Good afternoon, everybody. We are going to go ahead and get started on our call and webinar here in a few minutes. We are going to give everyone some time to log in. We are increasing the number of attendees. I have got my colleagues from Lantern and Open Medicine also on, Dr. Kishor Bhatia and Reed Bender. They will introduce themselves a little bit later in this call. As we give everyone some time to join in, I am going to remind everyone that we will be making forward-looking statements, and the presentation I am making today will contain forward-looking statements, and that I urge you guys to read our full disclosure that is in compliance with Section 27A of the Securities Act of 1933.
And that we may or may not update statements that we make today, and these statements will include potentially issues such as anticipation, revenue, product roadmaps, estimates, et cetera.
You can also access our annual and quarterly reports under the Investor SEC Filings tab of our website, and also at sec.gov. All forward-looking statements in this presentation represent our judgment as of today, the date, September 23rd, and we disclaim any obligation to update any forward-looking statements to conform to actual results or changes in our expectations. With that, I am going to go ahead and get started. Thank you guys again for joining us. I have got my colleagues, Dr. Kishor Bhatia and Reed Bender on as well, and they will help me with questions. I urge you guys to type in your questions.
We will be doing a demo talking about Open Medicine, where we are headed, the feedback that we have been getting so far, and also talking a little bit about how we see the roadmap evolving, both in terms of usage and in terms of growth of this amazing platform. Again, I urge you guys to type in your questions, or raise your hand when we get to the Q&A, and we will try to work through as many questions as we can today. Okay. Can everyone see the presentation?
Yeah, we can see it.
Great. We are going to talk about Open Medicine. It is a platform that has really been built internally by drug developers for the broader drug development community. Our vision for Open Medicine is really to become like a Bloomberg, but for medicine, for everyone involved, whether they be developer, service provider, investigator, individual, power groups, and large pharma companies. Our vision was initially to accelerate our own drug development because we had to. We were a small, emerging company. What we have found, which is really quite exciting and amazing, is that the entire world realizes the power of AI and has moved to very similar models. We have spun out or are spinning out Open Medicine into its own entity, which we accomplished earlier this year, and we have transferred certain assets and IP and capabilities into this new multi-agentic platform.
Again, I made the forward-looking statements earlier in the call, so I will not go through those. Let us talk a little bit about what we are going to cover. I am going to give you a view of the market size and how we view the market, which is a little bit different than how a lot of other companies and people approach the market. It is a large market, no doubt. As a former analyst myself and a consultant, I really try to size the opportunity bottom up and in terms of what is really addressable, where is the revenue, new or disruption coming from, and how will this actually shift to companies like ours? I am going to give feedback from customers.
I am going to talk about the roadmap toward the end, types of partnerships that we have gotten interest in and seem to be pursuing, and also a little bit on, of course, the demo, because that is worth more than any webinar alone, is actually seeing the differentiated capabilities in real time. Again, we are going to allot about 45 minutes. We are going to try to end our discussion and demo around 1:00 P.M., and then leave 10- 15 minutes for Q&A. Like I said earlier in the discussion, when we first started Open Medicine, back when I got to the company at the end of 2018, early 2019, we were not setting to build an AI platform company.
We were really building a drug development company that was using an AI platform, an industrialized AI platform, not one that you take off the shelf and put away, but something that was always on, always evolving, and really an AI-native drug development company is how we viewed ourselves. Our focus was really singular. Can we get cancer drugs faster and with less capital? We built the tools, and we built the methodologies, and we built the approach because that is what we needed. That was the way that we could do it. We built this on the back of clinical programs. At the time, we had no clinical programs. They were all preclinical. In fact, some of these concepts did not exist. The indications, the Fast Track designations, LP-284 as a molecule, all were ideas.
Today, we have LP-184 that is going into phase I-B, phase II trials, multiple trials, as well as has four Pediatric Disease Designations and two Fast Track Designations and multiple Orphan Designations. We have LP-300 that is in a phase II trial that has seen some great results so far in a very targeted population called L858. Mutation in non-small cell lung cancer, and we have also developed a whole new drug, LP-284, which is a stereoisomer of 184, which was optimized using our platform, and we brought it to GMP quality, launched at phase I, have three Orphan Drug Designations for that drug. Just yesterday, we talked about a new patent for that as well, and we have seen some good responses in the trial. We also focused on CNS cancers and developed a new subsidiary to focus on these devastating cancers, both in pediatric and adult brain cancers.
