MindWalk Holdings Corp. (HYFT) Q4 2026 Earnings Report, Transcript and Summary

Good afternoon, and welcome to the MindWalk Holdings Corp. Financial Results Conference Call for the Fiscal Year ended April 30, 2026. This call is being recorded. Before we begin, I would like to remind listeners that today's discussion contains forward-looking statements. These statements reflect management's current expectations and are subject to risks and uncertainties that could cause actual results to differ materially from those anticipated, including the company's history of net losses, the ability to convert platform adoptions into contracted recurring arrangements, market acceptance of ReefIQ and LensAI, intellectual property risks, competition and capital markets conditions. A fuller description of these risks appears in the company's annual report on Form 20-F and other filings available on SEDAR+ and EDGAR. All financial figures discussed during this call are in Canadian dollars. Financial statements and MD&A are available on the company's website at mindwalkai.com, sec.gov and on SEDAR+. A replay of this call will also be available following its conclusion. I will now turn the call over to Dr. Jennifer Bath, President and Chief Executive Officer of MindWalk Holdings. Please go ahead, Dr. Bath.

Thank you, and good afternoon, everyone. Fiscal 2026 was a defining year for MindWalk. The year our discovery engine and the bio native AI built on top of it came together as a single platform. That engine has a long record of producing molecules into the clinic. The AI layer makes that biology computable at scale. It was also the year the market thinking caught up to the thesis that we have held on to from the start that the durable value in AI for biology is not the model, but the data and the context layer beneath it, a layer that, in our case, rests on 20 years of curated biology, no competitor can stand up overnight. Independent research now describes that layer in language that we could not have written better ourselves. This was the year we began to convert that recognition into results. We grew revenue by 46%, expanded gross margin to 59% and narrowed our net loss by more than half, growth and discipline in the same year, not one at the expense of the other. We signed our first 2 contracted recurring platform agreements, the first recurring platform revenue in the company's history, and we launched ReefIQ, which organizes the client's own discovery data into governed biological context. We strengthened the balance sheet through the divestiture of a noncore business, and we regained NASDAQ compliance organically without a reverse split without dilution. Earning inclusion in the Russell 3000E and the Russell Microcap Indexes. The science underwrites all of it. Our discovery work for clients has now put more than 20 molecules into the clinic, and that figure reflects only the clients that we are free to name. Counting the programs that remain confidential, the true number is much more meaningfully higher. Among the clients we can name are Johnson & Johnson's Janssen vaccines, Sanofi Pasteur, argenx, IDEXX Laboratories and Xencor. And this year, one of those clinical molecules that our work supported became a commercial medicine approved by the FDA and launched. Those are headlines. Taken together, they mark a very different company than the one that stood here 1 year ago, and they change how the next several years should be underwritten. We've built the record. The market has arrived at the thesis, and we've signed the first recurring platform revenue. What's early is scale, and that is the opportunity that is in front of us now. For the past several years, the AI drug discovery sector has been consumed by a single question, whose model will win? That is the wrong question. And this year, the market has largely moved past it. In its place, a more durable conclusion has taken hold, articulated in strikingly consistent terms across independent research and industry commentary. It maps precisely onto the asset MindWalk has built, and it comes down to 3 points: adoption, not model capability is a constraint. Proprietary data is the moat and validation is the catalyst. The orchestration layer between AI and biology, the layer the market is now calling the moat is not something we pivoted into. It's the asset MindWalk has been building for years. HYFT provides the biological representation. ReefIQ organizes a client's own data inside of that representation, preserving provenance and program history across their entire estate. And LensAI is where the reasoning happens, target discovery, candidate diligence, portfolio decision support and agentic AI workflows that close the loop between in silico prediction and wet lab validation. A long-term advantage is not the model. Models are replicable. It's HYFT. It's 20 years of curated function annotated biology, encoding the sequence and the structure and the functional relationships that let any model reason over biology with context, the context that it otherwise lacks. That foundation took 2 decades to build, and it cannot be stood up overnight. And for each client, the value compounds in their own governed environment. Every program they run inside of ReefIQ makes their context richer and their work faster with their data remaining theirs. This is where the thesis begins translating into commercial execution. There's a validation ladder in this business and the rungs behind us and ahead of us are worth us being really clear about here. The first one is the application layer adoption. Clients engaging LensAI to solve a defined workflow problem on a contracted recurring basis. In the second half of fiscal year 2026, we signed our first enterprise layer SaaS agreement and then a second contract in the fourth quarter. Together, they marked the first recurring platform revenue in this company's history. That's the entry point. The client trusts the platform with a discrete application. They see the results and it builds conviction from there. The next rungs matters more. It's clients engaging ReefIQ to organize and manage their data across their broader estate. Well-organized context-rich data is where their competitive edge comes from. And at the Jones Trading AI Day fireside chat last month, when asked, I accordingly told institutional investors that this is the validation to watch for from MindWalk. The commercial engine we're building runs alongside that ladder from application layer adoption towards the deeper data management relationships above it. The most rigorous validation of a discovery capability is the medicines that reach the patients because of it. Behind that record of dozens of molecules that we have assisted with in producing for the clinic, 10 are in active Phase I through Phase III trials, 4 are first-in-class, and that work is backed by more than 400 peer-reviewed publications and issued patents. These