Ladies and gentlemen, the greatest opportunities are not found in predicting the future. They are found in building what's next. Ladies and gentlemen, good evening. Namaskar. I am Nipa, your host for today, and on behalf of Aurionpro, it is my absolute pleasure to welcome you all to our Investor Meet 2026. Thank you so much for joining us here today as we share our vision, our strategy, and the progress in building the next phase of growth for Aurionpro. Behind every number is a strategy. Behind every strategy is a vision. Today, we bring them both together. Our commitment to creating the best technologies, platforms, and solutions that will define the future of banking, transit and mobility, data center services, and Enterprise AI has been proven and is palpable.
Today, we are here once again, innovating with purpose and delivering sustainable value to our customers, to our partners, and our shareholders. Over the next few hours, you will get to hear directly from our leadership team as they tell you about the role Aurionpro is playing in transforming industries worldwide and the opportunities shaping our future. Now, before we move ahead, I'd like to give you a small glimpse of our safe harbor, which I'll have on screen for the next few seconds. Can I please have the slide up?
Passenger touchpoints.
Now, let me quickly give you a glimpse of the agenda for today. Let me bring to your attention that at the end of today, we will be having a Q&A session where you will get a chance to ask any questions that you have to our leadership team, who will be more than happy to address any queries that you have. For those of you who are new to Aurionpro, let me give you a glimpse of who we are and what we stand for. Can I please have the audio visual?
At Aurionpro, we are driven by a simple belief to lead the next. Not react to change, define it. We have built something real. Now, how far can we take it? We move forward by scaling with intent. Same strategy, turbocharged execution. Profitable, high growth, a record pipeline, strong visibility. We are aligned to a multi-year AI super cycle across infrastructure and applications. 4 focused growth platforms. AI data centers, enterprise AI, banking intelligence, transit, and smart mobility. We operate in the white space between legacy vendors and AI startups as a full stack AI native enterprise partner. Domain, product, AI, services, all under one roof. AI native is not a promise. It is already shipping. Fintra is live. AI native credit is coming. A portfolio-wide AI replatforming underway. Tier 1 wins across banking, transit, and data center infrastructure. AI is our lever for margin, speed, and scale.
We are evolving from a regional champion to a global contender with a culture anchored in AI native thinking, deep talent, and customer obsession. Core DNA unchanged. Product superiority. Never fail a client. That is how we build trust. That is how we scale globally. That is how we lead the next.
Yes, I do believe that requires a round of applause. Thank you. Now, ladies and gentlemen, to set the stage for today's discussions, please allow me to welcome on stage our Group CEO, Mr. Ashish Rai, who will share Aurionpro's vision, our priorities, and the roadmap ahead. Can we please put our hands together and welcome Mr. Rai up on stage?
Good afternoon, everyone. Welcome. This is the fourth Investor Day that I have been hosting. Thank you for taking the time out of your days to come and spend a couple of hours with us. On behalf of the management team of Aurionpro, with Shekhar, Sunil Sawhney, Sanjay Bali, Sanjay Varma, myself, warm welcome. We will spend the next couple of hours, maybe the first hour or so, walking through, like we always do, what we have been doing. Our point of view of what is happening in the world, our point of view on how we think we can seize the opportunity for us, for our shareholders, for everyone at Aurionpro who comes to work every day. It has been a bit of a journey the last 4 years. I will spend a little bit of time sharing where we got to so far. I see a lot of familiar faces in the room.
Of course, the number of people keep growing every year. I still remember four years back when we hosted this for the first time. I know, way smaller audience. I think we set our point of view that we'll pivot the business away into trying to build out something all of us can be proud of. Trying to build out a real scaled-up tech player Rooted in India, rooted in Asia, going global that builds up real IP, not really focused on selling ours. I think we've come some way. We'll share some of that, and we'll mix the format up a little bit.
I think between Sanjay, Shekhar and Sanjay and me, we agreed that we'll spend a little bit. I'll obviously do the overview of where we are, then we'll spend some time with the experts actually talking you through some of the key aspects of the business. For those who've been coming all this while, we've had the presidents usually presenting that. This time around, we'll actually have the actual experts come and talk through the various areas. Then, of course, the presidents and I will sit down and handle any Q&A. The deeper, the harder the question, the better for us. We'll try to keep it entertaining for you and hopefully useful. As we start, many of you have seen this before, I don't like reinventing the wheel again and again. What really drives us hasn't really changed.
The core idea was build a DNA of a real tech organization when we started off. Build the DNA of not so much of sourcing the spec, like diversified services business that we were before was doing, but to actually owning the IP, owning the solution that goes and creates value at the client, and five different levers. One, we'll not really look at the popular industry labels. You're a software, you're a hardware, you're a SaaS. We will write the whole stack from hardware to software that goes into delivering value to the customer. We will do that in a completely AI native way. We will prioritize the creation of value and the capture of value through creating real IP.
We will focus on what needs to exist that doesn't exist yet and go and invent it, we will help the industry, we'll help our customers really look and creating value well into the future, which means we will help lead the change, we'll help lead what is coming. Everything that we do is rooted and it's always focused on what is coming, not so much on what we do today or what we were doing yesterday. That is the lens through which I think you should see the rest of the presentation. What is it that a tech player needs to do in today's world to lead their clients, lead themselves into what is coming? Leading the next is what Aurionpro is about. That is essentially what drives us.
We've never paid much attention to the popular industry labels, I think we'll continue to do the same. What we care about is what is the value chain? How can I occupy every single point on that value chain where I get an economic leverage? That gives me the right to create value, that gives me the right to capture value. That's what we are about. Many of you have seen this many, many times. We've not even changed this slide for the last 4 years since we've been presenting, but this slide existed even before that. We started the pivot with many of you in this room, with most of the Aurionpro management team in this room 4 years back. The idea was to go from a very diversified services model into a much more sharply focused global products and platforms player.
Build up a real scaled-up global tech player, which meant build real products. Building real products, again, you've seen those three circles on the right, is about choosing spaces. How do you really go about choosing that space? We said three things are important to us. One, there has to be a demand runway that we see in the space, right? Demand runway obviously is very well abused in the tech world. Everyone wants to see a long demand runway. The idea is there has to be something fundamentally transformative happening in that space for us to get convinced that there is a long demand runway. For example, transit, closed loop to open loop sort of movement across payments. It's not just us, the entire industry believe over the next 10, 12 years, the entire world moves to open loop.
A 10, 12 year runway when we started. We continue to see that. Data center space in India, again, the same thing. When we started with the space, the core sort of hypothesis was India's always been under-capacitized on the traditional compute. Of course, since then, the AI cycle has taken over, and we are even more under-capacitized. That continues to power on. Corporate banking, again, the idea was banks have under-invested on the corporate stack. Go and prioritize that. AI, again, the same thing. We went and chose spaces where we saw demand runway. That was 3 years back. It has to be a space where we feel the global leadership is contested because if we go out and build products, if we go out and build IP, it should not be a Visa, Mastercard kind of situation where there's no space for a third guy.
It has to be a contested space where we can make a play for global leadership. The third thing is it has to be a space where Aurionpro brings a unique competitive advantage. There is some advantage we bring to the table. As long as those three circles intersect, we choose the space we bet. You all know those spaces. We went after that with discipline. We went after that with a lot of energy. We built out the whole stack. We are roughly 5 years in, that's kind of almost at the second layer, finishing up the second layer. We've built out very credible regional platforms. We've built out leadership position in a lot of markets that we are in. The idea is how do you go from there to really building out a global contender? Externally, this was always a qualitative view that we published.
We never actually published a number against it. Internally, it was always a quantitative view. It was always every business unit inside Aurionpro has a number carrying from 21 till 30. I'll report in a little while about how we are doing against that number. We've never really published the quantitative view externally. The idea is it all ends with getting to a top three global player in each of those spaces that we exist in. How did we get to where we got to? There are sort of those things at the bottom. Five important things for us. One is, what are the really hard problems to solve? One of the reasons we got on this journey is because we did not see many players in India, even outside elsewhere in Asia, really trying to solve the hard problems, doing the hard R&D. That's number one.
We've, across every single business line, scores of examples of we've gone in and invented things which did not exist, whether that's on the AI side with Lexi, whether that's transit, whether it's data center side that we're building out. Hard R&D is sort of core to really building anything. The second is being completely open-minded when we say there is a blueprint for a space and there is a gap in that space, we've both built products as well as gone and bought out businesses which had those products. Right? We've done a number of acquisitions along the way. Again, I think most of you have seen it. The third is get to the best talent in the industry.
One of the reasons we are where we are is because we've been able to attract the scale of talent, the quality of talent, some of which you'll see during the session later on, that our peers just don't have access to. I think if you want to build. This is a belief I had, right? For a long, long time. I spent most of my career in enterprise software. Enterprise software is a very level playing field. There is no advantage to building software in the Valley versus building it in Mumbai. The difference between teams that build out the best software in the world versus teams which build average is knowing what that best product looks like. It's not where you build that software. It's a very level playing field. Knowing is a function of the talent that you attract.
That's why we prioritized attracting talent from mostly global players in each of those spaces that we are in from the top universities. That really differentiates us pretty much everyone who are at least saying they're trying to do the same thing. The fourth thing is operate to a playbook that's global. One of the other sort of lessons we had on why we've not seen scaled-up tech players, this part of the world, is because there's a playbook to how you run global tech businesses, and you've not seen too many players do that. Which is, if you say software, how you build your software, how you maintain your code base, where do you keep it, the commercial framework that you follow, every single aspect.
There is a way to do it, which we said we'll operate to the highest standards in terms of the playbook that we operate in. The fifth thing is be very, very disciplined on where we put capital. Products is a very interesting game. Service is a straightforward game, as many of you know. I put 1,000 people, I make 1,000 people of revenue. I put 10,000 people, I make 10,000 people of revenue. 10,000 people, I start sponsoring some marathons and all, but the economics doesn't change. Products, you need to invest a lot right up front in building that product. When you're building that product, you don't know whether it'll succeed or not. If you succeed, you don't know how much you'll succeed.
Allocating capital wisely in a way that I am making the bet while still being able to load balance the risk, diversify my risk, is very, very important. We said we will go about it with a lot of methodical discipline. Those are the five things that really mattered to us, and that's how we got to where we got to. Where do we go from here? Like I said, we are roughly on that second part where we said we'll platformize the IP, we'll start scaling our talent, we start trying to go global. The journey for us from where we are now, which is a credible regional champion in most of our spaces, to becoming a top global contender in each of those spaces, and that's the journey that we are on. Quick recap on the numbers on Vision 2030, where we are.
That chart at the top left sort of tells you FY 2021, when we started looking at pivoting away from what we were to what we are. INR 375 crores. That 375 became INR 505 in 2022, INR 660 in 2023, INR 818 in 2024, INR 1,117 in 2025, INR 1,400 last year. That's a CAGR of 30.3% over the last five years, which is, I think, fairly unique in this industry and which is kind of what gives us the confidence in the model that we built out, the team and the talent that we built out. Last year, many of you have pointed this out, and I'm sure everyone noted, we thought we'll do between 25%-30%. We actually did 20. Lots of reasons for that.
I'm sure somebody will ask a question, we'll cover that. We grew 20% last year, which is really, given everything that is going on in the world, a testament to the quality of business that we built out and really the quality of the team that we built out. INR 1,400 crores, 20% margin, which is what we said we will do. I think the main sign that I look for, order book grew very strongly going into the year. That's a very strong position for us to enter this year with all the other things going on, continue to go on. The war in West Asia What will happen will happen. I think we need to maneuver around it, which we are. Then the general global cycles we'll need to play with.
I think when we enter the year, this is the strongest order book we've entered the year with. This is the strongest pipeline we've entered the year with, and we feel good about execution last year. Also, in terms of new logos, that is the largest new logo addition we've ever had in our history. A very solid year despite everything that was going on in the world. A little bit more characterization of what really happened as we went from 2021 to 2026. Below that 30% CAGR. I think the way to think about Aurionpro is, we of course have segments, but below segments, you have a number of businesses that each of the segments run under the presidents. The way to think about it is a fairly decentralized group of more or less independent P&Ls.
I'll spend, again, a little bit more time on it as I go. How did that group pan out? If you see even at the segment level, while the organization as a whole was growing 30%, from one year to the other, between the banking and fintech group and TIG group, you see that variation, 2021 to 2022, TIG grew 100%. Banking was not growing. Almost 2022 was when we were doing a massive build-out of the banking product stack. You see the 11% and 14%, not much growth happening on banking till then. TIG was powering most of the growth. As you got into 2023 and 2024, banking was growing very strongly, as we got the new product stack out across lending, across transaction banking, a lot of other sort of vertical growth businesses that we had, 37% in 2023, 34% in 2024.
TIG continued to power strongly. Overall last year, again, TIG and banking more or less grew at the same level. Over a period of time, you'll see the mix roughly for the last few years has been almost similar, but the growth rates would vary between businesses depending on what's going on in the sub-businesses underneath. Overall, the leadership in terms of growth will keep on changing. What was happening on some of the key things underneath? Customers. And by customers, we don't mean all the customers that we service, but essentially the key customers who have a certain level of revenue with the firm. The number of key customers really grew from 100 odd that we had when we started the pivot to almost 400 now.
The number of key customers working across businesses with us has basically quadrupled over the last four years, and that is a very powerful metric for us to see because when we acquire customers, typically customers sign up with us for very long durations of time. The key to running this business well is to really get the right customer and then get on the journey with them over time. That is what gives this business power. That has quadrupled over time. Order book, we were not publishing for some time, so I'm not getting into numbers there. We published for the first time in 2023, so it's gone from INR 820 crores that we published at the end of FY 2023 to INR 1,800 crores. That's again, roughly two and a half times, or thereabout. Number of employees, 1,300. That's gone to 3,000 Aurionites across the organization overall.
That's roughly, slightly more than doubled over time. Revenue per employee, which was at INR 29 lakhs, that's gone to almost INR 50 lakhs now. That's up 65% over the period of the pivot. That's, again, a very powerful metric on the productivity that we're driving across the organization. Right? A fair amount of work done overall. In that same period, look at number on the top right. We've invested INR 1,000 crores, slightly more than INR 1,000 crores in either building out product through R&D or buying a product business in a space that we wanted to be in. That is a very powerful metric in terms of the long-term power that we are creating in the business. A lot of analyst accolades along the way. We've gone from being a fairly small player to one of the most sort of awarded player in the industry.
What really gives me a lot of pride is if you look a lot of the coverage, it's not the volume of coverage, it's where are you in that coverage, right? If you look at RiskTech, for example, squarely in the leaders category globally, top north, as north as you can get. IDC MarketScape, again, as north as you can get. You're going to see a lot more of that this year as we are rolling out a completely new product stack, which is far ahead of anything this world has ever seen. A lot of analyst recognitions along the way. That's what sort of characterized our growth overall, right? The growth leadership switching between businesses as it goes, but both businesses have continued to power strongly, and they will continue to power strongly as we go.
A lot of that is because of the product build-out, the various firms that we've acquired, teams which have joined us. To those of you who really know our acquisition approach, we never bought for revenue. We don't buy for revenue. We buy for two things. We buy for talent, and we buy for product. For us, whenever Aurionpro wants to be in a space, like we said, four years back, we decided the spaces we want to be, right? AI, data center, banking software. The spaces don't really change, but within the space, we have a strategic blueprint. We have a lot of gaps on those blueprint. When we want to plug that gap in that blueprint, there is two things, two paths you can take. One is go and build the product.
Of course, that brings with it a lot of uncertainty, a lot of investment. You say, "I know a business which has already built the product out. We respect, we admire the team, we admire the founders." They come and join us, and we get into that space. It's available at a reasonable value, of course. We are very value-conscious when it comes to acquisitions. We've done six in the last, I would say two and a half years or so. In each of those cases, the founders, the CEOs continue to run their business. In each of those cases, their business has grown. In each of those cases, their business is a lot more valuable with Aurionpro than it was valuable without. Again, some of that you'll see during the session as we go.
Those are the businesses that we acquired over time, like I said, we'll continue to allocate some capital towards it as we go, where it makes sense to us. How do you think about Aurionpro in terms of what really is driving that economic power? Some of it I discussed last year as well. The way to think about it, like I said, is think of Aurionpro as a fairly decentralized group of independent, mostly product P&Ls, a couple of services P&Ls. For us, call it the operating engine. From the operating engine, any of those businesses, the ask is very simple. There's an example of that below. Depending on the business you are in and the demand environment, you expect the business to grow, let's say, between 25% and 30%.
We expect the business to give you an EBITDA of 20%, 25%, 30%, depending on the business you are in. Then you expect the business to convert 70%, 75% of that EBITDA to cash. You convert it to cash, you put that in the company's pocket. Company has a decision to make in terms of allocating that cash out. There are two primary ways we use that cash. One, build new product. As I said, we always have a gap in the blueprint where we want to build again. That's the R&D dollars, which is where over the last five years we've sunk in, I think, almost INR 500 crore in terms of R&D.
The second user for that is to go and do an acquisition of a product which we wanted to get in, which is again, we've done six of those over the last two and a half years. Each of those will continue to go. That is where the capital goes in. Over time, these businesses move from the right, which is the investing engine, to the operating engine. Again, the same cycle starts. We believe that is an extremely powerful model. We also believe that's a fairly unique model. Over time, if we keep selecting for the right R&D investments, not diversify too much, stick to the spaces we already are in, but keep building in the white space that exists, or in adjacency around us, or we continue to acquire wisely, I think this is a machine that cannot be stopped.
