I would now like to hand the conference over to Dr. Sam Hupert, CEO. Please go ahead.
Thanks very much. Good morning, everybody, and thanks for joining us. As most of you would know, we are a company in three jurisdictions: Melbourne, our corporate headquarters; Berlin, where we do all the R&D and support for the Visage product; and the U.S., which is our main market, where approximately 90% of our revenue is derived from. We have two product sets. Here in Australia, we develop the Visage RIS, which does all the scheduling, billing, back office, interface to payers for some of the largest radiology groups in Australia. The Visage 7 product, on the other hand, is a clinical product. It's a radiologist desktop, and it's the one that we are selling in the U.S.. In terms of the results, we felt all of our figures headed in the right direction.
Revenue was up 28.4%. Our underlying EBIT was up circa 30%. Our EBIT margins, period -on -period, increased. Our cash and investments went up by 5.3%, despite a buyback, increased dividends, and an investment of AUD 10 million that we made during the period. Our interim dividend went up 28% to AUD 0.32 per share, fully franked. So we felt that everything was, is moving in the right direction. In terms of the highlights, it was another record half. We won seven new contracts, totaling AUD 280 million at minimums. Putting that in perspective, just two years ago, that's what we would sell in an entire year. We renewed the FMOL contract.
We had our first very material sale of our cardiology suite to UC Colorado. We completed six cloud-based implementations. Our RSNA 2025 was the busiest to date. We have made significant progress with the other ologies, cardiology and pathology, to name two. And we believe we've formed a very strong base for growth in second half, FY 2026 and beyond. Some of the highlights, as I mentioned, UC Colorado was a AUD 170 million deal early in July. We followed that up with a very important client in University of Heidelberg, which I'll talk about a little later.
We won a AUD 44 million five-year contract with a group, ARM, which is their management name, but the name that they go by in the market is RAN, or Radiology Associates of North Texas, the largest private reading group in the U.S. Then in November, we had three contracts that together gave us nearly AUD 30 million at minimum TCV, and finished off the year with BayCare, adding the archive to their viewer workflow, which is a AUD 25 million deal, taking them to full stack. So a busy half. In terms of revenue splits, the salmon color is recurring transactional revenue. The blue is recurring again, which is more of our support licenses from our previous capital model. And again, all of our splits in the revenue went up as predicted in the first half.
In terms of the model, it is a highly scalable model. We think we're one of the most scalable on the market. There is no CapEx, hardware, or cloud. It's a software-only model. We charge for all our training and installation. We have a very highly contained cost base, and hence the reason the margin continues to grow as our footprint increases. Just putting that in perspective, we believe our margins are somewhere around triple our nearest competitor. In terms of the transaction model, most of you would be familiar with this. It is based on minimums. Our forward revenue for five years, based on minimums, has broken through the AUD 1 billion mark for the first time. There's a lot of upside as client examination volumes grow and our clients grow well above industry average.
So we feel it's an annuity sales stream with far greater predictability. In terms of our investment during the half, many of you will be familiar, we made a AUD 10 million dollar investment, which is a hybrid of debt and equity in another ASX-listed company in 4DMedical. It's a two-year term for the investment. It has a coupon rate of 12.5% per annum, but if the 4DMedical share price goes up from our entry point, if it doubles, we get back AUD 20 million, and if it goes beyond that, there is an equity component based on the share price of our entry and the then current share price. And as we sit at 31st of December, that has provided an unrealized gain of approximately AUD 150 million dollars.
Who are our market? Well, we are—we have the largest footprint in the academic medical center market in the U.S. We do 11 out of the top 20 hospitals, as rated by U.S. News. And we see that we're now expanding not only at that level, but also at the next level down. A lot of the regional academic medical centers like UI or U Kentucky, etc., that we picked up in the past 12-18 months. So very important space for us... But the biggest space is IDNs. It's about 40% plus of the U.S. market. We are increasing our footprint at IDNs of all sizes within that space.
And again, with things like UCHealth Colorado, which is a mixture of IDN and academic, that we saw good growth in, in the half. The private market's an interesting one. It was one that was previously dormant. There was a lot of M&A activity up till about 2023, when interest rates started to rise. But we have been quite successful there of late. RAN is, as I mentioned, the largest radiologist-owned private reading group in the U.S., and we secured a AUD 44 million, five-year deal at minimums. That adds to the two deals we made in the previous financial year in Lucid and Julie. So again, we're now looking at about AUD 115 million, at minimums in this market and growing. So we think this will become a material market for us, alongside the academic medical centers and IDNs.
In terms of Visage RIS, that continued its growth. We sell it mainly here in Australia and in Canada. We do have long-term contracts with some of the largest radiology players here in Australia, and we are seeing increasing uptake of the RIS, particularly as new groups are forming, as splinter groups are coming to us. So what makes us so special? Many of you have heard me say this many times before, but it still holds true. The three key elements of the system: speed, functionality, and scalability. We are number one in all these three areas, and it's the sum of all these parts that make Visage what it is. It, we've seen relentless increase in the size of datasets.
One of the areas, in particular, over the last 12 months, has been an area in CT called Photon Counting CT. The file sizes produced are orders of magnitude larger than standard CT, and it will become the new default standard. So we do see file sizes increasing relentlessly as new modalities and refinements of modalities are being reduced. That helps us because pretty much all of our competitors still use the compress-and-send methodology. The file's compressed, sent down the network, unpacked at the local radiologist's workstation, and all of the 3D manipulation enhancement is all done locally on that workstation. The problem with that model is the files are just getting too big. The gigabyte is the new megabyte, and it's just taking too long and clogging up the networks.
We, on the other hand, have a unique, proprietary streaming technology that we, that Visage developed in-house, and it's the basis of all of our, speed and flexibility. It allows even the largest datasets to be visualized sub two seconds, and in 95% of cases, sub -one. It is a huge advantage for us, particularly nowadays, where remote reading and home reading has become so prevalent around the world. The sales, again, going through them, some of you may have seen this before, but University of Colorado, it's our second-biggest deal. It's AUD 170 million, second-largest in company history. I call it full stack plus one because it includes cardiology, so all three core products, that's Worklist, Viewer, Archive, plus our new cardiology offering, which is so making this a material sale for that new product.
They are a highly respected hospital system. As I said, a mixture of IDN and academic, and that we are looking to implement that within the next few months, before financial year end. The second one is very, slightly different. Again, European, it's University of Heidelberg. It is a top German medical school and teaching hospital. Heidelberg is one of the oldest, if not the oldest, universities in Europe, possibly the world, still operating. The other thing is, its affiliated cancer center is the largest cancer research center in Europe. So very high profile. It will increase our footprint in Germany and in Europe, because it's of the network effect of the type of university and hospital that it is.
