Good morning. Welcome, everybody, to day two of our Life Science MedTech Healthcare Conference. I'm Scott Schoenhaus. I'm the Healthcare Technology Analyst at KeyBanc. Pleased to have Geoff Porges, CFO of Schrödinger, which is a tech-enabled drug discovery company. One of the unique tech-enabled drug discoveries in our coverage universe is it's dual-pronged with software licensing revenue streams combined with a pipeline. So with that, I will pass it along to Geoff, and he'll run through some slides, and then we'll do some fireside chat. Thanks, Geoff, for coming and attending.
Great. Thanks, Scott. And, thank you for including us in the conference. As I said, as Scott mentioned, I'll run through a few slides, just set some context about Schrödinger, and then we'll revert to Q&A. So let me advance. I will make some forward-looking statements. Please refer to our SEC filings for a full understanding of all of the risks associated with those statements and with our future outlook. So this slide tries to encapsulate what Schrödinger is really all about. It's a company that's built on more than 30 years of rigorous scientific research into how to model the behavior of chemicals in silico. The business is built on the industry-leading software to do that modeling.
This graphic on the right-hand side of the slide represents the heart of what we are doing in silico, which is modeling the conformation of a protein, which are the broad ribbons that you can see, and then ideating on different chemical structures that will bind at a specific site in that protein. And then the little white lines, different R groups or extensions of that molecule, that could confer different properties. And what we are doing computationally is trying to predict, in this case, the binding affinity of the molecule into this pocket in the protein. And we're doing this for, in this case, thousands, if not millions, of hypothetical structures to optimize the structure for the binding affinity. Now, our technology is not just about binding affinity into pockets in proteins.
We now have the ability to computationally predict all sorts of other attributes, such as not just potency and affinity now, but solubility, permeability, liability for hitting other targets, such as hERG, which is associated with QT prolongation, or CYP3A4 associated with drug-drug interactions. We're now even modeling different crystal structures or polymorphs that a molecule will exist in, to identify whether there's issues with formulations or manufacturing properties. The starting point is ideating millions of molecules to optimize for what will fit into a pocket. Then we go from there to predict multiple other attributes of molecules, to optimize the drug-like properties. Now, we do this first by, as I mentioned, calculating the properties of the binding affinity between the small molecule and the protein. This is de novo experimentation, but in silico. So we don't require a training set.
We're not using AI for those physics-based calculations. These are just very, very complex calculations that are enabled in software that is accessible to, to our customers. But using these physics-based methods, we can assess large, large universes of, of potential molecular structures relatively quickly. But we do accelerate our process by using machine learning or AI to take what we've learned from, for example, thousands of molecules that we model in this way and apply that to a very large universe of, of, in some cases, billions of hypothetical structures to select the best structures from that very large universe, for further development.
So, we think of our platform now as a hybrid combination between the first principles physics-based methods that generate a training set that we then apply to machine learning algorithms, bring those together to very quickly explore very broad chemical space during our development drug discovery process. With the goal, as ever, to come up with the very best molecule that optimizes for all these attributes from a very large universe of hypothetical molecules. And we believe that using computation, we can do this faster, cheaper, but most importantly, with a much higher success rate in future development, because from the beginning, we can eliminate or at least reduce to a very low risk many of the common liabilities that are associated with failure in drug development efforts. Now, having built this platform, we create value from it in three different ways.
First, we have a software business where all these calculations and many other attributes are predicted in software. It's licensed by 1,785 customers worldwide. Effectively, every company that's involved in drug design for small molecules is a customer, and most academic institutions around the world are also customers. Anyone who's trained at this stage in either computational chemistry and medicinal chemistry will be familiar with the Schrödinger technology. The scale of the customers I'll talk about, but the largest global pharma companies are spending many millions of dollars a year on the software. And then small academic institutions may be spending $5,000 or $10,000 a year on the software. And we certainly have a different price point for academic institutions or industrial software customers. We then have collaborations.
