Schrödinger, Inc. ($SDGR)

We'll kick things off. Thanks, everyone, for joining us. My name is Mike Ryskin. I'm on the Bank of America Life Science Tools and Diagnostics team, and I'm excited to host for our next session, Schrödinger. We're joined by Ramy Farid, Chief Executive Officer; Richie Jain, Chief Financial Officer; and Pat Lorton, EVP, CTO and COO of Software. Ramy, Richie, Pat, thank you so much for being here. Thanks for taking the time.

Thanks so much for having us. Looking forward to the discussion.

So the format will be a fireside chat with some Q&A. If you've got a burning question, raise your hand, and we'll incorporate you in. But maybe just to kick things off, Ramy, maybe one for you. Just you recently reported 1Q earnings, reiterated your full year guide. What are the key takeaways from the quarter, how it played out relative to your expectations? There's been a lot of moving pieces in the model the last couple of months. So let's talk through it a little bit.

Yes. I mean, of course, the performance itself was -- we were very pleased. We actually slightly exceeded the guidance that we gave. But I think beyond that, I mean, obviously, that's good. But beyond that, I think how broad the source of growth was. We saw existing customers, large customers scaling up their usage. We saw companies adding new products. That's always an exciting thing. When you introduce a new product into -- out there in the world, it's not a given that there'll be acceptance of it. So that was good. And then the other thing, which is encouraging, I think, maybe for all of us is that unlike last year, where we certainly saw quite a few companies losing their funding, and we were losing customers because they were just disappearing. We saw the opposite this quarter. That's encouraging. Now we're all hearing kind of things seem to be getting better, but it's nice to see some data to support it. We actually added new customers, new logos. And it sort of feels like we haven't seen that in a little while. So the broadness, I think, of the growth and the quality of the discussions, the interest in the platform and then again, particularly the interest in new products was quite encouraging. Is there anything else I think we should add...

I'd just add as we drive towards profitability over a 3-year time period, first quarter out on that path, but increasing revenue across software and drug discovery while maintaining expense discipline. OpEx was down 4% quarter-over-quarter. So just the start of that journey.

Yes. So pretty encouraging start to the year.

And then in terms of the 2Q guide and sort of the peak through the year, there's always -- you guys have some pretty seasonal quarters. Nothing unusual to call out there, sort of in line with historical trends and your prior expectations, right?

That's right. It's typically a smaller quarter -- well, not typically, always is. Most of our renewals are in Q4. But we're optimistic about the continued growth opportunities in Q2, and I think the guidance reflects that.

And just Q1, Q2, we're out of the gate strong on hosted conversion as well. Hosted revenue as a percentage of total software revenue was 34% for the quarter compared to 24% a year ago. So we're pleased with that performance out of the gate there.

Good. On the 2Q guide, when you're talking about ACV, you had a slide in your deck where you did a sort of a year-over-year comparison and you had sort of -- you have to adjust the last year comp. Could you just walk us through that, remind sort of what the adjustment is and why the year-over-year is what it should be?

Yes. We guided $19 million to $23 million of ACV for the quarter. A year ago it was $23 million total, of which $5 million was contribution to ACV related to the grant funding we have from the Gates Foundation on predictive toxicology. So on an apples-to-apples basis, the year ago comparison is really $18 million and our guidance is $19 million to $23 million compared to that. So 4% to 27% growth.

And in terms of 1Q, you saw effectively de minimis ACV contribution from that. So...

Correct. There was no contribution from -- we called it out in Q2 just to make sure that you're looking at the comparisons the right way.

Yes. Okay. Okay. Maybe let's talk about that ACV transition and sort of the focus on ACV as the KPI of the business going forward. You made a big change in the model on that in the prior earnings call. It's been about 3 months now where you've sort of been talking to investors about it. What's been the feedback on that? How has that gone over because it is a pretty meaningful disruption to the model.