In terms of, we are the value product, our highest charging product in the market, and we think that's where we should be. We provide the most value, the greatest and proven return on investment. Not only is it financial, but it is clinical. In other words, we allow radiologists to do what they otherwise would take too long to do with other systems or couldn't do with other systems, which is incredibly important for us. It's one thing to have the whole software, but it's another thing to put it in. We have a highly optimized fast-track implementation methodology. We believe we can complete large-scale projects in under 20% of the time of industry norm, and certainly of many of our competitors. It is a huge saving for the clients, and it's a huge saving for us.
We have used a highly optimized hybrid model with people on site for training and people remote, which gives us the flexibility, depending on whether the radiologists are at home or inside the hospitals in terms of training, and is a key differentiator. And I think it's becoming an even bigger differentiator now than it was maybe even two or three years ago, because there is. There are a number of our competitors that are having massive issues with implementations and timing of implementations, where sometimes they're delayed two to three years sometimes. So, so again, huge strategic advantage for us. We are also solving what's most probably the biggest issue facing radiologists, doctors in general, but radiologists in particular. The acute shortage with burnout is the new epidemic. We are able to address that.
We can increase productivity for radiologists out of the gates. Day one, they start using the system by north of 25%, sometimes as high as 40% or 50%. This has paid huge dividends for the systems that have implemented us. As I mentioned, the important thing for us, it's not just a matter of finances and efficiency, it's also the fact that we do move the needle clinically. Our growth strategy remains intact. We are expanding our footprint quite rapidly. We're now over 10% of the U.S. market. We are seeing above industry average transaction growth from existing clients because of what we enable. The new product offerings in cardiology and soon to be pathology are starting to pay material dividends.
We are extending our footprint in other markets, so Heidelberg, in Germany, for instance, and we are leveraging our R&D capability in terms of new add-on technologies, which include things around generative report generation and AI. So all of these principles of growth for us remain intact. North American market, we estimate roughly 670 million exams performed per annum, growing on average between 2%-3% a year. We believe we, from a product perspective, can address 100% of that, and we're unique in that regard.
In terms of when is a deal, deal too small for it to be commercially viable, we think that the tail is about 15% of the market or less, and that 15% is diminishing as the smaller players find it difficult to stand up stand alone and are absorbed into larger health systems because of security, governance, and the cost of doing business is just getting higher. So we're currently, as I mentioned, a bit over now, over 10% and growing, so we do have a very large addressable runway, which we're going after. In terms of the pipeline, it, we feel, it is very robust at the moment. There are a lot of opportunities in there. Those opportunities are across all three market segments and across all sizes of opportunity, which is important.
We think there's been a very strong network effect from implementations in all three sectors of the market, which gives us an advantage over others, and we have many prospects coming through on various stages of the cycle. In terms of the products, the first one was the viewer. We did release a number of years ago, a Visage 7 Open Archive. It's a key part of our cloud strategy, and I'm pleased to say that most clients nowadays take both the viewer, the archive, and also the workflow product, full stack. Not only does this increase the TCV for us, it makes the implementation so much easier because we don't have other third parties to deal with, and we are seeing more and more of that in our sales, and particularly in the last 24 months.
Cloud is huge for us. We have not put in an on-premise implementation in the U.S. in around about five years now. We believe we are the only truly cloud-native application currently in the market, and we think that has made a significant difference for us in terms of competition and our ability and speed at which we can implement. We are cloud vendor agnostic. We have large-scale implementations in AWS, Azure, and Google GCP, which gives us a lot of flexibility that others just don't have. In terms of our strategy, enterprise imaging, or one viewer for modalities, we are moving much, much closer to that with our new product suite. We now do radiology, cardiology, and soon to be released, pathology.
We also do reflective light and videos, which will cater for all the other ologies, like dermatology, ophthalmology. So we think we have moved a fair way down the track to being the only truly single platform for the entire healthcare enterprise, and we think that will increase our value proposition going forward, particularly since all of it is cloud. So the end game is what we see here on this slide. It is cloud. It's got Visage 7 as the center of everything for all clinicians in all ologies. It is also the repository of data for AI and generation of foundation models and AI algorithms, and it is the conduit for all image within the healthcare enterprise. And we are getting very close to delivering on that entire scenario.
Cardiology, as we mentioned, we've had our first really big sale with U Colorado. It is at the same code base as Visage 7. It's ultra-fast, and gives immediate access to the cardiologists for large data sets, which they need. It also has all the interoperability with third-party reporting systems, such as Epic Cupid, which is used by so many of the clients. We see this as, as I mentioned before, full stack plus one, so again, an ability to increase the total contract value. We have in the pipeline numerous opportunities, RFPs, that are for both radiology and cardiology, and we're seeing very strong interest from the existing user base in the cardiology offering. The last of the major ologies, digital pathology, we announced the release of that late last year.
We are going through the validation and FDA approval process for that. The important thing for us, again, is it is the same Visage platform, it is the same code base. It's not a whole separate development, it's part of our standard development. We've had an extreme amount of interest in that from a prospective and a user base already. So the last question that touches the topic du jour is, where I'm sure all of you've seen in the press of recent, of late, is AI, is it a disruptor or not? I thought this was an interesting quote from the ABC just last week, that it basically says, "Software is a sinkhole.
AI may revolutionize the way we live and work, could very well lay waste to business empires themselves, only recently overthrew the old world order. There's been a lot of headlines similar to this. Our view is that we think this is a gross generalization and overstatement of AI capability. Certainly, one of the concerns around massive infrastructure investment by some of the hyperscalers in AI and data centers does not apply to us. Quite the opposite. We are a capital light model. We're software, software only. We don't have any CapEx or data center build out pathway. Matter of fact, we think we'll be the beneficiaries of the build-out done by others. The other thing is, our software is proprietary. It wasn't developed on a you know, easily or readily available toolkit or platform.
It is deeply technical and highly specialized, and very domain-specific. So it's not a generic software stack, and it's not one that others know of the technology, because we have not only patented it, but we haven't published a roadmap of what we've done. And to date, no one has been able to emulate our technology stack, even though it's been in the market over 16 years. But having said that, our solution is more than just software. It's the systems and methods we've built around it, all the training, the rapid implementation that I mentioned. We often train thousands of techs, radiologists, and clinicians in a single go live, and it's all around how we distribute it and support it. Because healthcare is a highly regulated environment, you can't just whip up a program and there you go.
It has to go through a whole, regulatory cycle and be validated. Importantly, it is a mission-critical solution because patient safety is at stake. It's not a simple drag and drop, and you've got Visage 7. In terms of other modes that we have, unlike a lot of other SaaS companies, we have long-term contracts with our clients. They can't just cancel at a whim, or a month, or three months' notice. And as I mentioned before, we have over $1 billion at minimums of contracted revenue in the next five years, and we foresee that that will continue growing. The AI tools that are around tend to focus, you know, allows you to focus on system design rather than coding. That suits us because we use tools extensively throughout Visage and Pro Medicus.