That is the compounding engine that will go on for a very long time, regardless of the cycle that's going on in the world, regardless of what business is sort of the popular business of the hour. The idea is, are you in a space which has a large enough demand? Second, are you building out a product that's competitive in the world that adds value to the client? You get a right to capture value only if you create that value. As long as you're doing that and you keep allocating your cash wisely, this is a business that cannot be stopped, will continue to grow. That's essentially how you think about Aurionpro if you said you want to wear an economic lens on how that operates.
The key strength, if you said, I want to think about one thing in this business that's not product, that's not R&D, that's not people dependent, it is the installed base. Ultimately, this business is as powerful as the power of the installed base that we build out. Why is that? The reason that is the case is because in the spaces we are in, most customers select you for 10, 15, 20 years. Over that period of time, the economic power that you have, the leverage that you have, the ability to create value that you have changes over time. Whether it's banking software, no customer buys banking software for anything less than a 15, 20-year horizon. A lot of our customers have not left us for the last, I would say, 20 years. Same thing with transit.
You may sign a 10-year contract, but honestly, when you come in and build a payment stack around us, you don't want to go ever. That gives you a lot of time to create more value with the customer. Same thing with data center space. Almost everyone you work with, no one has a 2-year plan in that space. Any strategic partner that you're working with has a plan for the next 8, 10 years in terms of investments you grow with them. In each of our businesses, the real power in the business comes from the installed base that you build up over time, and that's what a large part of this game is. How do you keep increasing that circle?
For some of our main businesses, if you look at banks, we obviously work with, I know, 250+ banks now, almost 300 banks. The large regional banks that we work with, the large regional tier 1 banks, over the last 5 years has gone from 5 banks to 35 banks. That is a very powerful base to build out. Same thing with transit from 8 to 25 over the last 5 years. If you look at the hyperscale, neo-scale sort of players, again, gone to the strategic relationships that we have, the partners that we work with has gone from 1 to 4. That metric is very important. If you look at it, for most years, 75, 80% of your revenue would come from that existing base.
As that existing base grow, because most of the new business that you sell in the year doesn't really translate into revenue in that year, it translates into revenue in the future years. 75, 80% of your revenue, certainly north of 80% of your growth is coming from that existing base. Right? That is very important. That is sort of the power behind the machine. Really mission-critical systems, very high retention rates. Then once we get into the customer, the outcome is that chart that you see If you started at 1X today, which is INR 1 of revenue, that will almost double over the next 4 or 5 years from that customer, if you retain that customer. That is essentially what gives the economic strength to this business.
If there was one metric I had to track, which was not R&D, which was not people, it is the installed base and where that's going. That is going very solidly for pretty much all of our businesses. That's where we got to today. Many of you have been with us over the last four or five years, and I think you've seen us grow from across pretty much every bet that we made from being a reasonably good small player to someone acquiring scale, whether that's transit, data centers, a lot of aspects of banking software from Interact to lending to transaction banking. AI, where again with Arya setting up Lexi Labs and all, we are becoming one of these very, very unique players who can really bring our strengths to bear.
As we are doing that, as most of you would have noticed, there is obviously, and I think some of you were talking to me before this thing started, there's obviously a lot of change in the world. If I was to give what is my point of view on what really is changing, and I think all of us read the same news, so it's not very unique. There are four or five things to look at. If you look at the chart, the main chart on this, right?
Our point of view is if you think this is the parallel of some cycle we've seen in the past, I think the single important piece of information that we gathered, and we looked around and we looked at various sources overall, is that there's been no parallel to the CapEx cycle that we are seeing in the world today. There's been nothing like it. You look at the sum total of those two metrics, the software and data center, that's 10.5% of world GDP. I think the scale of CapEx that we are seeing, there is really no parallel to this scale of CapEx in the last 200 years or ever in terms of just as a percent of world GDP. That is bound to cause changes.
Even you go back to 1800s and the railroads boom and all that apparently peaked at 5% of the world's GDP, that took a very long time to get to that 5%. The difference here is not just the scale of investments happening, it's also the shape of the curve. It's much more vertical than it's been in the past, that's again bound to cause changes. Some of the spaces that we are in, AI native software is a category, I think we'll spend some time on it in the session after. It's a completely brand-new category that's getting created and getting built up. Has gone from almost nothing to almost a INR 30 billion category this year, INR 40 billion category this year almost.
A very big space, which is where we say, what does software look like in the age of AI as against what software is today or used to be? Data center CapEx, depending on which number you want to trust, anywhere between $5 trillion and $8 trillion from now to 2030. A massive build-out going on a global scale, a massive build-out going on in India. Again, a very, very large CapEx cycle to ride. Transit and smart mobility, almost $1 trillion of CapEx. That number, at least most of the reports, if you try to do on average, is somewhere between $900 billion or $1 trillion. Whichever way you look at it, almost 500 cities across the world building transit at the same time, as well as revamping the network. That's a massive spend going on across the world.
As I said, transformation of the payment platforms is a big key driver as well as the new lines coming up across. If you look at the Western world or the legacy transit thing, a lot of it is moving from legacy payment setups to the open loop setups that we built out. If you look at the emerging world, a lot of net new metro lines or transit lines coming up and infrastructure getting built out. Enterprise AI as a category, again, whose number you want to believe, but whichever way you look at it, the primary driver is you had software as a category for quite some time. The idea is if you had a lot of the operations pool, the services pool, the labor pool, join the technology category, the size itself becomes quite big, $3.7 trillion by 2030.
Again, depends on which number you want to believe. Within the banking software space, we believe banking software as a category is going to triple or more than that over the next five, six years because of the impact of Enterprise AI. A lot of change. The way to look at this is not really the size of the opportunity, but just look at the investments happening in the space and what are the changes that causes. If you look at what has happened historically when you've gotten into a super cycle, especially in the tech world, every super cycle has created a completely new category of players. If you go back and you say when you moved from mainframe to PC, you had a set of players who were large.
You had a completely brand-new set of players who came in and take share across semiconductors, across operating systems, across PCs, right? Same thing when internet came along, you had a set of players. You move to a new set of players. Same thing with mobile and cloud. Almost a $2 trillion category got created over a matter of years. The thing to look at is it's not the there's these really large players. For example, if you look at cloud and you look at mobile and you say, "Okay, there's a Salesforce and there's a ServiceNow, and there's very large players." It's not just the very large players that get created. There are literally hundreds of net new players in the category that created every time there is a super cycle.
With 100% conviction, I think we believe, we are in a brand-new super cycle, probably of a scale we've not seen. Which means a lot of the categories will shift in terms of the players who play those categories. We were preparing for this change for quite some time. As you can see, we've methodically gone and acquired the assets we needed, whether that's Arya.ai, whether that's built out the data center space, whether it's built out on the software side, or increase the surface area of software that we had because we wanted a bigger surface area of play inside the banks as you go in, and that's all the acquisitions that we talked about. We believe, the industry is ready for a change in terms of players, for a shift in terms of the leaderboard.
That is the change that we intend to play for across each of the segments that we've talked about. We've gone in over the last few years to go and really acquire the strengths that we thought we needed, the surface area that we thought we needed, the talent that we knew we needed. Go and invest the capital that we needed and that's brought us to this moment in time. If you look at the classic five-layer AI stack that we look at, the top three layers is where we play. Top three layers is where we believe most of the value will get created and captured.
The infrastructure that goes into delivering the compute, obviously, we leave out the SEMIs and the energy side of it, infrastructure that goes into powering that compute, the models and the agents and the application layer where we believe most of the value will get created. How do we play each of those spaces? On the application layer, which is the layer at which you touch the customer. Mobility and payments, a big play for us. How that gets transformed in terms of the tech stack. AI native banking software. Again, we'll spend more time on each of these. We believe smart mobility is a significant opportunity over the next decade in terms of the transformation need that exists. Sanjay Bali and Steve have built out an extremely high-quality global business.
We have probably the most integrated end-to-end offering in the transit space backed by extremely strong R&D, backed by very strong product ability, hardware to software to payment networks. We believe this is a space we can really build out a large global business in. AI native banking software, again, playing on the application layer. We believe the software space has changed. It's changing. As you move to the world of agentic execution, as you move to the world of AI, the need for software that the bank has does not change. What that software is needs to change. We had this definition of I think I hosted an event last year in December, and we kind of came out with our definition of what is banking software 2.0. I think we'll spend some time on it today.
We believe that is a completely net new category which, most of the legacy incumbents in that space will struggle to compete in. That is again a category, like I said, we believe will triple in size over the next five years. We believe we have one of the strongest games in town, together with Arya.ai and the product teams that we have to play that game. Enterprise AI, which is if I don't want to start on the application layer, I want to start on the intelligence layer in the enterprise. Again, with Arya.ai, with Lexi, we have a very strong play. What are the really hard problems to be solved to bring AI into production? Data to model is easy. Model to production is very hard. How can we solve those really hard problems? That is the focus on enterprise AI.
Again, a very large category to play in. Again, a very large category where you'll see a whole set of players emerge over time. Data center solutions, I spent time on it. Again, we believe we are in a super cycle of investment. Demand is like almost drinking from a hose pipe. The question is: where do you want to focus on in that business? How do you really create a very powerful business that generates value for the customers, for the shareholders? Then underlying all of that, how do you build out a very high-quality, specialized services set up across each of our domains to help our teams go and deploy solutions at the clients to handhold our clients into adapting to change that is coming? That's essentially the play that we have in those layers, as we go through this change.
Like I said, this time around, we've sort of changed the format, we don't have the presidents coming in and presenting each of the businesses. What we would like is I leave services out, but for the first four layers, I would call in some of the experts in this room who will come and talk through, specifically what is going on in that space for us, what are we creating? What gives us a right to win in that space? Then I'll sort of round off at the end. Getting into each of those things that I defined, right? Smart mobility, AI native software, enterprise AI and data center solutions. That is largely where we started, what did we do, where did we get to, where do we want to get to.
To cover transit and smart mobility to start with, I'll call in Puneet. Puneet's in this room. Puneet is one of the original founding team members of a company called SC Soft that we acquired. Like I said, Transit, when we got in, we started off essentially most of the business as a services player, then slowly we acquired IP, built IP across every single point on the value chain. Puneet has been one of the key architects of the whole journey and Sanjay Bali's team built up Oracle's business. I'll give the floor to Puneet to walk us through and maybe Puneet just help us understand the opportunity and
Sure
what the space is all about. You want to use this?
Yep. Thank you so much, Ashish sir. Good evening, everyone. Most of us in this room or anywhere in the world use transit or public transportation almost every day in different formats. It can be use of public, it can be use of private, or it can be even use of shared mobility. When we look at the stakeholders in this particular sector, whether these are government companies, public transport operators, or even system players like us, we all thrive to achieve three things. One is, can we have better access for mobility so that people can get onto the board? Second is, can we make operations more efficient? Third, and the most important is customer experience.
Over the decade that we have seen the growth in public transportation, these three have been pivotal in ensuring that the transport operators have profitability, as well as they are able to have as many people on public transportation that they can. At the heart of this is digital transformation. One of the pivots of the digital transformation is essentially smart mobility. Smart mobility can encompass many things. It can be related to payments, it can be related to having better assets. It can be related to better bus stops. It can be many other achievements that these transport operators want to have in public transportation. Aurionpro at its heart today is one of the key leading players that looks into payment for public transportation. When we say payment in public transportation, we essentially define it as automatic fare collection system.
In layman words, this is ticketing. Today, if we see it, what are the reasons for changing in the transit? As I said, digital infrastructure is at the heart of it, AI is definitely knocking at the door. We also see that when we talk about smart ticketing, it is the person, whether it is he or she, who wants to decide what kind of token they want to use in public transportation. It can be cash, it can be QR, it can be any card in your wallet, or it can be even your NFC on the mobile app. The third thing which most of the governments have declared is electrification or EV charging of transportation ecosystems. These are the three pivotal changes that we are seeing in public transportation.
Considering all of this, if you see the volume of the market that is available today, it is about close to $578 billion for smart mobility, and it is about $50 billion for automatic fare collection. Let's look at the world play, what is happening here in the decade. There are about 480+ metro projects which have been declared in 56 countries, which are going to use smart mobility and automatic fare collection. About $975 billion of CapEx investment is being done by governments for this digital transformation. We have about 27 Indian metro cities who are looking at having better ecosystems for public transportation. India is definitely at the heart of it, and Aurionpro is one of the leading players in the ecosystem, and perhaps the only player who is managing both bus as well as metro transit system.
We also see that the government is looking at about 1,000 additional network about railways and metros that they want to implement. Aurionpro is at the forefront of this ecosystem, what changes we are seeing in the payment side is that, first of all, worldwide, as I said, people want to use their own token to pay for the transit ticketing. It can be an open loop card, which is a credit debit card in your pocket. If you look at India as an ecosystem, it can be a national common mobility card, which is RuPay or UPI.
There is a new concept into the market called as account-based ticketing system, which most of the transit operators want us to implement, where any token that you choose is linked to your bank account and the entire play happens at the back office or the back-end software of this transit ecosystem. Something that we defined as SaaS-based transit ticketing system, essentially it moves away from the hardware that we deploy. The last one definitely is that mobile takes the precedence. If you look at any ecosystem today, whether it is transit, it is payments, everyone is looking at using wallets, looking at using NFC through their smartphones, and this is where the digitization is happening. We believe that by 2030, most of the ecosystems will move towards having mobile as the major token for public transportation.
When we look at Aurionpro's strength to deploy these transit technologies, Aurionpro essentially plays at the entire life cycle of this transit ticketing solutions. Essentially, what is happening is that Aurionpro is the only company who has capability to design, develop, manufacture, assemble, as well as customize the solutions that are required for urban mobility and public transportation. Aurionpro has its own manufacturing bases. We are one of the companies that do manufacture in India and are ready for Make in India. With this manufacturing capability, we are delivering to the world. We are locally manufacturing it, but delivering all the projects globally as well. Let's consider an ecosystem of how public transportation works. Imagine that you're going to a transit ecosystem, whether it is a metro or a bus.
You go to the system, you decide to take out a credit debit card from your pocket, or you may choose to pay using your mobile app with QR code or UPI payments. If you are a user of cash under the socioeconomic strata, you may go to a ticket vending machine, which is an interface at the metro station, and buy the ticket using cash. Aurionpro is doing this at this point of time. The second thing that you do is when you want to enter into the public transit system, you're going to go and tap this card. You're going to go and scan this card or validate this card, or you're going to go and also tap your mobile phone. Aurionpro also does the manufacturing and assembly of these gates and validators.
Once you tap the card, essentially, the entire linkage goes to the back office where the fare engine starts churning the entire journey that the passenger wants to take. This string is sent to a payment gateway, which is either a connection to an MPGS or a CyberSource based on Visa, because Aurionpro has partnerships with them. This goes to the acquiring switch, which is essentially the bank. Aurionpro is perhaps one of the only companies who owns five layers across this public transit ecosystem, which is rider interface, acceptance device, fare engine, back office, and the payment gateway. The only thing that Aurionpro doesn't do is, it is not a bank. Hence we have partners that are there into the banking ecosystem that connect with us on public transportation.
In terms of our proof in the market, today we have more than 22-plus projects globally implemented along with India. We have more than 20,000-plus devices on the field, which are running for about 10-plus years in operations. We have more than 16 countries where we have projects today established and running for almost more than five years. We have more than 120-plus transit experts into the system which look at different ecosystem, whether it is R&D or designing of the solutions. In the last few years, some of the major wins that we have achieved is Chennai Metro, Delhi Metro, Mumbai Metro, where we have been selected for station-level devices, in Chennai and Delhi. Mumbai is the only project where we are doing an end-to-end ecosystem, including 512 metro stations, where we are going to deliver the central clearing house for MMRDA as an authority.
Apart from this, globally, we have been also awarded projects in Saudi, U.S. and U.K. We are very proud to say that Aurionpro is only the second company to have a certified reader which the government can use in public transportation in Saudi Arabia at this point of time. To achieve all of this, we also have very sustainable partnerships across the transit ecosystem or the payment ecosystem and are certified by the likes of PCI, Mastercard, Visa. We are strategic alliance partner with Mastercard, where at this point of time, if any public transportation ecosystem has to go live with Mastercard, Aurionpro is the first choice for Mastercard today globally. What this does in total is it helps us to build an ecosystem worldwide where we have put up an illustrative model that how this is changing for us economically.
Look, I think overall, like Puneet explained, right? The stack itself, we built out probably one of the most powerful integrated end-to-end offerings in the world. Puneet is pointing out in terms of the economic model itself, why is it that you can't pick a space? You say, "I do software. I do AFC. I do back office software. I do the devices, the validators." Why really spend all the time and effort and investments building out the stack that Puneet walked us through? It essentially comes down to the economics. As the business matures over time, we believe there's a 30-point advantage to being a full stack provider versus being a point solution vendor in this space. The ability to really go across the chain in terms of need for the transit provider is what that 30 points is all about.
