My name is Mike Riskin. I'm on the Life Science Tools and Diagnostics team here at Bank of America, and we're excited to be joined by the Schrödinger team. Thanks so much for coming. I think we're going to kick things off with a quick presentation, and then we'll go into a fireside chat Q&A.
Okay. Thanks, Mike. I'm Geoff Porges, the CFO of Schrödinger, and I will give you a quick overview of the company, and then we'll get into a discussion about recent news and Q&A. Of course, I'll skip over the precautionary statement, but please refer to our SEC filings. Schrödinger is a company that combines the best of physics-based methods and fundamental scientific analysis and tools with artificial intelligence and machine learning to create technology that enables much faster discovery of novel chemical material for both life sciences and material sciences applications. This combination enables us to be very accurate, which comes from physics-based methods, and then also deploy the technology at huge scale and consider a vast number of potential development candidates by using AI. We harvest value from this platform in three ways. First, we license our software. We have about 1,800 customers globally.
That's customers spending over $1,000 a year. They're principally in the life sciences industry. Almost every academic institution in the world that studies chemistry is a customer. They pay a very nominal fee for the access to the software, but in return for that, they're training the future generations of industry chemists to use our technology. We have material sciences customers across many different verticals. We also engage in collaborations, principally drug design and discovery collaborations, but also materials collaborations, where we work closely with software customers on high-value projects, and we retain downstream participation in those projects. We have 10 drug discovery collaborators so far, and we have a number of active collaborations that Karen and Margaret can talk about. Lastly, in recent years, we've developed a proprietary pipeline. We now have more than eight active programs, including three in the clinic.
Those are wholly owned programs where we've deployed our technology to come up with innovative development candidates. The platform underlies all of these different sort of verticals of business. This is a summary of the financial results. You can see the software revenue has been the principal driver of our top line over the last four years. We reported $180 million of software revenue last year. That was growing at 13%. Drug discovery revenue was $27 million. That dropped from $57 million, where we had a large milestone that was paid to us by one of our collaboration partners in 2023. I'll talk a little bit about our guidance, but suffice to say, we think that both software and drug discovery revenue will grow significantly this year. Operating expenses have ramped up as we've invested in both the drug discovery programs, but also in the platform.
As I will highlight in a minute, we think we're more or less at scale in terms of our OpEx. Our cash burn has gone up and down. One of the great virtues of our business model is that we receive cash from collaborations and also cash from our equity investments in companies that we have co-founded that have been built on the basis of our technology. Those co-founded companies have gone public and been acquired, and that's been a significant source of financing. Recent highlights, we just reported Q1. We reported close to $60 million in revenue, up 63% year over year, $48.8 million in software revenue, 46% growth. That was principally driven by large customers, $10.7 million in drug discovery revenue, where we're starting to realize some of the value from our most recent collaboration.
We had a cash balance of $512 million at the end of Q1. Our guidance, we're forecasting total software revenue growth of 10%-15% this year, $45 million-$50 million in drug discovery revenue. Really a recovery there from where we were last year. Less than 5% operating expense growth and software revenue in Q2 of $38 million-$42 million, which is significant growth over last year. We have phase I data readouts from our three most advanced proprietary programs, 1505, which we'll talk about in a minute. We have posters being presented at EHA in June and then ICML a week later, also in June. We're expecting data from our next two clinical programs in the second half of the year. I'll skip over this abstract and put it back up in a minute.
Just to highlight our strategic priorities, we're continuing to drive increased adoption of our computational technology. That's the key to our revenue growth. It's not getting more customers. It's increasing the depth of utilization of our existing customers. We have a number of enhancements to the platform that will enable its use for biologics, for predicting toxicology risk, partly addressing some of the FDA's concerns about animal safety that have come out recently, and then also looking at other attributes of molecules for further development. I've highlighted the presentations of the clinical data, and of course, we're continuing to advance other programs in our proprietary portfolio and our collaboration. Maybe I'll just jump back to this slide, and then Karen and Margaret can talk about this.
Yes, go ahead.