I think people get it. I think they understand that -- and because we're not the first company to undergo this transition, right, from on-prem to hosted. And in the year that you facilitate that transformation, the revenue has to go down. I mean it's math. And I think companies are used to -- investors are used to, okay, let's look at the thing that actually measures the business, which is ACV. And really, it's okay to ignore the revenue. It will catch up. In the meantime, focus on the ACV. And it seems like people are focusing on the ACV, right? I mean it's not -- that's good. And it's -- we're in a funny situation where we're supposed to be bragging about lower revenue, right? The lower the revenue, the better the hosting transition is. So I know that sounds weird. But again, I think we're not the first company to be doing that. So we're happy about the lower revenue because it means that the transition to hosted is working. That's great. And of course, the faster you do the transition, the better so that you rip that Band-Aid off, you focus on ACV while you're doing that transition. And then soon enough, we'll be able to go back to revenue when most of the transitions occurred. So it's working. It's working quite well. The other thing that's great is we're not meeting with that much resistance. We even had some companies we reported switching from on-prem to hosted before the contract renewal, that's a good sign. So a few years ago, that wasn't the case. I mean everybody was all worked up about hosted. What do you mean? You're going to have access to our data, but it's just the normal thing now. So right? I mean it doesn't seem like there's that much resistance.

Yes, there's virtually no resistance. It's really gone from 5 years ago, it was almost impossible to convince the top 25 pharma to do anything hosted. And now it's almost the opposite where the CIOs are coming and saying, you guys have a hosted offering.

It's easier for them, right? I mean we're taking care of the things that we're good at and they don't have to deal with it. So it all works out for everybody. Now we should be clear, there'll be some regions. I think we said in Asia, they're a little bit behind, so that might not happen, but that's okay. It's progressing where we expected it to, very, very nice, very well.

So I mean on that point, you said, as you called out, 34% hosted in 1Q, 24% a year ago, so a 10-point jump. I can do that math in my head, I hope. When you first announced this 3 months ago, I think you talked about 75% in 3 years, kind of give us a road map. You're not guiding quarter-to-quarter on what hosted should be because that's kind of ridiculous. But still, could you put 34% in the first quarter in context of that? Are you a little bit ahead of schedule? Is it too early to read into that?

Good question.

I'd say we're right on track. The journey from 23% at the end of '25 to 75% at the end of '28. We've said many times that we don't expect it to be linear. I think the first year is going to be a little bit slower on the transition as you measure it in revenue. Our actual progress in ACV will be much stronger, but the conversion of that to revenue in the first year is just reduced because it's the math part of the accounting. So I think year 1 will be slower, year 2 and 3 should be much faster on that path to 75%. So I think we're right where we expect to be.

And is 75% like a terminal limit? Is that just a 3-year goal? I mean, do you expect it to get to 90%, 95% to 100%...

75% was really set. I'll start and Patrick should add to this. That was set on our evaluation of the customer set at the time. There -- how they embrace hosted solutions, our view of the geographical differences and customer differences. As this evolves, as we move forward, we'll keep you updated on it, but that was our framing around 75%.

Material science customers, right, are a little bit more resistant than life science. It was a consideration of all of those things, yes.

Yes. And it's worth -- as Richie just said, it's worth mentioning that the 75% estimate comes from a moment in time. Had you taken that moment in time 3 years ago, we would not have been saying 75%. So it's hard to say what the ceiling will be in 3 years, but 75% is definitely a reasonable target today. We should add to that a lot of the new products that we're rolling out are hosted solutions. So again, that's another variable that we're thinking about on -- is 75% the endpoint or is there some different number.

Okay. And as you said earlier, as you get higher and higher up and as the change in hosted year-over-year diminishes, then the gap between ACV and revenue should shrink.

Exactly. It'll converge.

Yes. Okay. In terms of what -- where you've seen adoption pick up in the next 3 to 6 months, is it broad-based? Is it more big pharma? Is it a little bit of everything sort of...

It really -- that's what I was trying to say at the beginning. It really is -- this is a great sign. It's a little bit of everything. New small biotech companies, material science companies, certainly, large companies are contributing in the way you'd expect them to, but it's broad.

One thing really very exciting and we mentioned about how new products are contributing is if we close a new product sale with a big pharma, it's almost always off cycle because they have multiyear big bundled contracts. But what that does mean is we've now -- by adding new products, the next refresh is going to be an even bigger bundle. And we're successfully selling our new products into places off cycle, which is really exciting.

Yes. I mean something else, Ramy, you mentioned earlier was the new logos in the quarter. It seems just sort of a more constructive tone on end market dynamics. It's just one quarter. It's been very choppy there for a while, but we track biotech funding, new biotech funding things like that. It certainly feels a little bit better. Does one quarter make a trend? How much more do you need to see until you really think...