That means the long-term return on capital that this business will deliver is dramatically different from most of the players in the space. The idea is create a really global powerhouse of a business that can play across the spectrum, but more than that, deliver a much more superior return on capital that we invest in this business. We have invested a lot, as you would have seen over the last few years.
When we look at AFC, which has been designed for India and built for the world, and we look at the market funnel that is available. The total accessible market for us is about INR 16 billion, which is there today in 2025. It's growing up to INR 51 billion by 2034. If we look at the SAM, which is the accessible market for us for open architecture and AFC open loop, it is coming to about INR 22 billion because of this digitization that I spoke about by 2034 again. The opportunistic market or the SOM which we really have is about near-term winnable opportunity, which is about INR 0.5 billion-INR 1 billion through 2030 as a cycle. This is majorly fueled by the Indian metro ecosystem and the bus ecosystem, as well as the international business that is coming through secured orders.
Why we are showing these figures is because today, Aurionpro has already implemented projects around 16 countries, and these vary from different markets as well as different ecosystems. If we look at the first tier, it is essentially the developed markets as well as the developing markets which have the largest opportunity stack present as of today. For us, Aurionpro is already present in these markets with existing projects. The second stack essentially are tier 2 developing markets where mostly we have the payment ecosystems like Mastercard and Visa which are investing into the projects and allowing the governments to digitize public transportation. The third and the fourth stack are essentially greenfield projects where government is trying to implement new age digital infrastructure for public transportation.
For us, the most important thing is that for all of these 16 economies, we are already certified, present, and available to deploy these solutions. A proof point of this is that we have already delivered a nationwide ticketing project in Maldives along with Mastercard, which is an account-based ticketing system for an interoperable solution where we manage buses, ferries, as well as high-speed boats. We are already ready to take over tier 2 markets and implement almost the same solution. Aurionpro is the only Southeast Asian, or an Indian company to have a master service agreement in the North American market at this point of time for public transportation. We have been awarded a master service agreement by California Transit Program or Caltrans and have delivered about eight to nine operators by now. One of the major operators is around the Disney park called as Anaheim.
We are also ready at this point of time for the U.K. market where our solutions are already certified for it. If we look at in the nutshell of what is happening today in public transportation, we essentially look at the next five years in a certain pattern. One is we want to scale international AFC wins, which is in the Middle East, Asia and Africa. We want to capitalize on our existing projects in the Indian ecosystem and build more metros and bus systems for Government of India, including digitization of payments. We want to move ahead on the eco-chain and the ecosystem where we have recurring transit revenue from a SaaS-based model, which we spoke about an account-based ticketing model.
Today, Aurionpro, which started as a transit ticketing and a fare collection company, wants to move to urban mobility where we want to diversify into more ecosystems which will help us to also gain the market in smart mobility avenue, which is airports, EV charging, and mobility as a service. As to what Aurionpro does globally and what it has achieved over a period of time, we have a small video that we want to showcase to you of what we have done globally for public transportation.
Cities move. Every journey begins with one simple moment. The payment. Tap. Go. Done. Fast boarding. Effortless flow. Scaling fast. We are transforming Indian transit from Noida and Delhi metros to Chennai's intelligent gates and Mumbai Metro's advanced systems. We digitize millions of journeys for Uttar Pradesh and Haryana Roadways. Our in-house tech clears bottlenecks, ensuring every journey is seamless. Our platform delivers real-time data and revenue insights for smarter networks. Our central clearinghouse unites agencies and financial institutions in one intelligent ecosystem. Partnering with Mastercard, Visa, and RuPay, we link banks to the gate, enabling seamless open-loop mobility and true account-based travel for passengers everywhere. To accelerate this global scale-up, we are redefining transit with an AI-powered enterprise platform. We bring intelligence to the ecosystem from AI-enabled ticketing kiosks and facial recognition to predictive maintenance.
We reduce passenger touch points while boosting efficiency, helping agencies scale seamlessly and manage massive networks with precision. Proven at scale globally. From Metro de Quito and CTtransit to Maldives ferries and California's open-loop systems. We power Sacramento's light rail and Nigeria's railway modernization. Our platforms manage extreme volumes worldwide. Trusted by agencies to move cities reliably. The future of urban mobility. From Tanzania and Canberra Rail to Costa Rica, Yucatan, and Lima Transit, our footprint expands. When fare collection just works, cities breathe easier. We power the infrastructure that keeps urban life moving invisibly, reliably every single day. We are Aurionpro.
That was all on Transit. Thank you so much for the opportunity, Ashish Sir.
Puneet and the team in Transit have built out an absolute powerhouse business. Not only is the span that Puneet talked about, it's also the. We built out a tech stack that can hold its own anywhere in the world. This is the time for Sanjay Bali and team, and you can ask him during Q&A, for us to really go from being a very strong tech solution with a lot of proof points of success to actually becoming a scale global player in what is a very large market. Thanks, Puneet. We'll move on to data center services. Like I said, a massive area of opportunity for us.
Bhaskar is one of the real veterans in the business, has built out this business from zero over the last five odd years that we built this business out. He runs enterprise business. Data center is a large part of it. Bhaskar to help you understand what the opportunity is in the business.
Thank you.
You want to use this? Want me to run for you?
No. That's okay.
I know how to run slides. Yeah.
Good evening, everybody. Most of us aware of the data center, that is market and how the people, everybody is talking about. I was just thinking before coming to stage, if I had been delivering this conversation a few years back, say three years back, I would've said data center is growing. Why it is growing? Because of digitization, localized post digitization, the adoption of digital with Government of India push, not only in India but also abroad, data localization, and adoption of cloud. This was the landscape that was there for three years back. The AI comes. Next evolution comes, AI comes in. Whole landscape changes. Landscape becomes exponential. To be frank, frankly, if you ask me, is it still we are able to confidently quantify what is the roadmap? Perhaps no.
All the big ticket, big companies, we are all guessing and maybe whatever we are talking about now, after one year, you may find the figures are totally different. That's where the evolution of AI is happening, and that's why it is redefining the complete landscape. Good thing is that the redefinition is on the upward scale. That is basically few of the pointers that what is the total $5.2 trillion AI CapEx that is going to happen. As far as India is concerned-- Sorry, I'll come this side so that I'm not obstructing the way. As far as the India growth is concerned, India, of course, when the world goes, India also basically is growing. Basically, if you see from the normal government sector reports, the power demands itself is going to 165% of current usage as far as India is concerned.
From one gigawatt that we are talking of installed base, we're talking of eight gigawatts, and round about what is that value comes to, it's around about $30 billion of opportunity comes in. The whole landscape has changed. AI came into picture, India also adopted quickly, and that is where the whole humongous amount of opportunity also got created. We are not left behind as far as the industrial progress is concerned. What are the typical pointers? I'll quickly go through. It's basically sovereign AI push that is coming from Ministry of Electronics or IndiaAI where they have committed around about INR 10,000 crores of budget for that, and basically 38,000, if the figures I remember correctly. Around about, yeah, 38,000 GPU that is basically going live. For whom? For the people to use this and for the companies to use.
This is a big push that's coming in because the data localization AI has to be in India rather than we depending on outside. That is one part of it. The second, of course, we are well aware that in the public domain, the hyperscalers are taking big land themselves, and data center operators are taking big land. This is one thing that it has never happened before for one industry taking and actually invading the land parcels that's available in Mumbai, Chennai, Hyderabad, and it's going to go to tier 2, tier 3 cities. That is one proof point that is coming in. The second is basically kind of five years, six years back or seven years back, typical rack densities was how much? 4MW-5 MW. That's how it was designed.
Cloud came into picture, that 5 went up to 10kW per 10kW to 15kW to 20kW . That is the second generation of change that happened. The third generation change that has happened, and is happening is right from 20, now it has gone to 70 to 150 to 200. There's a quantum jump on engineering. That engineering, when the IT changes, the rack changes, corresponding the engineering which is going to fuel that particular compute also changes. Most of the thing is that now the cooling patterns of the data center changes. There's more liquid cooling that is coming into picture. The power densities, how they transmit power there. We are talking of high density DC coming into picture because no more AC power can supply these kind of high density servers.
This is as far as AI coming and changing the data center requirement. At the end of the day, there's only obviously one constraint that will obviously come, though India has not yet reached to that stage. Very slowly India will also reach when basically you have the power, there would be a constraint in time very soon. That will be Government of India, everybody, we have to think that how do you supply that much of power? How do you have the capacity to generate that much of power? Obviously, there are ways where how the different countries across U.S., Europe, have circumvented the issue. This is something that eventually we will have to also circumvent this part with the lesson learned by the guys who are already ahead of our curve.
This is as far as the few pointers what is changing in India. Comes that what is the TAM that we are talking about? Globally, it is $1.9 trillion. That is basically excluding the IT hardware that I'm talking about. Just MEP or non-IT kind of a CapEx build. That is there. The SAM as far as India is concerned, taking from the various credible reports that is there, we're talking of $9 billion. The SOM is around about $5 billion, that we are talking about for us. With this, I'll move on to the next part. This particular exciting times are not only for us, it's for every data center player across. Each of the data center players have worked out a different strategy to grow and help the clients. Question is, we have thought our strategy.
What we have thought about and what we are working towards, it's not that this is one-day affair that we are talking, and just because we are presenting, this is the new thought. This is the path that we are already taking. Perhaps we are making it public today that how we are addressing and how we have gearing up for this big transformation in this industry. We have segmentized the whole life cycle of the data center. The first segment into 4 or 5 segments that we have segmentized. One is basically your data center design consultancy services, which actually starts from site selection, site assessment, to master planning, to your assessment of your concept design. Sorry. Again, my mistake. Assessment of the site, concept design, detail design, construction management.
This is one part of the segment that we operate on, and there are good amount of players who are also growing in segment 1. We are also growing in this particular segment. The second segment is basically your design and build segment, where design and project execution. In most of the terms, it is also called general contracting. This is the second segment that we work on. There are a good set of companies who are operating in India. We are also one of them, and we are also growing. They are also growing. These two segments covers this part of it. That comes in is the third segment where we are basically stepping into the manufacturing of in-house modular products. This has become our necessity.
If we have to be more agile, we have to actually be relevant to our client to service them well, we have to have those modular in-house capabilities to be designed at our factories and our manufacturing facilities in Aurionpro. What we have thought about it and what we are manufacturing, I will just come to it. What we are working on and we are on the path is that we are kind of redefining our existing transit manufacturing facilities in India and expanding into the data center products. One of those data center product that is typically that we have identified and that is basically helping us in the projects are basically DFMA.
DFMA is nothing but the modular structure or prefab structures of various utilities, chill water pipelines, your cable trays, your cables, your fiber runners and all having a modular prefab, and then put it across the data center for faster turnaround for the customers and less dependency on the skilled labor that the industry is currently struggling with. That is the first product. The second is, of course, liquid cooling. There is a huge demand of secondary fluid network. Anybody at all who is connected to data center industry will know. If we are not manufacturing all these things, slowly and slowly you will be written off, we will be also written off from the industry because your dependency on the outside market will be more where there is more demand, less supply. That is the thing.
The third is basically AI edge data center pods, which is just about to come and going to the tier 3, tier 4 as an edge data center, AI-ready pods, container-based data centers. These are the three things that we are currently. This is the third segment that we are working on, and we have started work on that. The fourth. I will cover the fifth segment and then come to fourth segment. With the advent of AI that has come in, it has created huge potential or threat of the cybersecurity. As such, cloud and normal cloud had already created a lot of headache as well as cybersecurity was concerned. AI has created another one. This is one segment that we have been focusing on for the long time now.
In the cybersecurity space, we are working with Government of India on certain, to identify and create a security framework for various financial institutions and also manufacturing institutions. How do they come out of this particular threats, and what are the remedial action to be taken? Cybersecurity is one part that is a focus area in our data centers segment, and followed by hybrid cloud services, which is a foundation. Basically, we give the IT transformation roadmap. We create the IT setup, help the clients to set up the IT private cloud setups in their on-premises, off-premises, managed cloud services for them. This is one segment that is the foundation segment for when you go from journey from non-IT to IT. Coming back to when the data and AI. Here is basically what is with the data center generating so much of data.
Now with this so much of data, our decision-making also has become complex. Now with this, what we have? We have our own homegrown applications which covers data lake houses, and from the data lake house with the data that is generated, we have our own AI offerings, which is again homegrown, which gives very specific use case basis LLMs which solves the customer problem. Or we give the AI platform to these guys, to our clients, so that they can test their LLMs and see for observability, security framework, traceability. All that is basically is the fourth segment. If you ask me only, where are we? We are actually growing in all the five segments. How does it help? One segment also helps the others, especially when it helps us to become a trusted advisor to our client. The question is this.
Perhaps we are only uniquely placed, one of the few only uniquely placed, I will not boast so much. We are perhaps only uniquely placed who understand the domain, who understand the business, we understand the complete 360 degrees, and when we talk to client, We actually are solving their business problem, not following what customer is giving. That relation, that traction, helps us to bond better with the customer. That is where it is. Few of the things, highlighters that we have done for the past few years. As far as the implementation is concerned, 200+ megawatt we are already pursuing as per the project implementation is concerned. We have so far for the past two years, we have already delivered 100 megawatt of IT load. These are the one, two segments.
That is first two segments that I'm talking about. In this, basically, our tier 3 and tier 4 certified data centers, there are enough approved points that we have delivered for the customers. The second part that comes in is when I say go to segment 4, now this is basically, we have that Arya.ai, which in any case we'll cover later on as per the agenda. There we have actually taken those offerings from Arya, clubbed it into our requirement, put it into the one data center offering, and giving it to the customer to solve their business problem. This is as far as this is concerned. One of the focus area, if you ask me honestly, this five segments requires very, very deep learning about their engineering skills.
Now what we have been doing over the past few years is basically, we have been focusing much on technical knowledge, engineering skills, and this is what have been focus area, and that is paying us their dividends. Then only you will be able to solve the problem when technical business leadership is what is required to solve business problems. Only technical knowledge doesn't help, only business knowledge doesn't help. It is when both of them get married, that is what it is. We have those own softwares, own tools as far as data center is concerned, own LLMs, own AI LLMs that we have. That is also there. Cybersecurity tools that we have in-house, and we use the other cloud services to support the clients and to solve their business needs.
Now this is what is right from your data center to IT, to AI data engineers. The pool of talent in Aurionpro is not less than anybody if you compare with any of the top bracketed companies. This I can confidently say, that we are at that level of technical knowledge as far as the industry knowledge is concerned. A few of the big-ticket items that I wanted to follow or tell. That recently we closed INR 350 crores of order, single AI data center ready that we are currently in the implementation. Our data center business actually grew from 70%-80% for the past three, four years. Four to five years, if you ask me honestly. This was our rate of growth. Today, when we are sitting here, the proof points are behind us.
That is why I am talking to you confidently that we are geared up. It is not a marketing that one year we do and fail next year. We are there. The data speaks for ourselves, we are basically proceeding in that direction. Of course, from HCS and cybersecurity, we are offering SOC as a service. Like I said, the Government of India, we are collaborating much in a deep way. Private cloud, CERT-In empaneled, we are CERT-In empaneled. That is basically one of the close point to give a traction with the customer to enter their data center. That's another thing. This is where it is. We are advanced partner to AWS, GCPs, Red Hat, VMware, and all these things.
This is in nutshell our answer and our preparedness to address the market reality and market opportunities so that we help clients and we grow ourselves. I think this is what I had to say, Ashish.
Thank you.
Thank you.
It's okay. Thanks. We can clap louder for Bhaskar. Look, I think Bhaskar has built out one of the highest quality businesses in this space, brick by brick, methodically. Today, if you look at the space and if you are an operator who's investing CapEx right now to build out a business, we are a very unique partner who can handhold the operator from the AI engineering stack that rides at the very top, right down to the bottom of the infrastructure layer, all the way through. This is one of the largest opportunities over the next decade. The data center build-out that this world is seeing is something we've never seen in our lives. No matter what we think the parallel to this is in the past, I don't think we've seen anything like the build-out that we've seen.
One of the other questions that I think I get asked a lot is this like a one-off thing that goes off in two years? I think as a nation, if you look at India as so woefully under-capacitized on the compute that we need over the next decade, that this has not even started yet. It is a massive area of opportunity over the next decade. A lot of the compute in the world went after training. As you move to models reaching a level of maturity, getting closer to the consumers, you need inference closer to where the customers are. It doesn't matter which frontier lab you talk about, the largest base of customers sits right here. You want compute closer to the customer. We are not even at single-digit % of what is needed. This is a massive build-out.
We are happy to be a part of it. Bhaskar has built out and Sanjay has built out one of the highest quality teams in this industry all across. This is our time to play this. That's the Data Center story. Trying to understand what happened to my slide. Can someone help with that? Now I moved ahead. I'm handling this. Coming on to the next part of what I talk about, you go to the top of the layer. Transit, huge opportunity for us over the next decade. Globally, we will build out the strongest powerhouse business we can build out in Transit. Data Center, massive opportunity. The biggest CapEx cycle world has ever seen. We built out a very high-quality business, and Bhaskar pointed out proof points of us reaching a stage where we're ready to scale.