Thank you. The abstract at EHA has been accepted, and it is a phase I dose escalation study in heavily pretreated patients with relapsed refractory B-cell malignancies. At the time of the data cutoff, SGR-1505, the MALT1 inhibitor, is safe and very well tolerated. There have been no dose-limiting toxicities observed, treatment-related serious adverse events, or AE-related deaths. Asymptomatic indirect bilirubin laboratory abnormalities were common. These were predominantly grade one and two, and there were no symptoms. Dose-dependent increases in exposure have been observed. Encouraging preliminary efficacy data has been seen in 23 evaluable patients, one PR, two PRLs, two MRs, 12 stable disease, and six PDs. In terms of the MR, that is an objective response rate for Waldenström's macroglobulinemia only, and the PRLs are an accepted response rate in CLLs.
PRs in the CLLs, we noted that these patients had been previously exposed to a BTK inhibitor and a BCL2 inhibitor, and there were two out of 10 in CLL and in Waldenström's, one out of three PRs. The poster presentations are in June at EHA and ICML. Thank you.
Great. That's a great introduction. We'll kick it off there. I mean, maybe just starting where you left off on SGR-1505, could you just walk us through next steps? I know this is just sort of the initial abstract, and you're going to have more data presented in June, but just lay out the roadmap from here for us.
As Margaret covered, we are pleased to have got to this point in the dose escalation. It is an ongoing study, so we will be finishing out that study and collecting both PK, PD, obviously safety. We do not mention PD on this slide, but we in the abstract do reveal that we have already hit 90% inhibition of IL-2, our biomarker. All of that data will be packaged up as part of our interactions with the agency around what we believe the recommended phase II dose will be. Through the rest of the year, that is really one of the goals, to understand that recommended phase II dose and have that interaction with the agency.
Okay. And then your other two lead compounds for CDC7 and WEE1, MYT1, you expect similar data readouts later this year in the second half?
Yeah, this study has been going on for a little bit. Starting with CDC7, that is an ongoing trial. Margaret can update you on that briefly here. Yes, the goal is to have ongoing phase I data update on the profile of the molecule in the second half of this year. WEE1/MYT1 was the most recently begun with an IND in the second half of last year. Not a huge package, but I would say some update on the profile in the second half of the year.
Okay.
Right. So just to state, as Karen has said, these two other studies are early in the dose escalation phase, so it won't be as robust data as you're seeing with the MALT1, but we expect to present the safety PK, preliminary efficacy data, and some PD data as well on both programs.
Okay. If we were thinking further out to phase II , would this be maybe one phase II initiated in 2026 and the others in 2027, or is it too early to say?
I think it's too early to say. The good news is, I think that, and you'll see some of this in June, we are now enrolling at higher doses, and we are looking at the more aggressive patients. We actually couldn't start dosing those until most recently. This is sort of a phase I dose escalation, but there's a sort of phase II feel to it. The question we have to figure out is, where do we go next in terms of tumor type, dose, and what would a phase II study look like? This is a drug, for example, here that would be appropriate for combinations, we believe. Thinking about monotherapy versus combination, all of that has to be worked out. It's too early to say on timing, I would say, for any of these programs.
Okay. All right. While we're talking about the pipeline, I think one thing we always talk about is Schrödinger's demonstrating ability to replenish the pipeline internally with additional compounds. You've got a number of other compounds internally in development that are in earlier stages of development. How do you think about backfilling as these progress further?
Great question. We had last year actually partnered some of our programs that had not yet been disclosed with Novartis. I think you are familiar with the Novartis deal we did. We do have other oncology programs ongoing that we are still considering what our options are. You know, this is a big year for us with three clinical programs. We want to make sure that we follow through on what we need to deliver here. Thinking about the next steps on the remaining oncology programs, as well as, as you know, we have got a neuro and an immunology program, that is something we will be looking at through the end of the year in deciding what our options are. We cannot have too many things going in the clinic.
Okay. I mean, to your point on the partnership and the out-licensing in a way, that's a great way to monetize the early pipeline as well. Could you give an update on some of those collaborations?
Yes. So it's been actually a very busy year for collaboration. As you know, we closed the Novartis deal, I think it was around November time. Realistically, it's taken a few months now to get going. In the first quarter of this year, that collaboration really took off with multiple programs ongoing. You saw the reflection of that in the revenue numbers that Geoff shared. In addition, we expanded our relationship with Lilly with a new program and also with Otsuka. The collaboration is going very well. A lot of expansion in the first half of this year. Capacity-wise, I mean, I think we have room over time for more because as programs complete, obviously that opens up capacity and allows us to either expand existing collaborations or form new ones.
Across the whole length of the portfolio, from very early, mid-stage through to clinical, as you know, we're always active in partnership discussions and considering our options.