We have, we think, a library of amazing work. But the most important thing is that all the molecules that we set out to develop not only showed themselves to be tolerable and got to a meaningful dosage that is therapeutically relevant in trials, but are actually seeing results, mechanistic results that were thought about using our AI and data-driven approach. We started this AI layer, this machine learning layer, to support these programs. Initially, RADR, which was a machine learning platform, which has a lot of accolades, hundreds of ML algorithms, 200 billion+ data points. But again, focus for us and for our collaborators. It was internally focused and continues to be used internally. It is a team effort. You have got to bring in people who are data scientists and molecular biologists and other multidisciplinary people.
It was really there to compress what traditionally takes place in early development that can take two, three, four, five years, and try to compress that to one or two years. As we saw that succeed, we thought, well, could we do something even more aggressive? Could we build this into a natural language system that you can prompt without the use of a data engineer, without the use of a multidisciplinary team, and still access a lot of the information and algorithms and set them off in an automated fashion? At the time, as natural language processing, ChatGPTs, and LLMs were growing, we thought, could we create this for rare cancers? Could we create an AI for good, where we take all of our knowledge, guardrail the architecture, and focus it on the entire litany of hundreds of rare cancers?
Initially, it actually started as an experiment internally with the team. We said, Wouldn't it be great if we could take a biomarker or mechanism and just fully address every single cancer and see where it made most sense and automatically rank it and do all the complicated pathway analysis that we do, and then actually verify it both at the RNA and protein level, and then have it come back to us with the results? That is great. It is not impossible to do, but it could take days or weeks to do that, enrich for pathways that are both known and maybe not known, enrich with your own proprietary data, look at results, look at different bioinformatic analysis, guardrail with existing published literature, and iterate. We thought there is got to be a better way, a more automated way, and that is why we created withZeta.
Our learnings in withZeta were fantastic, and Reed will talk a lot about the architecture later on. Each layer that we built was built to support and guide the programs. Kishor will talk about how we've now also used withZeta to develop whole new programs. We have an entire library of new molecules that we haven't even really talked about, which we'll be talking about later this year. Eventually we said the same architecture we can use not just in rare cancers, but we can use it for other therapeutic development purposes. We can expand it to go downstream. We can expand it to go into new disease categories. We can actually now use a subscription and freemium model. There's no reason we have to go door to door to every pharma and spend months convincing them.
Let's just open this platform up and let's disrupt the way drugs are being developed. Again, each layer for us was built to run, influence, guide, and validate the work that we've already done in the past. This is very important. We're not approaching this as some kind of data and AI agent building exercise. We're really approaching this as can we empower people to do drug development the way we would do it faster, cheaper, highly parallel, and at a level and pace that hasn't been done before. Where do we sit with Open Medicine? We see Open Medicine as a wonderful complement to our core drug discovery and development business in cancer. It allows our team to focus on a major new category and our investors to participate in the upside of this AI revolution.
Open Medicine does this not only just for us, but really for multiple categories. Some of the early users are not just biopharma researchers, but investment funds, academic centers, clinician scientists globally, and actually service providers and CROs. The way I view it is it's a platform that's serving not just us, not just cancer, not just small pharma companies, but really the whole industry that's invested in the success of drug development and making drug development faster, more precise, and more efficient. That's ultimately the real power of this kind of platform and these tools, is to allow us to produce new innovations in not only hypothesis development, but validation, allow us to evaluate, but also generate competing explanations in a way that just hasn't been possible before.
This to us is a complementary business, one that has a lot of legs beyond just our own capabilities, which is the goal of allowing it to be spun out and grow independently and be valued as an AI business. As we're developing this webinar, this is very timely. There was an article that came out, in fact, I think it was dated just two days ago, in Nature, volume 657. It was a technology feature, and again, this just came out. I just circulated this to my team. The title of this article is AI Co-Scientists Are Revolutionizing How Research Is Done. It's pretty amazing. It's even more amazing because when it was initially published, it was called Are Changing How Research Is Done. I looked at it again this morning, it altered from changing to revolutionizing.
The author, Elie Dolgin, sorry if I mispronounced her name, did a great job at highlighting one of the most important things, is that humans still need to decide what makes sense. Our ability to get to that point of making sense can be so compressed, and I will talk about how we are doing that. There are a couple of great quotes that are on the screen here, and these are exactly the issues that we set to solve in doing Open Medicine. We have had years of thinking about it already. The bottleneck is in reliable validation. Exactly. That is by Anton at Northeastern. That is why we built curated knowledge bases. That is why we guardrail architecture. That is why we created proprietary curated information. That is why we validated all the bio tools. Second, the most valuable part is now actually asking the question.