We have, you know, fewer smarter developers that have highly optimized the use of these tools. And then on top of that, we also have the in-house capability to develop AI, which we've done, previously and continue to do. And finally, the last comment I'll make is that AI was predicted to replace radiologists, and therefore, would there ever be a need for Pro Medicus Visage in the future? I think that's proven to be wildly overoptimistic. All manpower studies, as recently as ones published overnight, are predicting that AI will allow the catch-up of the backlog, but will not replace radiologists. If anything, it will help generate more work for radiologists. So we think we're ideally placed to benefit from AI, and I'm happy to take questions on that in the next few minutes. These are some of our developments.
We have breast cancer detection in FDA, the investment in Lucid, the investment in 4D and LungAI that I mentioned before, the collaborations through a number of top AMCs, the latest being UCSF, joins Mayo, Yale, and NYU, and we have a growing number of third-party AI integrations to offer to our user base. This is our leadership team. Malte Westerhoff and Detlev Stalling are the two co-developers of the Visage platform. They're both PhD scientists in healthcare and healthcare data analysis, and they lead a whole team that does not only our development, but our efforts in AI, both in terms of using it and in production of our own algorithms. I'll finally finish off with something that we are seeing more and more use cases for, that's the Apple Vision Pro, the three-dimensional immersive goggles.
So we were one of the first, if not the first, to have software ready for it. And this year, an unprecedented thing occurred. Apple did a joint development with us. Oh, joint session with us at their key, at their key Chicago Apple store. Apple are very protective of their brand, of their store, of the technology. It took 11 months to organize, but it was a wildly successful event. This is a picture of it. You see, less than half the audience. We estimate 300-400 people turned up to listen to three radiologists talk about how they actually use, in the real world, the Apple Vision Pro, and our technology. So it was quite futuristic. And finally, RSNA, it was our biggest and best.
This year we had 62 people, which we needed, full-time. And the standard you see in pretty much the entire photo is the footprint of our standard at RSNA. So it is growing every year, and paying significant dividends in terms of, new leads and new opportunities that we see. So just finishing up, it's been the most successful half in our company's history. Majority of sales are full stack, and we see that continuing. Our proven implementation and support capability, are really key strategic strengths, as is cloud. We think we have an unparalleled value proposition, and our North American footprint and pipeline continue to grow strongly. Cardiology, full stack plus one. We feel that we will see more of these deals, coming through the pipeline.
Pathology, again, very well received to those that we've shown it to. We believe we are ideally positioned to leverage AI. It is a plus rather than a threat, and we do see increasing use cases for Visage and Apple Vision Pro. I'll leave it there, and obviously happy to take any questions.
Thank you. If you wish to ask a question via the phones, please press star followed by the number one on your telephone keypad. To ask a question via the webcast, please type your question into the Ask a Question box and click Submit. We ask that questions be limited to two per person. Please rejoin the queue for any follow-up questions. The first question today comes from Annabel Li from Goldman Sachs. Please go ahead.
Morning, Sam and Clayton. Thanks for taking questions. I've just got two, and I'll ask them one by one. So first, on timing, we clearly have very strong visibility on your contracted revenues. But just from a timing perspective, are you able to give us some more detail around the drivers that we'll see into the second half, and potentially into 2027? And then just confirming that revenue growth in that second half should start to accelerate as more of these contracts come online. And then just to follow up on that one, with commentary that there are three more Trinity cohorts that will go live in that second half. Was that just in line with your prior expectations?
It is. And, you know, at our AGM, which was in November last year, our chairman said that we were ahead of our internal budgets, and we were and are. We knew that particularly Trinity, which is, you know, one of the biggest go-lives and one of the biggest contract wins in the history of our industry, we knew that the first cohort would only go live at the end of October. That was a time set by them, not by us. So we received, you know, there was only two months, and we knew that was going to be the case. There are five cohorts in the first tranche, and we have completed the second cohort in January. So all of that is totally on track in terms of our timing. Because they're such large.
Each cohort is as large as a huge academic institution. They're very, very large. So, each of those will contribute very significantly to the second half, as well as anything else that we put in during the second half. We've got scheduled for another seven, well, besides Trinity, another five implementation, not five, another three, three implementations on top of the Trinity ones. So, you know, there's plenty to go on in this half, but I think the main thing for us will be the step up in revenue from cohort one and cohort two in Trinity, both of which are completed.
Trinity, so just to say, to answer your question about Trinity, yes, there are four cohorts that have been completed by the end of the 30th of June, which was exactly how we thought that would be, another one early July. So, the five implementations, five cohorts by then, and fully completed by October this year, sorry.
Got it. Thank you. And then just my second question on AI. As you mentioned, we're seeing some very significant advancements in AI. How might this have changed your thinking around, you know, the opportunity for P&E, like ramping up investment now, versus a little bit later on? And I guess if you could share any key milestones that you might be thinking to achieve, in this business over the years?
Yeah. As I said, we, you know, we see AI through multiple prisms. One is we use it ourselves in our own development and have for quite a period of time. As we said, it does assist coding, so it's not like something someone else has, and we don't. Quite the opposite. But then we also see the market for AI in our industry evolving to being very material, and we think we're very well positioned to benefit from that. So AI will be used throughout radiology in various steps, not all at once, but I think it will become prevalent. But importantly, I think it will become prevalent as an aid to radiologists rather than a replacement. So I think the opportunity for us is to use it in our own development, which we have been doing.
We have highly optimized that. On the business side, for us, clearly, we're looking to position ourselves as a key player in that AI space in radiology as it rolls out.
Got it. Thank you.
Thank you. The next phone question comes from Garry Sherriff from RBC. Please go ahead.
Morning, Sam and Clayton. A couple of questions, one on AI and the other one on your Department of Defense opportunity.
Yeah.
If I start with AI, how should we think from a contract term perspective? Mainly because if I think about healthcare customers, as with everybody in this AI environment, are they seeking flexibility longer term, just given the advances in AI and PACS software? Do you think, I guess maybe not this current renewal period, but if I roll forward, maybe you know, three to five years, do you think there's a risk that some of your customers might seek shorter renewals rather than be locked in for a five to seven-year contract term? So that's the first question. If there's been any conversations around that. And I guess related to that is also price.
You know, your margins are triple your nearest competitor, which naturally attracts capital from other players. Do you think, again, that price could be potentially impacted longer term, given what's going on in terms of AI advances and competition?
