Good morning, everyone. I am so happy to be joined by the management team of Relay Therapeutics on the last day of the Morgan Stanley 24th Annual Global Healthcare Conference. I am joined here by President and CEO Sanjiv Patel, President of R&D Don Bergstrom, and Chief Corporate Development Officer Peter Rahmer. I will just read a quick disclaimer. Please visit morganstanley.com/researchdisclosures for our research disclaimer. Let us dive right into the Relay story. Maybe we can start with you, Sanjiv, and what has been 2026 has been a very eventful time for Relay with data disclosures across both breast cancer and vascular anomalies. What has been the company's focus over the past few months? First of all, thank you for the invitation. Thank you for all of you attending, both online and in person today. 2026 has been a very eventful year for us, exactly as you said. We have been pushing forward our lead program, zovegalisib, which is a PI3Kα mutant selective inhibitor, across three different indications. The first of those is second-line hormone receptor-positive, HER2-negative breast cancer, where in March this year, we showed data using our pivotal dose of 400 mg BID that showed a PFS in this population of 11 months. This compares very favorably with the current standard of care, capivasertib. Our heads down now executing a pivotal trial across the world, and we are very happy with how that is going. Later this year, we will share an update on when we believe the last patient in will be on that trial. In our first line hormone receptor-positive, HER2-negative breast cancer efforts, we declared in April of this year that we will use a regimen of zovegalisib with Pfizer's selective CDK4. This is a really important mechanism because using a selective approach, we believe we can minimize the side effects and maximize the tolerability of this regimen, which has been the challenge in this field. We know that adding a PI3K pathway inhibitor matters because we have seen that from the INAVO data. The challenge has really been the lack of tolerability and the lack of ability of these patients to take this regimen for up to three to four years. Using this mutant selective and CDK4 selective approach, we feel very comfortable that we will be able to provide the regimen that patients actually need. In May of this year, we showed data in vascular anomalies where we showed a very robust volumetric response rate of 60% across the doses that we were using. That compares, again, very favorably to the current standard of care, alpelisib. There, we are very much focused on the dose expansion portion of our study, establishing a dose, and later this year, we will do both a data update and a regulatory update on how we believe that we can get this therapy to patients rapidly. It has been a very eventful year, but it is still an eventful year to come. Absolutely. You mentioned that in the second-line breast cancer, you had an update from Zovega earlier this year. Maybe you could talk a little bit about what that data actually showed, and you referenced sort of the competition with the standard of care, capivasertib. Yeah. Maybe you can kind of profile that a little bit for us in terms of what you demonstrated. Yeah. This is a very large market, the second-line hormone receptor-positive, HER2-negative breast cancer market. The standard of care has evolved over the last few years. Alpelisib, which is a non-selective PI3K inhibitor, was approved a few years ago. Unfortunately, commercially was not that successful because although it showed a PFS in the kind of seven to eight month range, the tolerability challenges of hyperglycemia, diarrhea, rash, stomatitis, unfortunately mean that patients just cannot stay on this therapy. Although it's efficacious, it's unfortunately not tolerable. Then we saw the emergence of the AKT inhibitor from AstraZeneca, capivasertib. Again, this showed actually numerically a lower PFS than alpelisib but was perceived to be more tolerable in that it had in its label lower rates of hyperglycemia. But it did have still significant rash. This has been launched a couple of years ago, and it's been commercially very successful. It's kind of moving towards a billion dollars of run rate in its sales. It shows that if you could have what is perceived to be a more tolerable regimen, you can definitely be commercially successful to the tune of a billion dollars here. The data that we showed earlier this year in this patient population was in second-line, second-line plus patients using our 400 mg BID dose of Zovega. The premise of this program was if you can have a mutant selective inhibitor in this field, you can dial down the off-target toxicities of diarrhea, rash, hyperglycemia, stomatitis. That's exactly what we showed, which is low rates of all of these. That means that you can keep patients on therapy with a greater dose intensity, and that should translate into greater efficacy. That is exactly what the data showed. We showed 11 months of PFS in