100% true, and Kishor will talk about that. The quality of your answer is so much driven by the quality of your interaction, just like it is in any relationship. I do not believe in these one-shot, just ask anything, Hey, make me a presentation for Open Medicine. You are going to probably end up with junk. If you go through and say, These are the things I am thinking about. These are the must-haves. These are how slides should look. These are the library of content. This is the information that I have in my head. Same with any other thing. The most valuable part is in the asking of the questions, the guidance, just like any co-scientist relationship would be, whether it be human or, in this case, digital. One of the other observations from the article is once the protocols that are generated are not very good.
They are like recipes. I agree. I have worked a couple of other tools out there in the industry, and I will not name names. Some with big companies with fancy French-sounding names and some with generic tech names. They do seem very cookbookish. Again, the opportunity is, we were there, too. The opportunity is to make these a lot more specific and less cookbookish and more very targeted to the nature of the inquiry and research. That is what Open Medicine has accomplished. The field is really now at the area where Open Medicine is already built to solve. A co-scientist trained on very specific categories, a co-scientist built inside and from drug programs, and really understanding the prompts, the cost limits, and also being able to tell you when your hypothesis makes no sense. This is where Open Medicine is and what we have accomplished so far.
Let me tell you a little bit about since we have launched it and made this announcement in April. We have had excellent customer feedback loops. We have had multiple specific input from customers across multiple industries, ranging from hedge funds to CROs to clinician scientists to chief medical officers. We are putting together a world-class advisory board that we will be announcing in the coming weeks. We have had several people already say yes, several new invites go out, and that is really critical because the advisory board is going to surface market insights that we may not even know about, surface functionality, give us access to new talent, give us access to new people, users. Advisory boards are really critical, and again, as a small company, you need those.
You need those advisory boards to help you get to people you cannot get to help you miss and surface opportunities that you may not be thinking about. We have improved and strengthened the security and scalability of the core architecture. Reed and the team have been focused on expanding the functionality, but also, very importantly, running benchmarks to actually improve cost and scalability, which we believe is going to be absolutely critical. We have also gone fully mobile. That is one of the challenges that we had initially with the platform is how do we have that seamless experience to go from a desktop, which is perfect for it, but now to your phone or to a tablet where you may want to extend or do things quickly and follow up.
One of the largest recent updates of the platform was for mobile users, and we have seen significant increases in certain classes of users since launching the mobile module a few weeks ago. We think it is going to be a key ingredient in that seamless experience, but also very importantly in keeping people and making that experience sticky. We also have some great ideas. We have started some discussions outside of cancer with institutions and pharma companies where they will help us go well beyond oncology. We are in discussions and negotiations with an institution to go into neurodegenerative disease with some of their multimodal models, and we are also looking at inflammatory immune disease.
Again, there is only so many of those that we can take on at a time, so we will be very careful not to overextend ourselves, but to approach new diseases and new modalities where we think it absolutely makes sense and we have the right partners. How do we see the market developing, which is very critical. Again, we see these billion-dollar numbers all the time, and I want to talk about how this is a new category. This is kind of purposeful, but if you look at what I estimate to be about a $4.5 billion- $5 billion market spend now currently, and the agentic piece, the agentic AI is tiny. It is barely visible. But if you look at the other pieces, platform deals, about sub $1 billion in actual real platform revenue, and today there is a $2 billion deal or $1 billion deal announced.
But how that actually reaches the company, the company is not getting $1 billion. That is all on the come later if the drugs maybe work as a result of their platform tiered out over so many years. Out of that $1 billion, maybe they get $15 million up front. Okay, I count that $15 million. I do not count the $1 billion. In services and consulting, tons of companies offering services and consulting. We will build you a knowledge graph, we will do bioinformatics for you. We will take your data to show you how you can do a Bayesian model. We will do N-of-1 analysis. We will redo your trial. We will find and rank combination agents. Huge industry of services and consulting, whether it is data, but all focused on drug development. Then software, great software companies out there, whether it be Certara or Schrödinger or others.