are client-owned assets. The value of each molecule accrues to the company whose name is on it. What accrues to MindWalk is a foundation. A track record of molecules reaching the clinic is what earns a discovery company the standing to build on it, deeper partnerships, co-development milestones and royalty participation and the recurring revenue that comes from being embedded in how a program advances. The clinical record is the proof the partnerships and the recurring revenue built on top of it are where that value starts to compound. In June 2026, one of those programs reached patients, a therapeutic our client developed for a rare debilitating autoimmune disease was approved by the FDA and launched commercially. MindWalk's role wasn't the molecule itself. It was specialized, the anti-idiotypic reagents our client used to build and to validate the bioanalytical assays that their molecule had to clear on its way through clinical development, including in response to a direct request from the agency. That's a contribution that we're proud of because it points to something that I believe the market overlooks. Getting a molecule to patients takes far more than designing something that binds. It takes the specialized tools that measure a drug that validate the assays that are behind it and stand up to the regulatory rigor that an approval demands. Most of the field is so focused on that first step on discovering a molecule. We support molecules across the path to approval, and that breadth is far harder to replicate than a single binder. It's also how a discovery relationship becomes a deeper one, the kind of collaboration that compounds into partnership and recurring work. The track record is more than credibility. It's 2 decades of learning what it takes to move a molecule from discovery all the way through approval, now built into LensAI and being applied to our own programs for the first time. Our NK program has produced strong preclinical results, a conserved pan-serotype target identified computationally and then validated in vivo across 2 independent campaigns. The science speaks to exactly what the platform was built to do. GLP-1 shows the same engine pointed at a different problem. We designed the receptor agonist sequence in silico and then we tested it in vitro, where they activated GLP-1 receptor comparable or superior to semaglutide, one of the most competitive targets in medicine. It's early, and of course, we will update you as it matures, but early results on a target that hard are worth paying attention to. Influenza extends the pattern across more than 2,000 strains of influenza HYFT found a functional invariant that holds where the sequence itself keeps changing. That kind of target convention alignment simply is -- with conventional alignment, it simply cannot see. Behind those are more, our pandemic response platform deployed this year against Ebola and Hantavirus and a broader set of infectious disease and autoimmune programs that we'll bring forward as they earn it. On July 1, 2026, we filed a patent application extending our foundational HYFT patent, protecting our architecture behind how biological context is organized and made computable in our system. The context layer is where the long-term value in biological AI accrues, and this patent is built to protect that. Fiscal year 2026 also marked an important milestone in restoring institutional credibility. We regained NASDAQ compliance organically without a reverse split or other dilutive measures, reflecting genuine improvement in market support and investor confidence. Effective after U.S. market close on June 26, 2026, MindWalk was added to the Russell 3000E Index and the Russell Microcap Index. Russell inclusion expands institutional visibility, increases eligibility for index ownership and represents another important milestone in MindWalk's continued evolution as a publicly traded company. We completed the transition from ImmunoPrecise Antibodies to MindWalk Holdings Corp. in September of 2025, creating one unified commercial platform that simplifies how the company goes to market and supports more efficient growth. 9 months in, the metrics are consistent and directional. AI citations of our content have more than doubled above the legacy brand. The number of referring domains to our digital presence are up materially. Organic keyword coverage has expanded. Sustained daily dollar volume is a multiple of the pre-branded median and analyst coverage has expanded. These metrics all reflect the company being found by the audience that it is now serving. In October, Scott Areglado joined as Chief Financial Officer, bringing the finance and capital markets discipline that you'll hear from him shortly. On the commercial side, we deepened our enterprise business development team this year, and we've added 2 more senior professionals in both Boston and San Francisco to drive platform adoption in the 2 largest U.S. biopharma markets. In the first part of the fiscal year, we completed the sale of our Netherlands wet lab operations to AVS Bio, strengthening the balance sheet. Scott will cover the financial impact of that. More importantly, it sharpened our focus on a genuine structural advantage at MindWalk, computation and wet lab operate as one integrated system, not 2 functions handing work back and forth. That integration is why our programs compound data and biological context at a rate that conventional discovery cannot match. Starting today, we are in San Francisco at AMD Advancing AI 2026 AMD flagship event, a company AMD chose to build with. That choice is the signal. AMD is one of the largest companies in the world, and it does not put its engineering weight behind marketing relationships. It codesigns with a small number of companies whose work is demanding enough to push its hardware and it selected MindWalk. This is an engineering partnership. AMD's teams work alongside ours to optimize the infrastructure our AI platform runs on. Migrating our libraries off of a CUDA lock stack onto AMD's open ROCm stack was a joint build, production infrastructure that runs every single day at MindWalk. AMD's published case study put the results in hard numbers. On Instinct MI300X hardware, we screened roughly 170,000 antibody pairs in about 4.5 hours, work that had taken 145 days previously. Cost per million samples in literature mining fell about 39% and embedding throughput rose about 70%. Speaking alongside us at Jones Trading AI Day fireside in June, AMD described being highly selective about who it codesigns with and pointed to MindWalk as having the capabilities the market is looking for on the record in front of institutional investors. There is more from all of this and our interactions at this conference that we expect to share shortly. Let me turn this call over now to Scott to review our financial results, and then I'll make some closing comments before Q&A. Scott?

