Operator
Mr. Ninad Kelkar, Company Secretary.
Let me draw your attention to the fact that today's discussion may include certain forward-looking statements, which are predictions, projections, or other estimates about future events. These statements reflect management's current expectations about the future performance of the company and are subject to various risks and uncertainties that may cause actual results to differ materially.
As a reminder, all participant lines will be in the listen-only mode, and there will be an opportunity for you to ask questions after the presentation concludes. I now hand the conference over to Mr.
Ashish Rai, Group CEO and Vice Chairman at Aurionpro Solutions Limited. Following his comments, we will open the forum for question-and-answer session.
Thank you, and over to you, sir.
Ashish Rai
Thank you, and good afternoon, everyone, and welcome to our Q2 FY 2027 earnings call. This quarter marks an important operational step up for Aurionpro.
We delivered 415 crores in revenue, up 16% both year-on-year as well as sequentially. Crossing quarterly revenue threshold for the first time ever establishes a strong baseline for the quarters ahead.
We achieved this growth while preserving solid profitability, even as we absorbed input cost pressures and maintained an aggressive pace of R&D investment. For the quarter, EBITDA came in at 71 crores at a 17.2% margin, and PAT stood at 50 crores at a 12% margin.
Following a quiet start to the fiscal year, momentum is visibly building. When an industry undergoes a generational technology shift like we're going through, it calls for a disproportionate commitment to product innovation.
We leaned into that investment cycle early, and this quarter, we are beginning to see clear market validation of our roadmap. Let me walk you through what we are seeing across the two business units, starting with banking and fintech.
Global banking, as we all know, is in the middle of a fundamental transition from legacy process-driven software towards agentic AI-native systems. Over the recent quarters, the speed of this change and regional geopolitical headwinds as well prompted banks to take a longer look at their technology roadmaps.
While that stretched out some decision timelines and go-lives, the underlying shift plays directly to our strengths. Delayed implementations are largely back on schedule, and the near-term drag from those headwinds is behind us.
Because we committed the capital and engineering bandwidth especially to build a genuinely built-from-scratch, AI-native banking stack, we are meeting the market with ready products just as the demand is inflecting. Commercial activity has picked up meaningfully, with several large deals in advanced stages and multiple proofs of concept running across Asia, across Europe, led by our agentic AI workflows and AI-native platforms in transaction banking, trade finance.
With our foundational architecture in place, we will ship brand new AI-native software every single quarter this year, sometimes more than once in a quarter, helping clients turn this technology shift into measurable business results. Moving to our technology innovation group, our bets in both transit payments as well as data centers are paying off, driving record additions to our order book and pipeline, as you would have seen.
In transit, Aurionpro is firmly established as a premier partner powering India's automated fare collection ecosystem across flagship metro and mass transit systems. Having proven our full-stack execution at scale across domestic markets and selective global markets, we are taking this blueprint to international markets to capture the fast-growing global smart mobility opportunity.
In data centers, rapid AI workload growth is creating extraordinary demand for high-density compute, and we are deepening our footprint through new mandates with leading hyperscalers. Our ambition on the data center business is clear: we are building the preeminent independent, full-stack AI data center partner in the region, helping hyperscalers, helping AI clouds, and banks bring mission-critical capacity online faster with greater predictability.
What ties transit and data centers together is our atoms-to-bits strategy, which we've talked about before, the conviction that the next technology infrastructure cycle belongs to firms fusing intelligent software with precision hardware. Building on our proven transit manufacturing footprint, we are integrating backward to manufacture critical components for both next-generation data centers and mass transit networks.
Controlling the stack end-to-end gives us tighter command over quality, timelines, and unit economics, driving structural margin efficiencies at scale as India's high-tech manufacturing expands, and we become a key player in it. Across H1 FY 2027, our product investments are converting into commercial velocity, as you would have seen.
As global enterprises embark on history's largest technology cycle to rearchitect for the age of AI, our priority is capturing long-term growth rather than near-term margin maximization through disciplined investments across our businesses. Our H1 performance reflects this planned R&D front-loading and related development cost capitalization to launch platforms like WordAIk for onboarding, Integri-T OS, Lexileap's offerings, etc.
