Absci Corporation (ABSI) Earnings Call Transcript & Summary

Great. Thanks, everyone, for joining us at the 31st Annual Healthcare Conference Credit Suisse. Tiago Fauth, I'm the lead analyst in the biotech franchise. We're joined today by Absci. We have Sean and Greg from the company for a fireside chat. We can probably jump right in.

Generally, I'd like to open bigger picture for investors who don't know Absci. Can you kind of give us a brief overview of the business, where we currently stand, and I have several detailed follow-ups, so we can take it from there.

Yes, absolutely. Well, thanks for having us here, Tiago. As Tiago mentioned, I'm Sean McClain, Founder and CEO. I'm here with Greg Schiffman, our CFO. Absci is a generative AI drug creation company where we're really taking this paradigm of drug discovery, where you have either phage display or use display or traditional humanized mice where you end up generating these antibodies, but you end up like down selecting. You're able to explore what already exists in nature. And what we're doing is really combining synthetic biology, where we're able to generate mass amount of data on antibody functionality, on the developability and training generative AI models to explore a much larger search space. And this is what we call drug creation, where you're able to in-silico explore a much larger search space and really hone in on all the attributes that you ultimately want for a drug candidate, that the functionality, the developability, the -- ensuring low immunogenicity and getting that in one shot. One of the big issues that we have and why we have -- one of the reasons we have a 4% success rate throughout the clinic is that we're not able to get all of the drug attributes that we want on shot. We end up getting the functionality that we want but maybe not the developability. You changed the sequence. Once you change the sequence, you realize you don't have the functionality that you want anymore. It's this iterative process that ultimately gets suboptimal drug candidates into the clinic. But what if you could get it all in one shot? That's what's going to increase the success rate, and that's really what we're doing. Instead of finding the needle in the haystack, we're creating the needle. And ultimately, what this is allowing us with our own pipeline as well as our partners to do is really increase the success rate throughout the clinic and be able to shorten the amount of time it takes to get drugs into the clinic. With technologies like ours, we can reduce the time it takes to get drugs into the clinic by at least half, going from target to IND-enabling studies in 18 months. And really, this is like the new paradigm of drug discovery or drug creation as we like to say here. And so that's kind of a high-level overview of what Absci does.

Yes. So let's take a step back and unpack that a little bit further because you fall within a unique niche, right? So you have synthetic biology players out there, and a lot of those are more related to industrial applications. You also have antibody discovery companies that are based mostly on natural sources. So you kind of fall in between. Like what are some of the similarities and differences between those two extremes? And kind of where do you fall on that spectrum?

Yes, absolutely. Let's just take, again, phage display or yeast display like a company like Adimab. What you're doing there is you're taking a large library of different sequence variants and ultimately screening down to find the -- ultimately the antibody hit that has the functionality that you want, that binds with the affinity that you want and progressing that into the clinic. The other types of technology that are out there, technologies like AbCellera or Regeneron, where they have a humanized mouse model or -- and they're immunizing and ultimately generating the antibodies. And one of the issues that you have with these technologies is that you're only focusing on one specific parameter, in this case, finding an antibody that binds the target at the affinity that you want. And really what we're able to do is be able to specifically focus in on the exact attributes that we want for the biology. You no longer have to go through this iterative process. We can hone it. We can use our generative AI to, let's say, have an antibody that binds out a particular epitope that we want at a particular affinity and then has all of the developability profiles that we want and ensures that we have low immunogenicity. And we can go into the lab instantaneously and ultimately test that to ensure that, yes, that does achieve the biology that we want. But this is really -- it's precision biology. It's precision discovery, again, being able to hit all the different attributes that we want that you currently don't -- you're not able to get with existing drug discovery technologies. Like you can't tell a yeast display system or a humanized mouse to not to develop an antibody that has low immunogenicity, high developability. And again, this is what these next-gen platforms like ours are able to do, is really, again, hit all of the important attributes all at one time, ultimately increasing success rate.

