some number, and then we torture them back and forth. Harder to do that, understanding who your customers are, because your customers are literally over 1,700, the major pharmas, all the way down. There's also -- Ramy, remember, I asked you to put me on a call with someone on the industrial side of the business, which is relatively small. So not necessarily likely relatively to move the needle for the stock, but it's fascinating to see. It's like we're creating like smarter doors on airplane engines and all sorts of things like that. Just the capabilities are vast and can be extrapolated. You found your center of gravity in drug discovery and development. But getting back to this issue, can we come up with a couple of metrics or vernaculars, Geoff, that helps us get a sense for where things are shifting in terms of that customer base because there's kind of -- there's like the big spenders, there's the fat middle, there's the small innovative desperately trying to finance themselves, wannabes who are doing the free trials kind of thing. There's the recurring revenue stream. What's the vocabulary that we should start to have when we're talking to Geoff Porges, CFO of Schrödinger, about the software business and measuring that?

Yes. I think we're definitely focused on the metrics around the largest customers. And I think Ramy mentioned this many times, our use of our own technology to come up with our own medicines is an order of magnitude higher than our largest customer. So if you tease that apart a little bit, that means that a global pharma company with a $10 billion, $15 billion a year R&D budget is spending an order of magnitude less on the technology than we are in our internal sort of biotech drug discovery effort. That gives us an indication of the upside that we think there is with those large customers. And then what we've said is we have 4 customers spending $5 million a year or more, and that's up from 2 in the prior year. And the number of customers spending over $1 million is 18 compared to 15 in the prior year. Those are really important metrics to us because they are indicators of that headroom in terms of the incremental opportunity in those largest customers. And we, honestly, feel like we're just scratching the surface. Even in the large customers -- very largest customers. But then the difference between our largest customers and the smallest of the large pharma companies is a matter of order of magnitude. So household names, pharma companies that still haven't adopted computation. To a certain extent, we're going to look back at them and say they're in the dark ages. That's my personal view. But they will say, wow, how could we have thousands of medicinal chemists coming up with molecules like a cottage industry? It's like they spin out in a molt in 18th Britain, spinning done. That's going to change. And there are huge global companies that are still dominated by that mindset. So there's a lot of opportunity. So those are the metrics we look at.

Preach. Absolutely, Dr. Porges. Let's get concrete with numbers though. So at the end of the first quarter, we talked about second quarter commentary. You guided to flat to slightly lower growth in this current quarter, in part due to challenging biotechnology funding environment, and that's that segment of the customer base there that's very sort of mindful of their budgets and their capacity and the utilization as a consequence. Now that we're towards the end of the quarter, is there any sort of incremental detail you can point us toward with each progression point as we're going?

I think we're just really comfortable with our guidance for the quarter and the year. We don't have any inclination to sort of change it or add to it or modify it. What we said at the quarterly call was that we are having very productive discussions with those largest customers. We're having very proactive discussions with those largest customers. And the thesis about the headroom that exists in those accounts is still very true. And as we said at the time, the timing and the cadence and the kind of optimism about those discussions is it's earlier, and there's more urgency and there's more opportunity than there's been in prior years.

And in the back half of the year, you noted that there's some multiyear agreements with those larger customers that are coming up for renewal. How should we be thinking about that? Is that something that we should be concerned about?

Yes. Those are multiyear agreements that contributed significant step-ups in revenue in 2021, to a lesser extent, 2020, but a certain amount of that, so there's 2- and 3-year agreements. And in the nature of those renewals, we would expect there to be a big chunk of revenue recognized in the period in which those contracts are renewed. It's very hard to imagine that they would not be renewed, that a company that had committed 2 or 3 years to our technology and implemented it, in some cases, dozens or hundreds of chemists who have committed to hold one or more of their research sites, that they would not renew. And certainly, all the indications we're getting out that they will renew, and that's where some of the opportunities for expansion.

And then if we look at the reporting, drug discovery business is kind of the collaborations and partnerships that you have, it's a different segment. It's one in which you've provided explicit revenue guidance going back over the past 12 months or so. At one point, it was -- could exceed $100 million. It was then modified to $70 million, $90 million. Visibility around these things is constrained by the fact that you're not in control of the cadence of what's happening there. But talk to your confidence about the $70 million to $90 million that we've had updated. Any notable updates there?

Yes. We're pretty confident about the $70 million to $90 million in the same way that we have been throughout the year. I mean we -- as you've pointed out, we are challenged by a program that we discovered. We partnered with a company, maybe that's then being downstream. Partnered with another company, and we're knocking on the door of big pharma companies saying, "Hey, when do you think you're going to go into Phase II or get POC data and trigger a milestone?" And that's not always easy for a relatively small company. So that's kind of the basis for that variability. And there's still reported $25 million in milestone revenue in Q1. That's an indication of the opportunity that exists in our collaboration. That was with BMS. But again, we have to deliver those programs at that time point.