this patient population, and obviously that compares very favorably to the 5.5 months of capivasertib. We feel pretty confident now as we run this head-to-head pivotal trial against capivasertib that we could be successful in this patient population and create a very commercially meaningful therapy for patients. Sanjiv, can you talk a little bit more about the phase III trial design? What are the key timelines, and then how you think about zovegalisib ultimately fitting into the evolving second-line breast cancer landscape? Maybe I will hand that one over to Don to talk about the trial design and maybe then Peter can talk about the landscape. Yeah. It is a one to one randomized trial of fulvestrant plus zovegalisib versus fulvestrant plus capivasertib. The patients who we are enrolling have all previously been treated with endocrine therapy and a CDK4/6 inhibitor. The criteria for enrollment are a little bit tighter than what we have used in our phase I program to date. For example, we are not allowing patients to have multiple prior CDK4/6 inhibitors or an ADC. We are looking to enroll as close as we can to a true second-line patient population. The trial is powered to show clinically meaningful improvement in PFS. We also have OS as a key secondary endpoint. As Sanjiv mentioned, we opened the trial in the middle of last year, mid-2025. We are very encouraged with the way enrollment is going. Trials of similar size and a similar population have traditionally enrolled over the course of 30- 36 months. We just saw for one in second-line patients, this is the oral third trial that enrolled in 33 months. As Sanjiv mentioned, we will have an update at the end of the year on when we think we'll reach full enrollment. On the landscape in second-line PIK3CA-mutated patients, there are two pathway inhibitors that have full approval. We have mentioned both of them already. Alpelisib, marketed as Piqray, capivasertib, marketed as TRUQAP. The most recent data from both the assets that we have publicly puts their PFS in the same ballpark of about 5.5 months, six months. We feel very confident that with the data we have in hand today that Sanjiv has talked through, which is in almost the median third-line patients, that when we move into this phase III trial in more true second-line patients, we will have the ability to see at least close to a maintaining of the PFS we have seen so far in phase I/II now across 120 patients. As mentioned, the commercial market today across the class is probably in the neighborhood of about a $1 billion run rate and still growing, which is quite encouraging. If we can come with a next-generation profile, which we have in hand today, and prove that out in a phase III study, we do believe we should be able to rapidly take some of that market share and continue to grow the market. Just in terms of trial execution, we are very happy with how things are going. We started the trial in mid-2025. Precedent trial for full enrollment takes somewhere between 30 and 36 months. We just saw a recent benchmark for a similar size trial in this field taking close to 33 months. Our goal here is to go as fast as we possibly can, knowing that these are the benchmarks, and try to beat some of them if we can. Okay, great. Thanks to all three of you for laying that out. I want to turn the attention to the first-line opportunity now, and maybe you can spend a moment on the first-line landscape and where do you see the remaining unmet need there? Over to you Don. Yeah. In the current landscape in frontline trial, current standard of care for most patients is the combination of endocrine therapy plus a CDK4/6 inhibitor. What we've seen in retrospective analyses that have been run in a number of the trials testing these frontline regimens, is that patients with PI3K mutations tend to have a shorter PFS than the patients who are enrolled who have wild-type PIK3CA. In the case of ribociclib, the PFS that was observed for PIK3CA mutant patients in an endocrine-sensitive population was 19 months versus 31 months for PI3K wild-type population. About a year difference in PFS. We feel that the current CDK4/6 doublets can be improved upon, especially in PIK3CA mutant patients. I think that hypothesis was borne out in the results we saw for the INAVO120 trial of inavolisib, Roche's non-selective PI3K inhibitor in an endocrine-resistant subpopulation. These are patients who had shorter treatment-free interval between their adjuvant therapy and recurrence of metastatic disease. In that patient population, we saw inavolisib be able, in combination with fulvestrant and palbociclib, to deliver on both PFS and OS in a randomized phase III trial leading to approval. Now, the challenge here is that that was a very heavily selected patient population, both in terms of being an endocrine-resistant patient population, and the patients had to be very, very metabolically fit because of the hyperglycemia profile of inavolisib. That has translated into limited utilization after the