Yeah. So on contract length, we're seeing exactly the opposite. So we've, in recent terms, written contracts 10 years out. We don't know of any client that would even imagine writing their own product to replace us. What some will do is, and have always done, is develop things in-house that could be, you know, an adjunct to us. So they develop their own AI algorithms in very niche and bespoke areas, and that has always occurred. There's nothing new in that. So if we think of anything, my view is we're actually seeing longer term of alternatives in our contracts. So that was the first one and the second question, sorry, Garry?
Yeah, it might have been with price again.
Yeah.
You know-
Yeah. If anything-
Yeah.
I think it can, it will help us extend our price. Because we, we are building AI into our platform as part of the core offering, which allows us to make it, you know, even more automated, even further ahead of others. So I think AI, in terms of price, I think AI, in terms of its product, will all be incredibly positive for us.
Thanks, Sam. The next one on, just an update on that Department of Defense opportunity.
Yeah.
Anything you can provide, status, any updates that you're aware of, you can let us know about?
I think just that we are progressing. We're, we're nearly complete. Having our VISN23 in cloud, they did upgrade during the first, first half to Worklist, so we'll have full stack in cloud. We always envisage somewhere around February, early March, and that's on track for that. And we think that will be the poster child for all the others. So yes, there's something, we are close to the Department of Defense, looking at opportunities. You may have noticed that they, they had selected another vendor for a large tele, telehealth, teleradiology opportunity, and they're not progressing with that. So, you know, again, maybe that will come out to market too, which would be of interest to us.
So yes, we think that it's progressing well and all the steps we've taken with the FedRAMP authority to operate are all starting to come together in our favor.
Thanks, team. Thank you.
Thank you. The next phone question is from Josh Kannourakis from Barrenjoey. Please go ahead.
Hi, Sam and Clayton. Thank you for taking my calls. The first one's just on pipeline, and the second one, a bit of an extension on AI.
Yeah.
Firstly, with regard to pipeline, I think last time when we saw some of the major contracts, Trinity, they obviously kept a very close eye on the implementation and the success of the rollout with Baylor Scott & White.
Yeah.
When you look at the pipeline currently, and you've obviously progressed and done two of the big go-lives for Trinity, how do you see the opportunity opening up into some of those other top 20, sort of, you know, or top 20 or so hospitals in the U.S.? And maybe if you can give any feedback on where you're at in that domain.
Yeah, look, there's no doubt that, you know, success at Baylor, you know, even in early stages, the first two cohorts in Trinity, I mean, each cohort is bigger than any go-live, single go-live attempt by anyone. So we're, we're resetting the boundaries. All of that is resonating in the industry, and it's definitely been positive for us as a network effect because people go, "Well, if you can do Baylor, you can do us, or if you can do Trinity, and that you're somewhere in between Baylor and Trinity in size, we can do them, too." So yeah, they've, they've all been, you know, flagship implementations. No question about that.
Okay, that's really helpful. And you talked a little bit just on the pipeline around the, you know, the outpatient and the reading clinics.
Yeah.
That seems like that market is growing, and obviously with some of the shift in terms of reimbursement and things like that, that's becoming more important. And a lot of them are on older systems like Intelerad, which is being bought by GE. Just some more comment maybe on that, in terms of how you see the opportunity staging up in pipeline on that front.
Well, it's moving very quickly for us, and in a positive direction, simply with, you know, RAN is one of the most, highest profile of the fully private groups. It's incredibly well known. The thing about groups like RAN, they won't take our card because they're not the archive of record, the hospital is, so they don't need it. But their volume expansion is, can be considerable. So, you know, we think that, that, that particular contract, has a lot of upside in it. So we are seeing a lot more interest, in that, you know, the ambulatory market, pure ambulatory.
But having said that, a lot of our larger hospital clients are building, you know, centers or partnering with a provider for outpatient centers around their hospitals, and all of that has been good business for us as well. So, you know, from something where we really didn't have much of a look in about three years ago, that private space is growing very nicely for us, and we think that will continue.
Great. Thanks, Sam. And just on AI, so I'll take a different tack than the prior questions. You've obviously gone into a bit of detail on that.
Yeah.
In terms of your ability to monetize AI, I think one thing I often think about is the network across the academics and other major hospitals, where they're trying to, you know, bring their IP to the real world, and also using that network as a sort of validation rather than getting external, you know, validation outside the U.S. How important do you think that will be going forward in terms of segregating the different tiers, potentially, of algorithms? And have you seen much progress from those academics in terms of trying to bring some of these algorithms to market?
Yeah. Well, I think the whole thing about algorithms is that, you know, who, who created them, who curated the data, and how, how up-to-date they are. They're not a set and forget. And I think that's where the key academic centers, you know, have an enormous head start over others. And as you know, we tap into a number of them. So we, we see that they will serve multiple roles, like the breast cancer detection algorithm that we are releasing with NYU. Obviously, a lot of work in the background, a lot of clinical work, a lot of validation. We've had it validated in a second Tier 1 academic, so secondary validation, which we can do quicker than anyone.
I think all that will be, as you say, not only a source of algorithms, but a primary and secondary and possibly tertiary validation. The more validation you have by Tier 1 names, clearly, the more, you know, the higher the level of confidence in that algorithm by the other people who look to buy it. Yeah, I think all of that is working in our favor, and, you know, the first model of that will be the breast cancer detection that's currently in FDA.
Great. Thanks, Sam. Thanks, guys. I'll jump into the queue.
Thank you. The next question comes from David Low, from UBS. Please go ahead.
Thanks very much. Sam, just a quick first question. Just the model, the contract model that you have, as AI increasingly plays a role in assisting radiologists, what does that do to your per-click model? I mean, are there any challenges there? Is that built into the contract that it wouldn't have an impact?
No. Basically, you know, we look at AI for a radiologist in sort of two camps. One's under the bonnet, that is built into the application, makes things quicker and will allow us to maintain our pricing and possibly increase it due to the value that that gives. The second is around assisting a radiologist, second set of eyes, as we call it, and there we will look to charge on an opt-in basis. So if they decide to use, you know, the breast cancer detection algorithm that's currently in FDA, if a client decides to use it, there will be a cost based on a per click for using it.
So, you know, one is it embedded, and we think that will allow us to increase price. The other one is an opt-in, where they pay for it if they use it.
Yeah. I guess I was worried that AI might increasingly do the screening, you know, scan through a lot of images, and then draw attention to the ones that the radiologist should look at, and that might have implications for the way you set up the contracts. But I take it-
No.
You've thought of that? No. Okay.
Yeah, yeah. No, to be honest, at the moment, you know, some organizations talk about drowning in the workload, so anything that could even just get them square, which is what I think you're talking about, wouldn't take away from what we're currently doing in terms of volume.