The big thing, and when I spent several days modeling this a couple of weeks ago, one of the things that struck me is that services, consulting, and software are actually not going to grow. It's going to be disrupted. The growth rate net of inflation is pretty marginal in those businesses, and that's where AI is disrupting not only drug development, but actually if you look at any industry. The agentic AI is eating consulting and traditional software. That's why SaaS gets really affected. No one wants to pay $13,000 per seat for a rigid docking, maybe with some quasi-decent flexible docking and maybe some large-scale QSAR analysis. Who cares? To me, that's table stakes in the future. No one's going to pay that per seat. We do it inside of Open Medicine. We do it inside of Zeta.
Why not have it as part of a larger bundle for drug development? Same with consulting. Why do I need to go hire a bioinformatician or 20 of them when I can have my own bioinformatician use the withZeta bioinformatic toolkit and do the work of three or four or five people? This is changing how we see the market developing, and specifically, agentic AI will disrupt legacy software and services. About $3.8 billion, I estimate, of today's revenue that's eaten up by software and service companies will be actually flat to negative and close to $1.5 billion- $2 billion of it will move to agentic delivery. That's a big number, and it's pretty fast. If you look at even the rates of usage, even of people using these platforms, it will be very fast.
There was an analysis done by The Economist that, I think NVIDIA also was involved somehow, so obviously there may be a bias. But I thought it was very good work. The analysis suggested that agentic AI revenue in financial service institutions among the top users, of course, had grown from nearly zero three years ago to almost $7,000+ per user on average. I think that's probably right. Probably is even higher. In consulting companies, we've seen this across a lot of consulting companies. I sit across the street from one of the big consulting companies named after a major city in Massachusetts, and down the street from another one that kind of has a Scottish heritage name. But consulting is booming, but they're not hiring.
In fact, the word on the street is that it's flat to negative, because they're using agentic AI to do a lot of that work. The same thing's going to happen in drug development. So I see that market in legacy software, legacy services crumbling and all going toward agentic AI. The other thing that I see, and if you see this in the chart that I've shown, if you look at platform deals explode. I do think that platforms and providers of platforms like ourselves will take more on the come. If someone says to me that, Hey, we only want to pay you $1 million a year for your neurodegenerative module, and if one of our five candidates comes out as a result of it, and we'll give you 10%, 3%, 2%, I would take that.
So we see that companies that have these agentic platforms, because they can scale on their own and they offer a level of scale that you've never seen in the software industry, people are going to be more willing to take that risk. The risk that a traditional software and SaaS provider has not been able to take. That is going to cause more drug candidates sooner and faster. Agentic systems will hold a very important place in accelerating that because you can run multiple hypotheses in parallel, fail earlier, pivot faster, and advance more candidates. So that is how I see this, and this is very important to how we are architecting and thinking about Open Medicine.
So, going into what this causes, and this is very important because what you are hearing from me, and I think for most of people, the vanguard of agentic AI software, is that you are seeing simultaneous disruption and enablement at the same time. This is going to create a new generation of scientific work. Every prior computational wave made a single path cheaper. You do not have to go to the store and drive to Walmart to buy games. You can buy them online. Now they can be served online. Now they can stream online. But agentic changes the shape of that whole path itself and the execution, everything from the ideation and search to actually how it is accomplished. You can run multiple hypotheses in parallel at the cost of what it traditionally took only one.
The value is now going to be measured not just in hours provided or service or consulting fees, but in terms of the time to innovation and the encoded judgment. Very importantly, knowledge is going to be constantly created and encoded into the system. That has not been possible before. You can see these legacy tools, these legacy CRMs, these legacy software packages, these legacy chemistry packages. You create stuff, but it does not get re-encapsulated into the knowledge base, into the core workings of the software. That is changing now. Because of that, we can run these multiple programs, fail earlier, pivot faster, and the systems will now be valued not only on the cost, but really the speed and parallel enablement. Not just selling seats or butts in seats and licenses. You are going to be selling on your ability to execute and do things faster.
With that, we are going to dive into actually seeing it at work in a demo. Before I do that, I am going to ask my colleagues, Reed and Kishor, who are on with me, to also introduce themselves and give a little bit of background on what they do at the company and for Open Medicine as I pull up the live demo.
Kishor.