launch, just because the physicians that we talk to have a real challenge finding the appropriate patient who's going to be able to handle this regimen. The toxicity burden of the regimen when you consider the hyperglycemia, you consider rash, diarrhea, it's a heavy toxicity burden. Patients are taking multiple concomitant medications to actually manage the AEs that are associated with the drug. It is being used sparingly in very fit patients. Our hypothesis is we know that PIK3CA is a negative prognostic marker in these patients. We know inhibiting mutant PI3Kα can provide additional benefit, and we want to develop a regimen that has the tolerability profile, as Sanjiv mentioned, that will be able to be used broadly in a general patient population, and patients will be able to tolerate it and stay on therapy for three years or longer. Along those lines, we have really focused on atirmociclib and the triplet of endocrine therapy plus atirmociclib plus zovegalisib as providing that tolerability profile, as we showed in the data disclosure from earlier this year. We are focusing right now on the design of a randomized trial in endocrine-sensitive patients, so the more common patient population, where we would be testing the triplet of endocrine therapy, in this case, an aromatase inhibitor plus atirmociclib plus zovegalisib, versus endocrine therapy, aromatase inhibitor plus a CDK4/6 of an investigator's choice. Don, can you talk a little bit more about the decision to go in the endocrine-sensitive population? What is driving that? It sounds like a bigger population, bigger portion of patients. Yeah. It is commercially the larger opportunity. It is a place where we have the evidence that shows that the PIK3CA mutant patients do not derive full benefit from current CDK4/6-targeted therapies, so there is clear unmet need there as well. We are also, in the ongoing ReDiscover-2 trial, testing zovegalisib in patients who are post CDK4/6 inhibitor. With the endocrine-sensitive approach in the frontline, and then with the ongoing ReDiscover-2 trial in the second line, we envision that we could have labels that will capture the vast majority of the metastatic PIK3CA mutant breast cancer population. Okay. How do you think physicians will decide between using a PI3Kα alpha-targeting agent in first line versus second line, and when to engage that opportunity? Yeah. I think that will be the option that treating physicians will have with these two trials that we're running. I think, again, it's going to come down to the profile that we're building in our frontline trial and the opportunity to have this selective regimen for atirmociclib and zovegalisib that we feel the overall toxicity burden for patients with the selective approach, with the data we've generated so far, is not meaningfully worse than the existing CDK4/6 doublet therapies with current generation CDK4/6 inhibitors. I think, basically, the proposition we'll have is that you can have a meaningfully more efficacious frontline therapy with a triplet, with a safety profile that doesn't bring additional burden on patients, will be a very attractive option for physicians. There could be cases where you have patients who are maybe older or frailer, where the decision is made to go with a potentially milder therapy in the frontline and reserve zovegalisib for second-line patients. I think, again, what we're looking to do with this development program is really to be able to provide that optionality and to be able to have labels where we can capture patients where we think that the vast majority of patients with the profile we have would be good candidates for a triplet in frontline. Then for the population who maybe wouldn't be candidates for frontline triplets, there's the option to derive benefit from zovegalisib in later lines, too. When you think about the triplet with the Pfizer drug, what gives you confidence that the drug-drug interaction with the atirmociclib will not be an issue in this study? Yeah. I think there, we have a good mechanistic understanding of what's happening. What we have seen is that when we co-administer zovegalisib with atirmociclib actually increases the absorption of zovegalisib. We get the same blood exposure to zovegalisib in combination with atirmociclib, but with a lower dose than what we have in the doublet. The doses that we are focusing on that we reported out earlier this year was a 100 mg BID dose and 150 mg BID dose in triplet, compared to 400 mg BID in the doublet. This effect is consistent across patients. The inter-patient variability is the same in triplet as we see in doublet. If anything, it is maybe a little bit of a benefit in the sense that it has reduced our dose, and we are not concerned about there being new sources of variability or other complications that could make development more challenging. Just on the point around atirmociclib, is there any concern around whether atirmociclib is approved or has a positive phase III in its own first-line study, the