And there's a lag, right? It's going to take time, but it's encouraging, at least, right? It's certainly better. If we just look at the analysis of the quarter, it's better than Q1. That's last year. That's a positive thing.

I think it's obviously worth saying too that the lack of a headwind is better. When companies are going out of business -- that's net negative, and we have to make up for it in other places. And just that reduction is very useful for us.

Yes. I mean any ideas on what's certainly spurring this? Is it the pickup in M&A? Is it newer technology, sort of what's driving...

Funding environment, right? I mean, private funding, that certainly seems to -- all the stats point to that. I don't think to the point we were saying earlier, probably increasing IPOs, we're not going to see that right away. That's going to take time, right, to recognize that. I think it's that. It's actually fun. And it's kind of what Pat is saying it's sort of the lack of shutdowns, right?

The ones who've made it through the winter are hardier.

Right. Exactly. That's a good way of putting it.

Okay. And once those customers -- I mean, it's still -- it's early, but once those customers get their hands on money, are they spending it? Or are they kind of hoarding it? Because we've seen this dynamic play out a couple of times in prior years where you'll get a quarter or 2 of good funding, but you don't really see that come out because they're worried that it will go away, so they're hoarding...

Okay. I think that's an exciting thing to think about because I think -- and let me finish the thought because it's going to start sounding off not good, but it's going to end well. I think they're hoarding it, but that's good. They should be. And what does that mean? That means they should be doing -- what they should be doing is investing in things that are actually more efficient and that actually work. And that's obviously computation. And they can see that. I mean look at the success of Ajax just recently. And that's like one of many. So they look at that company and say, how did they accomplish that? They have 5 employees, right, and got to this huge valuation and this great exit. So I think the environment is still such that we need to all be efficient. It's not like all of a sudden, everybody is rolling in money. So you need to be doing things in a really efficient way. And obviously, using computation at scale is significantly more cost effective and has a higher success probability than doing drug discovery by trial and error.

And we're also very fortunate our track record of success with our collaborative partners as those companies go to the Nimbus's and the Morphic's and Ajax and all these companies go through exits, a lot of their employees end up reentering the life cycle of a biotech, and they have demonstrated success using our technology. And we see new companies essentially being seeded with these people whose success was demonstrated with our technology and who's the first person they call. They call us before they get their CRO because they know they can start getting results immediately in computation. So it's very promising. The ecosystem we've built from these successful biotechs is very promising.

You touched a couple of times during your earlier comments on new products from your end. I mean maybe let's start with predictive tox. Sort of what's been the uptake? What's been the feedback? What impact are you seeing on the rest of the portfolio?

Yes. As we said before, we were very pleased with the results of the method internally. We started to get this into the hands of other customers outside the company. And there's one very, very encouraging sort of development that we're pretty excited about. Our vision for this was the way tox screening is done, this is done generally very late in drug discovery, right? You make -- you try and optimize them since you're not able to predict those properties and you're not going to run some expensive animal tox study through the whole discovery. You do it towards the end when you think you're close. And that's a horrible way of doing things because, of course, you -- late failures are always bad, right? That's very expensive. And that's what's happening, right? You put a molecule that you think is clean, it goes into a GLP tox study and all of a sudden, it doesn't work, and that's it. You have to go back to the drawing board. And usually, that's when the projects get killed. Our vision is to move prediction of toxicology way earlier in the process. And we're really encouraged that our customers that are starting to get access to this have recognized this. That's how they want to use it. They want to use it early in what's called MPO right, the multiparameter optimization part of drug discovery, which is the very early stages. You want to be designing in safe molecules from the beginning so you don't have these kinds of surprises. Now the cool thing about that for us is that requires using the technology on a much larger scale than the way tox screening is used now, which is what we're just talking about, right, just a few molecules at the end. So if it's being used on that scale, obviously, that has a pretty big impact on the revenue that can be generated from it. So we're really encouraged that people are recognizing what we were hoping would happen is that this is getting moved up way earlier in discovery.

And it's probably worth reiterating, as we announced initially, we had a few targets, a handful of targets enabled at the initial release. Last year, we released even more. But this is part of the ongoing work at the grant, and we're getting more and more targets. And the broader the domains we cover, the more interest this is to people. because they might already know we're worried about this liability in the space. And unless we have that in our panel, predictive tox is interesting for that project yet. But virtually every day, we're adding more targets to the panel, and that's going to make it more and more interesting for every drug discovery project out there.