Got it. Okay. So the value proposition that you want to make is pretty clear. And again, you're collapsing a few of these steps into one. So therefore, if you're successful, you can reduce attrition. One of the questions that we always get from investors is validation, right? So it's kind of -- and sometimes it's an unfair question for companies that are relatively new just because drug development takes a really long time. So the question we get is like how many drugs they have that were approved, that were discovered by -- and again, it's -- so if you start with that as the ultimate validation of the platform, let's walk back from that, right? What's reasonable in terms of proxies for validation, right? So you have a few programs partnered already. You also had some really interesting publications recently that show what you can do with your discovery. Can we talk about what do you think establishes that proof-of-concept across a few of these attributes that you're talking about?

Yes, absolutely. I think there's different types of proof-of-concepts. I think there's technology proof-of-concepts. Does the technology actually do what you say it can do? And then there's the clinical proof-of-concept. Is the drugs that are developed with the platform are actually more efficacious, have higher probability of success? And are they shortening the time? And we recently, as you mentioned, came out with this really exciting manuscript that has gotten a lot of attention from large pharma, academics and a lot of the top AI talent where we were able to show from a technical standpoint that we could actually specify an affinity that we wanted in our models and generate a sequence that bound at that particular affinity that we wanted. It was all wet lab validated and really showed that this technology that we're talking about here actually works. No one had ever shown that before, that you could actually use generative AI to generate antibody sequences that have the affinity that you wanted and also had what we call the naturalness. And that was a huge breakthrough in the space and really put us as really the leader in kind of the generative AI for biologics. And the inbound interest that we got from this was actually pretty extraordinary. Over the course of a couple of months since we've released this, we've actually had inbound interest from 5 large pharmas wanting to talk about partnerships due to the breakthrough work that we've seen here. And additionally, we recently brought on Dr. Andreas Busch, who was the Head of R&D at Bayer and was the CSO of Shire prior to getting acquired by Takeda. And he joined our Board, and 4 months later, he ended up coming on full time in an operational role. And he came on because he sees this as the future, being able to dramatically decrease the times, increase the success rate and combining his expertise, and I would say that he's -- I'd argue he's one of the most renowned R&D heads with the number of drugs he's gotten approved under his leadership, going from bench to approval. He's gotten 10 drugs approved. And so combining his expertise of drug development with our platform, we really see this as huge success. And I think it's a huge validating just from having a large pharma executive come on full time on to the team to help us develop our own pipeline as well as develop it for others as well.

Got it. Perfect. That's helpful. Let's talk a little bit more about the business model. So we alluded to a couple of your partners. And again, you're not necessarily right now focused on developing a huge proprietary pipeline. You have some partnerships. So can you actually explain what the business model looks like? And then, Greg, perhaps you can put some numbers around this, on how to think about value creation for Absci.

You bet. So as we look at the business model, and Sean has talked about the value that we create, I mean, this is a very value-adding platform. We're not an outsourced R&D shop. We really work with partners, and we view them as partners. And we look to sharing the success those partners have. And so the deals that we structure in the discovery side will have milestones as we go through the discovery process, funding that we get for the headcount and the lab expenses associated with that discovery work. And when it moves to the clinic, we'll get milestones associated with the success as it progresses through the clinic. And so even if a drug doesn't get approved, there's a lot of revenue and a lot of milestones as it moves along the way and increases in value going to a Phase I, hopefully a Phase II and eventually a Phase III. And so that's how we structure the deals. I think Roche would -- I'm sorry, the Merck deal that we recently did is a reasonable proxy to sort of look at a discovery program. We received $610 million that would be upfront in milestones associated with that deal. That's for 3 molecules. And so it's uniform across 3, so a little over $200 million each. That does not include royalties. The royalties bring all these into sort of the typical if you want to go to BioBucket's concept that they're in the billions of dollars. When we think about it, we think about it on an NPV basis. And the discount factor is not time value of money. We do put that in, and that's certainly been increasing in terms of the discount rates that we're using. But that being -- pull that aside, the real big discount is the probability that, that drug is going to eventually get into the clinic. And for our models, we've used the standard data that's out there, which is about -- it's about 4% if a drug starts in a discovery process that it actually moves through approval. We hope to see that improve. And if that does, actually, the NPVs that we're quoting now are probably lower than the actual ones the company will see. But the NPVs for a discovery deal are $15 million to $20 million. If I take the low end of that at $15 million, and I sort of think about if I complete 4 to 5 programs, I'm basically earning the total amount that I am spending in running my operation today. And so on an NPV basis, we're actually essentially breakeven if we think about it. And once we've completed the program, there's no additional work that we ever do. It is one that just generates revenue over time, and it is a law of numbers. It's not any specific program that we're targeted to. If I sign 100 programs, I expect 4 of those 100 again generate billions of dollars of revenue that's going to come in. And it's not all long-term revenue because this quarter, we had $2.4 million of revenue came in. A good portion of that is associated with the success payment that we had on the Merck program that we're doing for the non-standard amino acids. We will see near-term revenues. It's just they have a potential to go and stretch on for up to about 25 years after we've completed our work.

