SDGR - NASDAQ

Thank you standing by. Welcome to Schrodinger's conference call to review fourth quarter and full year 2025 financial results. My name is Rob, and I'll be your operator for today's call. [Operator Instructions] After the speakers' remarks, there will be a question-and-answer session. [Operator Instructions] Please be advised that this call is being recorded at the company's request. Now I would like to introduce your host for today's conference, Ms. Jaren Madden, Chief Corporate Affairs Officer and Head of Investor Relations. Please go ahead.

Thank you, and good afternoon, everyone. Welcome to today's call during which we will provide an update on the company and review our fourth quarter and full year 2025 financial results. Earlier today, we issued a press release summarizing our financial results and progress across the company, which is available on our website at schrodinger.com. During today's call, management will make statements that are forward-looking and made pursuant to the Safe Harbor provisions of the Private Securities Litigation Reform Act of 1995, including without limitation, statements related to our outlook for the full year 2026, our medium-term objectives, our plans to accelerate the growth of our software business and advance our therapeutics portfolio, the timing of readouts from our clinical trials, the clinical potential and properties of our collaborators' compounds, the use of our cash resources, as well as our future expenses. These forward-looking statements reflect our current views about our plans, intentions, expectations, strategies and prospects, which are based on the information currently available to us and on assumptions we have made. Actual results may differ materially due to a number of important factors, including the considerations described in the Risk Factors section and elsewhere in the filings we make with the SEC, including our Form 10-K for the year ended December 31, 2025. These forward-looking statements represent our views only as of today, and we caution you that, except as required by law, we may not update them in the future whether as a result of new information, future events, or otherwise. Also included in today's call are certain non-GAAP financial measures. These non-GAAP financial measures are not prepared in accordance with generally accepted accounting principles and should be considered only in addition to, and not a substitute for, or superior to GAAP measures. Please refer to the tables at the end of our press release, which is available on our website, for reconciliations of these non-GAAP measures to the most directly comparable GAAP measures. This afternoon, Ramy Farid, our CEO, will review our recent progress and 2026 outlook. Then Richie Jain, Chief Financial Officer, will review our financial results and discuss our 2026 progress and 2028 objectives. Then Karen Akinsanya, President, Head of Therapeutics R&D and Chief Strategy Officer, Partnerships, will review our therapeutics portfolio. We'll then open the call for Q&A. With that, I will turn the call over to Ramy.

Thanks, Jaren, and thank you everyone for joining us today. Schrodinger is the leader in advancing and deploying computational methods for molecular discovery. We have developed a highly differentiated computational platform that integrates physics, AI and a scalable data infrastructure that is dramatically accelerating molecular discovery across life sciences and material science R&D globally. Our software business generated approximately $200 million in annual contract value in 2025. We also have a collaborative and internal therapeutics portfolio that leverages our computational platform at scale. We have an extensive and unmatched track record of delivering high-quality development candidates that has generated a high-value portfolio of future milestones and royalties on sales. Before highlighting our 2025 results and 2026 outlook, I would like to take a few moments to describe our vision for the future of molecular discovery. Our goal is to virtually generate all relevant molecules, computationally assay their key properties, and select the optimal molecule. We have made significant progress toward this ambitious vision, already achieving substantial reductions in the time and cost required to discover differentiated drugs and materials. This progress is powered by our unique computational engine that is an integration of the most advanced ground-truth physics and cutting-edge AI. Our platform's value is clearly validated by our high customer engagement, clinical success in our therapeutics portfolio, and the success in our research collaborations and the biotech companies we have co-founded. The AI landscape is expanding rapidly, but the fundamental challenge remains the same. AI models require massive amounts of highly accurate training data. Because chemical space is nearly infinite, experimental data alone cannot provide the scale needed, and therefore ground-truth physics simulations are often the only way to overcome this data scarcity problem. This physics-first strategy is already a proven standard in complex fields like autonomous vehicles, weather prediction, chip design, and aerospace. By leveraging this proven physics-first approach, we are pioneering the next frontier in drug and materials discovery. Schrodinger is the only company with a physics engine accurate and validated enough to power the next generation of AI for drugs and materials. We are uniquely positioned to lead the next era of molecular discovery by scaling our physics plus AI platform across a multi-billion-dollar growing market. Several industry tailwinds are accelerating demand for our technology, from deeper biology insights and a surge in available protein structures to the evolution of faster compute and agentic AI. At Schrodinger, we are leveraging these breakthroughs to provide the ground-truth calculations essential for training next-generation models. Additionally, as agentic AI drives higher utilization of high-compute calculations, our throughput-based licensing model ensures capture of that expanded consumption by our customers. Our strong performance in 2025, highlighted by 23% total revenue growth and a strong cash position of $402 million, provides significant momentum as we enter 2026. In 2025, we expanded our platform capabilities for biologics and materials, released the beta version of our predictive toxicology solution, and advanced our portfolio of proprietary and collaborative programs. We were also pleased to see molecules we co-invented advancing within our partners' clinical pipelines. Karen will discuss this in more detail shortly. Looking ahead, we are focused on achieving 10% to 15% software ACV growth, operating with expense discipline, and accelerating our transition to a primarily hosted model, which Richie will discuss shortly. We expect to continue to drive increased adoption of our platform through product innovation, addressing both our well-established and newer markets in life sciences and material science. Within our therapeutics portfolio, we plan to complete the Phase I studies for SGR-1505 and SGR-3515 while advancing our collaborative programs. Our software sales strategy is focused on scaling platform adoption and broadening our reach. Among our existing customers, we are focused on supporting them in leveraging computation at scale and have consistently observed customers increasing usage as they experience firsthand the outcomes of our predict-first approach. We are also targeting additional budgets within existing customers and unlocking opportunities in large markets such as biologics, toxicology, synthetic chemistry, drug formulations, consumer packaged goods, chemicals, energy capture and storage, and electronics. At Schrodinger, we do not just follow AI trends. We are leading the way with our gold-standard platform, where physics-based simulations provide the ground truth required for AI to navigate the endless universe of molecules with precision. We are therefore uniquely positioned to leverage AI in a way that cannot be replicated or replaced by a global AI model. We have set the standard for molecular discovery, one that delivers faster design cycles, higher success rates, lower costs and fundamentally better molecules. We have significant opportunities in the year ahead and look forward to updating you on our progress in the coming months. I will now turn the call over to Richie.