What, what has happened is that a number of the particularly large companies, but also small companies, have come to us and said, we'd really like to work closely with Schrödinger to, engage on a particular project to optimize the development candidate for that project. And, we now have 17. We've, we've had 17 drug discovery collaborations over time, and a number of programs are in the clinic. And then lastly, in the most recent five years or so, we've started deploying this computational platform on our own account, to identify and advance a proprietary portfolio. We now have two programs in the clinic, a third going into the clinic later this year, and a number of other programs in, in mid to late preclinical development, that we hope to become IND candidates in 2025 and 2026.
So the profile of our software business, we generated $159 million in revenue last year, grew 17.5%. There are 54 customers with annual contract value over $500,000, 27 with ACV over $1 million, four with ACV over $5 million. The retention rate for the customers who are spending more than $500,000 was 98%. Interestingly, the retention rate for all customers spending over $100,000 was ticked down to 92% from 96% last year. And the reason for that is that there was some flux in the small biotech customer portion of our customer mix that was associated with small companies either running out of capital, being merged into other companies, ending their drug discovery efforts and just focusing on drug development.
So, that weighed on the, on the revenue growth, but also the retention, particularly at that lower end, whereas the performance attributes were very positive at the high end. You can see the average ACV for the top five customers on the right is $6.7 million. That's up significantly compared to the prior year. And we think that will continue to grow, over time. It's noteworthy that if we take our top global pharma company and compare how much of our software they are buying to the 10th customer, not this is not in industry terms, but in our terms, there's a 20x difference in the ACV, between the top and the 10th. So there's still a lot of opportunity for us to move up those lower tier pharma companies up towards that higher echelon. So I mentioned that we have collaborations.
These are the collaboration programs that are in the clinic. FDA approved drugs. Agios two IDH inhibitors were originally discovered using Schrödinger software. We have a program, obviously, some rights to the residual rights to the Nimbus program that is now in Takeda's portfolio in phase III. There's a program at Gilead, the ACC inhibitor that was originally also licensed from Gilead, that's in phase II, and their Morphic program in phase II, and then a number of companies with programs in phase I. Then there are additional programs, with these collaborators and some others, in discovery and preclinical development. So a pretty broad portfolio of collaborations. Then a proprietary portfolio. Our leading program is MALT1 inhibitor. That's in phase I. This is for B-cell malignancies. Our second program is a CDC7 inhibitor, also a hematologic malignancy target in AML and MDS. That phase I trial is also enrolling
We've disclosed that we have a combined Wee1/Myt1 inhibitor that is in preclinical development. In fact, we'll be filing the IND and expect to be in the clinic in the second half of this year. Then behind those three leading programs, we have a PRMT5/MTA, EGFR, NLRP3, LRRK2, and then a number of other disclosed programs. The early mix of our portfolio was skewed towards oncology. That was really because that we saw promising targets that we could deploy our technology against in oncology. Now we see the portfolio pivoting more towards immunology and neurology. Really, we're pretty agnostic about the therapeutic area. We just need to have a target that's well characterized, and is biologically and, ideally therapeutically validated. Then we can deploy our technology against that target. So I mentioned the MALT1.
What's notable about this program is that we started the phase I B-cell malignancy study almost a year and a half ago. B-cell malignancy is a pretty challenging indication nowadays with a lot of crowding from CAR-Ts, bispecifics, you know, second and third generation BTK inhibitors, BCL2 inhibitors, etc. And so as we were getting that study site up, we had an opportunity to conduct a healthy volunteer study, which we did last year and in December presented results from 73 subjects who were treated for with 10 days of 1505. And the drug has a very clean safety profile. No drug-related grade 3 or 4 adverse events observed. No clinically significant bilirubin elevations or any other changes in LFTs. So, we were pretty happy with that.
And we've now confirmed that the profile we're seeing in patients, safety and tolerability profile, is similar to the safety and tolerability profile we saw in the healthy volunteer study. So, quite, quite encouraging in terms of the profile that we're seeing so far. CDC7, I mentioned this, the study in AML. This is a DNA damage repair target. And in preclinical models, this has shown pretty robust antiproliferative activity, a number of tumor types, including AML, samples that are resistant to standard of care medicines. We think that AML could be, is a very interesting test case for CDC7. But it's a target that if it works in AML, could have utility in a number of other tumor types, particularly tumors where there's very rapid turnover and a high rate of mutation in that replication cycle. And then lastly, we have a Wee1/ Myt1.