Yes. And one of the other items I'll mention is, with Andreas coming on board. We are looking to advance some of our own assets into the clinic. And the reason we're wanting to do this is because we can move much faster than our partners can. And with Andreas' success that he's had with our platform, I think that is going to allow us to get into the clinic, really be able to dose humans with the first AI discovered antibody and get proof-of-concept. And we're hoping to be at least with our first asset into the clinic first half of 2024, and being able to show proof-of-concept with our technology, I think, is going to be a really important milestone. And also too, that has value in and of itself. And so I think from a platform validation as well as creating value with some of our own assets, that's only going to drive more partnerships with large pharma, small and mid-cap companies. And I think it's going to be, yes, really exciting to be able to see that happen.

Got it. Okay. And I do want to touch upon that in a little bit more detail, but going back to the technology, again, what can you do with your platform that might be a little bit more difficult with other platforms? You did talk about the actual workflow, right, and collapsing several steps of that workflow into one or fewer steps and then have more of an in-silico predictive algorithm driving some of those properties. When I'm thinking about drug structure, when I'm thinking about targets, when I'm thinking -- you didn't mention about amino acids. Like what do you have in your suite capabilities that is kind of differentiated concert-wise relative to what you can find off-the-shelf with CROs or with other players in the drug discovery space?

Yes, absolutely. One of the exciting opportunities that our partners are really excited about is using our technology and using our generative AI to be able to design antibodies that hit a particular epitope at a particular affinity. With the existing drug discovery technologies, you don't know what epitope you're going to hit if you're going to hit a particular epitope at all. And with our generative AI models and where we're headed in the future is being able to specifically hit an epitope at a particular affinity or if you, let's say, you discover a brand-new target, being able to identify all the epitopes on a target and then be able to generate antibodies that can hit those at different affinities and then be able to go into the lab instantaneously and figure out what the -- which of those antibodies gives you the desired biology that you want. We've never been able to do that, and that's what's really going to accelerate our understanding of target biology, and it's going to allow us to get these cutting-edge therapies into the clinic much, much faster. And again, not only that, it's the multi-parametric modeling we can do. So it's, again, not only just hitting those targets but having the developability that you want ensuring that you can manufacture it at high titers, ensuring low immunogenicity. And so these are the areas that our partners are really excited about and what our platform can enable. And additionally, like another example is like being able to use this technology in the future to hit targets that have a point mutation and then conditional biologics, being able to have an antibody that combine in the tumor micro environment but doesn't bind to healthy tissue cells and having really high precision on that on/off switches. And since we're able to explore a much larger search space, you can really start to hone in and really get those very, very specific attributes that you want. And then additionally, we have the non-standard amino acid technology. There's been a lot of exciting developments in the ADC space. And what we're seeing even with our Merck partnership, they want to use our platform to develop the antibodies but additionally incorporate in our non-standard amino acid technology to specifically place ADC warheads on the molecules to develop next-in-class ADCs. And so I think all these different technologies really differentiate us and really show our partners that we are highly differentiated. No one else is able to provide these types of drug candidates, and therefore, really driving partnership discussions and being able to drive our upfront payments, drive milestones and royalties and continue to drive that platform.

And with that, maybe one just to add in, there's a lot of people doing AI. We're a large molecule AI. I don't think anybody else has published anything or shown any capabilities there. What we've been able to do in about a year's period of time is amazing, but what enables it is the technology. And this one is -- a lot of people can build out a wet lab. The equipment we use is not unique or specific. It's a great wet lab. That's world-class. But what really makes it unique is the technology that we have in the lab. And it's the proprietary E. Coli strain that Sean developed and the assays that we've developed that enable us to generate orders of magnitude more data than anybody can in an experiment. We generate millions of data points in an experiment instead of thousands. AI needs millions of data points out of an experiment to be able to do a generative model. That's what enables us to do what we do, and it also precludes anybody else from really being able to come in and compete directly with us in this area because nobody has the technology that we have that generates the data that feeds the model.