We believe the incremental intensity of this R&D ramp that we went through, it peaked in H1 and will moderate going forward. We head into the second half backed by an order book of more than 2,000 crores, a deep pipeline, and accelerating execution.
Our outlook for H2 remains strong, and we remain focused on executing towards the Vision 2030 and compounding long-term shareholder value. I want to thank Aurionpro teams around the world for their discipline, for customer focus, and hard work over the past six months.
With that, I will hand it back to the moderator, and let's begin the Q&A.
Operator
Thank you very much, sir. Ladies and gentlemen, we will now begin the question-and-answer session.
Participants who wish to ask questions through audio and video can do so by pressing the raise hand icon at the bottom of your screen and wait for your turn to speak. When prompted, you can accept the prompt on your screen, unmute your audio and video, introduce the firm you represent, and ask your questions or give comments.
Participants who wish to ask questions via chat can click on the Q&A icon at the bottom of your screen and post your questions. Ladies and gentlemen, we'll wait for a moment while the question queue assembles.
We have the first question from the line of Kunal Bajaj.
Yeah. Hi.
Am I audible?
Operator
Yes, you are audible.
Ashish Rai
Yeah.
Hi. Good morning and congratulations on the results.
I have broadly three questions. Firstly, that you indicated that incremental investments intensity kind of peaked in H1 and should moderate going forward.
What margin range should we underwrite for H2 FY 2027? How much of this improvement is expected to come from the lower R&D intensity versus the operating leverage and the business mix?
That is one. Secondly, on the point of the strong order book, obviously order book is north of 2,000-odd crores.
What proportion is executable over the next 12 months, and how do we see H2 in terms of conversion, and any quantification as to how we see H2 to be as compared to H1 and overall FY 2027? Lastly, a bookkeeping question.
Currently, we see that the current borrowings have increased to what, 150 crores from 50-odd crores in March, while cash also declined to around 83 crores. Any color on that would be helpful.
Thanks.
Ashish Rai
Okay. Hi.
Thanks, Kunal. Look, on margins, so, yeah, what I meant from the R&D intensity was obviously we continue to build a completely built-from-scratch AI-native version of every single Aurionpro product out there.
We continue to invest in the transit side as well as data center side. The build-out of software intensity will drop because we have obviously launched now our first offering on the trade finance side, we're very close to offering a couple of more, and bulk of the build is behind us, we're diverting that capacity to the revenue side, right?
I think that is how it's playing out on the software side. Overall margin is a function of not just R&D, but also the business mix.
I think this is what is happening if you look under the hood. On banking and fintech, I think we had a very strong H1.
In general, margins are more or less where we expect them to be or slightly better. On the transit side again, over the full year, they've been trending up even beyond this year, and they'll continue to trend up.
On the business mix side, of course, now the data center side is accelerating in pace and growing a lot faster than the other units. It'll probably continue to grow faster than the other units for quite some time, and then that does meaningfully change the mix sort of to a level where, and data center, of course, goes quite a bit below the enterprise margins at the moment till we start building out our products and bring some level of margin expansion on the business.
I think at a margin level, my expectation would be we more or less stay at the level where we are. Under the hood, the banking side will, of course, improve as we go into H2.
The transit side will also improve slightly as we go into H2. Then you'll see at the mixed change to a level where, because data center will probably grow 70%, 80%, if not more, this year, and if that happens, I think the mix will ensure that we're probably more or less the same level at the margin.
I think that's how it's playing out. Let's deliver Q3, Q4 and look at it.
I think increasingly I would say one will have to look at the underlying businesses as they get scale, and not worry too much about the enterprise margins, because whatever the margin percentage drop looks optically is more than made up by the actual volume expansion, right? I think that's how we expect the year to play out.
That would be my sort of commentary on margins. I would say don't plan any meaningful expansion on margin at enterprise level purely because of the revenue growth and the mix change.
H2 I think I probably covered both your questions in the first thing, right? We do, like I said last quarter, we'd been diverting a lot of capacity towards product build-out.