Like, a number of other Wee1s that have been advanced into the clinic, the art here is in designing a molecule that has the activity that we've seen from earlier entrants in the class, but doesn't have as much or has a facilitates a schedule and/or a profile that makes for better tolerability. We do think that with pretty exceptional potency, we have a chance of creating a treatment schedule that will allow for rapid recovery in the bone marrow suppression that's characteristically seen with Wee1 inhibitors, and may also limit some of the GI side effects. That's what we're going to need to explore in that phase I study. Quickly on some numbers, we recently reported Q4 results. Total revenue was $74 million, up 30%. There was nice growth in software.
We had a record software quarter of $69 million. Our drug discovery revenue was down because there were significant milestones contributions in the fourth quarter of 2022 that didn't recur in Q4 2023. The software gross margin increased nicely to 87% from 83%. OpEx did grow. We reported total operating expenses of $87 million, most of the growth coming in R&D, as you would expect. Net income was a loss of $31 million compared to $27 million the year before. We did end the year with cash $469 million. That was boosted by the $147 million in distribution from Nimbus that we received in the first part of last year, with that distribution exceeding the cash burn during the year. And for the full year, we reported total revenue of $217 million. Your drug discoveries, I mentioned $159. Sorry. Software $159, drug discovery revenue $58 million.
The software gross margin, 81%. OpEx did increase significantly to $318 million. As I alluded to, most of that growth, it was due to R&D. This is the long-term trend in our business. We, we think that this continues to be a healthy growth opportunity for software. The drug discovery revenue is, is clearly prone to bounce around depending upon the outlook, the development, commitment from our partners. We, we all we can do is design really good molecules. We can't predict which targets, our, our collaborators are going to, remain committed to. And, and that's, that's a real issue that we are, struggling with, as you'll see in the guidance for this year. Now, of course, the revenue that we are generating and that nice growth is, is supplemented by the distributions, from our equity positions.
Those distributions over the last five years have been $180 million, and we do have disclosed equity positions in companies such as Morphic and Structure that add another $80-$85 million of value to our balance sheet. Our guidance for the year, we're forecasting software revenue growth of 6%-13% compared to the 17.5% last year. That lower growth rate is because we're having to catch up after that very large Lilly deal that contributed a substantial portion of revenue in the fourth quarter of last year. Drug discovery revenue, the guidance is $30-$35 million. Software gross margin, we expect to be similar. OpEx growth has come down. It was 28% last year. That's going to be more in the range of 8%-12% this year.
Our cash used in operating activities will be higher last year, principally because of the growth in our R&D investment. So what are we looking forward to in terms of milestones? I mentioned the IND for a Wee1 inhibitor, 3515, that will go in the first half of the year and we'll start a phase I in the second half. We expect to have the data from the lead clinical programs late 2024 or 2025. We do believe one of the next wave of programs can become an IND in 2025. And then we're expecting to see significant improvements in the quality of our software for drug discovery and materials design being deployed this year. And we do have an ongoing research project funded by Gates Ventures into novel, into adapting our software for the discovery of novel materials for batteries that enhance the performance of batteries.
We're looking forward to some publications associated with that over the next year. So I'll hand it back to you now, Scott, for time for Q&A.
Well, thank you, Geoff. That was a really comprehensive overview of the business. So your Schrödinger's business model really stands out as, you know, from its multi-pronged platform. As you mentioned, you have software, licensing streams, you have partnerships, and you have internal pipeline. I guess let's start with software. So you've guided to 6%-13% growth for this year, coming off a record fourth quarter, strong bookings. What's building into that guidance range? What's that outlook? And remind people how your software licensing revenues can or contracts can be structured. That's, you know, sort of building into that guidance range.
Sure. Look, it is, it's a fairly wide guidance range.
A big part of it is, we have to try and account for the nature of the contracts and the renewals that we have this year. So, throughout last year, we mentioned that we had multiple, multi-year, multi-million-dollar contracts that we were discussing with our customers. And clearly we closed a number of those, but particularly the Lilly deal in the fourth quarter. And that gave us a really nice bonus of revenue. But now, as you know, what we have to do is overcome that and match that, match that and exceed that this year in order for the revenue to grow. And we do have ongoing discussions with a number of global companies about substantial step-ups in their revenue. Now, the wrinkle is, if a customer signs a one-year deal, then we get one year of revenue.