That's a great point. How broad is that database or knowledge that now?

Yes. So where...

What part of the portion of real...

No, it's a great question. So here's where we are. Internally, we've enabled approximately 100 targets. It's around 65 kinases and then the remaining are other very important, well-known targets that you want to avoid that you don't want to hit that are associated -- that are clearly associated with toxicity. It's on that order of magnitude. As Pat said, it's true. We are adding them on a very regular basis, almost on a daily basis. What's the end goal? It's probably a couple of hundred targets, I think, is -- would cover a very significant portion of the -- as some people now are referring to it as the AVOID-OME, the target you want to avoid, the AVOID-OME. And I think the fact that we've demonstrated that we have the infrastructure and the technology to enable these targets, it's become an engineering problem now, right? It isn't -- there's not a lot of uncertainty that we can get there. It's just a matter of enabling these targets, adding more targets to our tox panel, right? So that when you run it, you don't have any surprises that you're -- that there's a target -- an off-target that you're hitting, but that wasn't in our panel and we couldn't predict it. So that's our goal, 150 to 200. But the good news is I think the targets that we've enabled right now is enough to really get companies intrigued and starting to use it in production, and we're really -- we're actually seeing that. So I think it's one of these things that will keep getting better, but it's good enough right now to get started.

Yes, absolutely, yes.

Okay. When I think about the Schrödinger platform, it is a platform. It is a number of solutions. We rarely go sort of like -- or almost never go product by product...

Do you want to do that now? Are we going to...

Line by line. No. But I mean we've talked about FEP+ in the past. Now we're talking about predictive tox. Is there anything else that really jumps out as sort of like no single item, I think, really moves the needle, but what are the other important...

Yes. So FEP+ is very important because that's a workhorse for predicting affinity of molecules to proteins, which is kind of one of the most important things. I mean, if your molecule isn't bind tightly to the target, it isn't going to work. So that's kind of a pretty important step. Now that technology, the same exact technology, the same physics that underlies that technology can be used to predict another property that's really important, which is solubility. Obviously, if a molecule is not soluble, it isn't going to be a drug. It's not going to be bioavailable. There's a solution for predicting permeability. Again, if the molecule can't get through the cell membrane, it's not a drug. It's not getting to the target. It's not going to be bioavailable. So those are areas that we're excited about. The other really big area that we're focused on is what's required as input to these calculations. The input to these calculations is an actual structure of the protein. And it has to be an accurate structure. You got to get the atoms on the right place. Physics actually cares about where the atoms are and then they're in the right place. So there is a lot of work going into it, and we're making really, really cool progress here, really exciting progress on being able to determine the structure of a target to high resolution. That's enabling us to work on targets that were not possible to work on before. That's pretty important. And of course, it's the technology that enables the whole predictive tox initiative. So protein structure prediction, another really big area. Now we're also investing in our enterprise informatics system. This is essentially the platform where a scientist comes in the morning and work and sees everything that's going on in the project. Where -- what molecules have been made overnight, what were the predictions, what are the experimental results? And what am I going to make next? So they design in that interface. They collaborate. They collaboratively design molecules in this platform we've developed called LiveDesign. And it's getting used very heavily in the industry, a very large number of our customers have it. But that is essentially the interface to all of this technology. So it's fun to talk about all the science and these advances. But if you can't access the science, it isn't as valuable. So we put a big effort into making sure that the interface, the way you access these very sophisticated computational tools is also something that's highly robust, scalable, can deal with massive amounts of data, which we're now creating. So that's another area that we're really investing heavily in. The other area, just to focus on, just for a moment, I won't spend too much time in it, but a lot of these physics-based methods that we're developing that are aimed at designing drugs are immediately transferable to designing materials. So we do have lots of exciting technologies around, for example, designing batteries. That's obviously a pretty important area, polymer coatings for every application from aerospace to food packaging and lots of other things. I don't want to spend too much time on it given the focus of this conference. But it is -- those are very important areas, and we're very excited about them.