We had been focusing on getting ready for where the market was moving rather than keep on doing what we were doing earlier. We do expect every quarter in the year for us to get stronger as we go.
You already seen that meaningfully from Q1 to Q2, we expanded 16%. I think we may or may not expand at the same pace, but I think you'll see every quarter us getting stronger and converting that order book to revenue, right?
Now on the banking and fintech side, especially now that the product build-outs are lighting, there's a lot more capacity to divert towards revenue. On the borrowing side, I think it's a point-in-time pressure because of one or two very very specific deals, especially on the data center side, these are very large contracts, and we've not delivered contracts of this scale before at least, and it takes some basic retooling in the enterprise to be able to deliver these well.
Over the long run, I don't expect them to put any significant pressure on cash, but over the short run, you do see the borrowings expand because we do need to keep executing while we are ironing out sort of those issues, right? I think that is basically the thing.
By the end of the year, I think we'll be in a much better place both on cash, profits, on revenue, basically all the parameters, that's what I expect.
Sure. That's helpful.
Just a follow-up question from my side. Firstly, on the margins.
Obviously then in that case, is it fair to say that 17% to 18% margin band for '27 and '28 we can expect going forward?
Ashish Rai
I think that's what I would expect. If there is a meaningful change because some part of the business accelerates a lot more or a little bit less than what we expect, then we'll come back and sort of comment at the end of Q3.
I think that's what I would expect right now.
Any quantification on H2 in terms of growth?
Ashish Rai
No, we'll not quantify. You would expect us to sort of accelerate into H2.
I think H2 would be a lot stronger than H1.
Okay. Lastly, on the macro scenario, we had some pressure in the Middle East, how do we see that now?
Ashish Rai
Look, I don't think the Middle East thing is sort of as big a factor for us now. Obviously, it's unfortunate what's happening in the region, and that's affected decision cycles for the businesses.
As an impact on our own sort of results, I don't think it would be a big factor. We do see deal action coming back in the Middle East.
In fact, we do hope over the next quarter or two to announce some significant wins in the region. We do see the region sort of adjusting to the new normal, and deal action may or may not be back to exactly the level it was, let's say before Q4 last year, but we do see it meaningfully back, and we do hope to announce some significant wins in the region before the year is out.
Sure. Thanks.
All the best.
Ashish Rai
Thank you.
Operator
Thank you. The next question comes from the line of Anmol G熟rk.
Ashish Rai
Hi, Anmol.
Operator
Anmol Garg, I request you to accept the prompt on your screen, unmute your audio and video, and ask your question.
Hi, Ashish. Hi.
Ashish Rai
Hi, Anmol.
Couple of things that I wanted to understand. One is that we had signed two very large data center deals recently.
Now, has any one of them started contributing to revenue, or will it start in the second half for us? That would be my first question.
Given that there will be sort of higher share of data center revenues coming in towards the second half, would it be right to say that there could be further margin pressure in towards the 2H?
Ashish Rai
Hi, Anmol. Great question, right?
We are also navigating this there, and then obviously, there is a growth projection that we have penned in for the data center business for the full year. I think it's an interesting question if you instead of growing 70%, you end up growing 100%, what does that mean?
I think to some level data center business, the new contracts have started contributing. Right now, they have started contributing a little bit in terms of revenue, but more in terms of the cash pressures and I think like the previous question we discussed, some pressure on cash as well as borrowings.
Right now, obviously, the cost is ahead of the revenue on that business. Going into H2, you will see meaningful growth, and you will see meaningful conversion.
To a large extent, we have factored in that growth into our plans, I feel more or less, we can hold the margin level unless there is a massive sort of expansion that we cannot foresee at this moment. I think what's happening in that space is sort of very important to understand.
Speed is extremely important in that business. As you can see, we are also diverting a lot of our own R&D to try and solve those problems through a set of products that we are trying to build out.
We talked about the prefab stuff. I think there could be scenarios in which the business expands a lot more than we planned, and then that can put a little bit of pressure on the absolute percentage.