Thank you, Ramy, and good afternoon. Schrodinger had a strong 2025 delivering $256 million of revenue or 23% growth against a challenging backdrop of tight pharma budgets and challenging biotech capital markets. The software business generated $199.5 million of software revenue and $198.5 million of software ACV with strong growth from our commercial customers. The drug discovery business generated $56.4 million of revenue from our portfolio of collaborative programs. With over $400 million in cash, we have a strong balance sheet to invest in growth while targeting positive adjusted EBITDA by 2028. Our full year results demonstrate balanced growth. Software revenue increased 11% and drug discovery revenue more than doubled compared to the prior year. Software growth was primarily due to an increase in hosted and maintenance revenues, and drug discovery growth reflects the continued successful execution across our portfolio of first-in-class and best-in-class discovery collaboration programs and the continued progression of these molecules. Software gross margin was 74% compared to 80% in 2024, reflecting higher costs associated with contribution revenue from grants in 2025 compared to 2024. Total operating expenses were $310 million, a decrease of approximately 9% compared to 2024. This reflects rationalizations in R&D and G&A from our 2025 cost reduction initiatives, offset by a modest increase in sales and marketing to continue investing in long-term software growth. Total other income was a gain of $65 million compared to $24 million last year due to mark-to-market changes in our equity investments and currency fluctuations. We'd like to congratulate our partners at Structure Therapeutics, a company we co-founded, for their December announcements related to GLP-1 and their Phase I amylin program. Net loss for the year was $103 million versus a net loss of $187 million in 2024. The fully diluted share count was 73.4 million compared to 72.7 million in 2024. I will now turn briefly to our fourth quarter 2025 results. While the ACV for Q4 2025 and Q4 2024 were similar, software revenue for Q4 2025 was $69.3 million, a decrease of 13% compared to fourth quarter 2024. This is partly a function of the upfront recognition of revenue from a large multi-year on-premise deal signed in Q4 2024 compared to fourth quarter of 2025 in which portions of several multi-year deals were deployed as hosted, deferring most of the revenue recognition into 2026 and future years. This had the impact of reducing our software revenue recognition in Q4 2025, but reflects the accelerating transition of customers to hosted. The other line items for the fourth quarter reflect the same trends as described previously for full year results with a gross margin of 81% and operating expense discipline. Given our focus on driving software growth and the expected near-term volatility in reported revenue due to our accelerated transition to hosted, we are introducing a new set of software KPIs that better track our business objectives and enable the measurement of our progress. Total ACV increased to $198.5 million from $190.8 million in 2024, representing 4% overall growth. We continue expanding our top 20 pharma relationships, and ACV for this cohort grew by 15%. Commercial ACV, which includes the rest of life sciences and material science, grew 7% to $177.4 million. We continue to focus on growing existing commercial relationships with $1 million as a key threshold that demonstrates adoption of our platform at scale. Our average ACV in this cohort increased to $3.9 million from $3.3 million, or 16% growth. Within this cohort, 2 of our largest customers were acquired in 2025 by top 20 pharma customers. These acquisitions are a strong reflection of the impact of large-scale adoption of our platform. And while these acquisitions reduced our customer count by 2, their throughput and value were largely retained. From a retention point of view, we are shifting to dollar-based metrics focused on our commercial customers. Net dollar retention, which measures growth less churn from existing customers but excludes the growth from new customers, fell to 100% after several years averaging over 110%, reflecting the incredibly difficult environment for pharma and biotech in 2025 that impacted our ability to meaningfully expand relationships last year. We continue to see strong renewal performance as demonstrated by 96% gross dollar retention, which only measures churn from existing customers, underscoring the essential nature of our platform. In drug discovery, the successful expansion of our partnering activities since 2018 across 20 separate collaborators has increased the number of programs from 13 to 16 that are eligible for royalties on sales that mostly range from high-single-digits to low-double-digits. We believe there is significant embedded long-term value in the milestones and royalties associated with our portfolio that Karen will review later in the presentation. Overall, we remain pleased with our performance in 2025 looking across the composition of our ACV. As discussed previously, we achieved 7% growth across commercial customers and 15% within top 20 pharma. The rest of life sciences, including our biotech customers and government academic, reflected a well-understood challenging funding environment for 2025 that we were able to withstand. Our materials science business continues to scale up, growing from $15 million to $17 million as we introduce new capabilities. Finally, we continue to make great progress on the predictive toxicology and battery chemistry modeling initiatives that are partially funded by the Gates Foundation. Building upon several years of gradually increasing our hosted revenue mix and building out capabilities and processes