Mike, after the Lilly acquisition of Morphic, that became an active collaboration with Lilly as well.
That's right.
We were working closely with Morphic. We were really excited about those programs. Of course, it was nice to get the check associated with our equity position. Now those programs, both the lead and the earlier programs, are active collaborations between Schrödinger and a different part of Lilly. At a certain level, they all talk, but that directly broadened our Lilly collaboration significantly as well.
Geoff, maybe we all got you. Could ask you to walk us through some of the financials of those collaborations, both in terms of cash flow and revenue recognition, maybe using Novartis as an example?
Sure. Yeah, good question. The Novartis collaboration was complex. It had two parts. There was the software contract and the drug discovery collaboration. The upfront milestone payment of $150 million was all for the drug discovery collaboration. That revenue gets recognized over the expected period of our execution of the obligations under that collaboration. Now, there are multiple programs, and those programs each take a number of years. We think that revenue is going to be recognized over, let's just say, three or four years or something like that. It'll ramp up, plateau, and then taper down. There are significant milestones for each of the programs in that collaboration as we deliver development candidates to Novartis. Usually, those milestones are going to be large enough that we'll have to call them out when we report them in the quarter.
We have been saying we certainly would not expect one of those this year, as an example. That is on the drug discovery side. Now, on the software side, Novartis was a relatively small software customer. This is a great example of several things. First, the step-ups in scale that can occur in global pharma companies when they decide to go digital, in effect. Novartis said, "We have to modernize our drug discovery. We have to embrace computational approaches." We have all seen Novartis has been doing deals with a lot of different parties in the industry, but they said, "We need our scientists to all be using Schrödinger technology for drug discovery." That is including both our modeling technology and our Enterprise Informatics Platform. They scaled up to become one of our largest customers instantly compared to being, let us just say, not even in the top 10 previously.
They were new. What we've disclosed, obviously, is that they're in the top tier of our customers, which means north of $5 million a year. Now, that particular contract with Novartis, their original contract was running out at the end of 2024. When we renewed the contract, we recognized a big chunk of that revenue in Q4 for the full three years, but some of it is spilling over into this year and will be recognized ratably as well. There are multiple components, but suffice to say, we were really happy with the software agreement in addition to the collaboration. I remember what I was also going to say. The collaboration and the software agreement were effectively inseparable, meaning they sort of thought of them as seamless.
They wanted to learn how to use the software, to pay for using it at scale, to have it across the organization, and to see how we use the software on high-value projects that really meant a lot to their research management. These two components were really inseparable. We were really happy. We think, as an aside, the NPV of the software contract, at least at the start, is significantly higher than the NPV of the drug discovery collaboration because the drug discovery collaboration, it's like preclinical, early stage. All of that risk adjustment dilutes the NPV of that component. The NPV of the software agreement, given the size of it and of what we expect over time, is significantly higher.
That's on the revenue side. What about on the cash side?
Yeah. We build Novartis annually on the software side, so that rolls into cash, but we do not get the three years, so there is a deferred revenue component. On the drug discovery side, the upfront payment, we invoiced them at the end of the year. It was a receivable at the end of the year, and then we received the cash in Q1. That was a big component of why our cash went from, I think, $368 million at the end of last year to $512 million at the end of Q1. Huge step up, effectively, I hate to say it, but we did a financing courtesy of Novartis in terms of the impact on a balance sheet.
You have done something like that in the past. You have done partnerships like this every year, every other year, you mentioned. Bristol Myers was a big one a few years ago. That is a useful way, an easy way to replenish the cash balance, like you said.
Yep. Exactly.
It allows you to keep funding internal projects as well without needing to go to public capital markets.
Yep. We can reinvest in the platform, and we're also reinvesting in our proprietary portfolio.
Okay. I want to talk about some other recent news in the space. The FDA's policy announcement on animal testing, phasing that out, looking for ways to phase that out and replace that with technology, alternatives, things like that. You've talked about that a number of times, and you flagged it on your earnings call as well. Could you talk us through sort of how you can fit into that ecosystem and where Schrödinger can stand to benefit from some of these approaches?
Yeah, I can start. Over the last few years, actually, even in our own programs and our collaborations, we've been able to identify the structures for a lot of the off-targets that commonly plague programs. The FDA has a list of about 100 off-targets that have been identified as essentially dangerous to humans. One example of this is hERG. It's an ion channel in the heart that can cause sudden death. Most programs that go through the drug discovery industry at some point actually screen for hERG activity. We got the structure of the hERG ion channel and have been using that in our own programs. What it's allowed us to do now is create an in silico version, essentially, that we're building up over time of that panel of about 120 targets that are common off-targets and causes of safety concern. What does that mean?