Thank you, Jennifer, and good afternoon, everyone. I will take you through the fiscal 2026 financial results and the drivers behind them. As a reminder, all figures are in Canadian dollars unless otherwise specified and are preliminary pending the filing of our annual report on Form 20-F. The financial story of fiscal 2026 is straightforward: revenue growth, margin expansion, disciplined operating investment in the commercial engine and a stronger balance sheet. Revenue grew meaningfully year-over-year from $10.6 million to $15.6 million or a 46% increase in fiscal year 2026. Revenue for Q4 2026 was $4.1 million or a 50% increase as compared to $2.7 million for Q4 2025. The quality dimension of that revenue is as important as the volume. This was the year we booked the first contracted recurring platform revenue in MindWalk's history through the 2 enterprise LensAI agreements Jennifer described. Those are software layer engagements carrying different economics than project-based services, and they open a revenue model we intend to scale. Gross margin expanded from $5.7 million or 54% of revenue to $9.1 million or 59% of revenue. Gross margin for Q4 2026 was $2.5 million or 61% compared to $1.6 million or 58%. This reflects a shift in revenue mix toward platform and higher-margin service work together with structural efficiencies following the divestiture. Now on to operating expenses. Total operating expenses were $24.1 million as compared to $42.6 million in the prior year. However, the prior year included $21.2 million of intangible asset amortization and $1.5 million of impairment that did not recur. On a comparable basis, operating expenses increased by approximately $4.2 million, which reflects deliberate investments in the commercial infrastructure. Research and development expense for fiscal year 2026 was $4.9 million, up slightly from $4.2 million, demonstrating continued investment in our commercial infrastructure. Research and development expense for Q4 2026 was $1.6 million as compared to $0.8 million in Q4 2025. Sales and marketing expenses were $5.9 million in fiscal year 2026, up from $3.6 million for the same period in 2025, reflecting increased staffing and expanded North American business development capacity. Sales and marketing expenses for Q4 2026 were $1.5 million versus $0.9 million in Q4 2025. We expect operating expenses to increase as we continue to invest in our pipeline assets and ReefIQ. Net loss from continuing operations was $15.1 million in 2026 versus $33.1 million in the prior year. Net loss for the fiscal year 2026 was approximately $14 million as compared to $30.2 million in the prior year. Loss per share from continuing operations was $0.33 per share versus $0.99 in the prior year. For the quarter ended Q4 2026, net loss from continuing operations was $3.9 million versus $3.4 million in the prior year quarter. Net loss for the fourth quarter of 2026 was $3.9 million as compared to $2.2 million in the prior year. Now turning to the balance sheet. We ended the year with a cash balance of $11.5 million as of April 30, 2026, as compared to $10.8 million as of April 30, 2025. In summary, the company generated nondilutive capital by divesting our European operations for net proceeds of USD 10.3 million without sacrificing revenue growth. We continue to invest in the infrastructure that matters, integrating ReefIQ into our lab operations and advancing our engineering work with AMD, the foundation beneath both our own pipeline and our clients' data management. With that, I will turn the call back to Jennifer. Jennifer?