FourLight-3 study? Yeah, no, I do not think we have a concern. I think what we are looking for in atirmociclib is a tolerable safety profile. The whole premise here is to stack a mutant selective with a CDK4 selective inhibitor to try and minimize the Adverse Events. I think what we have seen in the data so far that is in the public domain is exactly that from atirmociclib, and that is what we want to use as our foundation. The FourLight-3 trial is a trial that is looking at superiority in terms of efficacy versus the standard of care CDK4/6. In a way, whether it is or it is not is irrelevant to what we are looking for. We are looking for a tolerable foundation to be the backbone of our regimen. Our focus is on, is atirmociclib a more tolerable mechanism than the CDK4/6? I think we feel very confident of that. It's kind of irrespective around whether FourLight-3 is successful or not. Okay. Just looking ahead to next year, you said publicly you'll actually start the phase III and the first line early 2027, pending regulatory feedback. Anything you'd contextualize for investors around the regulatory interaction that may be upcoming for that program or on just the plan for the phase III next year? No, it's just all the dull stuff that goes into making and developing a new medicine. We're out, obviously, getting all the sites and selection, CROs, and dose, and putting everything in place and going to the FDA. We don't believe that there's anything specific inside of that than just getting everything clarified. Obviously, we look to start this trial as rapidly as we possibly can. Okay, great. Good luck. Thanks. I want to turn to vascular anomalies and the market opportunity here. How would you size the opportunity, and what subpopulations are most useful to consider? Let me hand that to Peter. Yeah. What we've said is the current total addressable market, we believe, is $6 billion-$8 billion. What is behind that is in the U.S., there's about 170,000 of these patients with PIK3CA mutations. We don't believe all of those would need chronic systemic therapy to address their disease, so we've done some market research and talked to many investigators and physicians that treat these patients. What that funnels down into is within the subsets that we're currently testing, treating in the trial, which is PIK3CA-related overgrowth spectrum, that's 100% PIK3CA mutated. Lymphatic malformations is 80+% PIK3CA mutated, and then venous malformations, about 20%-30% PIK3CA mutated. Across those three subtypes, we think there's about 25,000 addressable patients that could be amenable to chronic systemic therapy to address their disease. We only need to capture a fraction of that to have a very large opportunity for ourselves. To put that in context, about every 2,500-3,000 patients, if you assume the current Vijoice pricings of the only drug that has accelerated approval at $36,000 a month, every 2,500-3,000 patients would get you a billion dollars of sales. We don't need to capture much of that 25,000 to be able to have a blockbuster product here. We do think we're generating the profile that will allow us to capture a good amount of that 25,000, maybe over time, able to grow that 25,000. I think where a lot of our confidence stems from is that it doesn't take a lot of patients to have a really meaningful opportunity here, and the early data we've shown has been dramatically better than what is out there for patients today. Yeah, let's go further into that, the data that you showed earlier this year. Maybe you can contextualize what you actually demonstrated versus the standard of care drugs. Yeah. So we brought zovegalisib into the clinic in vascular anomalies in the spring of 2025, and in just about a year of time of being in the clinic, we are able to show data that is demonstrating superiority on a cross-trial comparison basis to alpelisib. alpelisib has accelerated approval just in PROS today, so the small subset of PROS, and their data has shown somewhere between a 20%-30% volumetric response rate. That's the regulatory endpoint here is volumetric response rate, and a patient is considered a response if their target lesion is shrank by 20% or greater. So in the case of alpelisib, their data has shown roughly 20%-30% volumetric response rate. Our initial data across 20 patients demonstrated a 60% volumetric response rate, so a doubling or tripling of what alpelisib has been able to demonstrate. Most importantly, that comes with a much better tolerability profile than what we know about alpelisib today. I think these early data exceeded even our expectations, and around the time of the disclosure in May, we opened up expansion cohorts into 300 mg BID and 400 mg once daily. We've been prioritizing enrollment into the 400 mg once daily dose. We think moving to a once daily dose would be quite a good convenience advantage for these patients that would be on, intended to be on this drug chronically. So, very encouraged by the early data. We've guided to showing additional