to support our largest and most sophisticated customers, today we are announcing an accelerated transition to industry-standard hosted contracts that are increasingly preferred by our customers. Our business mix today is predominantly on-prem, resulting in lumpy revenue from mostly upfront recognition of contracts, in particular, multi-year contracts. Starting this year, we have begun prioritizing hosted deployments that support the continued trend toward cloud-based solutions and allows for faster deployment, enhanced renewals and licensing and support efficiencies. From a revenue recognition perspective, this will also result in more predictable revenue. Over the last several years, we have transitioned several of our largest customers from on-prem to hosted deployments, supporting their audit and service level requirements and resulting in 23% of our software revenue as hosted for 2025. The goal is to transition approximately 75% of our software revenue to hosted by 2028, factoring in that for some regions and some customers, a transition to hosted is not likely based on our current expectations. Given the accelerated transition, I will highlight the key accounting considerations for revenue recognition while noting it will have no impact on total ACV or cash flows. Hosted revenue is recognized ratably over the duration of the contract. So for deals booked later in the year or that have longer duration, this will result in reduced revenue recognition in the year the contract is booked with a corresponding increase in deferred revenue or backlog for future year revenue recognition. As a rule of thumb, we expect each 1% increase to hosted revenue percentage to result in a $2 million to $3 million reduction in current year revenue. The acceleration of this transition began in January, but because the majority of ACV is booked in Q4 of each year and our largest customers are on multiple-year agreements, many of which do not renew in 2026, we expect to see a modest increase in hosted revenue percentage for 2026 and greater acceleration for hosted revenue percentage for 2027 and 2028. We expect revenue to begin converging with ACV by 2028, but that ACV will generally be a leading indicator of revenue as the business continues to grow. Given the near-term reduction in expected revenue, we expect this will also compress gross margins and adjusted EBITDA without any impact to cost of goods sold or operating expenses. Reiterating that the accelerated transition to hosted has no impact to ACV or cash flows, we expect a more predictable financial profile as we target 75% hosted revenue by 2028. This hypothetical illustration captures the differences in revenue recognition for theoretical zero-growth $1 million annual contracts that vary in renewal quarter and duration between on-prem contracts on top and hosted contracts on the bottom. As you can see in the yellow box, $1 million in ACV will ultimately result in $1 million in recognized revenue regardless of the deployment. However, on-premise deals result in significant upfront revenue recognition with only maintenance for the remaining quarters, while hosted deals result in ratable $250,000 per quarter. This contrast becomes even more extreme for longer duration contracts, and the chart also demonstrates the revenue recognition straddling fiscal years for deals booked in Q4. Today, we are providing 2026 guidance as well as outlining our financial objectives that collectively result in a target of achieving positive adjusted EBITDA by the end of 2028. Given our accelerated transition to hosted revenue, we are providing ACV expectations rather than software revenue guidance for this year. While we will continue reporting software revenue, we believe ACV will provide important visibility into the performance of our business during a period where we expect revenue recognition to be highly variable. For full year 2026, we expect ACV to be in the range of $218 million to $228 million or 10% to 15% growth. Our expectation for Q1 is ACV of $24 million to $28 million compared to $25 million from Q1 2025, which implies $197 million to $201 million on a trailing 4-quarter basis. We anticipate drug discovery revenue between $55 million and $65 million for the year as we continue to advance our collaborative portfolio. As we have said previously, drug discovery revenue has quarterly variability due to the collaboration and milestone-driven nature of the business. Our operating expenses are expected to be less than 2025 as we fully realize the annualized impact of expense reductions and efficiencies announced in 2025 and maintain overall expense discipline. Looking over the next few years, we are targeting annual software ACV growth of 10% to 15%, substantially completing our transition to hosted contracts, and returning gross margin percentage to the high 70s. In drug discovery, we anticipate approximately $50 million of revenue annually, again, noting potential variability each year due to the collaboration and milestone-driven nature of the business. Our organization is aligned around these near-term and longer-term objectives. And our strong balance sheet with over $400 million in cash supports the path to positive adjusted EBITDA by 2028. We've taken deliberate actions to manage expenses, invest in the platform to fuel software growth, and prioritize discovery-focused therapeutics that drive our multi-year financial objectives of 10% to 15% software growth, $50 million of annual drug discovery revenue, an accelerated transition to hosted, maintaining expense discipline, and targeting positive adjusted EBITDA by 2028. We are really confident about our future and the opportunities ahead and look forward to updating you on our progress. Now I would like to hand the call over to Karen.