Thank you, Scott. As we enter fiscal year 2027, our strategy coalesces around 3 pillars: the same 3 we've communicated in our filings and the same 3 by which we would ask you to measure us. Pillar 1, build intelligence-driven recurring revenue. In fiscal year 2026, we integrated LensAI into every new client program. And clients began paying to add LensAI applications to their work, the first application layer revenue at scale. That adoption is deepening. Double-digit clients are now in the LensAI portal, accessing their data directly with dozens more of programs actively moving into that usage. This is our lab in a loop, outputs from our laboratory work that flow back through the portal where ReefIQ enriches each client's context with every run. And it's the feeder into what comes next, broader data management adoption and recurring SaaS licensing. That deepening is the leading indicator of the recurring revenue engine that we're building. The market is telling us engagement by engagement that this is the layer it needs. Pillar 2, advance and protect the internal asset portfolio. We are building a proprietary biologics and vaccine pipeline, dengue, influenza, GLP-1 and our pandemic response platform work and additional programs across infectious disease and autoimmunity that we will disclose as they mature. These are not service arrangements. They are assets we own built into our own platform. Our wet lab is one of the most productive antibody discovery engines in the industry. It has put dozens of molecules into the clinic for our clients, and it is doing that work right now. That is the discovery capability behind this pipeline. That -- what's new is the second level we've added to it, our AI, it encodes the relationships between the sequence and the structure and the function. So our models reason with biological context that others don't. Surfacing targets and designing molecules that conventional approaches never surface because they can't see that connection. Neither capability is borrowed and neither is new to us. Now they run as one engine. The AI points to the lab at the right experiments and the lab's results deepen the AI and every program compounds richer in context than either could reach alone. Pillar 3, deepen enterprise partnerships. The relationships that matter are the ones that change how the work goes or how the work gets done. AMD is the proof, joint engineering that made our platform faster, made our platform more scalable, running in production every single day, not a logo on a slide. We're building the same kind of relationship with pharmaceutical and biotech companies right now. In those relationships embedded in artificial intelligence and embedded on the capabilities running with MindWalk and AMD in combination. When those are far enough along to show results, that's when we'll announce them. And with that evidence, not just the name. The AI drug discovery sector is being repriced right now away from the model for the context layer beneath it. That is the layer we have spent 2 decades building, and it is the layer that we lead. This is no longer our view alone. Independent research has arrived at it. Our clients are arriving at it, one expanded engagement, one platform agreement, one medicine reaching patients at a time. Fiscal year 2026 was the year that the market caught up to what we built. Fiscal year 2027 is the year that we press the advantage. While we are in front, while the field is still catching up and while the assets we own are compounding on a foundation that no one else has. This week, we are on that stage at AMD Advancing AI in front of the entire industry. It's a fitting place to be because the company we're describing today is the one the market is now looking for. We thank our shareholders, our clients and our team for their confidence. Operator, please open the line for questions.