data before the end of the year, in addition to a regulatory update. Remain on track to be able to do so, and excited to see how fast we can get zovegalisib to the vascular anomalies patients. Okay, great. Maybe just a little bit more about the ReInspire trial that you're running, and you said there'll be an update later this year in vascular anomalies, but maybe talk a little bit more about the trial design and just the rationale here, what you're hoping to show and what would be sort of viewed as successful. Yeah. The study, because of our experience in oncology patients, we are able to go into dose randomization to do our dose finding. So we tested three doses in parallel, 100 mg BID, 300 mg BID, and then the top dose is 400 mg BID, which is our current second-line oncology dose. We enrolled 32 patients into that dose randomized portion of the study, and that was focused on patients that are 12 years and older. At the time that we disclosed the data at ISSVA in May, 20 of those patients were efficacy evaluable. These patients get MRI to evaluate efficacy every 12 weeks, and so by efficacy evaluable, it just means there's only 20 that had made it out to that 12-week point at the time in which we disclosed the data at ISSVA. By the time we make a disclosure before the end of the year, all 32 will now be efficacy evaluable, and we'll probably even have a good number of those patients that have made it out past multiple scans. So it'll be within that patient population, a bit more median follow-up than what we had at ISSVA. So that was about four months at ISSVA, so wherever we do the data cut, it'll be that much more median follow-up. So that'll be an encouraging early look at more mature data in this setting. In addition to that, we will show our early expansion data, but that'll be smaller N and certainly less follow-up. Nonetheless, with the data we've seen at 100 mg BID and 300 mg BID, we like the idea of what 400 mg QDay could look like in the context of those two experiences. It might be the optimal dose to bring forward, but we just have to enroll some more patients there and get some more experience before ultimately calling that a dose. Just thinking a little bit further ahead about the regulatory path for zovegalisib, what can you take from alpelisib's regulatory path, and how might it compare for zovegalisib? Yeah. To remind everybody, alpelisib had a very unusual accelerated approval. It was approved off of 37 patients that were treated under compassionate use, and it was a retrospective chart review of those patients. Not a traditional clinical trial by any means. But it really does speak to the level of unmet medical need in these patients that the FDA would allow such data to warrant accelerated approval there. That was 37 patients able to get accelerated approval. Alpelisib then ran a confirmatory study that failed, and they now went back to the drawing board to run another confirmatory study that started in the fall of last year. The second confirmatory study they are now running is a single-arm study, 104 patients. Again, that is for full approval. You have accelerated approval with 37 patients, an ongoing confirmatory study presumably negotiated with the FDA as being appropriate for full approval, single arm, 104 patients. As we think about our regulatory path forward, you have a high unmet medical need patient population with a very severe disease in the moderate to severe patients, and you have no medicines with full approval in those settings. We do believe that accelerated approval should be available to us, and that will be the nature of the discussion we go have with the agency before the end of the year. What we will propose is pooling of PROS and lymphatic malformation patients. So far, our data is showing good consistency across those populations. The disease biology is quite similar there, so we will propose pooling those two together, and pursuing accelerated approval inside the ongoing ReInspire study. The question really will be what is the requisite N to warrant accelerated approval? What would a confirmatory study look like if we were able to have accelerated approval? Again, we would propose for full approval, we just continue to enroll in a single-arm context inside of ReInspire, enroll more patients and just have more follow-up, and that would be what we propose to be the full approval study. But this is all subject to discussion with the FDA. It may be the decision is to split out PROS and LMs, which would be a fine outcome too. Our goal is just to get this as rapidly to patients as we can, to get it on the market as fast as possible. That is really helpful context, Peter. Maybe just one last question on zovegalisib before we look more organizationally. What are the pricing considerations for vascular anomalies versus breast cancer for the drug? Yeah. We had some precedence here, which is helpful. Alpelisib, it is approved in both settings, as I mentioned. Marketed as Piqray in breast cancer, and in breast