Turning to you, Karen, because I think when we think about Schrödinger as an enterprise and the companies in this space, there are often these many parts. It's not a chimera, or there's like 3 parts. There's like the service aspect of it. And then it's like, hey, let's [ feed ] our own cooking, right? Let's come up with our own proprietary pipeline. Let's put someone who's been there and done that. It's the leading pharma companies and your training, again from that Longform, just included so many fantastic mentors. Talk about your pipeline. We are at that actually really important for this audience, especially inflection point, where we're clinical, in the clinic, and that is really kind of a marginal line for kind of people to care, right? But maybe start with strategy because, again, there's different philosophies to have about going this. It's like, we have such a great tool. We can go after really difficult stuff or we can have derisk biology, and this is like a better one. What's the philosophical guidelines around how you're shaping your proprietary pipeline?

Yes, certainly. So when we describe our proprietary pipeline, it's really something that started about 5 years ago. And that has included some programs that we've already partnered. These would be precision oncology targets, some what we call modality switches. These are drugs we targets that had biologics for which we're now generating a small molecule, which provides potential for broader application, new indications, new populations. And so that proprietary pipeline and the initial phases, we actually partnered some of those programs with BMS as we've described or as Geoff has described. And we also began working on programs where we saw a really interesting opportunity to leverage initial validation of those targets that came from the clinic, but where the molecules potentially had liabilities that we viewed our platform as being extremely powerful to address. And that includes things like selectivity, time and target, potency, just overall pharmaceutical properties. And so we basically have identified a number of targets, some of which are disclosed, as you mentioned, are entering the clinic now, and others that we haven't disclosed where we see a very nice opportunity because of the human evidence and the validation around the target, but also the opportunity to design a fantastic molecule. And so over the last 3 years, those molecules have started to emerge from the platform. We declared an IND last year for our MALT1 program. That program is now in the clinic. We have 2 trials ongoing, a healthy volunteer trial and also a relapsed/refractory B-cell malignancy trial. This year, we are filing an IND on our second program, CDC7 for relapsed/refractory AML. And next year, we expect to file an IND on our WEE1 compound. Each one of those mechanisms, in our view, has great human evidence with validation. They just don't necessarily have great molecules. And we believe we now have best-in-class opportunities in all 3 cases.

Yes. No, I think the targets that we're going after here, MALT1, CDC7, WEE1, et cetera, are gradually becoming part of the macro, folks who like to pay attention to highly innovative, tough next-generation type of opportunities. And from that standpoint, you can actually get a little bit of a lift or this [ spike ] come in when some of the larger companies, I look to J&J, are also acknowledging, yes, this is a worthy target, and let's go after that. And so I always believe that, philosophically, that success amongst these difficult targets at these early points amongst anybody is a source of validation and it kind of like raises the floor here. But then there ultimately does come a point where you seek to have kind of the best or differentiation of it, so that you can think about how to position this and that you have some compass to guide how your clinical development strategy is going to map out here. So given all the intelligence and the capabilities and the precision of the design that you have, and typically, we're not -- you're designing something with some features in mind, you're idealizing to minimize this potentiality, maximize this efficacy profile, et cetera. So what is the hypothesis on the MALT1 profile that may ultimately enable you to join the crew with hopeful further validation, but ultimately come out in a real leadership position?

Yes, absolutely. So. MALT1, for those who aren't familiar, is a target that sits between the B-cell receptor, BTK inhibitors and NF-kappaB, which is essentially an important signaling mechanism for B cells, including other immune cells. The importance here is that you must cover this target 24 hours around the clock. And that requires, therefore, a profile that is rather potent, i.e., you can inhibit this target, as I said, around the clock. But you also need a compound with great pharmaceutical properties, great half-life, great selectivity. All the tools of our platform is very good at designing into molecules. We've recently become aware that the initiator compounds in this space that went into the clinic have some challenges in maintaining that profile with regard to, for example, the half-life of the molecules. We understand that there's a significant accumulation that's happening in the clinic of the leading molecule. And that scenario where our molecules don't appear to have that profile, we're very confident. And Ramy said it. Many, many compounds have been in the clinic now, really good at designing in these properties. Single-dose per day, selectivity, potency that allow us to cover this target and maximize the therapeutic index. There's initial data coming out now in MALT1. There is some potential for AEs at very high exposures, some of which may be compound related, so some of which we believe is related to the profile of the innovative molecules. And so that gives us a wonderful opportunity with a very selective, very potent, well-behaved molecule to really take a leading position, we believe, in the application of MALT1 to B-cell malignancies.