And you're spot on, on that. It's really the -- I call it the RPM. It's how quickly can you go from wet lab data of quality and quantity to then training your models than actually going into the lab and wet lab validating those. We can do that in a 6-week time period. And that's actually what attracts the best talent from Meta, from Tesla for at least on the AI side. Because you have those quick cycle times, you can iterate on your models. You can iterate on the hyper parameters, the architectures to really hone in on what is the best hyper-parameters architecture for the specific problem that you're trying to solve. And that integration against kind of that Synbio and AI and how quickly you can iterate on that is really what has led to all the success that we've had today.

That's an interesting point, because again, back in the day, there was some discussion about cell line development projects. And again, it sounds like you're kind of leveraging the same exact technology, but to build on top of that with the advantages in manufacturing and just the actual throughput that you can have, right? So to your point, how relevant is that in terms of actually having a short incubation period being easier to manufacture biologics in a different cell system? So how does that actually integrate into your current work stream?

Yes. No. I mean like the -- it's the same exact technology that we use for cell line development that we're using now, and that's the beautiful part. And like -- when I founded the company 11 years ago, I never would have thought that E. Coli was going to be the reason we got the data at the throughput and quality actually needed to train models and that would be extremely important for drug discovery. And that's really what gives us that 6-week cycle time. It's being able -- if you're not in a microbial host, you can't build billion member libraries or screen billions of different antibodies in a single experiment. And obviously, we have our ACE technology that then allows you to take those billions of different antibodies that you've produced and screen them for the affinity and the titer. And that's what gives us the ability to then go and train our models and be able to have that short cycle time. I mean we can screen billions of different cells that are producing individual antibodies in a given week and screen for affinity and titer. So that technology we've developed, we've never pivoted. It's the same technology, but we're now just using that to train our models and ultimately see our vision through of being able to design biologics at a click of a button. And that's really the future of where this industry is headed.

We joke, you can ask any company, AI is a buzzword that everybody is putting out there. You can ask anyone and they're all leaders in AI. And I think at some point, it's an investor really who is leading and why. I think if you look at the prepublication we put out and the attention that it drew, but on top of that, we've got Joseph Sirosh joined our Board. He was Head of AI at Microsoft and Amazon, an early pioneer in the space. We've attracted some of the top talent. We've got a great partnership with NVIDIA that gives us early access to the software, access to their systems if we need it. They see us as a leader in the space. I mean you've got a lot of third-parties that have actually validated it, not just ourselves saying that we're doing it.

Got it. No, that makes sense. And we just spent a lot of time talking about the technology and the actual craft of therapeutics. We haven't talked a lot about target discovery or target selection or however you want to point that. Again, it seems like you're pivoting for a business model where you're actually going to have some proprietary programs. But for most of the companies that have just partners, there is the selection bias that you only get hard targets or that you actually don't have the capabilities to find new targets that might be innovated and might unlock new biology. So how do you kind of square the target discovery, target identification across the platform? I don't know if Totient, if that acquisition kind of helped you enable that in some way shape or form. But again, I'm curious how that side of the business is evolving.

Yes. No, absolutely. So we about -- 18 months ago, we acquired Totient that really got us into target discovery. And that was one of the reasons we acquired them, was that for new target biology. And really what this technology was working off a TLS biology. Basically, there's a lot of research that's been going on in that. And that the B cells that are in tumors produce very different antibodies than those that are circulating in peripheral blood. And we are taking those tumor samples, extracting out the mRNA sequences, reconstructing those antibodies and then screening those in a proteome panel screen to find out where those antibodies are binding, what targets they're binding to, allowing us to actually start to discover brand-new novel targets. And the exciting part about this, too, from a technology standpoint, is not only are we discovering new targets to go after, but we're also getting data that can be fed back into our -- into our model on human antibodies to look at the naturalness, but also being able to look at the antibody-antigen pairing, again, to train models further. So it's another source of data for model training. And ultimately, where we want to go with this as well is being able to do this in-silico, where you can get the antibody and then instead of going antibody to target in-silico, you -- or sorry, a target to antibody, you're going antibody to target in-silico. And that's going to allow us to rapidly de orphan and discover new targets for new indications. And so we are -- we do have -- we are developing some of these and validating these targets internally for some of our own pipeline, but then also looking to partner out some of these really exciting targets and co-validating those with partners as well. So it's definitely an emerging part of our business model as well.