cancer it is priced at $25,000 a month. It is marketed as Vijoice in vascular anomalies, priced at $36,000 a month. Obviously, a bit premature for us to talk about pricing of zovegalisib, but suffice to say, it is a good precedence to see that you could have a very similar situation that we will find ourselves in, which is the same base API and two separate brands, and able to have differentiated pricing inside of these two very different indications. Okay. Maybe just last question on your pipeline. I do not want to gloss over the NRAS program, RLY-8161. Maybe you can just give a quick summary of that program and when we might expect to see some data. Yeah. NRAS is a very strong oncogenic driver. You see mutations most frequently in melanoma, colorectal cancer, non-small cell lung, thyroid, and some other malignancies. There are anecdotes of patients with NRAS-driven tumors benefiting from RAS inhibitors. Although, they do not get complete suppression of the pathway, and there is the toxicities associated with RAS inhibition. Historically, people have tried combinations of downstream signaling nodes, RAF inhibitors combined with MEK inhibitors, which have had, I would say, moderate efficacy, low to moderate efficacy, and high rates of toxicity. But everything we know preclinically about NRAS suggests it is a very strong driver oncogene, and everything we know about NRAS in terms of tolerability would suggest that it is largely dispensable in a postnatal organism. We think if you could have an NRAS selective inhibitor, you could see very powerful efficacy in NRAS-driven tumors, potentially with very clean safety profile. And we were able to use our platform to discover a novel pocket in NRAS, then to exploit that to be able to make an NRAS selective inhibitor. Very active as a single agent across preclinical models with melanoma, lung cancer, and colorectal cancer. And in non-clinical toxicology studies, almost no toxicity. The hypotheses have borne out through the preclinical testing. We brought 8161 into the clinic earlier this year, and we are now testing the hypothesis in humans. Okay, great. Just on sort of your cash balance, I know you completed a raise and now have $911 million of cash at the end of Q2, so you are well-funded into 2029. How are you thinking about just managing cash burn over the next several quarters, particularly as you get into later-stage trials and managing multiple different programs? I mean, as any company that has lived through the last year in biotech, we manage it very tightly. $911 million is precious, and we will focus the majority of that, obviously, on executing against the three registrational trials that we are focused on running, so the first-line, second-line, and the VAs. Around that, obviously, we start to think about preparing for commercialization and being ready to commercialize as rapidly as possible, as soon as we get the approvals for all three of those indications. Behind that, obviously, over the last few years, we have significantly reduced our research footprint, but it remains very productive, but it is obviously a much smaller part of our burn than it was back in 2021, 2022. As you know, with $911 million, we have got plenty of levers to move up and down as the programs move at different paces. We are confident that we can execute against all three initiatives. In the light of this cash balance, we think we can deliver top-line registrational data on the second-line trial and the VAs trial and be significantly through the frontline trial. We think there is lots of value we can create in this cash window. Maybe just as we wrap up, Sanjiv. Yeah. anything you want to leave investors with as you think about the next 12 months, what you're most excited for and what they should be focused on? Look, I think we sit now having kind of de-risked a lot of the questions that were sitting out there at the beginning of the year. Is there a meaningful opportunity to improve the standard of care in the second line? We believe there is, and we've shown data against that. Now we're going to rapidly executing towards getting that therapy to patients. On the other side, the selective approach, I think has a potential to be very large in a very big market. Then obviously opening up an entirely new therapeutic area in vascular anomalies. I think we've done a lot to educate the investor community around that. So I think that one of the few companies out there that sits with multiple large opportunities with a single asset that's now de-risked, and now it's all about execution. Obviously we have a significant cash balance against that. We think that in the life of the next 12 months, we have significant catalysts to generate value for investors. That's great. Sanjiv, Don, Peter thank you for joining us. We're really glad to have you here, and thanks for all your thoughtful answers, and we're wishing you luck. Thank you. Thank you to the Morgan Stanley team. Thank you for those of you in the presentation today. Thank you.