If they sign a three-year deal, then in some cases, if we sign that quarter, that's what we call an on-prem license. When we sign that deal, we record two-thirds to even four-fifths of that revenue at the time that we sign the contract. So the challenge we've had in providing the guidance this year is what is the nature of the agreements, those renewals? If we have a multi-year renewal, it'll contribute a bonus of revenue that will offset the Lilly deal. If we don't have a multi-year renewal, then we'll only get that one year. So we're trying to, that broad range, is to encompass the uncertainty about multi-year deals. I will say, we are optimistic that we will see multi-year deal opportunities this year, but it's a little too early for us to be certain and include them in the guidance.
So that's a principal thing. The other issue is, you mentioned bookings. We expect our bookings growth this year to be substantially above our revenue growth, whereas last year bookings growth was below revenue growth. Over time, those two have to more or less be the same. So last year was a year of very strong revenue and lower bookings growth. This year is going to be a year of very strong bookings growth and somewhat lower revenue growth just because of the normalization associated with the multi-year contracts that we signed in the fourth quarter.
Makes sense. Are you seeing, you know, you clearly have seen the size of your contract, average size of your contract increase, and your largest, pharma companies contributing, you know, continuing to, be your largest customers? Is that what you're continuing to see? Maybe comment on large pharma, the midsize market, and maybe the biotech, and markets.
Yeah. Yeah, I would, if I think through our mix, probably about 40% of our revenue is global companies. And then over the remaining 30% is probably SMID cap, SMID cap biotech and industrial companies, because we do have significant contracts with, you know, companies that are engaged in the design of materials for all sorts of industrial applications. And then, the remaining 30% is academic and small companies. And the small company piece has been where the softness has been, whereas, and that's been more than offset because obviously revenues continue to grow by the strength we've seen in the large global companies. And I think, the share of our revenue for the global companies actually trended up in 2023.
As far as we can see, is likely to continue to trend up. The reason that is happening is the sort of old guard of leadership in global pharma R&D organizations is turning over and their philosophy is changing. With every company that announces a digital initiative or that they're, you know, fully investing in computation, that's usually associated with a conversation with us about how do we deploy our technology at scale? And so, that is sort of rolling through the global pharma industry. We expect that to continue. We think that Lilly is a bellwether for the industry, but they aren't the only company that is going to be signing multi-million-dollar, multi-year contracts with us. Gradually, we think the whole industry will migrate there.
That's helpful, and Geoff, just as a reminder for investors, the 6%-13% software revenue growth for this year, what kind of assumptions are baked in on the, on the emerging or small biotech side, for that guidance?
Yes. You know, there we have not assumed any recovery in funding or in new customer creation on the small biotech side. We think that, there has been a fair amount of flux in that segment of our market, particularly in 2023, but also in 2022, and we think that we shouldn't be forecasting that that will bounce back. It did contribute a significant piece to the revenue growth that the company reported in 2021 and 2020. So, you know, if we get back to a frothy capital markets environment, then some portion of that may come back.
But we're not counting on that with the guidance.
Great. Sorry, I was trying to unmute myself. Okay, let's switch over to your partnerships. Maybe talk about your strategy when it comes to signing a partnership. I think it's evolved over the years and, and even more recently. So maybe just talk about your partnership. I mean, you've already talked about your partnership programs, but maybe talk about the strategy on signing on partnerships.
Yeah, look, the strategy is pretty simple, which is to sort of, minimize what we have to invest in the partnership and then create layers of future value, with the future value being realized through milestones and then, then ultimately royalties.
In the collaborations, for example, Lilly and BMS, and some of the others that were on that slide, we have single-digit to low double-digit royalties, and we have hundreds of millions of dollars of milestones, depending upon the commercial outlook for the program and indications, etc., so that's the strategy is to do that. But I think what we are reconsidering is, can we achieve that goal without having to take the program all the way, for example, to IND. And so we're becoming, I think, more disciplined about the scope of our engagement in these collaborations. So the time that we're engaged with them is less, the costs and our investment in them is less.