Hey, thanks, Reed. Thanks, Panna. I am Kishor Bhatia. I am the Chief Scientific Officer for Lantern Pharma. I have been with Lantern Pharma for about five years. Began mostly with focusing on the preclinical aspects of our drug pipeline. It was very exciting but slow work. Here is the advent of RADR, and things kind of jumped up several magnitudes. Perhaps, what I would like to do in the next five minutes is just share with you some recent interactions with Open Medicine to kind of give you a sense of what this enhanced excitement of discovery is about. In trying to have that dialogue with you, I thought perhaps what I will do is share a couple of case studies. One where we engaged Open Medicine to design first-in-class molecules to the areas of uncovering synthetic lethal pathways or identifying rational drug repurposing combinations for rare cancers.
Understand that each of these interactions involves a different persona within Open Medicine. As you probably understand, Open Medicine platform has several personas including medicinal chemists, translational biologists, so on and so forth. So one of the more recent challenges that I posed to Open Medicine came from a question that arose in my mind after a recent paper in Nature talking about ferroptosis, which is a pathway that causes cancer cells to die, particularly certain specific cancer cells. I wanted Open Medicine to synthesize a drug that uses this pathway. To summarize, Open Medicine rapidly synthesized insights across several different parts of this puzzle. Across the ferroptosis, across lysosomal targeting, across medicinal chemistry, and then proposed a bold design strategy that created a hybrid molecule which combines lysosomal accumulation and iron mobilization in a single structure.
When working with this chemist persona, it identified a scaffold, a known scaffold, artemisinin, and added a tertiary amine, which resulted in a novel molecule. Turns out that based upon other analysis, this is a first-in-class iron activator capable of triggering a synergistic ferroptosis cascade that bypasses conventional resistance mechanisms. So within an hour, this platform delivered a fully synthetic route for the drug, which has physiochemical drug-like properties, a translational development plan, and identified the right indications where such a drug could be used. Now, clearly all this needs still further validation which we are doing, but I think the point I am trying to make is that trying to pull together all this data and putting together a hypothesis that can be validated, a drug that can be synthesized within a short time is just not feasible if one were to do without such a platform.
To give you another example, I think I will take the example of understanding how I will try to extend further indications for which LP-184 or a drug can be used. So, I was interacting with Open Medicine to ask similar questions, and it turns out that Open Medicine identified a very specific gene called STK19 as a potential pathway to use to expand the indications of LP-184. This, to me, was very surprising because my recollection was that STK19 is a kinase and why would Open Medicine connect STK19 with LP-184, which is a DNA-damaging drug that requires a specific DNA repair pathway? Turns out that I had missed a critical flaw that was published very recently. The fact that Open Medicine could pull out the more recent information which showed that STK19 is also a DNA repair protein was quite surprising.
Nonetheless, what Open Medicine then did was identify a convergent pathway that expanded the synthetic lethal mechanisms of LP-184 to cancers that we were not thinking about, to cancers that had biomarkers that we were not thinking about. Clearly, I think both those case studies demonstrate how the utilization of the right questions and the right mechanisms of interrogating with Open Medicine allows to open doors that would be difficult, if not impossible, for a small team to get into.
Thank you.
I'll introduce myself real quick and then hand it back over to you, Panna, for the demo. My name is Reed Bender. I'm the lead platform architect behind withZeta and Open Medicine. My role has been in developing the infrastructure and the agent itself behind withZeta. Before that, I was working as a data engineer, and my role was in compiling all of the data sets within the RADR team that ultimately became the foundational knowledge base for withZeta. I've really enjoyed working across the whole stack and beginning with the data itself, building out the core knowledge bases for Zeta, the rare cancers knowledge base, our ontology of the rare cancers, and then expanding it beyond a literature search or simple lookup, but adding in really complex analytical tools like computational biology and pathway enrichment tools.
Those have all been fantastic, as well as chemical structure tools. We have Ether Zero integrated for generating real chemical structures and reasoning over SMILES strings, which normal LLMs are generally pretty bad about. It has been quite an adventure and very fun to develop this platform with the great team that we have, and it's cool to get to see it in demo here.
Great. I've started the demo in parallel. Kishor had mentioned ferroptosis, and we got a great answer ranking by creating a kind of a ferroptosis therapeutic index.
Yeah.
Not only did it think about a mechanism, but it thought like a scientist. It said in order to do a ranking, it created sensitivity times clinical need times feasibility. Just the way a scientist or consultant would think, and then gave us an answer of the top five cancers to think about, and then also the mechanistic convergence on the therapy-resistant mesenchymal states that unify why these cancers are being ranked. Very importantly, as it does this, it creates a knowledge graph, which is one of the areas of feedback that we got that people really love, is the knowledge graph. I'll show you the knowledge graph quickly. I know we're running out of time, but just like any knowledge worker, inside your head, you're going to create a knowledge graph that associates things.