That brings me on to AI-native software, which is, again, we talked about, we expressed a point of view last year on where we think the world of software needs to go. Alban will help us walk through some of that. Alban has this unique, one of the most interesting roles in the world today. He's an enterprise software person, and been all his career, which is the Software 1.0, and he's in charge of trying to lead us towards what is Software 2.0. He's also in charge of leading us towards using the AI stack to build our own products. We'll hear from him on what goes on. Chloe heads consulting for us in Alban's team, and she moved on to try and help us build out what is the first of the AI-native stack that we built out.
Again, I'll let her talk through it. With that, Alban, the clicker.
Thanks, Ashish. Hi, everyone.
Let's go, Alban.
I hope you're going to cope with the French accent. We're going to change things a bit. Ashish told my intro, but I spent basically most my career in what we call now banking software 1.0. In fact, we work very closely with Chloe. We wanted in fact to come back of what is banking software 2.0, what it means for us. In fact, it's not a new definition for those that were with us last year. We in fact come up with that definition, that diagram. We just polish it back with a new color and dip a bit more into what we mean by that. What is the banking software 1.0? It's basically a system of record.
I think if you go to any banks, if you are working in the banks today or some asset manager, you probably are interacting with those system. It's built with traditional code. You pick your Python, Java, whatever it is, it doesn't matter. At the end of the day, it's a codification of a process. It's only deterministic. If something happen, it will give you an answer. If the codification doesn't happen, the system just stop. In fact, it's built for an operator, a human to use it. It will have a lot of click, menu, and button. How do you usually in the system of record, those system 1.0, they are good when they have a lot of menu and a lot of operations. This will change, in fact, and we expect the future to become from a system of record to a system of action.
What we mean by that, it mean that the build itself will change. The software 1.0 still exist in the 2.0, but at the base, there is an AI-native basically capacity that will change the process to be codified by basically being an outcome-driven software and to be as well probabilistic. I think now we all understand what models and LLM can do. We understand that they will never say no. They will always come up with an answer. That's exactly what a system of action will do. Those system are as well defined not for human anymore. They are defined more for agent to agent. Of course, human is still in the loop. The interaction is of course very different. We are not clicking everywhere. We are basically talking to the system, whether we are typing or with voice. What does it means?
It means that the value will totally change. The creation of the value over the time. We see today that we are at that inflection point where the S-curve will bring us to the top of that value. Why? It is basically the two things that you have at the bottom. The first one is our addressable market will increase. I think we have seen that at the beginning. The investment going into AI software, we believe that in the banking software business will quadruple by 2030. Today, we are at around INR 100 billion-INR 115 billion. That will go into the INR 400 billion-INR 460 billion. More than the software TAM, we expect basically that software TAM to collapse with the labor TAM. Think about it.
If we have a system that now replaces some of the action or some of the process done by an operator before, we can now address that TAM. This is of course very big. We are talking about multiple trillion. Some of the sources say 4 trillion-5 trillion, but that gives us basically a very big place to play. The nature of the software being different, that composite AI, it is a totally different way. The way you build the software, the way you deliver the software is totally different. We expect the unique cost of the transaction cost to decrease around 50%-70%, bringing again, more value to our clients and higher velocity for us. The time to ship is totally different. In the old world, you go to a bank, you spend some time, you go live, the value is unlocked.
Here, we are talking about sprints, maybe weeks, maybe we will go to days. We are looking maybe less value but more often. That will bring us basically to that S-curve. What it means for us. Again, domain is everything we have. The banking software will be totally rebuilt and repriced. Where we play today, we have shared that before. We are in the lending business with our Integro brand. We have trade finance. Transaction banking, of course, is one of our biggest franchises in terms of banking software, but we are as well in customer communication, capital markets. That is today where we play. Again, with our engine that we have, we may increase the surface area. It means that INR 460 billion TAM that I was mentioning before will become an addressable market of around INR 5 billion today, but likely to be INR 7 billion.
You should get a yield of INR 435 million. What we want to do basically, we will build for all of those domains basically a new banking 2.0. What we want to do today is I want to hand over to Chloe, and we want to share what we have done on the trade finance, which is the first live banking software 2.0.
Thanks, Alban. A change from French accent to Chinese accent. Show of hand who sort of know about trade finance. Okay. Two. Yes. Thank you. What is trade finance? I have to do this myself. Yes. Picture a seller in Mumbai selling goods to a buyer in Paris. The seller doesn't want to ship until they are paid. The buyer doesn't want to pay until they receive the good. There's generally a trust gap there, and that's where the bank step in to bridge that trust gap by offering instruments like letter of credit or guarantees. 80%-90% of that INR 10 trillion global trade depends on trade finance, and that is the space this AI-native trade finance solution is serving. What is Fintra? What is our AI-native trade finance? Essentially, it does three things.
One, it provides that operational backbone to support banks to run their trade finance operations. It is essentially what Alban described the banking software 1.0. It serves as a system of record or system of action. All the swift messages, the different type of transactions or instruments, the type of services banks offer to their corporate customers, it provides all that operational continuity for the banks. Second, it provides a suite of AI agents that does the intelligence and judgment that human operator used to perform. It offer out of the box a suite of AI agents that are already delivering certain tasks, and on top, it provides that AI-first architecture for new agents to keep getting built.
Lastly, and very importantly, it provides the governance to bring AI to actually working in production in a highly regulated banking industry without the assurance and capability to answer to the regulator, how do you know your AI decisions are the correct ones? You will never be able to realize the true ROI there. That's the third thing that Fintra does. Since the launch of Fintra in April this year, we've had many conversations with banks across the globe. India, Southeast Asia, Middle East, Europe. We've received really positive feedbacks from all these banks and customer conversations. There are cases where it's a new market, we were not invited to participate in a trade finance RFP. We got the opportunity to showcase the system the next day we received the RFP.
There are cases where the conversation was a little on hold or stalled, and we got the chance to showcase Fintra, and a stalled conversation turned into a leadership meet with Ashish in Singapore and turned into strategic conversations, not just on trade finance, but across many product lines within Aurionpro. Why do banks love Fintra? It is something that is quite unique. It brings together the traditional software and AI intelligence. It blurs the lines between what software used to do and what intelligence used to do. I think maybe the best example to give you is this. We provide a conversational interface for corporate and for the bank trade ops team.
Imagine a ChatGPT or a Claude experience where you just type in what you want and the system reacts and completes that action, be it inquiry the database, or you want to initiate a letter of credit. This is powerful in a sense, and this is unique in the sense that if you just add intelligence to a 10, 20, 30-year-old software, you won't be able to bring together this capability. First, you need to interpret what the person is asking. Second, you need to trigger the corresponding actions accordingly. Be it query the database, be it trigger a workflow or call a machine learning model or just call an agent. You can argue this whole thing is like a super agent that helps with the bank's trade finance operations.
Yeah, I can keep talking about trade finance for the whole afternoon, and I think I may enjoy it more than you do. Instead of doing that, let me end with a video that give you a flavor of how the system actually looks like.
Global trade is a INR 10 trillion market. Payments got fast, but trade finance stayed slow, not because of volume, but because of the old system still running it. 70% of trade documents are rejected on first presentation, and the cost isn't just time, it's trust. Trade finance doesn't need another layer on old systems. It needs to be rebuilt, AI native from the core. Introducing Fintra, the world's first AI native trade finance platform built from the ground up for the agentic era with humans in the loop. End-to-end coverage for both corporate customers and banks. Documentary credits, guarantees, collections, supply chain, structured trade, and trade loans. At its core is a suite of AI agents, autonomous specialists that read documents, screen every party, score risk, examine compliance, and recommend clauses, all coordinated under one governance envelope. Lastly, the conversational experience.
A revolutionary interface for both corporate clients and bank operators. Agents don't get autonomy by default. They earn it by proving reliability over time. Every decision is gated, auditable, and reversible via CGHP, governed autonomy protocol for regulated banking with a system-wide kill switch in human hands. Since launch, Fintra has drawn attention across global trade finance media and broadcast, and it's already in live trials with leading trade finance banks across multiple regions. Fintra, trade finance built for the agentic era.
Thanks. Thanks, Chloe. This is just one example and then more to come. I think
I want to come back on what we mean by AI native and why the software moat is getting stronger. Previously, system of record, my data is in the system of record. This is how I defend my moat against maybe my competitor. I think the game is totally different and is way more powerful today. In fact, the compounding agent flywheel is what's going to be my moat. Briefly, if we look at that loop, again, what I was saying in my first slide, the time to ship, why is it so important? The first thing is we will deploy, like the trade finance that we have seen, we'll deploy an agent to cover trade finance, lending, payments, operations. Those agents will start to be used by the banker. They will approve. Of course, human-in-the-loop may not agree, they may override some of the decision.
They may decide otherwise. That's fine, because what we're going to do here is we're going to capture every single decision, every single correction that is happening in the system. Remind as well that we are working here with a non-deterministic system. That is gold, because that data allow us basically to fine-tune and basically to create an even more powerful agent. That will compound. The next agent that we'll deploy will get that data. What does it mean? It means four things for us that you have on the right on the screen here. Embedded, as I mentioned, we are not anymore a process. We are part of the cooperation of the bank. It means that we are really inside and embedded.
It's very difficult to displace unless you spend a lot of money before moving from one system to a system of record was probably quite easy because the process was outside. It mean by FY 2030, the moat will be really structural for us, not contestable. We are working in a bank environment, everything we do has to be basically regulatory grade. This is something we are very good at. In fact, everything will be audit ready . We need to understand the provenance, we have a specific platform for that, which is Lexi, that my colleague, Deekshith, will spend a bit more time to explain. Ultimately, 100% of the decision will have to be explainable and traceable. As I mentioned, the goal is basically be context rich or data rich, if you prefer. I think now everyone understand what context mean, right?
Those decision that we capture, whether it's been correct or overwrite, will become basically something that will be very difficult for a peer to get access to. As we deploy more and more agent to more and more clients, this will basically compound. If you bring that to the scale of event that we process today or in our system, 1.2 billion, you understand that that compounding effect will make our moat very, very strong. Where does it bring us? I think it bring us to become a new category leader. If you look at our peers today where Chloe and I used to work, the banking 1.0, they have the domain expertise, but they are not AI native. They are only deterministic. They are not able to capture a decision. Software 1.0, basically. The frontier lab, the LLM provider, Yes, they are AI native.
They have that capacity of reasoning, but they don't have the domain expertise. It's only basically the combination of the two, which is Aurionpro today, and there's a brand Aurion AI, that will increase our capacity to win. If I have to summarize it basically with four things, which is what you have at the bottom, our domain expertise, 20 years of banking in the model, our data flywheel that I was presenting just now. Of course, the trust of our existing client. I think Ashish was mentioning that what is really our moat today, if I just apply it to the banking, we have 25 of the tier 1 banks. That's trust that we can use and we can leverage, which is that 1.2 billion of event. The architecture, which is being truly AI native. That's about us.
I think in the next session, we'll deep dive a bit more about what Aurion AI is.
I think the sort of important thing to understand about the AI native software space is in front of our eyes, a completely new category is getting created right now. Banking IT, if you look at it, is give or take a $400 billion space, everything put together. Banking operations is a $4 trillion space. What Alban was talking about are the TAMs sort of collapsing together a lot. How does the AI native software space or the banking software space, we think, go from $100 billion space to almost a $400 billion space? How does it increase three times in size or four times in size over the next five, six years is essentially because of these TAMs crashing. You need to be a very special type of vendor to really be able to play that game.
The need for software does not go away, but what that software needs to be changes quite dramatically. What Chloe showed through Fintra, I think it's important to understand this is today, and I have zero doubt, the most advanced trade finance software in the world by a mile. No one can even touch it. It's that special. The idea about where does the agent start, where does the software end, it's very hard to tell. The whole thing is like a trade finance agent. It's completely AI native. Every bank we have talked to wants to work with us on it. It's a question of how can you scale that business methodically. That's number one. Number two, we've obviously, as many of you would have seen, invested massively over the last six months in terms of building our AI native stack.
Every single domain we are in, we will have a brand new build from scratch, AI native stack coming out, and you see a lot of that as you go. This is a massive prize to play for. We methodically put together pieces we needed to play this game, and the time for it is now. I think we have one of the best teams in the market today in terms of building this stack out as the world needs to change. That is on the AI native software space, and that brings me on to the enterprise AI side, which was the last part we said we'll highlight. Some of you know Deekshith. Deekshith is the Founder of Arya.ai. We acquired Arya, I think maybe two and a half years back, thereabouts.
Yeah.
I think the thing to understand about the AI space is the entire world has two years of experience on AI. Really the number of people, especially in this country, who've done only AI, only in the financial services space for the last 15 years, you can count on your fingers. Deekshith is one of those very rare veterans in the space, given what the space is, and he has his own unique point of view of where the space is going and obviously how we want to play that game. With that intro, I would hand it over to Deekshith to.
Thanks, Ashish
to take it from here.
All right. A lot of AI discussions from right from the beginning, right from the data center, transit, and we have seen what the latest Aurion AI, we have been doing it. Again, we'll cover a lot more details what has been powering in all those sections. Before going into the slide, I want to just play what our latest platform has been all about.
It starts with imagination. A simple thought, this could be better. All enterprises, small to large, run on systems of record, engagement, and action, and tying them all together are workflows. Somewhere inside those workflows is a moment where you know there has to be a better way. A way that could read the documents itself, check the data automatically, bring in systems without waiting on manual steps. A workflow that could think. Introducing Infinity OS, where your natural language prompts transform into workflows. Create task agents. Connect systems and tools. Infuse intelligence into every step. Build powerful AI workflows on a single canvas, and bring them to life in minutes. Your workflows can finally think, decide, and act, because the workflows you wished existed can finally be built.
Thank you. All right. That you have seen the small glimpse of what are the new platforms. We have built on, basically it's an agentic-based workflow automation tool. I'll talk about that in the later phase. Before that, I want to set a context what's happening in the whole industry. As you know, the race of enterprise AI is moving rapidly, very fast. You might have been following the whole news, what's latest models coming up with the latest technology and latest accuracies, and it has become a full-fledged commodity now. Every phone, every app, I think every day definitely you'll be using at least more than 50% of time is using conversations through an AI. That's exactly what's the biggest problem here with an enterprise also here that we've been thinking that the best model will be ruling the whole enterprise AI.
That has totally changed. Now, the edge of making a best AI is not just the only solution now. When we talk about enterprise, we need to make sure that enterprise plus the deep dive domain information has to be infused both together so that the enterprise wherever you've been working, it has to be collated into those things. Just having the best AI model would not be helpful at all. Just, I want to give you a kind of a what trend is happening. Alban has already covered some of the aspects, but I just want to give you overview. If you see the whole last 40 years of the whole banking industry, the whole banking technology, whatever the kind of INR we have spent on, has been spent on these two different systems.
One is a system of records, where you have your payments information, balances, what kind of transactions happen. All has been developed here, like your core or like a ledger system. The different system which has been evolved over the period of time is the system of engagements, where it's more of a touch point, where you have your mobile applications, consumer portals, or even the teller screens, et cetera. These two, if you see here, these are more like making the operator do a lot of more work easy. The new trend which is evolving is now the latest, which is system of actions. There's a huge trend which is almost like taking away the operations work from the operator to the system automated itself.
What exactly it does, it tries to understand the situation, it tries to understand what it has to do, probably act and take a decision, and also a teller decision. That's exactly is a major chunk which we need to be done in the enterprises, and that's been getting evolved. As we've been discussing, we have seen best AI models and we have seen a lot of pilots being run over the last two, three years that we have seen recent reports also. A lot of enterprises have been running a lot of pilots, but what is the value? Where is exactly the value which has come up now? Recent report says that only 67% of the value has been moved from pilots to production.
Have you observed why exactly that kind of success when the people are investing a lot and the models are getting more and more better? The only problem which is definitely different from the consumer point of view to the enterprise point of view is the domain. That's exactly, these are the major gaps which we're going to talk about more, where just to give a simple example, what exactly I wanted to take a reference of what AI native or AI model is all about. Assume that a freshly graduate guy, probably you've been hiring him for your own internal processes. Would you expect him to do your daily task in expert manner? Definitely not, right? Because he has just come to your job and he still has to understand your own processes. He has to understand your workflows. He has to understand your systems.
That's exactly the problem. You don't have to have a base model with having any kind of expertise of what the bank is going through. That's exactly we need to have an expertise and the AI system working together so that we build a proper AI native system. Of course, you know about the banks or any financial institutions, it's a tons of different system. I'm pretty sure every financial institution should have at least 300 to 400 different systems. When you drop a tool on top of these particular systems. Talk about a financial system or talk about anything or taking a simple example of credit assessment. Probably a new application comes into the picture, and then the particular application has been sent to a model.
Model will say that, "Oh, I've read all the application details, and I feel that the particular profile is just having a limited information, and I would reject this particular case." If you give the same particular application to a credit officer, he says that, "Oh, this particular person, I know him well. He's been doing business with us for the last 15 years, and he's been also having a corporate account with us and also have a family banking." Why exactly that the particular assessor has this information? Because he knows that every day he uses multiple different systems. That's exactly what happens in the bank. Every night, the reconciliation happens. Every different system, and there's a layer of reconciliation. You can't have a model to just run it in an isolated manner and expect the results.