Okay, great. And then the other question, just on competitive dynamics. It, we're not really seeing your competitor, sort of. Well, they're also winning plenty of contracts. The way you describe things is pretty compelling. Do you think the reputation of Visage and Pro Medicus's, well, versus any of your peers, and there's an obvious one there, has moved at all in the last 12 months? Do you think there's still more education required? I mean, I heard what you said about network effects, but I guess I'm trying to understand those dynamics and whether, you know, there is a clear differentiation there.
I think there is. I think there's a clear differentiation, and I think those that actually, you know, try before they buy will see the difference being far bigger than they imagined, and we've seen more of that. And I think, you know, it's also reflected in how we can implement. You know, the technology works, the cloud works. I mean, we know of others, I won't name names, but, you know, they've signed contracts, and three and a bit years later, nothing's been implemented. That's not us. So I think there is a difference. Not everybody, you know, you can't win them all, but I think, as we put more and more in, the delta between us and others gets bigger, rather than smaller.
If I could squeeze just one more in on what you've said there. I mean, is there any opportunities there where, you know, other implementations are clearly running behind schedule for Visage to replace those, you know, those who won the contract?
We hope so, yes.
If they come back out to market, yes.
We've seen that before a number of times where people have gone for something else, something cheaper, something they thought would work, and it hasn't, and then we win that contract, but-
Treat it as one case in point, where it had been won previously, by another group and clearly they weren't satisfied with the implementation and came to us.
There are, depending on how the contracts are written, if groups are unable to deliver, eventually, one would hope that you know, the institution will go, "Enough's enough. We've waited long enough." But, you know, all I know is that we can, and we know that others are having difficulty, and I think that's telling.
Understood. Thank you very much.
Thank you. The next question comes from Andrew Paine, from CLSA. Please go ahead.
Yeah. Morning, Sam and Clayton. Thanks for taking my questions. Just on FX, just wondering if there was any impact in the first half, and also the outlook into second half 2026, and then just also how that's assumed in terms of your forward revenue guidance over the next five years?
Yeah. In terms of the half, that's been not a lot. It's an average of the six-month period, and although the Australian dollar has appreciated since December, the average was pretty similar to the prior periods. In terms of forward-looking, well, if I could predict FX, I wouldn't be here. But clearly, as we stand today, it's higher than what we did in the half. Our five-year forward revenue takes into effect FX as of today, pretty much.
Okay. So that's got, you know, AUD 0.70-AUD 0.71 in that-
Yep.
1080?
Yep.
Okay. That's great. And then just also, look, just staying on the kinda competition dynamic at the moment, obviously, there's been the acquisition by GE of Intelerad. Are you seeing any, you know, evolution of competition within the space post that acquisition and, you know, changes in the dynamics at all?
Well, we think, to be frank, we think that that acquisition is gonna, is, has been already and will be good for us. You know, there'll be some that feel, which is my feeling, that a software company that's meant to be agile and nimble doesn't sit well inside a modality vendor that's a lot more structured. So, yeah, I think the dynamic will suit us, and some of the opportunities that we're seeing in our pipeline sort of reflect that.
Okay. So in terms of the Intelerad by themselves would be a stronger competitor 'cause they're more nimble and versus going into a larger company like GE?
Correct. But I think they've also... You know, they originally found a lead. They were four founders, but they sold out to private equity quite a number of years ago, and since that, I think their potency has decreased in the market considerably. But this, I think, will be a, you know, an even bigger, you know, an even bigger leg up. But, you know, that, that's a personal view, and, you know, we've seen the first window of it. So it'll be interesting to see the, how the, that dynamic plays out.
That, that's great. Thanks for answering my questions.
Thanks.
Thank you. The next question comes from Paul Mason, from E&P. Please go ahead.
Hey, team. I've got two related to AI. So the first one, I was just interested to understand a bit more, like, with, with all the AI diagnostic tool partnerships that you've got underway,
Yep.
Yeah, what, what proportion of, like, radiologists' diagnostic problems that they have to sort of think about... Like, I, you know, my understanding is there's, like, maybe, like, 10,000 different potential diagnoses a radiologist might make. Like, what-
Yep.
What proportion of that would you have covered, like, with the pipeline of your development? And then the second one was just, if you could maybe flesh out for us a little bit, like, currently, you know, Pro Medicus has this very big advantage around how much radiologist productivity comes out of using your tool because of how fast images load. What you think that would look like if you had like complete coverage of like, you know, diagnostic assistance in inside the platform?
Yeah. So, look, the number of, you're right, I think there are about 2,400-something differential diagnoses, and that maybe there's even more when you look at the various body parts. So I think, you know, AI traditionally has been more single diagnosis, single body part. We're seeing a broadening of that in the market, where some AI will look at a body part and have multiple diagnoses, so they cover multiple bases. The problem with that is, the FDA needs to clear each one, you know, at, at, you know, independently. So if it does 20 things, they independently clear 20 different algorithms. Now, that, that is slowly changing. So I think we're a bit at the beginning of that journey, not, not us, the whole industry.
But we are starting to see real-world application of AI as an aid to diagnosis. Some, you know, specific cases have reimbursement, which obviously is easier for organizations to implement because they get paid for it. So that helps too. But we're at the beginning of that journey. In terms of efficiency, I think anything that the AI gives sits on top of our 25% +. So it doesn't... You know, you don't get 20x percent out of Visage, and then you use AI, and it soaks it, it soaks up some of the advantage we have. It's all additive, absolutely 100% additive. So, yeah, so I think we're a fair way away from every body part being covered.
But, you know, the industry is moving quickly, and all of that we see will benefit us, but certainly not detract from the speed that we give as a platform. It'll always sit on top of.
And maybe just like a follow-up on that then. Just in terms of your ability to actually drive that versus it needing to come out of, like, the academic institutions that you've sort of talked about. Like, do you, do you have an ability to sort of forge ahead with that, or do you, do you need like a, you know, a partner hospital to actually decide they want to develop something before you can go and do it because of where the data's sitting?
Oh, no, you know, if you look at what we did with Elucid, we're just in cardiac CT AI. We'll be partnering with them. We don't need anything from the academics. We're taking an FDA-approved product from a third party and, you know, taking a pass-through by putting it through the platform, and you'll see more and more of that. Where we use the academics is in the areas where we develop or co-develop. In other words, it's a build or partner-type scenario with every algorithm in it. So we have a mixture, and the mixture is always dynamic, depending on who's doing what. But we always feel we'll be the point in the middle. So whether something we do with UCSF or it's something we partner with another third party, that'll all be on the cards.
Okay, all right. Thank you.
Thank you. The next question comes from Christine Trinh from Macquarie Capital. Please go ahead.
Hi, guys. Have you got me?
Yep.
Great. Thanks, guys, for taking my question. Just a couple. Firstly, on the margin piece, slightly better year-over-year, but less of an expansion than what we saw in the first half, and I guess, less than what the market was anticipating. But can we just get a bit of your thoughts on ongoing margin improvements and moving pieces in there? I know you mentioned RSNA, but a bit of a breakdown in terms of what are your expectations for sales force or other cost contributions, would be great. That's my first question.