And to be more specific, some of the adverse event profiles that were being mindful of include, I believe, hyperbilirubinemia in a Phase I/Ib study, in particular, because the exposure duration is not going to be as robust in terms of informing that kind of exposure risk. Nonetheless, with Phase Ib, we're all junkies for catalysts and cards turning over, et cetera, what can we expect to learn from the 1b, in particular? And when might you be able to share that?

Yes, absolutely. So you're correct to say that the leading market in the clinic did show about 15% hyperbilirubinemia greater than Grade 3. We think that, that's related to UGT1A1 activity of the compound. That's a space where we think we have significant margin for that type of [ compound ]. We've also got a cryo-EM structure of UGT1A1. It's one of the ways in which we dial out off-targets for our programs. Now in terms of delta, we , as I said, are in the Phase I study. We expect to go from that dose escalation, safety, tolerability, PK/PD study into a cohort expansion study where we're actually looking at identify a recommended Phase II dose and exploring combinations. What we know about the data that's been released so far on the [ advanced ] molecule is that you see monotherapy activity, which is great. But in combination with BTK, you're seeing really quite nice, around 60% ORR effects. And so we believe that covering the target more profoundly, let's say, with our molecules could potentially even lead to greater activity. And so that Phase Ib study will be all about demonstrating monotherapy as well as combination activity with a BTK inhibitor. And our focus will, of course, be on the third generation now emerging standard of care BTK inhibitors.

And the real work that happens actually involves getting these trials done. Overall, oncology trials have not been a simple task. Pandemic has been not anybody's friend. Just finding folks who need to make the doughnuts and help with enrollment of patients and gathering data, et cetera, has been challenging. What's your sort of perspective with these early oncology clinical trials, in particular?

Yes. It's a really great question. As you said, the pandemic had an impact on the ability to enroll trials. But in addition, we know B-cell malignancies have seen a number of new products come out. So we've got CAR Ts, bispecifics, a lot of competition for those patients. However, we are seeing now, I think, a revival post pandemic, where it's actually possible to get sites activated. And beyond that, instead of focusing just on the U.S. we're going global. We're actually looking at sites around the world where perhaps there isn't quite as much competition, CAR Ts are not quite fully available. You have patients that are more accessible for these new small molecule approaches. And so we expect to be global quite thin with our trial.

How far along do you want to carry this and do it on your own? I mean you're certainly friends and have relationships with everybody who's out there. So the [ Rolobox ] is just like half a breath away. How far do you want to go?

Yes. At the very first instance, our goal is to generate a very robust data package for our program. As I said, safety tolerability, PK/PD and some efficacy. As I mentioned, we believe that MALT1 is going to be most powerful in combination with a BTK inhibitor. Now we don't have one of our own. So we expect to partner with a company that has what I would call best-in-class BTK inhibitor coverage, market product. And so we expect to partner, at least initially, to study the combination but also to maximize the potential of this mechanism, and including V1 and CDC7, which all have combination opportunity with marketed products, standard of care products, we see ourselves partnering. Now we do have options as a company in the face of a very significant package that really positions us well, for example, to go to accelerated approval. That discussion that we're going to have to have. That will be an interesting day for the company, I think.

Yes. Absolutely. And is Geoff giving you enough money to be able to like grow your pantry? Or how are you fueling for those things? You can say it now. Just [ announcements ], whisper it in my ear.

Well, I think that as what have already been described, the diverse value creation opportunities, the collaborations, the software business, it does allow us to now invest in our own pipeline. We want to be careful though. We want to pick the right programs, have the right balance in our portfolio. We see a lot of value in collaborations. And so I think we'll continue to push some of our programs forward and collaborate in other cases. I don't know if you want to add to that.

Perfect. No.

Yes. No, that totally makes sense. And like versus comping you or contrasting you with just a molecule and a half dozen people in an overly priced shared workspace in Campbell, Florida, you guys actually have a relevant business with a very sizable bank account, et cetera. And so there's this self-generative, there's that word again. But nonetheless, ability to create this composite. And I think it's going to be fascinating to watch particularly as the MALT1 continues to make progress clinically, it's going to have this rotational pool over the therapeutic specialists, many of whom that we interact with, especially at this conference or whatnot. So it's going to be an interesting shift. I think the story has the potential to appeal very broadly across folks. And that's actually a bit of a risk mitigation hedge, right? Because that there are times when people just like, "Oh, my God, I don't want to take any science risk." Well, we can talk about well-established signature software business and the recurring revenues of that. So a fascinating juncture for the company, a lot of maturing transition points here. So I really appreciate all of you coming out here telling story, [indiscernible] with my way of having this discussion. But you [ were a trio ], especially, I think. It's just very valuable to be able to get these combined insights, especially because it's not a singular story or a singular business. It is the engagement and interaction of this composite. So thank you very much for joining us, Ramy, Geoff and Karen.

Thank you, Chris.

Thank you.

Appreciate it.

Thanks.