You need to have that particular context being switched from one place to the other. You know, as we've discussed, like why is the need of a Software 2.0? Some of the systems are almost a legacy system. We can't expect that the same value to be coming up from these models when you top it up on that. If you are expecting a regulator to sign off any kind of a new particular model to run or any kind of enterprise software to run. We expect it to be audited. You can't expect this legacy system to produce the same kind of outcome to be coming up in this particular auditable manner. The last and the main important aspect is implementation muscle . Like I said, model is just like giving a just prediction.
It's almost like giving a value, what kind of input it has to give. That's just a 10%. The value of what it provides is the remaining 90%. What kind of information it has to bring, which system it has to look at, what kind of a regulations, what kind of a workload it has to look into that, what kind of a next actions it has to do. That's exactly the whole environment which makes it work. These things are the major legacy problems which I've been looking at. I mean, every enterprise has been trying to solve these things, but unless and until they solve these kind of major gaps, they won't be able to come up with a proper AI native or enterprise software.
We discuss about the movement of the shift from legacy systems or system of record, system of statements , or system of actions. Let's take an example how the system of actions look like. The same example of a credit assessment. Now, the application come to the model or a proper system which is having a full-fledged context. It knows that where exactly it has to look at. Which knows that if the customer has come up, it has to look into different system wherever this particular references are there. It now understands the full-fledged context of the bank. It gives a prediction, says, "Oh, now I need to know what I have to do." It says that I have to decide whether a particular model has to give a prediction in a rightful manner or not. That's another major aspect. Then you know that, right?
It knows the workflows. It knows the risk. It knows what next it has to happen. Like, it has been connected to almost, like, different systems. It knows that it has to drop a email to a customer for asking any kind of a missing information or probably create a CAMs report for them so that it gets approved to the credit assessor. Every particular action is there inside the system. Every transaction, whatever kind of a decision it is made, it has to give a report. It has to give an answer that why that particular decision has been made. That's exactly what it is. Of course, there is a human in the loop here. Ultimately, every decision, whatever it is right or wrong, that is fed into the proper feedback loop, these transactions are made sure that we train the model more fine-tuned and all.
That's exactly what we bring Aurion AI, which does all these particular things in a different manner. Let me talk about Aurion AI. I think Ashish was mentioning about last 6 months, we launched this particular new vertical altogether. Why exactly is that? Because we have seen there's a need for to do this. It can't be just like having a Software 2.0, but we need to make sure that the whole set of problems, like the expertise, the kind of problems which are like model gap, all this have to be solved in a different approach. You know, the whole Aurion AI, under one umbrella, we have 4 different verticals altogether. Arya.ai, you must been hearing a lot about this. We've been doing this business from the last 13 years. Before that, the founders and the core team have been doing AI from much, much before.
We've been running live almost like with 100 plus customers in financial institution. More than 80 plus models are already live, doing more than 350 million-400 million transactions every day. You see what's happening here. We know what is a kind of patterns emerging. We know what kind of regulations are changing. We know what the requirements are coming up from financial institution. This is the learning where a legacy Or a kind of a frontier models weren't able to come up, because making the model refined on the public data weren't able to come up with this particular inside the particular databases allowing in this enterprise weren't able to come up. That's exactly our strength. Coming to our Lexi.ai, that's our governance layer. You must be hearing a lot about the research work which we've been publishing.
This is that kind of a trust layer. This is exactly what the regulators would be asking about in any financial institution or any kind of a regulated nation. Unless they until they see that the results or whatever the kind of prediction the model has given, it needs to be audited. It needs to be proven why exactly that particular decision has been happened. You can't have a model, whatever, we bring in a model and then have that particular trust to be done in the one go. That's exactly what our governance layer does. You have been hearing about Aurion Trade, which is one of our Aurionics product. We have already built Aurion Trade, and we have also built Aurion Credit. We'll be building a lot more number of products of our other portfolios. Just to give out 2 minutes on this here.
If you have seen the kind of a difference between what Aurion earlier products versus Aurion this product, every particular layer in this particular product is being infused through an AI. Every particular touchpoint, like Chloe was mentioning about the kind of a conversation it has. It can ask any kind of information you want at any point, either from the customer portals or from the compliance officers, or it can be even admin person. It can ask any information. It gives you the natural language process because this has been backed by our own banking LLMs. Lastly, wherever we have been doing business till now with our flagship models of iCashpro+ or SmartLender, we've been doing an add-ons over there. This is not a rip and tear. This is more like giving an add-on payment. Give you an example of iCashpro+.
We've been doing almost more than 20-plus agents for that particular product itself, like payment initiation agent or cash flow forecasting, all these things on top of that, so that we make sure that we just not just have a full-fledged product of Aurionics itself. We just want to give them the flavor of how the agentic looks like so that they keep on having an upgraded version of Aurionics future. The major idea of what I wanted to put up in this Aurion AI, the whole fusion if you see, we don't have any kind of a bias or We don't mind where the particular enterprise is being placed, how mature the enterprise is all about. We wanted to land with any of our product and then expand.
Either it can be Arya.ai's product of credit assessment with looking at a simple bank statement and giving a prediction, or it can be Lexi just talking about governance layer, or it can talk about fully native approach. With our install base, we can install any kind of a multiple agents which we're talking about. That's the kind of a flavor which we're trying to do with Aurionics, Aurintelligence, et cetera. Just want to highlight you what's happening in our applied AI engine. As you have already seen our Infinity OS, that's one of the new approach which we have taken. With a simple natural language query, you can build your own workflow tools. This particular system sits inside the enterprises connecting to more than 200 particular different connectors, where you can directly say that I want to do a whole workflow automation.
We wanted not just to have an AI agent just do one step and then give a handover to the particular person. We want to make sure that the whole workflow is automated. That's exactly what we wanted the AI to be envisioned. That's our Infinity OS, I've been talking about the whole conversation, which is our banking-specific LLM. We have two specific models. We have a seven billion and 12 billion model. We are trying to make a couple of more domain-specific models, which we are trying to get more data into that. Not just into banking, et cetera, we also have launched recently Launch AI, which you might have heard. We'll soon hear from the public announcements there. This is more of a public announcement system where you can hear it from airports, et cetera.
That's the kind of work which we're doing with TIG team also. The other kind of things which we have already discussed with our install base, we are doing almost all the kind of different agents. On Lexi, this is one of the core offerings from the regulator perspective. We have already established three different research hubs: Mumbai, Paris, and London. We wanted to get talent from all the global particular pieces so that we don't want to miss out what's happening in different verticals. Recently, we also launched DL Backtrace. We optimize in such a way that any particular verdict it takes, it's almost like near real-time. Every particular at every stage, the particular decision is being audited. It tells you why the particular decision has been made. That's the explainability model.
We have released Aurion TFM models because this is one of the only unique propositions what we have from Aurion AI. Everyone knows how the particular data is stored in the banking and financial institutions. It's not just like free text over there, how the LLM looks like. LLMs cannot read any kind of system which are there in rows and columns. That's exactly what Aurion MSP will get feedback into it. Basically, it looks into your structured information, which is rows and columns, and gives you a prediction. It wanted to do a fraud analysis. You want to do any kind of forecasting directly connecting into your data records. We also have our different framework on the guardrails, which is a group. You have in-built guardrails, and you can also have external guardrails in-built into it.
That brings me to talk about where we are heading and what we're trying to achieve. Whenever we talk to any kind of investors, the whole TAM keeps increasing. At least we have done our research, we've been looking at it almost like an INR 3.5 trillion to INR 4 trillion enterprise market. Where we function, which is BFSI market, it's almost like around INR 820 to INR 850 particular point. The more interesting aspect is: What are we trying to achieve here? Where are we trying to kind of focus on? The focus point is, where our customer bases are. The total sum comes into almost INR 350 million plus. Again, this is a kind of point where I can also tell you that why this share belongs to us is because we are already inside banks.
We have been working with one of the top-tier banks in every particular different regions. You talk about Southeast Asia, you talk about India, you talk about Middle East, or you can talk about any other particular region, we have been working with the banks, financial institutions. Of course, like I said, the four verticals which I have seen it in the AurionAI, not just only talking about agent integration now, we wanted to kind of build any agents, existing systems on top of that, or talk about RAI modules, or talk about the governance modules. That places us in a safe manner to kind of have this particular share in a rightful manner. Lastly, before I end the talk, I wanted to kind of leave with you two different points. AI that acts. That exactly is what we are trying to achieve.
We don't want to have a kind of an AI which kind of converse with you, which gives you a kind of a good summarization, which kind of gives you a rightful manner, then you do the remaining of the work. We don't want to do that. We want to make sure that AI which totally acts and completes the whole workflow so that you take the whole final decision. That's the only thing you have to do. While doing this, we want to make sure that it is accountable, because unless and until this is accountable, it cannot be kind of accepted by any regulation. That's exactly where we are heading towards the AurionAI part, and we are being rightly placed by doing all these particular points. Yeah.
Okay. Deekshith today, together with his co-founder, Vinay, I think they have built out one of the highest quality AI labs across anywhere in the spaces that we specialize in. I think the thing to understand about enterprise AI as it rolls forward is, it's one thing to look at the output coming out of the frontier labs and what the consumer models can do. It's another thing to make it work in the enterprise, where a business is trying to run a real business process that creates value and you say, "Okay, how do I really bring AI into this process?" It's exceptionally hard if you said the business was a regulated business like a banking or insurer. Right? This is a long journey over a fairly longish period of time before you start seeing businesses realize value from AI.
Of course, there's no one who's not experimenting with it. You'll see that business after business, this is again, one of the largest opportunities of our lifetimes. I think Deekshith and Vinay have built out one of the highest quality businesses to go after that. We looked at some of the opportunities in front of us. I think the important thing to look at it is we have actually not really changed our strategy for the last four or five years since we had the advantage of planning that pivot. I stood in front of many of you four years back when we did our first Investor Day and we said, "We need to pivot away from running a services style business." I think one of my opening line was, we see services margins going to zero, partly because of AI. This is all pre-ChatGPT.
Many of you were in the room then as well. Partly because of AI, partly it's not a game for a small player. We want to focus sharply on building out products. We want to focus on a few spaces. Honestly, last 4 years, we've not changed those spaces. It's very same spaces, whether that's data center, whether that's transit, whether that's banking software, whether that's AI. We've gone about methodically putting the pieces in place. Of course, the demand environment overall has changed quite dramatically over the last 4 years, while we were preparing to build out businesses in each of those spaces. Today we are a very, very credible player across each of these spaces. Where do we really go from here? To me, we are in a space where, 1, the demand is clearly established.
Second, the team is in place, at the scale and the quality that we needed. Third, the IP is in place, right? We went in 2 niches. Each of these niches is going to explode over the next 10 years, whether that's transit, whether that's data center, whether that's AI, in terms of the sheer size of demand in those space. The thing to understand about AI is as the value increases, it doesn't increase linearly the size of the space. I mean, you're talking real compounding taking effect as the value gets captured. The second is, as we operate in this space, we'll stay true to our atoms and bits strategy. We're not a software, we're not a SaaS, we're not a hardware. We will ride the whole chain. Third, we are used to operating in regulated industries, especially banking.
We know what it means to create an AI native stack that is trustable, that is explainable, that has observability, that can operate in high stakes industries. That is the whole purpose of setting out research, that what Deekshith was talking about, setting up research labs in Paris and London and Mumbai, and going after what does it take to really build out a stack that can operate in the high stakes industries like banking. Then stay true to creating our IP, stay true to creating this engine over time. Those 2 things, I've talked about it before, I think I talked about it probably quite a bit. The core DNA of Aurionpro does not change as we go into it. If you said you go to Aurionpro as a fairly decentralized group of, let's say, 14 or 15, mostly product businesses, a couple of services businesses.
What is my ask from each of the businesses? There is only 2 asks. Number 1, product superiority. When Aurionpro builds a product, we need to know very clearly what makes us superior to every other option that the client has in the market, every other product who thinks they are doing the same thing as we are doing. We need to know very clearly why what we build is superior. I am the best transaction banking product in the world because A, B, C. It doesn't need to be 100 reasons. It needs to be 2, 3, 4 reasons. They need to be very clear reasons. That clarity goes into the roadmap. That roadmap goes into the product. That's why we win. That's number 1. The second ask is never fail a client.
When a client decides to trust Aurionpro with the business, it does not matter what they bought, it does not matter what their team is, it does not matter what contract we sign, we make sure the client succeeds. Reputation is everything in this business, and we earn it one day at a time. We've built out a stellar reputation in the industry. There's a reason why every institution wants to work with us, and that is what gives us the competitive edge in the market, right? There is no third ask that I have to any of those businesses. Product superiority, never fail a client. That's core to the Aurionpro DNA. That's what makes us succeed in the market. That is what will give us the right to win in each of these spaces as we go. Where do you really go from here?
I think before I press clicker on this and go into it, you saw a lot of numbers. You saw the size of the market. You say, "Okay, what is the serviceable available market?" For definition, when everyone says SAM, the way we do it inside Aurionpro, Chloe runs the consulting set up. You saw her today. We've got a bunch of people in the team. We go in and actually look at it account by account. These are the markets we are going to be in or we are in. These are the specific named accounts in that market. This is the size of business for that product in the market, and that is the size of the serviceable market.
You go from there into what is actually obtainable for us, which means what is the business which is up for grabs over, let's say, the next five years, six years, what part of it can I take, right? When you say TAM, that is the total market. When you say SAM, that's I am in the market and I can service it. When you say SOM, that's sort of what we think is obtainable over the next four, five, six years. We don't give forward guidance four or five years out in any case. Everything that I see in terms of what the potential value of businesses could be is all directional and what it could be. If you look at it, the size of the market that we are going after for each of those spaces is a very large market.
It says XDC because I took out the data center TAM from it. There's no point in talking trillion U.S. dollar numbers when you say target market sizes. We took that out. It's a very large space. The specifically sort of serviceable addressable market for us, again, a very large number, again, excluding the data center space, just because of the noise it causes in terms of numbers. A very large space. We have built out a lot of very powerful compounding engines over time. We did not talk about specific software because you've heard it. The last few years, lending, we are a very strong business in Asia. We are the number 1 corporate loan origination software provider in Asia. Very strong across retail, SME, auto as well. You'll have Shekhar and Sanjay during Q&A. You can ask them that. Very strong business.
You got transaction banking. We are probably one of the strongest vendors in Asia in that space. Poonam is somewhere in the room, runs that business. You can ask her about why over the last three years, we've gone from a credible player to basically the number one player in terms of win rates in that market, in the markets that we play in. V`ery large space. Mobility, capital markets. Capital markets, again, we built out the business with Fenixys over time. Data center, we talked about. Enterprise AI. And if you look at the value for each of the segments going into 2030, that is what we feel. When we do our planning, we look at low end to high end. That is what that we feel is the potential value of the obtainable market for each of those spaces.
If you look at it, overall adds up to a pretty large number. That's a very wide range you'd look at it. I think the point that we are looking at is you're sitting in a number of spaces which are going through a fundamental shift. When you talk banking software such as lending, transaction banking, a lot of the other things that we do, the whole category is getting reinvented as we speak. Over the next decade, you will see fundamental shift in the market which. To be honest, we were always out to build a very strong banking software business. We knew there is change coming. The scale of the change is very different from what we see. Banking industry has had this pecking order for a long time. You got the really large players, they're worth $1 billion.
Maybe halfway from the bottom, you got an Aurionpro. We keep doing our 25%, 30% every year. It'll take us 15 years to get to the top. I think for the first time, you see. That's because banking software has not gone through any major change for a long time. You can probably crash through that pecking order. What would have taken you 15 years will probably take you five years. I think that is the scale of opportunity in front of us when we look at the banking software units. We talked about some of the others. If we run this business directionally through 2030, depending on how you want to model it, overall, a very large space, anywhere between half a billion to a one and a half billion dollars space.
What do we really, from an aspirational standpoint, aspire to when the business gets to maturity? Probably it won't get to maturity over the next decade. As you get to a mature number, I think long run, can you get to what we say, the rule 50 plus, which is sum total of your growth plus margin exceeds 50 as you get to maturity. Right now, obviously, we are in an investment phase of the business, and I think we probably still get to 50 plus. The idea is, as the growth rate slows down, let's say a decade from now, will the EBITDAs go up? I think that's what a good product business should get into. Can you get to a very high level of ARR mix?
I think right now where the sort of recurring mix stands is probably something between 50%-60%, because you're still growing at 25%-30% of the business. There's a lot more new business that keeps coming on. As again, you get to maturity over the next decade, this is a business which has a very high recurring potential across every space that we play in, whether that's transit, where you're talking very large and very long relationships over a period of time and globally, is the way the business behaves. Whether you're talking banking software, where banks normally buy tech stacks for 10, 15, 20 years, or you're talking data center space, where again, you're talking very long-range plans for pretty much all the operators. Extremely high level of recurring as the business matures over time.
Can you continue to deliver a very good return on invested capital in the business? I think the investing engine that I talked about, whether that's investment in product or that's in acquisitions, we will stay disciplined and go after generating a good medium to long-term return when it comes to investing that dollar in the business. I think the thing that really happens when you're looking at the P&L from one year to the next and when you are in the phase of growth that we've been the last five years, where you've been CAGR-ing at 30%, is that you are building an awful lot of product at the same time. You're sinking in a lot of R&D dollars. You're expanding into markets where it takes time for the teams to get productive.