I don't think that really should be the lens with which anyone should be looking. I think more and more where Aurionpro is going, it's important for shareholders to really look at business as an owner of multiple businesses and stop looking at this quarterly number.
Because I think if the business works the way it works, it will consume very little capital, and even at a slightly lower margin, the return on that capital is exceptionally high. What will happen if you say, let's say you grew a lot more than what you're planning, is still a net addition to the absolute number.
Whatever margin dilution may look like, it is basically an optical number because the way we run our businesses is, some businesses need a very high margin to drive the return on capital that we need, some businesses need a much lower margin to drive a return on capital because they consume a lot less capital. From us, right?
Data center as a whole, of course, is the biggest capex cycle the world has ever seen. The question is as a business are we committing capital to it or not?
I think broadly, what we see playing out is a strong contribution to revenue growth, the effect on margin more or less I think is already showing, and you should not see serious deviation unless we have to come back and say the revenue growth is somehow a lot more than what we're projecting or what we're thinking of right now.
Right, Ashish, we used to give revenue growth guidance in the previous sort of years, and now that the revenue stability is coming more and more, and we are much more confident towards the second half, would we start to give the guidance again?
Ashish Rai
No, I don't think so. I don't think there is any player of our size, you can educate me if you or any of the investors can educate me if you need to, but we've looked at this across the market.
At a board level, we discussed this as well. We don't think a player of our size should be giving guidance at this stage.
It's not even got to do with all the uncertainty that we're seeing, right? I think the important thing to sort of understand is we will invest where we see the best return on capital to meaningfully accelerate growth across the segments, and there will be a variability across segments given everything the world is going through.
There's too many uncertain factors. I just don't think it is fair to start throwing a number out there.
It just puts unnecessary pressure on the business to sort of compromise on the long-term objectives of the business, which is to build out a global leader that can play at global scale that we can all be proud of. I think we feel very good about where we are.
We feel we will accelerate through the year, as you're already seeing. I don't think anything is gained by putting a number to it.
I would say we'll come back, and we'll keep talking transparently about the trends we see in the business and sort of run it like that.
Understood. No, fair enough, Ashish.
Ashish, just one more thing that I wanted to understand that on the cash side of things, what we see is that there has been sort of OCF for the first half is also kind of negative right now, and we have increased our borrowings, our receivable looks a bit high. From the overall standpoint, by as we end the year, do we think that we can go up to 60%, 70% of the OCF to EBITDA, given that the second half is usually stronger for us in terms of the cash conversion?
Or you think that for this year the OCF is going to be relatively weak?
Ashish Rai
No, I don't think so. I think we can manage it.
I think so basically for us, everything comes down to H2 execution. Typically, I think you know this very well, the way the cycle in the business works is most of the collections get skewed towards Q3 and Q4.
We will stay sharply focused on execution. I think something between 60% and 70% is doable.
I think that's what we'll sort of stay focused on. The wild card essentially is the growth opportunities available to us right now, right?
I think I have personally never seen such a large, meaningful growth opportunity in front of us for each of our businesses, right? Banking if you look at it, the AI-native stack, it's becoming very obvious that we are clearly ahead of the market, not just sort of our so-called Asian peers, but globally.
The more we take the stack out to Europe, to Middle East, even to US now increasingly, we see a very strong reception. The opportunity to grow the AI-native stack is something I've never seen.
The data center business, of course, the opportunity that exists in front of us, not just as Aurionpro as a firm, but as India as a nation, industry as a whole, is just of a scale and size which was probably what you could say IT services was 30 years back. It's a massive opportunity for us overall as a nation and as an enterprise.
Transit, we are now going global. It is a very very large segment, and we feel we've got every single tool we need to compete globally.
You would have seen the win in Finland. Increasingly, we'll go after European markets and US, and we feel we've got the partnerships, we've got the IP, we've got the capability now to really scale that business globally.
Honestly, the opportunity set in front of us is so humongous at the moment that I think the wild card could be sort of maximizing growth in a particular segment, which may need investment. If it needs, we'll come back and transparently talk about it, right?