Thank you, Richie. Our therapeutics activities comprise 4 key value generation opportunities. Equity stakes in companies we co-founded over the past decade and a half have led to M&A transactions and cash distributions from Nimbus, Relay, Morphic, and Petra. We maintain equity positions in Nimbus, Ajax and Structure Therapeutics. Licensing of our proprietary discovery programs and collaborations with biotech and pharma companies generate upfront payments and milestones. As Richie noted, the $56 million of drug discovery revenue for 2025 reflects the progress of our collaborative programs. Potential future milestones of up to $5 billion and royalties on 16 programs, as well as future value creation from wholly-owned discovery programs represent longer-term opportunities. Our diverse ecosystem of collaborations continues to expand. This includes multi-target collaborations with pharma and biotech companies that we co-founded and were later acquired. This is resounding evidence of the impact of our platform, and successful drug discovery and financial outcomes for the molecules we co-invented. The impact of our predict-first approach to drug discovery, pursued by Schrodinger's design and discovery experts and our partners, has resulted in a growing portfolio of optimized compounds. There are more than 25 active programs across the combined portfolio. Multiple oncology drugs across the portfolio have received Fast Track and Orphan Drug Designations, and approximately $650 million in cash, upfronts and milestones have been generated to date. In addition to best-in-class compounds like Takeda's Phase III Tyk2 compound, zasocitinib, we have successfully discovered oral versions of injectable antibodies and peptides. We refer to these as modality switches, and they represent a powerful application of Schrodinger's platform and expertise in structure-based drug design. Modality switch programs represent large market opportunities, and are associated with lower clinical translation risk, given that the mechanisms have already been validated through to commercialization. Oral drugs have a lower cost of goods, greater ease of access, and importantly, enable combinations of incretins or combinations of immunomodulatory mechanisms to enhance responses and achieve durability in patients with cardiometabolic and inflammatory diseases. The portfolio of programs that are eligible for royalties continues to grow and advance, with 7 clinical programs from Phase I through Phase III. The programs span a diverse range of disease areas and many undisclosed and disclosed modality switch programs, including oral Entyvio, the small molecule alpha 4 beta 7 compound being developed by Lilly after the acquisition of Morphic, and first-in-class oral amylin being developed by Structure Therapeutics. We believe there is an increased probability of success for modality switch programs to achieve the royalties for these collaborative programs. In 2025, we were very pleased with the release of the Phase 3 data for zasocitinib, a best-in-class Tyk2 inhibitor co-invented with Nimbus, prior to its sale to Takeda for $4 billion. Takeda reported that they expect to launch the product in 2027. We are eligible to receive up to approximately $100 million in additional future cash distributions from the payments made to Nimbus based on global sales milestones. In addition to the progressing collaboration programs, we are finalizing the Phase I studies for SGR-1505 and SGR-3515. We expect to report initial Phase I data for our Wee1/Myt1 co-inhibitor at a medical meeting in the second quarter. Finally, our inflammation and immunology programs, including our best-in-class brain-penetrant and peripheral NLRP3 inhibitors and several modality switch I&I programs, continue to demonstrate promising profiles. I'll now turn the call back to Ramy.