If you look at yearly P&L as an indicator of the return on the investment, that's not a great indicator. The investment returns in most product businesses would play out over four, five, seven, 10 years as the product acquires scale in the market and as the operating leverage starts coming in product by product, business by business. Right? I think that is the way to look at return on capital. At least that's the way we look at return on capital when we look at planning our investments. As you get to that, the whole point I was making about installed base, you get to a net revenue retention which is very high, certainly not of 100%, but can get to very high levels as we go and mature that base. Net, every single business that we are in has something fundamentally transformative happening.
At least a number of our large businesses have. We feel we have built out the business, built out the teams, built out the products that can go and compete now on a global scale. Transit, very successful here. Can we make it a big global business? Banking software, very strong in Asia. Can we become a big global player? AI, again, strong in at least India and parts of Middle East. Can we make it a big business across Europe and Southeast Asia? This is the time for us to go and build that out. That is largely what we had in terms of content. I'm looking forward to some very interesting questions coming through on the Q&A. If you want to beat us up on Q4 and all that stuff, feel free to do that. Happy to answer those questions.
The nutshell, if I was to go and sum up where we are today. One, we really thought through where we wanted to compete in, even we were not expecting to get it so right in terms of the change that we were anticipating. I think we got it unfairly right almost in terms of the spaces that we chose to be in. We pivoted, we recruited teams, we brought in talent from outside. We methodically prepared every single aspect that we needed to compete. We did not have AI. We brought in Arya.ai. We did not have the right surface area of products. We went and did five acquisitions to increase the surface area of products. We did not have IP around transit and all. We went and did acquisitions. We built out our own R&D team to start building out stuff which just did not exist.
We really prepared for this time. While we were doing it, we also built out very strong businesses in each of these spaces, very credible regional champions. 90% of our business is still Asia. Of course, that sort of limits the size of the pie you can go after. In each of those deliberately chosen spaces in the markets we are in, we are a very credible competitor. This is our time to go and now make this business global. To really go at it with speed, with scale, make the investments that we need to make in the business to really build the businesses of tomorrow. That's the mission that we set out on four years back when we started this with many of you in this room. That's still the mission that we go after. There is really no change in what we are going after.
The difference is the size of opportunity in front of us is something we've not seen in our lifetimes, we intend to play this to really build out a business we can all be proud of. That's the game, that's the mission, hopefully you found this presentation useful. We will go to-
Q&A
Q&A after that.
Yes.
Thank you.
Thank you, Mr. Rai, for summing up today's sessions. Ladies and gentlemen, before we move ahead, allow me to take a moment to thank all of our incredible leaders who've graced us with their presence here today. We truly appreciate the valuable insights that all of you have shared and given us a lot of food for thought. While I get my stage ready for the Q&A, allow me to just state a few ground rules. If you have any questions, which I'm sure by now you do have, please raise your hand and I'll have my volunteers come to you with the mic. Yeah, just let me state out the rules and then we can start with the Q&A.
Once you have the question and you get the mic, I request you to stand up, introduce yourself, introduce the organization you're from, and I request you to humbly keep one question per person in the interest of being fair and giving opportunity to as many people as possible to ask questions. We're almost ready. Speakers now up on stage. Can we have the mics as well, please? Mics. All right. Please put your hands together and welcome our speakers up on stage. Once again, Mr. Ashish Rai, our Group CEO. Can we also welcome Mr. Shekhar Mullatti, President, Banking Solutions Group. Please welcome Mr. Sanjay Bali, President, Tech Innovation Group, and Mr. Sanjay Varma, President, Fintech Solutions Group. Welcome, gentlemen. As requested, please raise your hand if you have a question.
I'll have the mic brought to you, then introduce yourself and the organization you're from, our leaders will be more than happy to address any queries that you have. Let's begin. Yes, the gentleman right there. Can I have the mic brought, please? Yes.
opportunity. My first question is, what we've been able
Could you please introduce yourself and your organization, sir?
Yeah, sure. I am Anuj Lunawat. I am a retail investor, my question is, whatever I've been able to pick from the earnings call commentary is, we started as a project business. From there we understood the customer pain points, the requirements, from that, whatever insights we've been able to gather, we have built a product. That product is embedded in the customer's workflows, from that, the more and more products we add to the customer's workflow, we go from product to platform type business. My question was, can you help me understand the customer journey, wherein the customer goes from one product to multiple products, is it a push-based sale from one product to multiple other products, or is it where the customer insists us, "More and more products is what I would like to want from Aurionpro?
Shekhar, want to begin?
Okay. Let me take that in the context of banking anyway. Banking, each product is very large. Each product, like Ashish mentioned, when the bank takes it's to run a core business process. It runs for 8 to 10 years. Yeah, the journey of moving to another product is a combination of push and pull. As the bank experiences the ROI from the first product that it acquires, implements, starts operating, then they say, "Okay, here's a good vendor, here's a good provider. What other capabilities can they bring to help us lift our efficiencies, help us increase our market penetration?" That's the pull aspect, and the push aspect is that all of us are team members who are engaging with that customer at that point of time, are also having active conversations saying that, "You know what?
We can also do this, this." "Here are some logical extensions to whatever you're doing in this area, which can solve another business problem." It's a combination of that push and pull which translates to more products embedded within our customer. It's a very conscious strategy from us. It's a land and expand kind of strategy, because in the banking world anyways, the effort involved in acquiring a new customer is humongous. Sales cycles are long really, at the time of entry, it's often competitive. Pricing can be disruptive. All of these things make acquiring new customers quite challenging. We are fortunate, we have a large base of customers who have pretty decent-sized wallets, and we use that land and expand strategy very consciously.
Okay. Do we have any particular product which is helping us expand more in adjacent categories or like loan or transaction? How do I say? The loan management software helping us expand into transactions or something like that?
All of them. Look, again, in banking, lending is on the asset side of the balance sheet. Transaction banking is on the liability side of the balance sheet. While these two might not per se be complementary, there are areas of intersection like the financial supply chain, so on. It's not a formula thing. Every customer is a very big specific customer. You have to understand their problem statement. You have to understand what is their business priorities and position the suitable product to them accordingly. Yeah?
I think just to add to that, I think one of the things to understand about Aurionpro and how we think about a space is, regardless of what it is, we'll normally very closely look at the whole value chain, and we'll typically find many points on that chain that we occupy. So one is the footprint in terms of functional solutions and how much of the domain do we cover. Second is footprint in terms of value chain and how many things we can do on that value chain. So even when I provide software, can I extend to deploying that software? Can I extend to doing other things around that software? Can I extend to implementing that software, the cloud side, et cetera.
The utility that we have to a strategic customer is very large compared to most commodity partners because you're not a pure services player. You can bring in real IP of real value, real software in the enterprise, and then you can also go up and down the value chain depending on the Aurionpro product that we're talking about. That makes us very sticky and that makes us very useful to the customer, and that's what make the dollar, like I said, over four or five years double almost from the same customer. If you started at one INR, you go to two, over a four or five year period with the same account. That's the beauty of how we have been building out the stack. I think move on to the next question.
Okay. Do we have more questions? Yes, please.
Yeah, hi. This is Anmol from.
Could I please ask you to rise, sir?
Yeah.
That our speakers can see. Thank you.
Hi. This is Anmol from DAM Capital. A couple of things. Firstly, wanted to understand that when you say 25%-30% growth over the next five years, CAGR, what is the organic growth that we are talking about?
Okay.
Also in that, you said that 110% net retention ratio. In that, are we saying that our existing customers, we are expecting them to grow at 10% and 20% is what we are trying to add from the new customers?
You'll have to ask the second part of the question again.
What you said was that net revenue retention is nearly about 110%.
At maturity in the business. I think the point was long range, that's where you want the business to get to.
Okay. In the next five years, what kind of organic growth are we targeting right now?
Thanks, Anmol. Look, one is we obviously would not publish a number going out. The reason we normally try to plan around those numbers, which is can I grow 20%, 25%, 30%, is we've always been acutely conscious of not overheating any side of the business in terms of really trying to grow it faster than either your ability to invest or your reputation can support. I think reputation is everything in this business, so you want to grow consciously at a pace that we can support. There's some parts of business which can scale faster, and of course, we'll try to scale it faster because it's probably not capacity dependent. Some parts of business have become easier to scale with the new tooling that's come out. Again, you can do that. Broadly, that's where that longer run ambition sits.
How much of that is organic? We typically plan to grow organically. We've done some acquisitions over time. We've done quite a few acquisitions over time, they've actually never been for revenue. I would say most of the growth that we've gotten in the past has been organic growth. Wherever there's been some inorganic contribution in the year, that's been certainly low single digits. Going forward, probably you don't know what you don't know. If a fantastic opportunity comes across that is a large business and that's at the right value and very good for the shareholders, we'll probably obviously still pull the trigger, and we have the cash for it. We would typically not go on and do, let's say, a levered acquisition for revenue, that sort of stuff. Organically, I would say we'll continue to grow strongly. Most of the number we planned is organic.
We don't plan for inorganic growth numbers. Inorganic for us is I have a blueprint for the business. Instead of building the product, I got an opportunity to buy a very high-quality team with a high-quality product, which is already built out. That sort of de-risks that investment for me, we'll opportunistically do it. We do not programmatically sort of plan for inorganic number. I would say any numbers we throw out, typically just expect most of it to be organic.
Ashish, in addition to that, for the next five years, we are expecting our banking to grow faster or TIG to grow faster organically. At least what we have seen is that last couple of years, our TIG business has grown much faster. Probably, that have also led to some lower OCF generation for us as well. For the next five years per se, how are we seeing this as?
Yeah. Look, you don't know what you don't know. That's why I specifically presented the numbers because of that question, right? I'll contest that observation a little bit. Last year it was basically neck and neck, 19 versus 22-
Yeah
in terms of growth. Year before, 2023 and 2024, banking grew faster. 2021 and 2022, TIG grew faster. I think 2025, TIG grew faster. It's not that it's been one way. It depends on the year and the opportunity in front, and what does it take to scale. I think the reason that varies partly also is because from time to time, we put our heads down and start rebuilding the product. The reason banking was slow in 2022 and going into 2023 is because we were rebuilding the whole stack. We said, "No point to sell you grew." You still grew 14% without selling any material new logos, right? When we came out of it, then you logged multiple years of 30% plus growth for banking. Which was 33% one year, 37% the other year.
I think that is what will cause the rates to vary, but over the long run, I think it's a wash. I don't think either of the businesses can really grow strongly or slowly. It will come down to execution. The demand environment for both the businesses is like we've never seen. The size of the pipeline is something we've never seen. The size of the order book that we're walking into the year with is very strong for both the businesses. I think in terms of growth opportunity, I think it's a wash. I don't think one or the other. The OCF point is a function of maturity of the business in a high growth phase of the business. The advice that I would have in terms of how to look about it, yes, different businesses have different characteristics.
Part of it is also just the growth. It's very hard to deliver sort of a mature company economics and a startup style growth in the same place.
If you want to grow at 30%, 35%, 40% to sort of capitalize on the opportunity in front, there will be pressure on the cash flow just because the sheer number of invoices you're sending out, the number of products you're building out and all that stuff. I think, on the TIG side of the business, and Sanjay can comment on it, the businesses are maturing quite fast as well. I don't think the characterization is something that will hold going into the future. I don't know, Sanjay, if you want to comment on?
I think you covered up everything, Ashish, I think so. Yes.
Each of the businesses, if you look at it, whether it's Transit, it's maturing very fast. Whether it's data center, I think consumes almost no capital. Bhaskar's running a very high-quality business. Over time, I think the characteristic of the TIG side of the business is also extremely efficient businesses. Now the focus that we have is to go global with Transit in a very big way, which again, gives very different economics and a very different level of visibility and predictability on the cash side as well.
Thanks.
All right. Do we have more questions? Yes, sir. If you could just pass on the mic and request you to rise and introduce yourself.
Hi, this is Ashish from Goldman Sachs. Ashish, I had one question around data center. I think in the slide you mentioned about modular products. If you could just double-click in terms of how exactly do you determine which are the areas where you want to get into manufacturing? I think the case in Transit was a bit more simpler in terms of you wanting to own the full stack, which delivers a higher gross margin and ROC. From the data center perspective, not very clear on that in terms of what's the roadmap, if you could just clarify a little bit on that.
Also just the factors as to how big in your view does manufacturing become when Transit has a piece of it, when data center has a piece of it, in a business which obviously on the other side, you're investing a whole lot on the services stack as well around AI, and then there is this manufacturing. If you could just help us with some clarity on that.
Okay. Thanks for that, Ashish. The way I'll do it is I'll probably philosophically just set a little bit of thing. I'll move on to Sanjay to explain what we mean by sort of manufacturing on the DC side, and then I'll come back and address the last part of your question on how big the whole thing can get. Just philosophically, I think the way to look at data center service business is essentially, not dissimilar from how we approach Transit. Transit, of course, now we've got integrated end-to-end across the chain. When we got into that business, it was largely a services business and we said, "Okay, we'll understand every point of the value chain." We went and acquired something. We built our own IP. We got the whole chain. Data center, probably a couple of years behind.
I think we've built up one of the highest quality teams, understands the business really, really well. Now it's a question of what are the points at which we can really deliver better economics and capture value. It's not a question of owning the whole chain because also the size of the space is very different.
Yeah.
The whole chain is different. The thing is, can you find points of value where, one, the Transit strengths we built out are leverageable. One is the revenue becomes nonlinear. Second, you improve the economics on that point, right? That is the whole idea of starting to build our own products on the data center side. Slowly you'll see them mature as well. I think where we want to do it, maybe Sanjay will comment on it, I'll come to the larger manufacturing point.
Thanks, Ashish. As we know that we have built a complete stack on the transit side of business. We have our manufacturing capacity and capabilities built out. If we see the last 3 years when we started building the manufacturing capabilities, we had 1 unit of this, then when the orders started coming in, we from 1 to we are almost 4 units. The kind of machinery and tools which we have, it was brainstormed, and we saw that some of the components on data center, as Bhaskar pointed out, especially on the cooling side of the thing or it's kind of modular solutions on when we are putting it all fiber, the electrical components and everything, and building it out. We have everything with us, and we are just leveraging those machineries and tools which can build up.
We are not going absolutely on the entire value chain of data center as of now. We have picked up a few of the very important where the lead time is more, as the manufacturing requirements are more for others, that's not only for the data center part. The logistic time is taking more, we wanted to have more control and more faster approach towards building the data centers, which are required especially for the AI side, which are more modular and faster. That's why we are moving into using those particular 4 or 5 components on the manufacturing side as of now.
All right. The gentleman behind the first row had raised his hand before. Yes.
Hello.
Sir, I'll get to you in a moment. Yeah.
Nipa, can I just address one last part of
Sure
Ashish's question on Look, I think when we are expanding, I think the question was, where does this whole manufacturing thing go in terms of how large it gets and stuff? Look, the idea is. First of all, we've approached it so far in a fairly capital-light way. Right? I think by and large, our initial instinct usually is to stay capital-light when we get into spaces. The other thing is we also learn, then we decide what is the best way to do it. Like on the electronic side, validators and all, we were doing our own. At some point, once I understood the whole thing, we say, okay, we outsource the actual manufacturing to someone in Malaysia who can do it a lot better. We still do the whole R&D ourselves. The product designs are ours, then outsource the manufacturing, right?
There is no one set answer to how you do it. I think the way the product business works is, it's not that you're trying to become a contract manufacturer for someone else. You need to understand the product really well. You need to understand the economics of that product really well. You need to understand what goes into building it very well. Then if there was a choice to be made, you said, okay, someone else can do it better. We'll very happily do that, right? Until you really get to that point, it's very important to sort of really crack the whole nut because it's not that you're building manufacturing or something else. That's how we look at building this capacity out. Our initial instinct is to not go CapEx heavy on this side.
We will be very stringent in terms of what we expect the team to do and how we expand. Right now, as Sanjay was mentioning, by and large, it is using the current sort of production capacity that we already built out to try and do a bit of a line expansion to get in on Bhaskar's side of the business, right. Over the next year, not significant, but I think we'll probably come back as we concretize this a little bit more, right. That's roughly how we're thinking about it right now.
All right. Yes, sir. If I can please ask you to rise and ask your question and introduce yourself.
Good evening. This is Ramesh Pojwani from Mehta and-
Sir, could you hold the mic closer, please?
Good evening. This is Ramesh Pojwani from Mehta & Vakil. First and foremost, many congratulations for an all-encompassing presentation of all your verticals. The most striking feature was that the revenue per employee has gone up 64%, which stands now at INR 0.47 lakh. It speaks volumes of the fructification of the hard work and effort put in by all the verticals, creating the products, creating the software, and now it is fructifying. Going ahead, you have shown by 2030, you can be a $1.7 billion company, which I believe is a given. Specifically, you have mentioned that banking software 2.0, you have referred, saying that banking software is being rebuilt and repriced. Would like your comments or expansion of this line. The second thing is about AI-native software. Ms. Wang explained very beautifully about Fintra.
The AI trade finance is a software which will tackle or which will attempt to address a $10 trillion global trade. These three thoughts.
Okay.
Thank you.
Sanjay, you want to take on the AI software-
Yeah
banking software side?