Yeah, clearly, you know, revenue increased and helped with the margins, so they did go up. We did say, you know, six months ago, we think the margins can increase, but don't expect them, you know, to shoot through it too much higher. We might be able to eke out a few percentage here or there. Costs in the first half, clearly RSNA's in there, so it's always, you know, lower than the second half. The employee benefits was also a little higher this period. We had new employees that have started. We had, you know, some mark-to-market LGI and SGIs that had to increase. So they were in this half. That won't be repeated in the second half, but as I said, pleasingly, the EBIT margins did increase.
Great, thanks. And then just on AI developments, one of your competitors came out with a set of tools, essentially looking at reporting and diagnostic support. Is that something kind of you're building into the base Visage platform, or is it in the pipeline? Like, things like error checking and AI, kind of voice recordings.
Yeah. Well, we've already done that. We showed at RSNA, you know, the generative AI tools that we have that can look at, look through and do error checking of reports. And also, if you dictate just key findings, generate the clinical summary. So, we're already there with that, and we're extending that capability in the reporting side. So, yeah, they may have announced it, but we've already done it.
Perfect. And then just one last one from me, on the Trinity contracts. Apologies if I'm going over anything obvious, but just so I'm clear, on the first tranche, your five cohorts will be complete by July. So just how much that represents in terms of, like, the full contract, and then just an update on the following tranches, timing, contributions, and whatnot?
Yeah, that, that'll be about 80% done, maybe a bit more, 85%. So we'll hit the, you know, the new financial year, which will still be under their one-year minimum, at about 80%,: 85%. So that'll be good going into FY 2027. The last two cohorts, the Trinity, go live in August and then October. So, that'll be fully completed within 12 months of go-live, which, again, as Sam mentioned, is unheard of in the industry. Groups of that size in the past, we've had other, you know, competitors take three or four years even to start. So, while I think it has been misread a little bit in the market about all the timing of it, you know, we'll set us up obviously for a better second half and then a better, you know, going into FY 2027.
Perfect. Thank you.
Thank you. The next question comes from Sarah Mann, from Moelis Australia. Please go ahead.
Morning, guys. First question from me is just on the pipeline. Obviously, you've called out that, you've had a couple of cardiology wins in there, but today, just curious, when you look forward, like, what percent of the pipeline, or just a rough idea, any color around, how many of the clients are looking for your full stack plus one?
Yeah, well, it's actually material. I don't know the exact number.
And it can change 'cause sometimes people don't want it in the RFP, but then they look at it. So it can, it is fluid, it can move around. These RFPs, as you know, can be 18 months, you know, to 24 months, so things can be added.
Yeah, but we know that there's some already where it is in the RFP. And, you know, we're seeing a lot more of that than we would have seen maybe even 12 months ago. So look, we're not saying every single deal will have it, but, you know, we do believe we will see more of it, not less, that's for sure.
Got it. And then on the pathology opportunity as well, can you give us a feel for approximate timeline of when you think you're gonna be able to launch that? And I guess curious as well, in terms of the discussions you're having now that you've, I guess, kind of prototyped it at RSNA, how many of the people in the tender pipeline are interested in the pathology aspect as well?
So the launch is imminent. We have one client that we know we will be able to put it into, which they're doing their own internal validation, therefore, it doesn't need regulatory. So that's within the next few months or less. What we have seen is not so much people putting RFPs for pathology, or, you know, with pathology included, but it's an important decision maker, when they're looking at a vendor. So they'll say, "We're not looking at it for 12 months or 18 months, but we need to know that you'll have it," sort of thing. And I think the fact that we've announced it, they've seen it, that's, you know, that's validation enough for them. So I think you will see more and more throughout the year.
But it's not that people are just... You know, it's not like cardiology, where it's here and now. I think you'll see more and more looking at it within, you know, the 12-18 month period, but it's important for them to know you have it now. So clearly, it's a very important thing for us strategically.
Excellent. That's all from me. Thank you.
Thank you. The next question comes from Peter Meichelboeck, from Select Equities. Please go ahead.
Hi, guys. Just a couple of questions. Firstly, in relation to the buyback, obviously, the share price come back a fair bit since the first part of the buyback. Now that the result's out of the way, just wondering what your thoughts are on the buyback going forward are?
Well, clearly, that's a decision for the board. And, you know, we obviously would look at that, if we thought that the buyback made sense at a higher price, that I think it's telling you something. But again, I can't comment fully on that till we, as a board, sit together, which I'm sure we will.
Great. And just second one, in terms of renewals, you called out the renewal that you had in the first half with FMOL. Anything else coming up for renewal in the next 12 months?
Yeah, yeah, we have a few, and I think we've, you know, mentioned them before, you know, at Mayo Clinic and, Mass General and a few others, Yale, are coming up. So there have been a few that we're working on.
OSU.
OSU is another one. So yeah, all, all moving forward.
Are you in sort of discussions or, you know, active sort of discussions there on those ones or?
Oh, yeah, absolutely.
Yep.
Okay, great. Thank you.
Thank you. Moving to the webcast questions. The first webcast question is: Do we have a sense of the likely uptake of PME's cardiology add-on, and how potentially significant is it longer term?
I think it's gonna be very significant for us. One is in its own right, so we, as we've just mentioned, won University of Colorado, which is a big one with cardiology, and we are seeing a trend in the RFPs. I think the second thing is it's important strategically because people are looking, you know, at the broader enterprise and the fact that we can do cardiology and soon to do pathology, on the same platform, and I stress that 'cause no one else can do that. That, I think strategically, they're very important for us.
Thank you. The next question is: Does Pro Medicus see its relationship with 4DMedical growing, and what does the future hold there?
Well, as you know, or most of you would know, I was on the 4D advisory board, so I'm familiar with their technology. We have made the investment, which today, as it sits, is a good investment. And we also said when we made the original deal that there was potential to put their technology on our platform. So again, it's something we're looking at in amongst other algorithms. As we mentioned before, third-party algorithms on the platform and their VQ could potentially be one of those.
Thank you. The next webcast question is: How do you interpret the PME share price over the last full year? Ignorance regarding AI, misunderstanding future income commitments, competition potential changing?
Well, I can't, you know, we, we don't control the share price, but certainly we do believe there has been a lot of generalizations and misapprehensions around AI in general and its role in terms of software disruption. I think it's been totally oversimplified, and to be honest, exaggerated, and we've, we've put our view forward on that, as have others. We also think we have the most future commitments of any company and certainly the most in our own history, which we've now announced over AUD 1 billion. So take that for what it's worth. We do think we're sitting in a good position with known future contracts at minimum, with that increasing day by day.