We gave it a very strict ontology to think about disease, drug, gene, molecule. These are the things that you as a drug developer or scientist think about. Of course, there may be things out there that aren't associated yet with your knowledge graph. You can see there are concepts like this specific gene, VAMP8. It wants to connect it to something, but it hasn't yet. That's always going to be important. It's always any part of knowledge. As it grows, you'll see the knowledge graph grow as well. After we looked at ferroptosis, we then asked Zeta to give us a design of some ferroptosis inducers. As you see, it's working on this right now. It's working on this. It's going to take a few more minutes.
It gives itself a design strategy with key benchmarks, and it's going to work to try to optimize against those benchmarks. Either it's going to get there or it's not. If it decides not to get there, it's going to ask you, Can we sacrifice this? Just like a scientist would. It's not going to give you a hallucination and get there. Again, it starts by going out and using tools. You can see which tools it's using, like ChEMBL, molecular descriptive tools, validating SMILES strings. Again, these are all tools that we've given each of these personas access to. One of the important things to think about, and I know we've got a couple of questions piled in, I'll make sure we get to the questions and then go back to sharing my screen in the presentation mode.
If I can find my Zoom screen. Give me a second. Okay. Can you guys see my screen properly?
Yeah. Yeah, we're looking at the PowerPoint.
Great. Like I said, this new generation of using these tools, and I urge you guys to definitely use the tool because you can spend hours. Again, use the code WITHZETA14 for those of you on the call, and that way you'll be able to access all the professional tools as well. Let's keep going through this. The key thing that you'll see is that in using it, you have one question that's carried end to end, not a series of disjointed questions like you'll see in a lot of other tools. You're going to have a question a drug developer would actually ask. The handoff goes through all the different personas automatically and to the levels of recursive thinking that are required. Then it's going to bias toward action.
Zeta's always going to try to get you to go to the next stage, to try to develop the drug. Also, as you see it in action, it'll actually refuse comparisons that are going to generate a false discovery. That's very important. It's also going to prioritize druggable biomarkers or things that are in high patient need and involve you in the cycle or the loop. This is an important point of Open Medicine, what differentiates it from other platforms. It generates, and proposes, and reasons, and then allows you to continue being in the decision loop. Let's talk about who's using the platform, what we've learned, which is critical. We've seen that biopharma R&D teams are asking questions largely about translational analysis, and their big feedback is they want to get ability to upload proprietary data.
Academic and translational labs are using it to build cohorts, do bioinformatic analysis, because that's, again, expensive and hard to get to. They also want to do group sharing. They also want to get alerts and new research. Service providers are some of the most extensive users. They were early adopters, actually. These are CROs, regulatory groups, bioinformatic groups, and they want to embed these co-scientists as part of their client work. They want to create project codes. They want to save into different formats and starts going into what we think of as enterprise features. Then we have independent developers, which are very exciting, that are trying to run investigations end to end, that previously, where they needed a full department. They want to upload their own documents, and some of these developers, if they're part of larger teams, they want to actually do on-premise deployment.
We'll talk a little bit about some of the service deals that we're looking at doing as well. Power users, I took a couple of quotes from conversations I've had over the last few months. We've had users that want to bring their own methods, upload some of their own models that they've been working on, but take advantage of the recursive thinking and the deployment that we have. Others that want to bring their own data. We have groups that want to do multiple scientists and agents, not just be limited to one work stream. So when they ask a bioinformatic question, they want to launch a series of bioinformatic streams in parallel, multiple approaches. Not just ferroptosis induction, but multiple different mechanisms, which we're looking at deploying in what we think of as a swarm technology.
We've had people ask about, and this is recurrent, My team needs to share and build on work. I want to see what Reed is doing or Kishor and have a work folder. We want to take it further downstream and do regulatory type work. Then we've had larger groups, including some bigger centers, ask for onsite deployment. We would love to deploy this across all of our groups or work teams that are in Europe and in Cambridge and in San Diego, but can we do it on-prem? We don't want to do it in the cloud. So we've got over 200 users, and it's growing each month. Super users are very sticky, is what we're finding. Several dozen interactions and questions every week. Most usage is in investigator mode, which gives us some thoughts in terms of how to make it more efficient.