Yeah. See, at the heart of all software, the business model hasn't changed. The business model, for example, in transaction banking is still cash management and cash movement. The business model in lending is still giving loans, getting interest back. What has changed is the methodology in which that is being done. That methodology resides around three components that artificial intelligence brings in. One, it automates redundant tasks, like if I scan an Aadhaar card or if I scan a document and it can automate the filling of fields. This is a task which would take a data entry person a little bit of time to do. It can do it in seconds.
The second, it provides insights, which means it gathers information over a whole variety of information, over a whole variety of media, over a whole variety of areas, brings it back and presents it to you in a meaningful way, which you otherwise would struggle to do. The third is it uses its analytical ability to generate decisions which you would require experts or you would require people with very high caliber and experience to do. It does that in seconds, and it does that instantly. What AI native software basically does is it takes a combination of these three traits using AI and applies them to the way you construct and you deliver software.
You have moved away from software, which is a system of record and a system where you are doing the action, to a system where the intelligence itself does the necessary entry, insights, analysis for you. You become pretty much like the checker, the human on the side who understands what has happened, who understands where the information is coming from, and from there, either you take the actions forward. Whether that be approving a loan, whether that be approving a transaction, so on and so forth. That is the fundamental shift that has happened. You have stopped making software, or we have stopped making software in the traditional way in which it was just, I create a lot of forms, a lot of data fields, a lot of databases, and you fill in information.
Now we create agents, we create the fundamental framework which allows you to automate redundant tasks, to go with insights, to analyze intellectually, provide back information, and allow you to take decisions. That's the fundamental shift around AI native software.
Look, the whole banking software 2.0 that Alban was talking about, what Chloe showed was our first example of what we built out from zero, is clearly way ahead of the rest of the industry product. You'll see a number of them come out. I think Deekshith was showing some of that in his sense. Aurion Credit, again, we are ready with it, we got a whole sequence of products coming out. That's the reason why the R&D spend is extremely elevated for the last 6 months, as you would have noted. What does this banking software 2.0 do? If I was to put it in very simple terms, it does three things. One, it should do everything that the software 1.0 was doing. What is software? At the end of the day, all enterprise software. You got a business.
That business runs a business process, which creates some value. The software is a codification of that business process in terms of code. You basically put that business process in terms of code, someone is using it. Your banking software 2.0 should do everything software 1.0 was doing, which is that codification of a business process. It should do two more things. One, it should provide the rails, the tools that the agents need to become first-class operators on the banking workflow. The agent can act like an operator, and the software needs to provide this agent the tools that it needs to do its job, which is very different. You can't do that easily on a legacy software because you just don't have the architecture or the rails for it.
The third thing that it needs to do is give the bank, the enterprise, the human operators in the bank, the tools that they need to govern these agents, which is the whole CGHP that Deekshith was talking about. We've got a very unique, very innovative governance framework for that to happen. Your banking software 2.0 essentially brings these three things together. That's what it needs to do. That is very different from what exists in the market today. That is why it's not just us saying it, the entire industry believes it. That's why it's a net new, much larger category than what we've been used to seeing so far.
Thanks. Hi, this is Namit Arora from Ingrowth Capital. Thank you very much, Ashish, and all the business leaders. It's been very helpful last few hours. I just have one basic question for Ashish. Ashish, as you rightly said, you have been as a company very early to all the right spaces, and as you mentioned about where you started four years ago. Given that as you reflect back over the last four years, have there been any learnings in terms of actually translating that early mover advantage into, let's say, client wins or market share or cash flows? Any learnings or takeaways from the last four years so that as we look at the next three to five years, you're able to translate that early mover advantage more into, let's say, financials? Thank you.
Thanks, Namit. Great question. Look, most of the learnings come from mistakes, and we've made literally hundreds of those. Yeah, that is true. Look, I think if you look back, the spaces that we chose, one, it was not accidental. We said, okay, it has to be a space where we see there is something fundamentally transformative happening. We went looking for it. You had that much of an advantage you said in terms of looking at finding the right spaces. What we learned over time, one of the comments that I was making right now, and especially for the investor set, product businesses are hard. Services is a game that is easier to understand and it's easier to play because what I say is, I want a skilled man-hour.
I want to understand the cost of that man-hour, depending on who I am in the business in the pecking order, I want to charge a premium on that man-hour. Right? Now, if I am a strategic consultant like the Big Fours and all, I put a certain premium. If I'm a IT services company, I put some other premium. If I'm a, I don't know, an accounting firm or something, which is local, it puts another premium. Ultimately the game is, I understand the cost of my man-hour, I put some premium on it, and I do it. I don't necessarily put my balance sheet or too much capital on it. Even at lower margins, my return on capital is pretty decent because I'm not putting any capital. Product is a very different game, right?
If you started with a core thesis that tech is a net value add to the world, you've not seen scaled-up technology players in this, at least enterprise tech players in this country or even this continent before. You really want to build out a scaled-up global tech player rooted in India, which can impact on the global scale, you have to prepare to sort of take the hard path, right?
That was the whole thing. Which means, I think roughly half the R&D spend that we've done has gone to waste, more or less. Because when you're making that spend, you don't know whether that's going to work or not, right? A lot of the R&D we continue to do right now will go to waste, right? That has been one big learning on how do you really continue to make the bets while staying responsible in terms of also diversifying that risk out so that you are in enough buckets. One of the reasons why we have a slightly wide-ish portfolio for ourselves is also to diversify that risk, right? Because you know some of it will fail. That is the nature of the game, right?
The other thing, again, that I hinted at earlier, at least I personally learned, is just do not expect a startup-style growth and a mature company economics in the same place. The reason varies from one year to the other, but when we say cash is super important to the business, right? Businesses die because of cash. We know that. We sit on a little bit of a cushion. We are happy with that, right? If we start expecting the cash flow or the DSOs of an INR 10 billion IT services company without naming any names in the forum, you're just not going to get it. Right? You want to grow at 30%-35%. You want a startup-style growth. You want to see a shape of a curve which is near vertical. Don't expect mature company economics to come.
You want to get the cash flows of a mature company, then be prepared for a 4%-5% growth, right? You drop the growth to 4%-5%, all cash problems get solved, right? Because it's impossible then to not have positive cash, right? If you want to grow stronger, and right now, where we are at the stage of our life, at the point of time in history, we will aggressively go and build product. This is the year for us to build. You prepared very hard to come to this moment in time. You are in the right spaces. The opportunity is in front of us. It would be wrong to not play this game to the best of our ability, to the hilt, right? We are sure we can create value, space after space that we are in, right?
We learn from some of these things. We try to be responsible as a business. This year, we'll focus very heavily on the balance sheet and cash side of it. We prepared for the hard path. That's sort of roughly I don't know if that's useful, but yeah.
No, that's very useful, Ashish. Thank you for your very candid and very detailed thoughts. All the best to the entire team.
Thank you.
Thank you very much.
Thank you, sir. This gentleman right here.
Yeah. Hi, this is Avi-
Could you please hold the mic closer?
Yeah. Hi, this is Avi from Choice Institutional Equities. My question is with respect to EBITDA margins. Once the accelerated investments are over in FY 2027, do we expect EBITDA margins to expand from FY 2028 onwards? Also while we see a better growth in the TIG segment.
Sorry, what was the second part?
While we are expecting a better growth in the TIG segment that can dilute the EBITDA margins on a consolidated level. Do we expect EBITDA margins to expand from FY 2028 onwards, while the investment phase is over this year?
Look, I think the problem with the investment phase getting over is, honestly, we were, let's say five years back, an INR 375 crore player. Now you're roughly an INR 1,500 crore player. If you really think about it, an INR 1,500 crore player in the spaces that we defined right now, you're talking a $140, $150 million player in what is literally trillions of dollar TAM. I think you have to at least have some humility about the size. You are still a very tiny mosquito in a very giant industry. We are a very fierce, very competitive mosquito, but you're a mosquito nevertheless. Right? I think investment phase getting over is probably not two years away, right? When you look at the FY 2028 question, I think the investment phase getting over is kind of a long way away. Till that time, what do we do?
The goal is to stick to some sort of a reasonable EBITDA number, which is why we sort of stick at 20%. As the business scales, as the economics improves, as the operating leverage kicks in, obviously the latent EBITDA is going up, right? You almost cannot stop EBITDA from going up in this business. The excess goes back into R&D, gets invested back into the R&D, right? At the end of the day, the core EBITDA of the business, do you expect it to go up? It has been steadily going up, right? Where do you see that EBITDA going up? That's why in the strategic architecture slide, I was trying to segregate the R&D dollar from the M&A dollar. If you look at the R&D dollar, five years back it was near zero. Four years back it was INR 40 crore.
Last year it was almost nudging now close to INR 130 crore, INR 140 crore. Right? If you are not investing that INR 140 crore behind the product, that goes into your EBITDA. Right? If you said, "I'm going to only start selling from today everything I've already built up, and I'm not going to build any more new product," that is EBITDA. Do you want to do that today? Absolutely not. Because you're still a very small player in a very large space. I would build more product, get into more pockets, actually sink that R&D dollar today to be a much bigger player down the line, right? I think that is a way to think about EBITDA. In the product business you cannot almost not have an improving EBITDA as the business scales. The question is, what do you want to do with that dollar?
For a long time, as we continue to grow at the pace we've been growing, you will see that EBITDA masked because of the scale of investments that we are doing. Will the investment phase get over by FY 2028? No way. I just don't see how. Right? I think we'll continue to press the lever on growth and we'll continue to invest.
All right, the next person who has the mic. I think there's a lady behind.
Yeah. Hi there.
I request the gentleman in purple to please stand up. He had the mic before. Yes, sir. Please.
Okay. Thank you.
May I request you to please rise and introduce yourself?
Shall I go ahead? Yeah.
Yes, sir.
Thanks. Just continuing the same discussion. In trade finance, what is the cumulative spend-
Sorry.
The mic is not working well.
Sir, if you could just stand up, we'll be more able to see you. Thank you.
Yeah. Hi.
We can hear you now.
You can hear? Yeah. With reference to trade finance product, what is the cumulative spend which we have done for development, what about the other verticals? What are your plans for that? When you think of this kind of development, what's a payback or some criteria which you have mind? Corollary to that, the question is that, in one of the slide, competitor profiling was done. Where do we stand with their pricing and their size of business, whom we are benchmarking, and what's the timeframe to reach those kind of a business standards? Thank you.
Thanks. Good question, sir.
I have two more questions, I'll come back.
Thanks for that. Look, I think the way to look about the investment made behind trade finance is actually, the way I would look at it is that there's two sort of legs to it. One is we went and acquired Fintra as a business, I may be getting the timing wrong, but six months back. Right? That was to get the core domain expertise that we needed to build the business, and then we build the AI native product from scratch over a period of five, six months. Right? I would say the net total investment is probably mid-single digit million dollars. Without throwing an exact number there, but roughly single digit mid-million dollars. Where do we stand, let's say, in the market?
If you look at the trade finance software market, I think give or take, you're talking a couple of billion dollars. Obviously, the point that Alban was making about the operator TAM or the services side of the business, the operations side of the business merging with the software TAM, that is obviously a much larger pie. If you say this is a AI native software stack that goes after also the operations spend and the software spend, you're probably talking a double-digit billion dollar space. Where are the competitors? The reason we are so excited about trade right now is, one, it's sort of dominated by a few vendors who've been in the business for the last 30, 40 years. The software stack in trade finance overall is very old.
Second, trade is ripe to be disrupted and is very attractive to AI because it's a very document heavy, very rule heavy space, right? On one side you've got a largely legacy stack. On the other side, you've got a very document heavy, rule heavy space, which is ripe for AI. Right? Hence the strong reception we've had in the market. We're actually negligible in dollar terms from a size of the business standpoint right now, but we feel that will rapidly climb because there is net new demand in the market, but we can also take share. This can actually escalate pretty quickly, and that is what our plan would be overall. Right now it's negligible, but we will deliver an extremely high return on the investment we made there. I think we feel very good about it.
Can you make-
All right. Sir, I'll have to ask others, give others an opportunity to ask questions as well. There's a lady behind who had stood up or is it the gentleman who wants to ask? Yes. Thank you. Please, I request everyone to refrain to one question only. Yes, sir.
Yeah. Thank you for the opportunity. I'm Arpit. I just want to understand, you have done about INR 500 crore of investments on the R&D in the products. We are seeing some non-linearity coming up in the business. When do we see when the employee growth is maybe lower or much lower than the revenue growth, j-curve, where in the cycle we will see that? I'm assuming that a lot of your own processes will also be now driven by AI. To that extent, employee intensity. One more question is on you have two very different businesses in one company. How is the working capital intensity in each of the businesses? That would help us to decipher, because your working capital is higher than in terms of the cash generation is lower than the competitors in the software business.
How do we break that up? That would be all. Thank you so much.
Okay. Thanks for that. I do see that everyone asks two questions. I think you asked only one question because you were waiting for the second one. I think we'll come back with Mike. The two questions. First one was the dollar-
Per employee
revenue per employee. I published that metric, right? You already see the differentiation between the employee growth and the revenue growth, right? Revenue per employee over the last five years has gone up 65%, which means the revenue growth is outpacing the employee growth even without all the AI code generation and everything else that people expect to deliver productivity. That is the nature of this business. Even now, all of this year, our headcount actually hasn't increased meaningfully at all. Maybe one or 2%. You already are at a place where revenue has been climbing, and the employee count other than a few parts of the business, it's not been climbing because we're not that people intensive as a business.
On the software side of the business, we honestly, for the current rate of growth, we do not expect any meaningful increase in employee count going forward. If it's an exceptional level of growth, then we'll see. I think on the current pace of growth, we feel quite good about the capacity that we built out over time. On the TIG side of the business, given just the sheer scale of orders that we are picking up right now, there'll be some climb up in headcount, but I think that will also be non-linear. The employee headcount growth is not going to be totally proportionate to the revenue growth. I think that's sort of how I'll put it. Even now, it's not exactly proportionate.
The revenue has been growing quite meaningfully compared to the employee headcount growth, but that will only get sharper as we go forward. On the difference between the two businesses, you really have to sort of slice it down to the exact business, because, across the segments, yes, the software side has been more efficient from time to time, but also the software side has consumed R&D dollars in a meaningful way if you go back three years back. It goes through cycles. TIG, again, goes through its own set of cycles. Some part of the business has consumed working capital from time to time. On the other hand, if you look at data center, which has been a huge growth driver, doesn't really consume any material working capital. Even some of the newer contracts that we're picking up as Transit gets more global.
As you go global and you sell Transit, the economics behaves much more closer to the software side of the business than anything else. Ultimately, both businesses have their own cycles. I know we are forced to look at things from quarter-to-quarter and year-to-year, but the way I would sort of like to look at it is what is my return on capital that I invested over, let's say, a medium-term period of time. Our goal would be for most of our businesses to more or less get to the same or similar number. The problem is from one year to the next, you can't tell, but over a three to five year period, you can tell.
Transit will become an extremely strong economic driver, especially now that Sanjay's business now owns spread IP across the chain, including the back office software and all that we expanded with Mumbai. The economics change. On the services side, data center side, and all, the goal would be to grow in a very capital-light way. Even if you said I do a little bit of lower margin, if I don't consume any capital, again, the return on capital would be very good. The way I would look at it is what is the return on capital that we are able to drive over a three to five year period? We would tend to stay in spaces where they more or less end up at the same number. We don't have differing asks over a medium term on that.
I think that's the way I would like to look at it from one year to the other, you're right. Some working capital here more, banking R&D in some year more. Banking is going towards a place where we're investing a lot of dollars now into R&D. This again, will be more. That's the way I would kind of look at it.
All right. There's a gentleman right behind. Yes, sir. Okay. I'll come next to you, sir.
Okay.
Yes, you. Please stand up.
Thank you for the opportunity. My name is Ganesh Shetty. I am an individual investor. I have been tracking Aurionpro for more than a decade now. It has been a phenomenal growth for the last three years. The actions the management has taken is also phenomenal. I would congratulate you for that purpose. My question is regarding the total addressable market, what we are having now. Be it TIG or be it banking product, we have got phenomenal road ahead to improve ourselves and I can say, conquer the world.
In these circumstances, I would ask you one thing, whether is it any possibility that we are going for QIP for raising funds, or in that way, promoters stepping in and putting more capital and raise their stake in the company so that we will have a better balance sheet as well as we will have war chest for future opportunities and building our capabilities. Thank you very much, sir. All the best.
Thank you. No, look, I think we feel fairly good from a capital standpoint. Unless you come up with some super exceptional opportunity, there is really no need to raise any capital that we can think of. Obviously, the cash cushion has sort of worn down a bit. This year the focus would be to kind of widen it again and make the cushion a bit wider. That is normal operating cycle of the business. It's not something we can't manage. There is no debt in the business. There is some cash cushion. Businesses are generating enough cash for us to invest. Like I said, over the last four years, despite everything, the massive growth, how much you get stretched because of the growth.
The thing about 30%-35% growth is, last quarter of the year, you're pushing out a massive number of invoices out which the customer has not paid you for. You need to go and run and get those collections, so you will get stretched. Yet, we've invested INR 1,000 crores plus behind R&D and behind acquisitions that we've done. Right? Business has the capacity to generate cash and invest, and I think we'll continue to do that. I don't really see, for the current plans, any significant need for capital. Now I'm forgetting the second part of that question.
Partner or QIP.
No QIP.
I mean, owner or QIP. The source of funding.