So congratulations on all the developments, obviously, having the second or 2 recurring revenue licenses is very good. But I just have a few questions, if I may. So to start off, of the $15.5 million that you recorded as revenue for 2026 -- fiscal year 2026, what percentage of that revenue comes from the recurring revenue the first license you have in the books now?

Yes, RK, I mean, our revenue is almost primarily still fee-for-service work. We signed this agreement early in the middle of this fiscal year. It was a modest amount. And obviously, I don't disclose individual revenue amounts, but it wasn't significant to our overall revenue for fiscal year 2026.

Okay, perfect. So...

I wouldn't mind adding just a little something to that, RK. Because, yes, I would say is it material to the $15.5 million? No, I wouldn't say it's material. I also wouldn't say it's nothing. It's actually pretty decent, and we're just rev rec'ing it every month as we go, right? So we provided an invoice. They paid that invoice, and now we're rev rec'ing in equal distributions over the course of the time frame that they have agreed upon in that agreement. And then I would like to follow-up on that with probably the most important part here is we provided that SaaS model contract to a company who saw the value of the applications that were provided in that contract. They have been using that contract actively against that invoice and that payment. And 100%, the most important part was they worked with us. They saw the value. We were doing fee-for-service work first. Then they turned around and said, we want to actually take the SaaS model license to these applications and take a deeper dive in this, and they did. So of course, going back to everything that I'm pointing to today, that is our primary focus now is, yes, great, you're seeing it, you're using it. We're rev rec'ing that. We're watching them use the applications. Maybe most important part here to focus on is, of course, now we're targeting that exact profile for ReefIQ for data management engagement. And so kind of back to that thesis, that's really the next step I'd ask people to watch for. And that's exactly what we're doing with the dozens of clients who are also utilizing our SaaS model really the portal with the LensAI applications as we turn data back as well. And that's also our focus for the other client with the newer agreement. So I just want to make sure I put that in some full context there.

Thanks for that explanation. Yes, now I can -- I understand that since it becomes part of already the programs that the clients have signed up for, it's really difficult to tease out specifically the dollar amount. Probably that's -- I'm assuming that's what you're telling me and that's understandable. So has -- maybe another way to look at this is as you continue to increase utilization of the platform in any or all of these programs, I'm trying to think about at least margins. So can we talk through the margins and say, how -- as you continue to increase the utilization, how the margins could expand from here, not only at the gross margin level, but also at the operational level?

Yes. Look, the SaaS -- one of the reasons people like the SaaS model subscription is because they do generate high margins. I don't have an exact number because we still need to understand what the compute cost that goes with that would be as they start to engage more with it, which would end up in the costs. But I would still expect it to be a high-margin business. And I think once we get a couple of more clients and I have a sense of where that will fall, I could be able to talk a little bit more about it. But I would expect it to be very healthy margins.

Okay. Good. And then on the -- 2 things happened over the last 6 weeks or 8 weeks now, not only launching of the ReefIQ, but also the IP that was filed, the application was filed. So as you go forward, how do you plan to pack this together? And is there a way to invite new customers such that they use that as a separate package? Or would that always be included within your services that you already offer to any of your current customers?