I would say we are going through a point in time where we've not seen such large opportunities in front of us, and that may need us to react, obviously with a very sharp return on capital lens and what is best for the Aurionpro shareholder in terms of investing it. If that happens, we'll come and talk about it, right?
By and large, as we go through execution for this year, I think we made the investments, we are executing it, and if we go for the growth number that we have planned right now, we should end up in a much stronger position on every single parameter by the time we finish this year, right? If we need to make incremental investments, we'll come and talk about it.
Thanks, Ashish. Just small thing on the data center business, we have spoken about that we are manufacturing some of the key components.
Directionally, when the data center revenue will have more of our own manufactured product revenue or it will be a mix as of now?
Ashish Rai
Yeah. No, I think so that's exactly the point, Anmol, right?
That is where our R&D focus is right now. On the current span, this is very similar to where transit was 4 years back.
On the current span, we occupy a fairly strong position in the value chain in terms of being able to see all the components. Now the question is, can you now start building in your own products, and every product kind of meaningfully pushes that margin up, right?
I think that is what the focus over the next couple of years will be. The more products we can bring into the same contracts, the better the margin profile should get, and which is we've done this in exactly the same way on the transit side, we know how this works.
We've got the engineering and the R&D strength to do it. We've now increasingly got the manufacturing strength to do it.
It's a question of progressively sort of doing it. It's not an immediate thing, though.
This is several quarters to several years as we slowly progressively build out our strengths in the area.
Sure, Ashish. Thank you so much for answering my questions, and all the best.
Ashish Rai
Thank you.
Operator
Thank you. The next question comes from the line of Vinay Menon.
I request Vinay Menon to accept the prompt on your screen, unmute your audio and video, and ask your question. Vinay Menon, can you please unmute your audio and video and ask your question?
Hello, you can hear me now?
Operator
Yeah. You are audible, sir.
Hi, sir.
Hi. Couple of questions from my side.
One, in the order book, can you just give out how much would the data center part be right now, because since we are executing a couple of large orders, just want to understand what it's contributing right now?
Ashish Rai
It has picked up. I would say it's north of 30% now.
Overall, I think the split is probably 2/3, 1/3 now between the TIG side and the banking and fintech side. Of course, the 1/3 on the banking and fintech gets executed much faster and will deliver the revenue growth that we are planning for, but the other side is skewed quite a bit.
The data center I think the overall order book size is pretty large right now, and there's still some pipeline as well, so, yeah.
Got it. In terms of the capitalization, I saw that in H1, the capitalization has come down drastically, and you have mentioned that the R&D which you have done, major of the investments are kind of through.
Can we expect in H2 and going ahead in FY 2028 that most of the R&D is going to be expensed out, and this is something as a practice as a company we're going to follow ahead?
Ashish Rai
Look, I think the way especially the software side normally works is typically we will build software. These are obviously very large pieces of software, each is several million lines of code.
They are complex pieces that take some time to build. Once you build it, then typically they will last fairly long, and then you will need to kind of incrementally keep on servicing that, which usually is very easy to do that, make that piece client funded, and do that from within the projects, right?
That's how typically we've done it. That's how the software business would normally work.
The capitalization essentially is a function of a point in time I need to do really fresh build-outs, right? Which we decided to do sometime around I would say Q1 of last year or Q2 of last year, and then it slowly accelerated through the year last year where we said we will build out a brand-new, AI-native version of every single Aurionpro product.
As the recent launch on the trade finance side and transaction banking is proving, it was the right bet to make. We will announce wins against that stack I think very very soon, and not just in India, but across Middle East, across Southeast Asia, hopefully Europe as well.
We are clearly ahead of the market on that. I think basically when you've got to build fresh product, then there is no choice.
There is no opportunity to get it client-funded. There is no opportunity to expense it.
It's something that needs to be done. That is what I meant when I said bulk of it is done, because we were really focused on it in H2.
We are still focused on it, but you are reaching the tail end of it for our major products, and you will see those launches come in as well, I suppose We already launched one, and then we'll probably launch depending on the timing and Christmas holidays and all that stuff, one in this quarter or early next quarter. You will see that come through.