Oh, yeah. Look, I think, yeah, there's not a lot of us. The promoter holding around. Across the board of directors table, there is, I know, what? 44%, 45% holding. You can check with investor relations where that is. We feel that is the right level of holding. There's no need to sort of take it above that. Part of it is promoter holding, part of it is obviously mine and other directors. I think that's the sort of right level to be at, because then you also need to leave enough float in the market, right? There is no need to sort of worry about such things, or at least, I don't think we spend any time thinking about such things.
All right. The gentleman here. I'm sorry there are too many hands going up at the same time. Whoever my eyes are falling on first, I'm just allowing them to get as many questions as possible. Yes, sir.
Hello. This is Yash from VI Invest. The question is regarding, given the rapidly underlying foundational AI models are evolving, how is Aurionpro structuring its product suite to remain model agnostic? Like specifically for platforms like iCashpro+, SmartLender, and the new Fintra engine. How does our engineering roadmap ensure that the live bank deployments can seamlessly absorb newer, more efficient model iterations without causing costly core-banking disruptions or high migration overhead for our clients? Coming to the banking part, security and strict regulatory compliance are traditionally viewed as massive entry barriers for AI adoption. Like with our specialized focus on Arya.ai and Lexi AI for explainability and alignment, do we view this banking security constraints as a major commercial moat?
How aggressively are we looking to monetize these defensive AI capabilities, like autonomous fraud detection and risk scoring, to capture market share from traditional players and in major growth markets like India and U.S.?
Okay. Thanks, Yash. Look, we try to present at least our point of view on where we think the whole stack is going. On one side, obviously, you bring the point about frontier models. There is obviously a lot of investment going in that space. There is a leaderboard. There is lots of changes in that leaderboard.
No, this one
They are net providers of intelligence. At some point, some of it is almost commoditized. Some of it would be valuable. The point is, model providers have their space, but they are not necessarily a replacement for the software stack that a bank needs or any enterprise needs to deliver its business process. I have seen no signs of anyone trying to get in on the application space. Of course, from time to time, Anthropic releases a SKILLS.md and the market comes down 5%, and all that is fine. The reality is you need software to run a business process inside the bank, right? The need for the software does not go away. If someone knows of a logical way to run a business process using a model, someone needs to show it. There is no way you can run in any enterprise, a business process without the software stack.
Our point of view is that software stack needs to evolve to account for the intelligence that's coming in. It needs to be a slightly different software stack, and that was the whole point of the banking software 2.0 point of view that we had. We intend to lead in terms of building this architecture out. We believe the size of opportunity as the software stack gets transformed and displaced is massive. Like Chloe showed on the Fintra sort of trade finance side, we will do that domain after domain after domain. We believe this is the time for us to go and build out that stack and take space. We feel very excited about the opportunity for that to happen, and we already see signs of that happening. I think that's how we generally look at the AI opportunity overall.
Alban was kind of summarizing that. We believe the size of opportunity for banking software is going to triple over the next five years. It's not less software or model doing something. We believe the size of software that the banks will buy will triple over the next five to six years. Right. That's essentially the scale of opportunity in front of us, and I think we intend to play and get our fair share of that game. Right. That's roughly how we look at it.
Yes, sir. Please.
Thank you, sir. This is Gurvinder Juneja. I work with Fortuna PMS. Only one question about your transit business. I haven't understood the point you made on why will the ROCE go up as the business builds. I didn't understand the principle behind it. Not asking for numbers, but why would the ROC go up in that business?
Oh, why would the return on capital. No. That was not the ROC growing. ROC sort of is a point in time thing. When I'm making the investment, what do I expect that capital to return back to me as the business grows and matures, right. It was not a one year to the next year point. The point was compared to if you are a point solution provider versus owning the full stack, you deliver much better economics and hence your return on capital improves. Right. If you sunk in, let's say INR 100 behind just making validators, or you sunk in INR 100 behind owning the full stack, that INR 100 will deliver a much higher return, more than two and a half times the return if you were not a point solution vendor and a player across the whole stack. Right.
Sanjay also has this point around us being a lot more competitive as. Other thing is, if you own the whole stack, it also makes you strategically very agile. I think, Sanjay, you want to make that point.
Yeah
how it helps you win more.
I'll just give you an example. The numbers aside. Right now, the entire war was going on. It was a logistic nightmare. Normally, to deliver a 24 station, it takes 180 to 200 days. Because we were having the complete stack with us, we did 77. Flat in 77 days, we actually made all the metro stations up and running. When you have a complete stack with you, right from a gate, having its mechanism. You have the electronics of the gate, you have the validators, you have the back end software, the front end software. You have the complete control on the entire value chain and the logistic requirements of it. That's how when we go and we deliver the project, and that's how the return of your invoicing comes faster.
This is what Ashish wanted to tell you when we have the complete value chain with us. I hope I answered the question well. Right. Thanks.
All right. We have a question here. Yes.
Hi. Hello. Varun Gandhi from Finavenue. Ashish, I really acknowledge and I share the optimism you have about this space, especially the banking software segment. I see AI, we can harness intelligence, and that really improves our value proposition. In fact, I'm not a technologist per se, yet it appears that the product suite is quite competitive. The story has been very consistent from your end. If I may play the devil's advocate, this optimism has not translated into tangible growth in your BFS segment. Right? My question to you is, what exactly is holding us back? Is it delayed client spending? Is it challenges in displacing incumbents? Is it some sort of implementation bottleneck or anything else? More importantly, I wish to understand what differentiated strategies would you adopt to tackle these challenges, yeah?
It's easy for me to just stand here, ask question to you. Business isn't always a smooth ride. I acknowledge the fact that BFS clients are hard to deal with, especially on technology, complex decision making, very long sales life cycle, et cetera. Would like to hear your thoughts over there. When exactly would you expect the revenue to tangibly convert? What differentiated strategies are you adopting?
Okay. I don't know of any player who sells to the banking world in this part of the world, I mean any, who was an INR 200 crore banking software business four years back and is an INR 650 crore, INR 670 crore business. You're like 3.5 times the business. If you think it's not growing, I don't know what your definition of growth is. Right? Basically, there is no one in the industry who's grown at the pace we've grown for the last five years. I don't mean we are one of the top five. I mean no one. Right? I would love to be educated on some other player who's been growing faster than we've been growing. Whether it's banking or TIG, it doesn't even matter. It's a wash. Both of them grew about the same over the period of time, right?
I don't know how to answer that question, right? Nothing is stopping us from growing. It's just we grow. We grew 20% last year. We grew 37% the year before. Only banking software, right? Banking software grew a little bit over 19% last year. It grew probably 37% the year before, 33% the year before that, 14% the year before that, 11% the year before that. Right? Yes, you will have ups and downs. The problem is businesses don't run on an Excel sheet. They need their breathing space. They grow. What is our strategy to grow the business? We have a very, very clear point of view on where the space is going. We have built out products which are extremely competitive in the market.
If you look at the lending stack, there is almost no one who has the same share that we have on corporate loan origination amongst the large banks in Asia. If you look at transaction banking, in the markets we've been in, we've won more than any competitor, but we've probably won more than all the competitors put together. Right? We've won every single deal that's come in India over the last two and a half years in the transaction banking space, with the exception of one deal, and the only reason we did not win it is because we did not bid. Right? I don't know. We find no problems growing. We will continue to grow. Banking is, you are right, a space where the quality of product matters a lot. Your delivery reputation matters even more. Local references in the market matter a lot.
As we go into newer and newer markets, as we push hard into Europe, push hard into Middle East, we need to continue doing what has made us successful in the markets we are in. Which is one, go to market with a product that is better than everyone else. Chloe showed that with Fintra. We can show that product after product after product. Lending, transaction banking, trade finance, Interact, whichever way you look at it, these are not one of the top products in the market. These are clearly the number one products in their space. Then you go and build the local references, walk the hard mile, slowly go and acquire share, right? Our focus is now push very hard. Transaction banking stack, push very hard in Southeast Asia, which is traditionally the stronghold of the lending stack.
Push very hard with both lending and transaction banking in Europe. Double down on the new AI-native stack, which we feel very confident will win share. I have no doubts. We'll be talking in one year time with a number of wins behind that. Then double down on the enterprise AI space, right? Which is where we can hand-hold a lot of banks in terms of scaling their stack, right? That is the growth strategy. That's not to say we've not grown in the past, and that's what we'll do. That is to say that is what we've done in the past to grow faster than the rest of the industry, and that is what we'll continue to do in the future.
Go ahead.
All right. The gentleman there, I believe. If you can just pass on the mic.
Hello?
Hello.
I'd request you to please stand up, sir.
Thank you for the opportunity. Last few years, we've seen data center come on to its own, and we also saw banking and transit. Just want to know, what is the vision with Arya.ai? Is it going to be its own fourth part of our growth story, or is it more on lacing it over all the other products which we have? Just a follow-up that, have we exited the smart city business? Any clarity, last year we had that in our presentation. This year it's not there. Just to get an update on that. Thank you.
Okay. Good question. Look, Arya.ai, it will continue to ride both the legs. It is a key driver of growth on the entire software stack because like we said, the AI native stack as well as adding intelligence layer to our existing solution base is very important, and that comes through the AI capability that Arya.ai has. A lot of the growth on the software side over the last two years has happened because of our leadership on the AI side. That will continue to happen. Arya.ai will continue to grow on both sides. I do see it as a fairly sizable business by itself as we progress in time.
You will see it grow through the banking stack, but you'll also see it grow standalone as a business as we push hard around the AI engineering stack, around various parts that Deekshith was talking about on the overall chain. Right? It will grow very strongly on both sides.
All right. I have time for two final questions only. Sir, I'll come to you. There's a lady here who's been raising her hand. Yes.
Hi, sir. Krupa Desai from Electrum PMS. My first question was on the data center thing. Now data center is a large part of our order book. Have we got any advances there? Because in some call, you have said that it's a negative working capital business.
I did not say negative, I said very small, in the past than when we looked at the business.
True.
I think now that we are going after larger projects, I know Sanjay can comment, but I think there will be from time to time some need for capital in the business. The way we really look at the capital demand from that business is. We will go after larger and larger projects now that we have the capability to do it as well as capability to execute. Then you need to put the return on capital lens in terms of what do you actually get, if at all your capital gets held up for some time in that business. Right? It's not that you're investing for long term in that business.
It is from time to time, given the scale of the project, you'll get caught out with some scale of capital, then is the margin allowing you to deliver the right level of return, right? I think to some extent, given the sheer scale of demand in the market, we can sort of get the payment terms. We can command the sort of margins to deliver that return. Right? I think that is the sort of slight change in shift of the nature of the business.
Right now we are not getting any advances?
We do get advances from the customer time to time. Am I audible? Okay. We do get advances on the data center businesses when we get projects. As what Ashish said, there are different milestone payments required, so there's not much capital required. That's a life cycle of any project or any implementation we are doing. We time to time get advances from project to project depending on the size of the project.
Typically we will.
Yeah.
Yeah, it depends on the specific project, typically we will.
Okay. Going ahead, we always used to spend 9%-10% of our revenue in R&D. Do we expect to capitalize any of that going ahead?
We are at a point in time where the R&D spend would be elevated for, I don't know, a few quarters at least. As we build out the sort of the new AI native stack for pretty much every single part we are in. In the past where we've capitalized is one, on the most of the banking stack buildout, in the past we did not capitalize, but there were some regulated businesses for which we did. Then slowly we stepped it up. I think to some extent it's a decision of how can we fund the build which needs to be done. We feel that there is a growing sort of demand.
For Fintra, for example, if you look at it, right now the stack built out, but we see a massive opportunity to go and take revenue right now. To a large extent, I think it comes down to can you client fund the build? We've got to do the build anyways. Can you client fund the build or will you need to build first and then go and sell? In the case of Fintra or in the case of some of the work on the AI side, we had to go and do the build ourselves rather than client fund it. Increasingly, we see a lot more opportunity to client fund. In which case, the level of R&D or other stuff you need to capitalize should be much lesser.
The capitalization would continue, right?
Ma'am, that would be your final question. We'll have to move on to others.
Yeah. Good evening.
Maybe in some shape, yes. In some shape, yes.
What would be the amortization policy then?
Okay. I'll need to get back to you on that.
All right. Yeah.
Yeah. Good evening. Hi. Yeah. This is regarding your lending and credit risk management platform. You're using a lot of artificial intelligence in that. Vis-à-vis your competitors, what is the USP you all have? Why would a fintech, NBFC or a bank take your product? Secondly, once the client is empaneled, what is the structure for the revenue? Is it an annuity-based model, three years, five years, or is it a monthly retainership, or is there a license agreement?
I'll take it.
Sanjay.
I'll just start with the initial part and then hand over to Shekhar. When we talk about lending and when we talk about credit risk, right, a lot of people are using AI as added plugins. Whereas for us, AI is native. Right? We are not using AI as glorified RPA or as something which sits outside the software. For us, when we talk about AI, especially you mentioned credit risk management, it is native because we have Arya, we use our own LLMs, we use our own models and RAG, and we are also training our own models around the data that we have. You're getting a software which is AI inside out, not a software which has AI on the side as a plugin, which differentiates very distinctly because it then evolves.
What that also means is that your software, by nature, becomes API-first, by nature becomes a engine which can then feed onto others. That is a big distinction that we are able to see today, right? Competition doesn't, as to the best of my knowledge, do that. They use AI most as plugins. They use RPA in some places, and they use engines from the side, right? Before I hand over to Shekhar, the models that we have are both subscription-based as well as license-based. So both upfront license as well as pay-per-use or subscription, depending on whether it's assets under management or whether it is users, and we offer them both across this. Because the software is modular, it's a combination of what modules you take, what processes you use that will define this. Shekhar? Yeah.
The differentiation to the customer, right, is not just at the software level, but the overall experience and the efficiency lift that they can get. It's that combination together with the domain expertise that we have, right, in learning the best practices across various banks, across various countries, that is packaged into that. That's where the differentiation is.
Sorry?
Annuity?
Three years.
For the annuity.
Three years.
Our term, the subscription, we are flexible customer to customer.
Flexible.
I think five years is our term of choice. Yeah. For the right customer, we will negotiate plus minus a bit.
All right, final question of the day to this lady.
Thank you. Good evening, all. I had one question on the data center part of it. We are very much clear on the design and engineering, wherein comps also benchmark you in the sense Aurionpro in terms of design engineering plus data AI. My question would be more on the MEP side, which is segment 2 on slide 24, plus your modular products, right? One question on the comp side, are we comparing it to Vertiv, ABB, Delta in terms of modular products? Can you do Sterling and Wilson in terms of the second part, which is DC MEP, right? One on the comp benchmarking. Second, where are the margins more? Is it on the data engineering plus AI versus the DC MEP and in-house modular? Also since we mentioned CapEx light.
Is my understanding correct, is CapEx light is not to where we could see DC MEP and the modular, because it is CapEx heavy to be really sure. It is on comp margins and
You are asking me to give you an entire P&L, right? Of course, why we are going into the modular side on manufacturing is that if Bhaskar pointed out was not because of the factor that we are going to have more margins. Of course we would have. It is going to the market faster than what is required. If you see, these days, the customer wants the implementation to be faster. Second, the reason is that we have the skilled resources. Shortage of skilled resources because the training and the market requires them is more. Avoid that to have much faster return to the things. That is the reason we are going to the modular side of the business. Of course, engineering is required, a lot of other factors are required.
The most important reason is to be more faster, need less resources, less skilled resources at that site. That is one of the reason. It is a logistic nightmare these days because these kind of products just not goes into the data center, but the manufacturing boom, which is coming in India. Everybody is talking about manufacturing. There was a question of manufacturing out here as well, right? As the manufacturing boom is coming up, especially on the pharmaceutical side. We have a lot of liquid cooling solutions which are required. Similar the things are required on the data center sites. There are other some of the components which are required on the data center site as well and in a few of the manufacturing components. A lot of the volume goes into both side of the businesses.
Just to have a control and to deliver the project faster, that's the reason we are going for few of the projects which are on the both side of the requirements. Of course, the margin will increase because we will save some money on the infrastructure side. Yeah, of course. I hope. Can you give her the mic, please?
Tell her to catch.
She will take away all the P&L from me.
We'd also be happy to take any question you have offline, by the way. Just in case.
Oh, you can take.
Yeah.
We'll let you know.
We are running out of time, we'll have to wrap the Q&A session. If you do have questions, which I know that you do, please feel free to write to us at investorrelations@aurionpro.com. I'll have it up on screen. Can I please have the email ID up on screen?
Sure.
Of course, our leaders will also be outside, and I'm sure they'll be more than happy to take any questions offline as well.
Showing something?
All right. Till the time they put it up, thank you to our leadership. There you go. That's investorrelations@aurionpro.com. If you want, you can take a screenshot. Thank you so much to our leaders. Thank you so much for gracing us with your presence here on stage.
Thank you.
Graciously answering all the questions that our guests had. Thank you. We appreciate it.
Thank you.
Ladies and gentlemen, thank you so much for joining us today here at the Aurionpro Investor Meet 2026. It is such an honor for all of us to have you here with us. The future rewards builders, not spectators. A big thank you to everyone for your questions and engagement. Please do join us for high tea outside. Thank you, ladies and gentlemen. Have a wonderful evening ahead.