Thank you, Karen. As you have heard, 2025 was a year of significant progress across our business. We delivered strong financial performance despite a challenging macro environment and took decisive steps to position the company for long-term success. We are confident the strategic pivot we initiated last year, combined with the 2028 targets we've shared today, puts us on a clear path to achieving our financial and operational goals. As always, I want to thank our dedicated employees for their exceptional work and accomplishments. We are energized by the momentum heading into 2026 and look forward to sharing our progress with you throughout the year. At this time, we are happy to take your questions.

[Operator Instructions] Your first question comes from the line of Mani Foroohar from Leerink Partners.

Guys, thanks for walking us through some of the impact of that transition, which is one that many software companies have to go through and does require a little remodeling work on our part. I want to hop over to think about strategically how you think about potentially the ongoing process of partnering out assets in your own pipeline, continuing to rationalize across your own pipeline programs as necessary or appropriate, and how to think about what impact that could have on potentially pulling forward the 2028 profit metric?

Mani, so with respect to partnering, as you know, this is an ongoing and active component of our business. Over the last 5 years, we have essentially done collaborations and licensed programs from our wholly-owned portfolio on a fairly regular basis, so it remains a very important part of the business. And as you know, we're constantly in conversation with potential pharma partners and others. So that will remain something that we will be very active in. As you've seen from the portfolio, there are a number of assets that are available for us to partner, both now across oncology and immunology. I'll leave it to Ramy and Richie to talk about the potential impact, but obviously, we're not guiding to any BD today. So Richie, I mean, do you have any...

Yes. Mani, just on the second part of the question, I think the goal for profitability on an adjusted EBITDA basis by 2028 is a function of growth across the software business of about 10% to 15% a year and drug discovery revenue of about $50 million a year, and then continued operating expense discipline over the 3-year time period. As we have further updates on the portfolio, we'll keep you updated on those goals as well.

That's helpful. And as a quick follow-up, when we think about the ACV guidance, which we're reaching out all the way to 2028 now on that growth rate, I recognize one of the things that could drive additional growth, whether on revenue or ACV, however one wants to think about it, is additional tools flowing into the products such as the predictive toxicology platform. Is there any value baked into that growth for monetizing the value of predictive toxicology, or is that entirely incremental? Should that become a meaningful driver, that would be incremental to the 10% to 15%?

Yes, thanks Mani for the question. We certainly expect to generate additional growth from the new products that we're releasing this year. As you mentioned, predictive tox is one of them, but there are a number of other new products that we mentioned in our prepared remarks. In particular predictive tox, we're very excited about it. The feedback so far from the beta has been very positive. In fact, it's actually outperformed our expectations. So we're definitely looking forward to that. I mean, the launch is underway, so we're excited about that and we're looking forward to realizing growth again from predictive tox as well as new products that we're releasing this year and that we will be releasing in the future.

Well, the right interpretation is that there is at least a little bit of value for incremental growth from predictive tox baked into that 10% to 15%, or any incremental value -- or any value would be incremental to 10% to 15%. Just to clarify in case I missed it.

Yes, Mani, it's Richie. As in the slides that we presented today which we will make available on our website, we listed out a number of new products that are launching. And our growth expectations over the 3-year time period include the impact of those new products.

Your next question comes from the line of Scott Schoenhaus from KeyBanc.

And yes, thanks for all that color as we make this transition in our models. I guess I want to start on the first quarter ACV versus the full year ACV on the software side, 10% -- a little bit above 10%. And I understand you're still going through the transition and that 4Q is still your big renewal season quarter. But maybe walk us through the dynamics of the first quarter ACV for this year relative to the full year, please.

Richie?