Got it. So I guess in the last few minutes, I just want to try to be everything together. So we do have a few partner programs already. So again, you're eligible for milestones and in the future, potentially royalty payments on those. You're investing in your own pipeline, and I'm assuming in your technology platform, and it sounds like you also have a few assets that you might be able to partner in the future. So a few different moving parts there. So balance sheet, cash position, how do you actually make that come through? You have a goal for potentially getting an asset in clinic in 2024. How well capitalized are you to delivering some of these ambitions?

You bet. So I think on that side, I feel very good as CFO that we have $198 million in our balance sheet. We do have -- $181 million of that is unrestricted. There is an amount that's restricted mostly associated with the payout when we have our first successful Totient program. When we look at the spend, we did $19 million in cash that was spent associated with operations this quarter. I think there was $5 million on capital. That completes any large purchases of capital that we'd expect over the next several years. And so we think we have a cash balance that carries us through late 2025, which just expected. We will cap some receipts of cash coming in over the next couple of years. And so if you put those together, so we feel really good about the cash balance that we have, the runway that we've got. We have an ability to execute, hopefully, see the capital markets' two sides. One, I think you get very comfortable with us as a business, which we signed an additional partners. And I think we just have an ability to see the good stock appreciation by execution there, and then hopefully, a recovery overall in the market because everybody in our space growth companies have seen a lot of pressure on their stocks over the last 6 to 9 months.

That makes sense. Any final thoughts on what investors might be missing, and the final value proposition for Absci as an investment idea?

Yes. I would say, again, it's really this concept of going from drug discovery to drug creation, sort of finding the needle in the haystack. You're able to actually create the needle, being able to design antibodies and biologics that have all of the attributes that you want the affinity, the functionality, the manufacturability, developability, being able to hit that in one shot, being able to rapidly discover new biologies faster than you could have ever done before. And all of this is going to increase success rates throughout the clinic and really shorten the amount of time it takes to get these into the clinic and really help cut down on that attrition rate, getting kind of through that valley of death. And really, this is the future of where the industry is headed. And the reason why we're able to do this is because we're able to combine our wet lab SynBio technology that generates the data, the throughput and quality needed to train the models, but then we can go and validate those models. How accurate are they? And that's all done in a 6-week time period. And that allows us to iterate faster than anybody else. I mean most other companies, that iteration cycle with the amount of data that we train and can validate, I mean, it takes years. And we can do it in 6 weeks. That's it's recruiting the best talent from an AI perspective and has what has brought on Andreas Busch on to our team, large pharma executives, and it's really creating the future of biopharma.

It is. What I mean, if you look the last 25, 30 years, you haven't seen improvements in terms of drugs, probabilities of success, the amount is $2.6 billion for every drug that's approved. If you can just go from 4% to 8%, you're pulling out $1.3 billion of costs. You've also got a large number of indications that we've never seen any progress over the last 30 years. I think the breakthrough next-generation biologics and AI has transformed so many industries. I think you're going to see it have a substantial impact and really things that we've been thinking about and talking about for 30 years coming to fruition over the next 5 years to 10 years. And I think we're going to drive a good portion of that.

And one last thing I'll actually bring up that I think is actually an important piece that gets overlooked is actually the IP that actually is generated from this. The sequence variance and the sequence diversity that we get out of these models is extremely large. And it's allowing us to actually get very broad claims to targets as well as the actual antibodies themselves. And think of a world if you could actually take out like PD-L1 was completely taken out from an IP perspective. That would be absolutely huge, and the breadth of diversity that we're getting the ability to actually validate those in the lab is going to allow us to actually generate IP claims that I think will be highly attractive for us moving forward. And so I think you have the business aspect, but then also there's some really exciting opportunities from an IP perspective as well, giving us a huge competitive advantage.

Understood. We're at time actually. This is great. I really appreciate you guys joining us at the conference, and I'm sure you're around for any follow-up questions. So again, thanks a lot for joining us.

Yes. Thanks so much, Tiago.

Thank you.