It means that over time, based on what the FDA is saying in their replace and reduce initiative, these types of in silico assays are going to be used much more systematically in the way that people now systematically do this in wet screens. They will start to adopt this, is our view, for de-risking the safety profile of their compounds. Not only will they do this as has been done traditionally late in the day in a drug discovery program, because this is in silico and will be available to the whole project team, we think this will happen a lot earlier on, and that will be an important piece of our platform going forward. It aligns beautifully with the announcements that we've been hearing from the FDA.
Our platform has many ways in which you can interact with the predictive tox announcement, looking at antibody structure across species, for example. There are existing parts of our platform that allow you to do that as well. Yeah, we started this initiative last year. It's just very timely that it aligns with some of the ambitions and goals of the FDA.
What's early adoption on that been like? Who are your customers, and what are they using it for?
We had funding from the Gates Foundation to advance this, and that's been in place for about a year and a half now. That's a $20 million grant that we're recognizing starting in the middle of last year till the middle of next year, probably. We're executing on that, which really means qualifying all of those off-target proteins for deployment of our core technology, as Karen described. We're planning to, we've been using that now in our internal proprietary programs to optimize the selection of development candidates for our proprietary programs. We're planning to, in fact, we're in discussions with customers now about them coming on board for beta testing in the second half of the year. We think that we should be able to see a significant commercial opportunity emerging next year and beyond that.
The way to think about it in terms of the use of our core technology, this is going from sort of just modeling any molecule for optimization against one target to optimization against that one target, but also avoiding dozens, maybe hundreds of other targets. It really is a large potential incremental opportunity for our core technology.
Okay. I kind of want to pivot back to software for a little bit and talk about just the underlying health of the market, your customer groups. When I think about that, the part of the business, it's broadly exposed to the drug discovery community. There's some academic exposure there, but a lot of it's going to be pharma biotech. Can you talk about spending decisions among that cohort of customers by basket, especially in light of some of the policy changes and announcements in the last couple of months?
Have there been any policy changes?
I almost said years, but yeah.
It has been noisy. As we said on our earnings call, we had a great first quarter. We think we're very confident about the outlook for the second quarter. We maintained our full-year guidance for 10-15% growth in our software. We think that the software has a very long kind of growth runway and potentially accelerated by things like predictive tox. Now, we frequently ask ourselves, okay, what's our exposure to all of these kind of macro policy headwinds? Tariffs, not much. Right now, China can impose tariffs or is imposing tariffs on software imports to China. Our China revenue is low single digits. Contribution. So not much exposure there, but that's a potential risk. Who knows where that's going to land? No exposure to kind of U.S. tariffs being imposed. We don't think that's going outwards, but that's an issue.
In terms of things like IRA, we actually think that if the landscape is evened out between biologics and small molecules, that is slightly positive for us because we have had some large customers who swung their portfolios so heavily to biologics that it has resulted in them not increasing the use of our software because they said, "We are going to biologics." We think that if that swings back, that is helpful. We should get a little bit of benefit from currency in the second half of the year, depending upon how that plays out. We have talked about predictive tox. There are some tailwinds that will help us. What we cannot anticipate is whether pharma really goes through with the idea of cutting back on drug discovery, cutting back on R&D.
We have seen no evidence of that in our customer discussions so far in the large companies. Small companies, where everyone in this room knows they're in kind of terrible distress, yeah, we're seeing all sorts of noise, and we're sort of maintaining level ground there, which we think is a win right now in the small and emerging companies just because there's so much uncertainty there. In the large companies, none of them are saying, "Because of tariffs, IRA, MFN, etc., etc., we're cutting back on the use of our software." I think it reflects how absolutely indispensable our software becomes once the drug discovery organization has adopted it and embraced it. I mean, our retention is 99.9% in the large customers over $500,000 a year.
To help us understand that, maybe can you talk through your top cohort of customers, your top five, top ten, sort of like how is Schrödinger embedded into their workflows? Where is it being used? How is it being used? I'm kind of getting at penetration rate, use case, and then the following question would be, how do you drive that higher?