Thanks, Scott. So as you noted, Q1 does tend to be one of our smaller quarters following the Q4 season, which is customary for software companies and a reflection of our customers' budgeting cycles and our contract renewal dates. ACV, as we shift to that as our metric, it only reflects the value of the deals that we closed in the quarter, whereas revenue, where we have guided to in the past as a financial metric, reflects deals closed in the quarter but also revenue from prior quarters. So that's why we've guided to that range, thinking through those considerations. Also, given it is Q1 and it's a smaller quarter, the achievement of it is much more sensitive to each individual contract. So we do expect to grow in each quarter, but that's some of the thinking that went into the range that we set for Q1. I would reiterate that over the course of the year, inclusive of Q4, which is where the majority of the business is booked, we are expecting that $218 million to $228 million of ACV or 10% to 15% growth.

That's helpful. So any upside utilization would drive revenues above that ACV number reported in the first quarter. How are your customers dealing with this transition? We've always talked about in the past that you're never forcing this model, that customers sometimes preferred this, but now it's a clear adoption of a hosted platform. Can you just give us how your customers are engaging with this or reacting to this strategy now?

Yes. Hosted cloud-based solutions are an industry standard and allow us to deploy faster and also support in an enhanced fashion. Customers are actually increasingly preferring hosted deployments, and we've had a track record over the last several years of being able to support our largest and most sophisticated customers with hosted deployments. From an investor point of view, 23% of software revenue today is hosted, which we've been gradually increasing over the past several years. Because the revenue is ratable versus upfront or on-prem contracts, this will also result in a smoother and more predictable revenue profile as we target 75% by 2028.

Your next question comes from the line of Michael Ryskin from Bank of America.

This is Alex on for Mike. My first question is on AI. So there's been the focus of AI playing a bigger and bigger role in R&D processes and Schrodinger fits neatly into that. Have you noticed any changes in your conversations with pharma customers in recent months regarding this? Are they open to engaging and leveraging your solutions to drive efficiencies and become more computational? So in other words, is Schrodinger an AI winner in the R&D space? Or is pharma shifting funds and focus away from these traditional methods and prioritizing new solutions pioneered by maybe Anthropic or OpenAI? And then I have a follow-up question.

Yes. Thanks for the question. Yes, we view AI and the sort of new revolution of agentic AI absolutely as a tailwind. It's very clear that, that is -- the adoption of AI and the scale-up that it results in is actually increasing the demand for our software. The fact that we license our software using a throughput-based licensing model obviously means that we benefit from that scale-up. The other thing that we've said, and you've heard us talk about this for years, is that one of the barriers to adoption of our platform at scale is know-how and just availability of humans to run the software at scale, and obviously, agentic AI solutions address that. We've actually -- we're working with Anthropic directly to explore ways to integrate agentic AI with our computational solutions. And we're -- and this is now implied in your question, we're actually excited about the broader adoption of these types of methods, again, because it increases the demand for our technology, not the other way around.

Great. Thank you. That's helpful. And my second question is about -- you talked about 2 largest customers being acquired. Can these customers kind of disappear in terms of usage over time? Or could this be a beachhead in terms that maybe it'll lead to more major pharmas like, learning about Schrodinger and helping to grow the business with pharma acquirers?

Yes. We -- despite the fact that it reduces the number by 2 on a KPI slide, this is actually a great fact pattern for us. It's a recognition of the predict-first approach and using computation at scale in the molecules that they're able to develop and resulting in M&A traits. It's no surprise that the companies that acquire them also have the same approach. So it is -- actually, we view it to be, despite the interim loss of a customer, we were able to retain the throughput and the relationship there and view it as a positive long-term signal.

Your next question comes from the line of Brendan Smith from TD Cowen.

I wanted to ask just a little bit more actually about your go-to-market strategy for this predictive tox launch this year. I hear you on the 10% to 15% kind of blended growth, including existing and upcoming product launches. But just curious if you're thinking that predictive tox will largely be an add-on within existing customers or if it will honestly require new touch points at some of those accounts or even if reps will largely come from exposure to brand new customer accounts altogether? Just kind of wondering how you're thinking about that initial sales outreach and potential impact on SG&A through that lens.

Yes, it's a very good question. Actually, it's both. One of the use cases for predictive tox is using it very early in discovery projects and to the extent that that is the case, we expect existing customers to acquire the technology. Of course, it is an add-on, so that requires them to pay more for it. And by the way, that's also a throughput-based hosted web service, which is the way it's being delivered. So from that respect, we'll see growth that way. But it also will be used by researchers later in discovery and preclinical development by toxicology groups who we currently do not -- or we don't -- they're not currently among our customers. So it will result in growth from tapping into those new budgets. So it's really both.