I love that question, RK, because we didn't talk a lot about that on the call. And so I think it is important for us to make sure that it's extremely clear that, yes, of course, it is something that can be packaged. It's something that I think is more tempting to people who have used our applications, who have used our fee-for-service who see what a difference that having the connectivity and the context and the meaning have made to the programs. And I think in large part, if we wanted to point to a few specific but anonymous examples, they have seen that most definitely when we've been able to solve problems that other people couldn't. And that's what draw them deeper -- drew them deeper into using our platforms into taking a SaaS model license or into contracting us to do that type of LensAI work with them. Just as a quick note, that type of fee-for-service work has actually come to an end. We are now asking people to contract us either in a partnership model or they can take a SaaS license to do some of that work on their own. We're no longer offering piecemeal fee-for-service for people to get introduced to the platform because the platform in and of itself is more mature than it was when we began offering that. But can people come to the table and use these platforms without going through the channels of like right now where we've distributed SaaS model subscriptions or we've distributed data back to dozens of clients within LensAI so they can see the applications or trial the applications. That's just one direction, and that's a really unique way because we have these loyal clients. We have clients just day in and day out where we're turning data back. But I think the most important part here that we need to emphasize is the fact that utilizing ReefIQ to provide context and to provide meaning isn't limited not only to our clients, it's not limited in almost every other way that you would look at software or applications or access to technology and potentially being limited. It is -- ReefIQ is something that everyone can access no matter what AI model they're using, right? And that's what we really see out there. We see a lot of people in the drug discovery business, people in just areas of research and development, developing AI models in order to try to do things faster, be more accurate, save money, whatever it might be. And what we're finding is they're all running into that same difficulty. That same difficulty being that all of a sudden, they've been training their model, they're training their model. When their model sees something it's never seen before, it hasn't trained on it before, it doesn't have the answers. And unfortunately, that's where these models start to hallucinate. And it's not like using ChatGPT where you just get back an answer that might be a bit ridiculous. You're probably going to flag it and you're going to see it, but it was hallucinating its answer to it, and it might cost you 15 minutes to get it back on track or maybe something you don't recognize is wrong and it costs you an afternoon or a couple of days. When this -- when hallucination happens at this level in drug discovery, it can cost hundreds of millions of dollars. It can waste years. And if people have never experienced it, it's a new target. It's a new indication. It's a first-in-class molecule. They don't know the mechanism of action. When their data -- when the data is being analyzed and then the information coming back isn't accurate, it can completely destroy and kill a program. And that's where ReefIQ definitively comes into play. Anyone can use it, no matter what their AI model is, they can literally snap on. It can be used most definitively with agentic AI. It can be queried directly with generative AI programs, and it's compatible with any other platform that anyone has built, and it doesn't have competition in the space to do this. So it elevates everyone else's AI models, and it is perfectly fine for someone to come to us and say, ReefIQ is the orchestration layer that we need and to purchase access to ReefIQ and then to feel confident that there also can -- this can be utilized in a secure environment without jeopardizing or threatening the security of their data or the systems that their programs are running within. So absolutely, and that's a very important message that we will continue to talk about this year so people can understand the breadth in which this ReefIQ technology can be utilized.

Okay. So one last question from me before I step back into the queue. In addition to this, you also have 3 assets that you're working on, the dengue virus, the influenza and the GLP-1. Of these 3 over the next year, which ones do you think could advance enough that either you can attract somebody to help you out in developing further or you would be able to attract enough funding such that you can take this forward itself?