One of the most surprising things is we've had hundreds of new molecules generated by a handful of customers. Not just ourselves, but people who are really using this for a lot of small molecule and are asking for now other modalities as well. So what are some of the discussions that we're having? Then we'll get into some of the Q&A. I know that's piling up. Biopharma co-development. A lot of patient and disease groups actually have, and we think that's a great way to get sticky new customers. We've actually had a couple of funds and financial institutions that want to use it for decision making, diligence, and want to do it in a proprietary compliant manner. We've had service providers want to integrate it and embed it into their own workflow.
And we've had academic groups that want to use it with their own proprietary literature or with their own models, or share their own libraries that they have. More and more recently, technology groups. Groups that are at the frontier models, that want to use this to power their own specialized tools or incorporate their specialized tools. Think of a unique group working on peptides or a unique group working on bispecific protein constructs that it wants to make available. Then interestingly enough, governmental and regulatory agencies, because of the rare cancer focus, we've actually had a lot of interest from groups in rare disease and rare cancer to use it for fact-checking, knowledge base, proposal review, and creation of knowledge graphs. So a lot of excitement around where this is heading.
And where it goes next, I want to talk about four specific categories that are really very important. New disease categories, which we'll do with partners and knowledge groups. Neurodegenerative, immune, and I think we'll talk about some of these in the coming weeks. We've got some deals that we're working on. Novel therapeutic modalities, trial and regulatory filings. These are new functionalities like trial benchmarking, support for FDA filings, trial design analysis cohorts, and then proactive portfolio intelligence. This is important because all this is not just knowledge synopsis and knowledge briefing, it's actually creating new knowledge, and it's creating new forms of analysis or running analysis in parallel.
And when the marginal cost of analysis approaches zero and the constraint moves to hypothesis quality, decision judgment, and time to innovation, I think that's exactly what we created Open Medicine for, is to handle the. And marginal cost is going to continue approaching zero. Then the tools and technologies that are going to win are going to be those that offer this massive improvement in time to innovation and decision judgment. So this is really important, we think, for the future. This is a great example for AI for good. Because every person on this call has the potential to be a patient at some point in the future. That's the reason why this company exists. It's the reason why these tools were built in the first place.
I think Open Medicine AI is how we make this available to everyone who wants to get a therapy to a patient. It's not just pharma companies and big pharma companies, it's clinicians, scientists, research groups, the empowered citizen scientists, the small biotech that wants to do more. We all want to get therapies and therapeutic programs across the next stage faster and with greater validation, and these are the kind of tools that'll empower that. We think this is going to be a quiet but very important revolution in AI and in medicine. With that, I'm going to turn to a few of the questions. I think we had a question from Michael King at Rodman. Is that right? Michael, I think your line is open.
I'll read it because it's in chat.
Okay.
Oh, go ahead. Go ahead, Michael.
I was just going to ask, since most everybody would be developing a pharmaceutical and biotech AI, LLMs or sourcing their data from public sources, what do you say Open Medicine differs? How does it differ from those that clients can access from other AI providers, and what would you say are the top one or two advantages that you would highlight for Open Medicine?
Yeah, I think the biggest one is harnessing insights from the disease. That is not trivial to generate. So that's a key advantage. Also, having tools that are guardrail from producing false results, and that requires a lot of disease specialization and looking at competing explanations, which we've worked through. The models that sit on top of that to enhance the answer. Again, those are things that you have to come from inside the drug development, not just gather data. Gathering data alone is not going to make you a better scientist. In fact, it'd probably create a worse scientist. It's generating knowledge from judgment, and that's what we've trained our models and Open Medicine's models to do, and that's differentiated. The next one, which is very important, which you've asked about, is its recursive ability to think and generate new knowledge that it puts back into its thinking process.
That kind of recursion is very expensive using a lot of the frontier models, and we've made it a log or two different in terms of cost, which is critical. Speed. If you compare the speed of our platform to other platforms, there's a big difference as well. I think there are a lot of technical deployment differences, and I'll let Reed expand on that, but also I think going beyond just summarizing existing literature to actually generating new and competing innovative thought and structuring that back into the ontology of reasoning is something that's unique. Reed, do you want to talk a little bit about the technology aspects as well?