And we are doing what we do really, really well, which is design novel molecules with unique properties, but we don't have to get engaged in, in any other expensive preclinical development activities. So that's shifting. And then I would say the interesting thing is that, that, we've, we've been affected by changes in prioritization in, in a number of our partner, collaboration companies. And we're starting to think, okay, we need to understand the value of this target. In addition to it having value to our partner, which is something that perhaps we didn't originally contemplate doing. If a partner said, we want to do a collaboration on target X, we'd say, okay, we'll do the work and we'll get downstream participation. But now we're realizing that that part, that collaboration partner's commitment to target X may be fickle.
So we might have to independently assess whether we want to be committing ourselves to that target. So it has changed a little bit. The other thing that I should say about partnerships and collaborations is that they are enabled by high quality protein structures. As more and more companies doing what we're doing, which is adapting AlphaFold and then using cryo-EM, crystallography, all sorts of other techniques to come up with high resolution structures, that is enabling more and more programs to be amenable to our technology and hence creating opportunities for collaborations.
That's a really good point to end on there for the partnerships. On your pipeline, clearly you have a lot of, you know, IND phase I readouts coming up, which is exciting. I guess my first question is, what excites you? What internal molecule or asset excites you guys the most? And then is there a shift in strategy here, like other players in this field, to really bring the assets further along the clinical stage, to better realize better monetization of those assets?
Yeah. Look, we love all of our children equally. I think, ironically, perhaps the CDC7 is a molecule that may be more amenable to us carrying it further because the scope of the indication is smaller. So it's not; at least AML is something that a company of our size could manage and execute. Conversely, 1505, the MALT1 inhibitor, that's a very challenging competitive landscape in B-cell malignancy. Now you have multiple other modalities, combinations are required. It takes a long time to move up to frontline treatment.
That probably is beyond the scope of a company like ours to execute. So that makes much more sense to be a partner molecule, even though it has a very attractive profile and a lot of breadth of opportunity. And then the case of the Wee1/ Myt1, that is a really exciting molecule in terms of its profile. It has exquisite potency. And the fact that it has both Wee1 and Myt1 inhibition means that it is effectively blocking the escape clause. If someone escapes from Wee1, a tumor escapes from Wee1 control, we frequently would do that through Myt1. So the fact that it's hitting both of those, with some degree of activity is very exciting. But again, that means this could have very broad utility across multiple tumor types.
And so again, is that the sort of program we're going to take all the way to phase III? Probably not given that scope of activity. So I like them all, but they're slightly, slightly different in terms of their profile. And then in terms of just partnering, I mean, I mentioned that already, but we think that there is a very steep part to the value inflection curve that occurs as you de-risk the program for each program for their principal liabilities. So we need to show safety for the CDC7, and we now need to show some reduction in the number of blast counts in AML. If we show that, then we know it's a viable drug. So we don't need to have hundreds and hundreds of subjects, and then it could be partnerable.
So each one of these programs has an inflection point somewhere in the phase I, phase II development cycle that is likely to, we think, be the basis for partnering and/or further investment on our part.
Great. I'll end it there with my fireside questions. Audience members, if you have a question, there's a chat box below your screen to submit a question. So please, if you have a question, please submit it and I'll see it in my end. I have an emailed question for you, Geoff, on software.
Okay.
Thinking past 2024, and the 6%-13% revenue growth for this year, how should we think about longer term, mid to longer term revenue growth for the software business?
Yeah. We are very excited about the growth potential of the software business. We think this is a long-term, mid-teens or higher growth business.
There are some very exciting new capabilities that we see on the horizon for our software that can dramatically broaden the use of it and its utility for drug development projects. So, where we go from not just predicting what the right molecule is, but eliminating failure risks associated with off-target toxicity. So, we emphatically think that this is a growth business, with a growth profile that is more like that we reported last year than the guidance that we provided this year.
Great. Well, we're coming up on time. Thank you so much, Geoff, for participating and looking forward to following your story more. If you have any questions, feel free to reach out to me and I can connect you to Geoff and the Schrödinger team. Thanks, Geoff.
Thanks so much for your time today, Scott. Appreciate the questions.