That is an interesting question. So I think there's a number of points, and I'll try to be succinct here in my answers to them. The first one being the honest answer is I don't know which one will move the furthest. What's really interesting is we continue to find and build like new context around these particular molecules and the diseases that they cause. And so there's an aspect around that of when do we move something forward a little bit more and when do we take it and take a look at maybe formulation or presentation of that molecule. We're looking at some of the minute details as we go through. So we kind of on the surface make it sound relatively simplistic, right? We're immunizing these animals. We're collecting the serum. We're doing this. We're doing that. We're looking for neutralization, et cetera. Reality is we're also taking a fine-tuned look at what arms of the immune system are we actually amplifying and are we stimulating, and to what extent would we -- would we predict MHC Class I or Class II responses and in what geographies, what people are likely to give us these responses. And to what extent do we understand the type of response that we need as we move through looking at a desired mechanism of action, a desired type of response that's going to provide lasting immunity or lasting therapy depending on the product that we're making. So not to get too wordy about that, but it's a little deeper than kind of what we've already provided. And the assets right now as they're moving along, we are quite happy with the way they're moving, but there's another caveat here, which is like what we've shared is not the full extent of, of course, what we're working on. And so that makes it even more difficult to choose one of those 3 assets because certainly, some are ahead of others, some are just in validation or in repetition because we'll never take a single data point or study and lean into that to move it forward. And others have different aspects of how far we plan to move them. Some of them have prospective partners already lined up and waiting, which means we aren't going to move them all the way through if they continue to be successful. And when we look at some of the other assets that we're moving along that are not -- we haven't really talked about and we haven't at all talked about in the public domain. Who's to know if some of those might even leapfrog and move faster. So what I can say, RK, with certainty is we are focused on these every day. We've built a team, and we're continuing to add on to that team in subject matter experts, whether it is in the engineering, whether it is -- it is biologics experts in immunology and design. We're continuing to build that team to make sure that as our assets grow, we can continue to keep this a priority. And so it is -- as I mentioned, it's one of our major 3 pillars of focus this year. And we are also focused on getting that nondilutive funding in the door for those assets, making sure that we have a ring-fenced structure that makes piping that money directly in as easy as possible for investors, and that's a lot of work we have been doing over the last quarter, in particular, in making sure that, that's designed in a way that is optimized for MindWalk and for those assets to move forward. So all of that is kind of front and center in what we're doing and things are moving all along likely and quite well. And I think over the -- over this fiscal year, we don't know which is going to move the fastest, but we are probably just as excited as you and some of our investors are in seeing how far these go and which ones are moving most quickly to these milestones.

Perfect.

Congrats on the progress. First one is, what should we expect for the operating expenses trajectory over the next few quarters?

So I think our operating expenses, as I indicated in my prepared remarks, I expect them to increase. The work that we're doing on the pipeline assets, the continued R&D in ReefIQ are all commercial investments we think are worthy. I don't have a specific comment on like the percent that's going to increase, but I would expect them to increase. We're trying to run as tight a ship as we can. And we're going to make commercial investments where we think they're going to generate a return and obviously trying to be mindful of capital allocation and expenses.

Yes. And for the pipeline, what should we expect? Can you give any color or guidance on clinical data updates, publications, any updates on IND-enabling studies on any of the programs?

We do not have any updates to prepare or to provide here that we haven't already provided with the exception that we don't have any clinical data for pipeline programs that we have built because we don't currently have pipeline products. And so there's no update to be had there. I think we've provided our most recent update on dengue kind of reinforced it here. And we're getting the cross-reactivity from that platform that we desire. That is one that we continue to take a close look at the exact details of what that immune response looks like. We're quite happy in being able to establish that cross-reactive immunity. The next step in that, where we expect to finish -- we don't have an exact date. We do have a third-party partner that's doing some further analysis on that, but it will definitely be over the next couple of quarters. I don't think we'll be doing any publications on these molecules just for the sake of the fact that any publications that we would do on an internal product at this point in time would jeopardize our IP protection of those molecules and none of these molecules are all the way through IP production and any data that does come out wouldn't have been something that would be included anyway. So what I can say is it's not impossible for other pipeline products depending on the partner that some of those could be in peer-reviewed publications, again, just depending on the partner and the level of disclosure that might have already occurred around those particular molecules historically. So what I will say is as these molecules are moving forward, when there are material updates, we will release them. But one thing that we're considering quite important is we get a lot of -- not from our analysts, but we do get a lot of requests to just release as much information as possible or release updates as often as possible. And oftentimes, I don't think there's something that like Dania, you or myself would consider to be material updates. And that's something that giving the updates in between the material updates is something we want to back away from because it doesn't seem -- it's not -- they're not meaningful, right, for people like ourselves who understand like you and I, who understand what really adds value to a molecule as it's moving forward, the touch by touch and play by play, little things that move in very small increments, but don't represent true milestones with regard to the end game of the product are an area that we're going to watch closely and to back down from just to preserve energy and professionality when it comes to announcing these. So every material milestone that we hit, and we will still consider things like strong in vitro readouts and in vivo readouts, anything that's moving us toward IND application as a material update, we will keep everyone posted on those.

And everyone, at this time, there are no further questions. That does conclude our question-and-answer session. It also concludes our conference for today. We would like to thank you all for your participation, and you may now disconnect.