And Brendan, I'll just add quickly to that. A number of the new products that we're launching fit the same profile, which is they're reaching new end customers, new touch points within existing customers. This obviously will have the impact of adding to our total addressable market. You had touched upon how we operationalize this in your question; we have made some investments in sales and marketing. You'll notice it in our 2025 results to account for this and be able to execute against getting these additional products into the hands of customers.

Great. And if I could just stick with a quick follow-on, just when you're talking about the transition to hosted contracts and flagging the impact on margins, if we maybe just zoom out a bit from just the software margins, how should we think about the concurrent impact of that transition against overall blended margins kind of versus the backdrop of winding down the internal pipeline and presumably some of the OpEx savings you could see from that?

Yes, there's -- I'll tackle that. There's kind of 2 elements of the question. I just want to reiterate that the transition to hosted has no impact on ACV or cash flows. So while we expect there to be some interim variability in revenue because of the accounting recognition, and that will have an impact on margins, in particular with gross margins and adjusted EBITDA. The actual cost of goods sold and the actual operating expenses are -- have not changed at all by this impact. So we did want to give a longer-term view of gross margins returning to the high 70s, which is where we've been prior to the last few years when we took on the predictive toxicology grant that temporarily lowered gross margins and also give you a view of the adjusted EBITDA opportunity for 2028. When we're looking at it from that lens, it's not only the growth in the software business and the transition to hosted, but it is also reflecting some of the operating expense reductions and efficiencies across the whole business, including the software business, the therapeutics business, and the overall enterprise that gets you to that adjusted EBITDA profitability benchmark out in 3 years.

Your next question comes from the line of Matt Hewitt from Craig Hallum.

This is Toph on for Matt. So just getting an idea on your 2028 goal of adjusted EBITDA break-even. What assumptions do you have baked in there about a biotech rebound? And then, do you consider for capital allocation any buybacks?

I'll take the second one first. We always think about capital allocation. We're very pleased to have the balance sheet to support our growth opportunity and also support our pathway to adjusted EBITDA. We -- from a buyback point of view - I mean certainly, the stock price is not one that we think reflects our intrinsic value, but we see a long growth pathway ahead of us and would rather invest our cash into growth in the business, as opposed to buying back shares. In terms of the 3-year outlook on growth, biotech has been a challenge obviously for the last few years. We are assuming over that 3-year window to see a recovery there to normalize levels and growth within that segment.

Your next question comes from the line of Michael Yee from UBS Financial.

This is Kyle Yang for Michael Yee from UBS. So you guided software ACV to $218 million to $228 million in 2026 or roughly 9% to 14% year-over-year growth. It would be helpful if you could remind us and the Street, how do you define ACV? And how does it differ from reported revenue? Help us understand how this ACV guidance could potentially translate to reported software revenue in 2026. And so given the transition from on-prem to hosted, and that's -- we would assume that's not going to be fully completed in 2026, would you expect this reported revenue could come higher than software ACV guidance? And finally, just help us understand the puts and takes that would ultimately determine whether your revenue could land towards the lower or upper end of the revenue guidance range?

Okay. There's a lot of questions in there that I'm going to try to take one at a time. So ACV, annual contract value, it is reflecting the value of a contract. If it's a 1-year deal and we book it in Q1, for example, the ACV will be fully reflected in Q1. If it is a multiple-year deal, we will reflect the annual bookings of that deal for each year of the deal. So that is the definition of ACV. ACV and revenue are actually quite different. We will post our deck to the website, but if you look at Slide 19, we have a pretty detailed visual explanation of the difference. You can see from a revenue recognition point of view in on-prem deals, there is a significant acceleration in the quarter the deal is booked, and then very little revenue in the quarters that follow. Whereas in hosted revenue deals, it's ratable across the 4 quarters. So you had asked about expectations for 2026. We do not expect that revenue will be greater than ACV this year, given that we are very busy at work trying to transition our contracts over to hosted. For all the deals that renew this year, and for all new customers, our goal is to move them over to hosted. Our goal over the 3-year time period is to get to 75%, not 100%, because there are customers in certain regions and certain sectors where they have not been as aggressive in embracing cloud-based solutions, and so we don't think that they will convert over to hosted. But given that the majority of our ACV is booked in Q4, and we intend to transition those over to hosted, it will result in reduced revenue recognition for that quarter, but you will see an increase in deferred revenue and backlog to capture the amount of revenue that will be recognized in the following year. We actually have not guided to revenue for this year, so I don't know how to respond to that comment, but we are guiding to ACV for the year of $218 million to $228 million.