Yeah. So our largest customers have mostly been on a long kind of journey with us that's 10 to 15 years now. Karen reminded me this morning of a couple of exceptions. For example, Novartis were not a major customer. And the reason that we've made such a big deal about the software contract last year is because they literally went up virtually tenfold in their scale of adoption. And that was because of a change in management at Novartis and a change in attitude. They pivoted away from traditional old-school craft-based chemistry to embracing computational and kind of operational solutions to come out with drugs. That pivot resulted in that scale-up. Typically, it's a multi-year. They go from $500,000 to $1 million or $2 million. Then they go to $5 million. They go up from there. That kind of progression is occurring.
As you go up those stages, the companies transition from kind of post-hoc analysis, "Okay, the traditional chemists came up with a molecule. Let's just check that it's okay with the computer before we take it into the clinic." That's what a company that's spending $500,000 or $1 million is doing. The company that is spending $5 million or $10 million is doing what we do, which is using these computational tools to design and select the molecules from the beginning and then having medicinal chemists, in effect, managing that process or coordinating it. It upends the traditional drug discovery model. That's what we usually see as being required to get to that high level of adoption. Karen, do you want to add to that?
No, I mean, I think that's well said that there's a spectrum of how people use the platform. I think there's also this sort of snowball effect as the more access they have, the more they see the impact. If you think about the way the internal team and also our collaborators working with Morphic and Lilly and others, they've seen how this works at full scale, where there is no throttle, if you like, on the chemist's ability to predict any number of properties in parallel. Once they start to see that and they start to see that impact, I think that the demand locally from the teams to get more and more access, even mid-year before renewals, increases. It's really that curve of impact, experience, and scale-up that takes, in some companies, a short time, but in other companies, longer.
Michael, we think of this like computational tools being used, CAD/CAM, for example, in industrial categories like aerospace or autos. Once companies have gone away from hand-drawn cars and physical models to doing it on a computer, they do not go back. The same thing is true in movies. Once they have moved away from trying to actually create all of the computerized graphics and do it all on a machine, they are not going to go back. We think the same thing is happening with the use of our technology.
When you think about the customers we have sort of the highest penetration with, is there a way to quantify what it could be, what it should be in terms of, you talked about Novartis greater than $5 million, right? We've always had this debate of sort of what full utilization would look like because it's still not far from being broadly utilized within the organization. Could you just give us sort of a bridge to get to that?
We calculated recently that if we were pricing and using the software internally, if Karen's STG Group was an outside customer, their use of it would be larger than our top 8-10 customers. Another way of saying that is our internal use is an order of magnitude higher than our largest customer. We think that's the model for embracing computational approaches to this enterprise of novel drug discovery. Interestingly, we have another data point, which is there are private biotech companies that are using, I'll say, millions, but of dollars' worth of software because they start from the approach of saying, "We're not going to hire a bunch of medicinal chemists. We're not going to have a full drug discovery organization.
We're going to do this using the Schrödinger computational approach and shortcut our way to having development candidates." They are using the technology at the same scale as we are. Those two data points.
Maybe just a couple of minutes left. Maybe I'll end with this. There's been a lot of hype and interest in AI over the last year or two. A lot of companies have cropped up that are arguing that they leverage AI or have developed their own AI models to improve drug discovery and sort of using AI in healthcare is the next big thing. Schrödinger has been at this for many, many, many years. How would you say the competitive landscape looks to you and what sets you apart and why is your approach different or better than others?
I mean, first of all, I think there's many different verticals within this AI sort of boom that's going on in drug discovery. There's the AI to discover, to validate, to categorize targets. There's AI to use existing data to come up with molecules. There's AI to kind of predict protein structures. There's very many verticals here. The one vertical where we sit is in this highly accurate physics-based methods combined with machine learning to give you scale. We don't see a lot of competition in that space, to be honest with you. We're happy about the boom in the other verticals because, frankly, more structures means more structure-based drug design means more use of our platform, more validated targets, more understanding of human biology. We think we'll channel people to doing more structure-based drug design as well. In all honesty, the boom, I think, is ongoing.
I think the biology piece is less validated. It will take time, but it's important that we're doing it because, frankly, biology is the bottleneck in drug discovery and development, picking the right targets that are matched to the right disease. We do not see a ton of competition for what we're doing. If anything, we see a sort of pipeline of activity that will only lead to more of what we're doing.
Okay. On that, I think we're going to have to end it. We're out of time. Thanks, everyone, for joining us. Thank you.