Thank you. The next question is: What is the view on realizing the AUD 149 million 4DX gains, given the huge volatility in its share price and markets? What is the thoughts around keeping and why versus realizing? With the contract pipeline, what percentage is in negotiation phase by value?
Well, in terms of realizing the 40 gains, it's a two-year loan proposition, debt and equity. So, it's only realized after two years. We can't break that before then, so, there's no, no chance of realizing that beforehand.
Yeah, it's purely we, through accounting standards, have to show its value as of 31st of December, but clearly, the ultimate value will depend on the share price in end of July 2027.
Thank you. The next question is: What is the fast track implementation cycle in days? It seems that it is not translated in ramp-up of use or sales.
Yeah, I certainly think in terms of days, we don't have a one track. All we know is we've implemented things within three to six months, where others have taken years. Clearly, the bigger ones, like Trinity Health, UCHealth Colorado, we specified it in the announcement when we signed them, that would they would be longer than normal, and that's generally based on them, not us, which is getting their ducks in a row, making sure they have radiologists available to train, and just the large organization. So it, it's still a fast-track implementation compared to the competition. It's just sometimes for others, it's, you know, it takes longer for our customers to get, to get organized.
Yeah, and also on translation to sales, I actually disagree with that. Quite the opposite. We have, in this half, sold more than just two years ago, we did in a whole year. So it is translating to increased sales. We've had, you know, record last year with Trinity and others, and we're well on the track this year to just shy of $300 million in new sales. And that's at minimum. So it has definitely translated to improved sales, no question on that.
Thank you. The next question is: Regarding the VISN23 extension, is it for a five-year term? What is the total per annum minimum revenue from this contract now? What is the potential in expanding to other VISN?
Yes, it is a five-year deal. When we originally signed them in 2013, it was for a bit under, you know, $3-million-$4 million over five years , and it's now around $11 million. So you can see it's more than doubled, a bit over $2 million per annum. So it's more than doubled in the, you know, 10 years that we've had them. They've taken additional products, plus, you know, they have more licenses than they've ever had. In terms of expanding into other VISNs, I think Sam spoke about, you know, the opportunity there, and we are looking to move VISN23 into the GovCloud, you know, secure government cloud with AWS. And that should be completed by the end of March. And then we have a reference site for other opportunities in that space.
So we think, you know, there's you know, good potential because the government has said everything needs to go into this GovCloud, and we feel we're one of the few, if not the only one, that has been able to show that we can do it. So clearly, we are speaking to the other VISNs, and we are going to be using VISN23 as the example or poster child of how we think it should be done.
Thank you. The next question is: Given the significant share price decline we've seen today, despite what appears to be a solid operational result, can you help us understand what it is you believe is driving the market's reaction, and specifically from your perspective, what individual investors should do now? Should long-term holders stay on board, and what key upcoming catalysts or indicators should reassure us of future growth?
I comment on share price. My view has always been share price has a mind of its own, and I let the markets decide that. But I think our view is that we have shown, we've grown 30%, with our biggest implementation only coming at the end of the period, and 30% off a big base. We've had our second highest dollar value of sales in a half ever, and the year is not finished, and we know that we will have completed—we've already completed twon major cohorts of Trinity, our biggest client. So, you know, we think that the base has been well set up. We have said at our AGM that we expect the second half buyers to be stronger, and they're the reasons why.
We're on track for that because the other three cohorts of Trinity are all measuring, you know, to exactly the plan. Yeah, we think the pipeline is strong, and so therefore, you know, we're hopeful there will be more opportunities to add to that $1.08 billion that's sitting there over the next five years as a minimum.
Our second biggest contract, also in UCHealth Colorado, goes live in April, which includes both radiology and cardiology. So again, sets us up for a good reference site for any other cardiology opportunities, but also is out of the... given the size.
Thank you. The next question is: Does cardiology suffer the same issues as radiology?
It, it has the same sort of issues around data size. Cardiology with ultrasound and video clips now, the studies are regularly over 1 GB each. So what makes us good for radiology in terms of speed, the scalability, applies equally well to cardiology. So that's why we think it, it will be a significant contributor for us going forward.
Thank you. The next webcast question is: Has the company lost any tenders to the competitors since August of last year? And if so, why?
We have. We, I wish we could win every single RFP, but that's never really been the case. We have lost a few. Invariably, it's been around price, which I think is a false economy because, you know, radiologists generate so much income for the institution that you really want to give your radiologists the best tools. But, yes, we've lost a few, but it's always been around price.
Thank you. The next question is: Would you be able to please describe in a little bit more detail what you mean by third-party AI integrations? Furthermore, is there a good usage of those integrations in production?
Yeah. So, something like with Elucid. So Elucid have AI for cardiac CT, and what we would do is not just sell it as a standalone. We would sell it so that the output of it could be viewed within Visage as part of looking at that particular study. So it would just make it more seamless. Any of our integrations would make it far more seamless and intuitive for the radiologists to use. And that, you know, has been one of the big problems with AI in general, that some radiologists feel it's not well integrated and therefore takes additional steps and actually slows them down rather than, you know, speeds them up, and we think integration is the key there.
Thank you. The next question is: How will AI be implemented on Visage? Will it help triage images, flagging scans with potential critical findings, intracranial bleeds, and move it up the stack for the radiologist? How will revenue be generated, subscription or per-use model?
Yeah, I think that sort of prioritization we can already support. You know, there are third-party AI algorithms that do exactly that. They prioritize, you know, CTs if there's signs of bleeding. That's what we think of low-hanging fruit. I think the integration that we're looking at is even more sophisticated, like, if there is a bleed, where in the brain is it? Which images inside the stack have the bleed, and can we open the images to exactly that point? That's an example of a much higher degree of integration, and it's the latter that we're looking at. So as I said, the more integrated, the more seamless it is, the more it'll be adopted.
And in terms of the model, the fee model, more likely fee per study, so similar to what we do, you know, already. So if it's a discrete AI that we sell on top of our current offering, it'll be in that methodology. If it's within our product, it will just become part of the core stack.
Thank you. The next question is: Is there a pipeline of investments along the lines of the 4DX model?
We have been approached by a number of companies, as you would expect, over the years. We have a person, Tim Kern, who's joined us recently as our Chief Strategy Officer to look at exactly those sorts of opportunities. But look, again, we're very selective and would only look at opportunities where we thought our involvement could add value. But yes, we potentially see that there could be others going forward.
Thank you. The next question is: Could you see 4DMedical as a potential bolt-on to Pro Medicus? Would it complement your current offering?
I think, you know, I referred to that a little earlier, but basically we would see it as a third-party algorithm as we do, others, like what we're talking about Elucid and Cardiac CT. And there's a whole raft of other third parties in and around both the cardiac and non-cardiac space, and 4D could potentially be one of those.