And so now let's -- I'd love to highlight 2 of the big new things that we're focusing on. So Ramy talked about LiveDesign, and it is used at the overwhelming majority of top pharma as well as many, many biotechs, but it is used primarily for small molecule design right now. In the last year, we did our initial release of LiveDesign Biologics, which targets the other 50% of R&D out there. And we're seeing really great interest. We have several big pharma that have signed up for pilots. We have a couple of big pharma who have entered in co-development because they're going to integrate in their entire system. So we see this as an opportunity to extend from our base of small molecule strength into the biologics space where there's an equal sized R&D budget. In addition, what we announced on our last earnings call, Bunsen is the new Agentic AI that we released. We are extremely excited about this. We've been using it internally since late last year. It has the ability to take a computational chemist who is completely overloaded and can only do a fraction of the work they'd like to do and give them a dramatic speed up similar to what we're seeing with cloud code and developers. We might have referenced it before, but internally, when we're running programs, the reason we have the success rate is we have something like 4 computational chemists per discovery program in the industry that -- it's the inverse of that. It's something like 1/2 to 1/4 for discovery program. So we're putting 8x to 16x the computational resources, just human being wise, not even GPUs. We think Bunsen can help close that gap. We can take somebody who only has half of an FTE to put on a project and allow them to accomplish 2x to 5x as much and in the process, use 2x to 5x as much of a throughput-based licensing. So it fits perfectly with our commercial strategy and our scientific strategy. So we're super excited about Bunsen and LiveDesign Biologics.

A bunch of that I want to follow up on real quick. First, Ramy, your point on protein structure elucidation. Just to be clear, is that entirely computational? Or is there an experimental component -- we've used cryo-EM in the past.

Yes, that's a great question. There are many methods for producing low-resolution structures, which are an input to determining high-resolution structures. So experimental methods like X-ray crystallography, cryo-EM and then applications like AlphaFold. They all produce sort of a good starting point, but they're not the end result. Those structures are simply not accurate enough to use in sophisticated physics-based methods. So then there's another computational layer on top of that, that relies very heavily on rigorous physics-based methods to take those kind of low-resolution approximate structures and get them right. And so we're developing those technologies, the latter. But of course, we will -- I mean, thank goodness for those other methods, right? Otherwise, there would be nothing to apply those sophisticated physics-based methods. So it's highly collaborative and you need all of that, the experimental methods, the low-resolution computational methods and then the high-resolution computational methods that we're developing.

Okay. Okay. And on Bunsen, yes, I mean, I think you just announced it a couple of weeks ago, I think early access this summer, right?

Yes. On track for that. Yes.

How do you expect the initial commercialization of that to go? Are you targeting specific accounts? Sort of what's that road map going to...

Yes. As always, when we initially release something like this, we go to our close partners, and we look for feedback. We need -- we -- of course, with all new technology, I mean everything, there'll be things that work really well and things that need to be improved. And it's nice to know what those things are, so we can prioritize them and continue to develop them. So we have a set of our partners that we generally work with. We've already demoed the product to some of those. They're blown away. So far, it seems like the reaction is universally positive. This is something they've been waiting for because they're struggling, right? They're struggling. They know that if they use the technology on a larger scale, they're going to have a bigger impact. They're seeing that happening. They see that happen at Ajax. They see it happening at Schrödinger, and they want part of that. And I think they see this as an answer to that. And we'll get -- do the usual thing. We'll get that feedback. We'll incorporate that, have another release, and we'll do that in a pretty rapid pace. We're pretty excited about it.

One additional point I'd say that really does differentiate us from a lot of people in this space is we already have a couple of hundred drug discovery people using this. So we're -- because we have our own internal drug discovery and we can dog food with them, we're not going to release something that we think, oh, this might work, the customer feedback is a polishing phase. We already have computational chemists in-house that are telling us they are multiple times more efficient. We have medicinal chemists accessing technology that was impossible before. We know this works. It's really just getting into that final stage because there is a gap between an internal platform and a commercially sold platform. So we're really getting that final polished done to the point where there's a commercial quality platform.

Yes, good. I mean just on that point, just play devil's advocate a little bit, though. You're not the only ones launching some of these agentic solutions, right? Just in the last 3, 6 months, a lot of companies come out of the woodwork, both traditional tools vendors and more traditional AI or software vendors. I don't want to say that the competitive landscape is crowded, but again, it's not the only platform hitting the market. Do you have the name recognition? Do you have the relationships to sort of give you a leg up there? Have you been asked to do like head-to-head demos or sort of like what's that market look like?