And just to correct some math, probably there's a rounding problem, that corresponds to 10% to 15% growth not 9% to 14%, I think as you stated in your question. I just want to add one more thing just because it's sort of a math thing. If we are successful in transitioning our customers to hosted on this path to getting to 75% transitioned by 2028, mathematically the revenue has to be lower in '26 than it was in '25. That's just a mathematical consequence and again that table reflects that. I hope that's clear.

Great. And to the last question, could you please help us understand what could be some key puts and takes that could determine whether your revenue is going to land toward the lower or upper end of the guidance?

Okay. So just to reiterate, we are not providing revenue guidance for this year. We are providing ACV guidance. In terms of our ACV outlook for the year, it is based on expanding relationships within our existing customers and embracing the AI workflows and the demand that is generated out of that for our platform. It is based on rolling out new products that address different workflows for our customers and different budgets. And it's based on expansion within our material science business. As a rule of thumb, which may help address your question, we are at 23% hosted software as a percentage of software revenue. In any given year, as we increase that percentage, each 1% increase will roughly result in about a $2 million to $3 million reduction in revenue for that year. So as we've put out a 3-year view to getting to 75% hosted revenue percent, we will make progress towards that. I expect in 2026 the progress will be rather limited on that basis because of the amount of business that's booked in Q4. But that should help give you a rough guide on a way to measure our progress. But the -- what this is revealing is why we have focused the guidance this year on ACV, which is an operating metric, and hosted revenue percentage; we think those 2 metrics will help evaluate our performance this year.

Your next question comes from the line of Evan Seigerman from BMO Capital Markets.

This is Conor MacKay on for Evan. We found some of the color that you provided on Slide 16 as it relates to ACV composition pretty helpful. We're just wondering if maybe you can share more on how you're thinking about the evolution of customer split as you look towards your annual software ACV growth goal of 10% to 15% by 2028. Maybe how might a gradually improving biotech funding environment and the addition of some of your newer products impact your customer split?

Yes. I'll take a crack at that. I think you're sort of asking about ACV growth of each of those segments, and we're not guiding to that. So -- I mean, as Richie just said, we do expect to see some recovery in biotech. We're seeing tailwinds from a lot -- bigger adoption of agentic AI. We have new products that are being released this year that we expect to see revenue from that should impact all of those segments, both in life sciences and in material science. So hopefully that answers your question without getting into details of each individual segment. Richie, do you want to add anything to that?

No, that's perfect.

[Operator Instructions] Your next question comes from the line of Sean Laaman from Morgan Stanley.

This is [ Morgan ] on for Sean. I had a question related to the last one, but more so dialing into the $1 million and above customers. So we saw this year that the ACV on average went from $3.6 million (sic) [ $3.3 million ] to $3.9 million. What were some of the key actions or key measures that resulted in that increase and what can you do to continue to increase that average ACV exponentially at this point?

Yes. So within that customer cohort, we think that $1 million threshold we think is a reflection of a customer adopting our technology at scale and embracing a predict-first approach. We grew that segment -- we grew customers within that segment from an average relationship of $3.3 million to $3.9 million, or 16% growth. And this customer cohort, you can say, includes a lot of our top 20 pharma customers, as well as a handful of biotechs that are adopting the technology at scale. And so within those 2 groups, top 20 pharma customers have been our longest-standing customers, that's where we are introducing new products and growing relationships, as well as increasing the adoption, closing some of the adoption gaps between our largest customers in top 20 pharma and our smallest customers in top 20 pharma. Outside of top 20 pharma, there's a handful of customers that have embraced this predict-first approach. Fortunately or unfortunately, 2 of them got acquired last year, but there are many companies that understand the power of computation and are fully embracing it across their organizations to predict fully optimized molecules.

Another encouraging thing in that cohort is, actually, there's a fair amount of variance within the cohort of customers spending over $1 million. So that's an exciting opportunity that even within that group there's still significant potential for the customers that, obviously there are ones that are spending under $3.9 million, otherwise that wouldn't be the average to spend significantly more, which there already are customers in that cohort that are. So we're encouraged by that as well. I hope that makes sense.