Thank you. The next question is: UC Health will implement in a few months before year-end, so how do we look at UC Health revenue contribution in 2H 2026? In general, do we think of how the implementation at this stage, does cardiology impact the speed of implementation?
In terms of the contribution by U Colorado, yes, it starts in April and finishes around July. There's three phases. So it'll have a small blush of revenue in the second half of 2026, but clearly, you know, pretty much full revenue for 2027. In terms of the impact on speed, it doesn't affect it in this case because we're implementing both at the same time, so both radiology and cardiology. We're not doing radiology and then six months later coming back and doing cardiology. It's all at the same time. So I think the implementations are exactly where we thought they'd be at this stage, which again, I think it may have been misinterpreted in the market.
Thank you. The next question is: Great AI investments in third parties. Can you provide more color related to those in-house capabilities to develop AI? What is the perspective of in-house AI development?
We have the capabilities to do it in-house or with our clinical partners, and we've shown that in the past. It really depends on the opportunity and the AI itself and whether, you know, someone's already got something out there in the market and therefore, why just replicate that? So it really depends on the opportunity. We do see most of the applications being third-party, just because they're literally gonna be thousands of them. But, you know, that won't be static either. There'll be times when we develop things. There'll be time when we develop it in a year or two and use third party in between. It really is what suits the client best. That's how we would approach it.
The next question is: You've locked in contracts at 80% of planned volumes and have had substantial overage in the past. Are you seeing the same level or growing overage from your contracts? Can you give any additional color to the percentage attributed to cardiology in the recent two contracts announced? Over a medium, long term, how do you see revenue segmentation based on radiology, cardiology, pathology? Will radiology always be 80% of revenue?
Yeah, we are still seeing overages in our volumes, so we still see them on a quarterly basis, and we always get overages from the contracts. If the planned minimum is 80%, they tend to do their 100% within the first quarter, and then grow from there. So we're seeing larger than industry standard in terms of growth. In terms of the contribution from cardiology, we mentioned with U Colorado, it was around 13%-15% of the total contract value, so significant, you know, material, in that way. Vancouver Clinic, it was pretty similar, around 10%-12%. So yeah, it is a, you know, fair size of it. The revenue segmentation, I think radiology has a massive head start. We're already 10% of the market, so clearly it's a bigger base.
So will it be 80% of revenue? No, it'll be greater than that because it's not based on the minimums, it's based on what they've previously done. But cardiology and pathology. Well, cardiology is already starting to grow and pathology will follow, so.
Yeah. So in any particular deal, you know, we think that radiology will be between 70%+. If you add cardiology and pathology, that may take, you know, between 20%-30%. But clearly, that, that's as we see it today, it could become bigger if pathology becomes bigger.
Thank you. The next question is: Stripping out the $149 million non-cash gain from 4DMedical, does that NPAT figure mean we are seeing the lengthening of sales cycle in North America?
No, I think all it's showing is that we reached 30%, knowing 30% NPAT. Knowing that one of our biggest implementations, there would only be two months worth in the first half. That's it in a nutshell. Obviously, it will contribute six months worth in the second half, and the second cohort of Trinity will contribute a bit over five months. So that's why we think there'll be a second half bias.
Just on that, the AUD 149 million is a pre-tax amount, so it's not a case of taking the NPAT of AUD 171 million and take a AUD 149 million from that. But I've seen some commentary around that with the AUD 20 million profits, only AUD 20 million. That, that's not the case. That's obviously got a tax effect on it as well. Just to be clear, the AUD 149 million is a pre-tax, non-cash gain.
Thank you. The next question is: With the share price reacting, reacting negatively today, what specific business metrics or milestones should individual investors track over the next 12-24 months that you believe would naturally translate into share price growth if delivered?
I think the things we've talked about today. You know, the fact that we've been able to grow first half by 30%, and we know there's a stronger second half, and we've talked about that. The fact that all our implementations are large ones, you know. The Trinity Health and UCHealth Colorado, two of our biggest deals ever in our history, they're all on track to, you know, get implemented within the next few months and deliver material revenue step up. And then the fact that we've sold, you know, in, in six months, what we used to sell in 10 years, and all of that revenue is ahead of us as well. I think they're all, they're all the, you know, signposts along the road.
Thank you. The next question is: Can I clarify regarding the volume-based pricing model, please? The volume is the total images going through the whole workflow, including what AI tools screen initially and what human radiologists look at without the AI's initial assessment.
When we talk about volumes, we're talking about our exam volumes that we get for radiology. Clearly, for cardiology, pathology, AI, that's separate to that, so we don't include that when we're talking about our volume-based pricing. If we have a model that's for AI or something else, we would charge that separately on a more likely volume-based methodology. But when we talk about our volume of numbers, it's clearly, you know, what we do for radiology.
And the other thing is, at the moment, we don't know of any autonomous, fully autonomous, reading. That even if they're exams that are deemed to be normal by AI, they're usually looked at by a radiologist. But our volumes are the ones looked at by a radiologist, which is the volumes created.
Thank you. The next question is: Revenue split in the last three years has been 46%-54%, 1H to 2H. Do you expect the same, larger or smaller split in FY 2026, excluding the impacts of FX?
Yeah, we you know, we spoke about it at the AGM, and I think we mentioned it on this call, that we expect it to be bigger than prior periods. So how much that'll be, we'll have to wait and see, but we certainly think the second half will be bigger because of, you know, we keep saying it, the implementation of Trinity, and the different phases of that.
Thank you. The next question is: Did any of the cost lines grow greater than expected?
Not really, no. I mentioned around the salaries. Advertising is already, you know, a larger cost in this period because of RSNA, but in terms of salaries and new employee costs, it grew in line with what we expected. And pleasingly, like I said, our EBIT margins still continue to grow.
Thank you. The next question is: Can AI usage drive even more efficiency in a radiologist's workflow if the viewer and reporter are connected?
Certainly, integration does do that. It really depends on the integration, although we have seen that even in unintegrated viewer and reporting, that we can eke out quite material efficiency. So it really depends on the applications themselves.
Thank you. The next question is: What is the average price per scan, and how has it trended over the last 12 months? How do you forecast that average price per exam scan growing as you get into these other pathologies?
We don't actually give out the average price per scan other than to say that the cost or the price has been going up over the years, we think because we can prove additional value. So those supporting years ago would have bought it at a lower price point than today. And again, you just... You can't keep raising price forever, but we've been able to incrementally raise it year on year, and I think we don't see that per se changing in any material way.
Thank you. At this time, we're showing no further questions. I'll hand the conference back to Dr. Hupert for any closing remarks.
Just to say thanks very much, everybody, for joining us, and appreciate the interest. Thank you.