Yeah, I would just mention the cost of it, first of all. We've done a lot of work to make sure if you use an open-ended agent, we've done benchmarking on this as well. These tools like OpenClaude, OpenCode, Claude Code, these types of things. Oftentimes, you will get to a half decent answer, but it's going to take a lot of time and a lot of tokens churning through that and searching through the internet, searching through just open-ended sources. Whereas largely our goal with Zeta was to, for the disease that we're particularly interested in this case, which being rare cancers, having all of that pre-indexed, pre-curated in Postgres accessible databases to the agent, and then building the computational biology tools on top of that.
Not just so that any agent can go out to the internet, download a dataset, and start doing work on it, but having all of that data right there with the methods for processing it immediately available to it. That is something that none of the other platforms right now have available to it.
Great. Thanks.
No problem. Another question, I think this is from Mac, is, How will Open Medicine AI's roadmap be funded? We are pursuing a separate round of financing at the OMAI level, and proceeds from that would obviously flow some through Lantern for the IP license and some ongoing services support and agreement. We may sell part of Lantern's stake as the company grows. We think it can be very lucrative, especially as the AI companies and AI revolution continues. We would be very open to selling the stake. It's 100% owned by Lantern today, so the first round of funding, I'm sure we'll sell a piece of that, but still be a pretty significant owner and then continue selling, as we need to. Second part, we do not have an ATM that's active right now, so I'm not sure, but that's kind of separate.
But, yes, thank you for that. Any other questions? Feel free to type them, or if you want to raise your hand, I can also I think there is a question. Go ahead. Question from Michael at Rodman.
Yeah. Hey, Panna. Thanks for taking the follow-up. The question we have over here is how does Lantern benefit from all of this other than the equity stake? Will this economically benefit Lantern as well as Open Medicine, or how do you envision the economic split?
Yeah. Again, I don't want to talk in front of some of the agreements that are being made, but we do expect to have—
Yeah
—downstream access to a revenue share, revenue split, and services to do some ongoing improvement to the models, especially in our area of expertise, which is cancer. We expect part of Open Medicine's unique business model will be that it will enter into agreements with experts in certain disease categories so that its focus can be on the deployment and engineering and making all these tools available to the entire industry. So we see revenue sharing, licensing. Then most importantly, we do see that we know we will sell our stakes as we need to. The other thing that has been very important for us is that we have a whole new generation of molecules. We have not talked about it, but we will talk about it in the coming months.
As Kishor pointed to, we have a number of very unique first and class mechanisms, and we think a lot of those will have a lot of interest in the pharma community. Of course, those belong to Lantern. We will develop them and sell them off. It took us a couple of years to take our molecules into trials and obviously to validate that they made sense. I think the next wave of our molecules will be highly compressed. We can do it faster and cheaper, and we are already down that path, and we are actually already validating a couple of these molecules, to make sure that they are actually synthesizable in a very doable way.
So making sure they are synthetically feasible, deployable, and we have got some exciting new molecules that we will be putting on into our pipeline as a result of the work with Open Medicine.
It also helps us expand on newer indications, that we would never have thought about for existing molecules, too.
Yeah. Another question, I think this is kind of in separate formats. We do expect to announce additional deals around Open Medicine. Again, our focus has largely been on getting this up and deployed and scalable, usable in the mobile setting, ready to pass scrutiny for security and all the traceability and stuff of that nature. Our next generation of focus is on larger pharma and enterprise type deals. We do have a base of users that is continuing to grow on the subscription basis. We do expect subscription prices to be increased, too, in the next quarter. Yes, we are pursuing actively now, additional sources of revenue generation with larger groups and institutions. With that, I am getting a flag that we have gone over by about 10 minutes. Any other questions? Okay. With that, I am going to stop sharing.
But again, as I mentioned, if anyone is interested in Open Medicine, we are happy to take on additional calls or conferences. We are out and beginning to have discussions with, not only institutions, but institutions that both would be users and financial partners. Larger pharma companies and technology companies that see this as a critical, unique, and differentiated position in the landscape of where medicine is going using agentic architecture. Again, we are at the forefront of this, and everything that we have built is actually deployed, scalable. This is not kind of, we are looking at deploying this.
I urge you guys to go look and use the tool to really understand how different it is and how our agents divide up these complex drug development problems and are able to give meaningful and useful answers that we think will be very valued by the pharma and drug development community.
Thank you everyone for joining me. Sorry for running 10 minutes over, but I think this was very useful. Again, I urge you guys to contact us for more questions, more discussions, or, most importantly, looking at the platform even closer. Thank you very much for your time today.