So it's really interesting. We haven't been asked to do head-to-head demos, but we -- so I just actually demoed this to 14 of our top customers last week and got 100% interest. The -- we haven't asked to do head-to-head demos because when they've tested the generic solution so far, they just fall flat. They don't have the specialist knowledge they need. One thing we've been able to do is through constructing specialized harness and skills that know exactly how to use our software, exactly how computational chemists work, we can build something that a genericist just cannot do. There is not enough information on how this works out there. So our core advantage is kind of what it's always been. We understand computational chemistry better than anyone else, and we're embedding that into the agent. That said, if one of these other agents work, the only way it's going to work is if it has access to our tools because the agent can't make up for the fact that the tool isn't accurate enough. So if another agent beats us at one of these customers and it's using our tools, we will still capture the throughput-based usage of our tools. That's still a win for us. We do think to do everything right, at least at this point from what we've seen, we have to own that because the skills of these agents not developed by us is just simply not there.

Okay. We've only got a couple of minutes left, still a couple of other topics I want to touch on. Maybe, Richie, a couple for you. You talked about profitability, operating leverage in the quarter, just sort of tightening up that cost control a little bit. Could you just talk about the ramp from where you are now towards EBITDA positive? And then maybe I'll just throw on it at the same time, Ajax, parts of the therapeutics collaboration portfolio and maybe some of the incoming cash flows. You haven't seen the cash come in yet, but how does that change your approach on where you're investing and how much?

Sure. I'll take those in reverse order. So we ended the quarter at $402 million of cash. We should expect that to increase once we receive our 6% portion of Ajax's upfront with Lilly. That amount was not disclosed, but that will be an increase in cash for us. In terms of the use of our cash, this is mostly investing in operating burn over the next few years and investing in R&D and science to support a growing software business and drug discovery business. So the extra use of cash, once we receive it, we'll evaluate what to do with it. I think we've said in many different forums. Our focus right now is investing in growth. If there are adjacent nearby adjacencies for M&A, we'll consider that as well. The outlook over a 3-year time period is 10% to 15% growth in software ACV while transitioning over to hosted revenue and having the revenue converge with ACV. $150 million of revenue in drug discovery over 3 years. So roughly $50 million a year. The timing of that can be a little variable, but $150 million over the 3-year period. And this year, we're guiding to operating expenses being lower than last year and then a disciplined framework on that going forward. So the entire package there is what leads us to the profitability goal on adjusted EBITDA in a 3-year time period.

Okay. Okay. Maybe just with the last 30 seconds, Ramy, I'll toss it up to you. Any concluding remarks, any last points you want to leave us with? What's been some of the more interesting feedback you've got or maybe some questions you had with investors you'd like to address?

Sure. Yes. I think, first of all, obviously, what we were saying before, I think it is really worth pointing out, it's very encouraging to see that things are improving. We're happy with the sort of broad-based growth that we're seeing. One of the things that I also really want to highlight that I think sometimes gets lost, there's a lot of noise out there, right? There's a lot of companies talking about computation and AI. And I hope that investors recognize that these companies aren't all the same. And I think our track record, which is pretty extraordinary. I mean, Ajax is just the latest example of that, but starting with Nimbus and Morphic and Relay and Structure and Petra, these -- and the nearly 100% retention rate of our customer base and the tens of thousands of users using the platform and even we were talking about earlier today, companies, biotech companies that are using the software on this really large scale, seeing these really great successes. This is really an actual track record. We're delivering development candidates. We're delivering molecules that are making it through the clinic. There's, by the way, nontrivial royalties on sales associated with those. So we look forward to realizing the value of those in the future. And I hope there's a recognition that there's a track record here of delivering with science that actually works when it's used at scale, and that's differentiated from some of the other things that you might be seeing. I hope that doesn't sound arrogant and it's not -- I don't like to sound critical of what other people are doing, but I think it's worth mentioning because it's a pretty big difference. I hope that was okay to throw in there...

Yes. That's a good place to end it.

I appreciate the opportunity to say it. Thank you.

Of course, of course. All right. With that, thank you very much. Thanks, everyone. Appreciate it.

Thanks.