You will see the wins start coming in on that stack, which will sort of report, and which means the products are now out in the market, and the incremental work we need to do can get funded from the client-funded projects across projects, right? Those tend to get expensed.
I think that's how the whole thing played. It's not gone to zero, but it's tapered off, and I expect unless we see something really large in front of us for which we need to build something new, I think by and large, we are at the level where most of those are getting finished or finishing off or already finished.
That's how the whole thing plays, right? When you're building new product, you capitalize.
When you are incrementally sort of adding on top of it, there should normally not be a need to capitalize. Then the exception to that would also be, for example, on the payments business where the payments cycle and the sort of return cycle is fairly long, but I need to have the whole capability built right upfront, right, because it's a regulated business.
There is no sort of moving around from that, right? Those kind of things are what would sort of contribute to the capitalization, and then typically in the normal course of the business, you'll get expensed out.
In a long way, now coming back to your question, I think right now we are at a fairly reduced level. The second factor which I was probably not clear about last quarter, which again is something to see, we're not really adding to the development capacity inside Aurionpro for the last several quarters because we are beginning to see the productivity gains from the usage of new tools.
I think it will stay that way as well, right? What really happens as the R&D projects reduce is that I've got more capacity available to divert towards the revenue-earning projects.
That will also help the banking side because it helps us accelerate projects and convert revenue faster now that the capacity is available. Actually, we had that question last quarter.
We're discussing why don't we hire 300 more people, and I don't think that's the right thing to do right now, right? Now we have more revenue-earning capacity available on the banking side, which again you'll see in H2 sort of beginning to start deliver, right?
Sort of positives on both sides, I think.
Okay. Couple of more questions.
One is, one of our peers mentioned that
Ashish Rai
Your line is not clear,
Vinay.
Ashish Rai
Line is not clear.
Some large RFPs have started coming in, in the Indian market also on the
Operator
Vinay Menon sir, I am very sorry to interrupt you, your voice is not audible. Breaking badly.
not present over the last few quarters. Hello?
Operator
Vinay sir, your voice is completely breaking. We are not able to hear you clearly.
Okay. I'll get back in the line.
Thank you.
Operator
Yeah. Thank you.
I'll get back in the line. Thank you.
The next question comes from
Operator
Darshil Jhaveri.
Yeah. Hi.
Good morning, sir. Thank you so much for taking my question.
Hopefully, I'm audible.
Ashish Rai
Hi, Darshil.
Yeah. Hi, sir.
Sir, just wanted a few data points. What's the data center revenue that we've had till now in H1 FY 2027 and last year, sir?
Ashish Rai
I think we don't declare subsegment revenues right now. Let me try to do a sort of the relative thing that I do normally.
I think data center business was a third of the TIG business last year. This year, it would probably be something between I don't know, 40% and 50% thereabouts this year, right?
That's how the mix is sort of changing on that side. The challenge there is the build-out of capacity, and which we mentioned last year as well.
We are really trying to ramp capacity on that side, and that may be the sort of variable there on how quickly can you grow capacity to convert revenue. These are all obviously time-bound projects, you don't have much of a choice.
You need to build capacity in time to deliver it. I think that's the sort of variable out there, right?
I think overall it should end up somewhere between 40% and 50% of the TIG business, and TIG business as a part of the mix is also climbing up.
Correct, sir. Just if my understanding's right, the data center business is usually more intent towards the H2, right?
In H1, if we would not have a higher contribution from it, right?
Ashish Rai
It's grown in H1 as well, but you're right. I think it'll grow a lot more strongly in H2.
Okay. If possible from next quarter or next year, we could have a data center separate segment, because data center and TIG, I could understand, are somewhat distinct, if that would also be helpful.
As it's growing, as you think, 70%, 80%, so that's also be meaningfully easier to get those numbers, if that's possible.
Ashish Rai
No, we've taken note of that. I think it's a sort of a progressive judgment in terms of when we think the business has acquired scale and predictability for sort of it to be a meaningful guide to the shareholder, right?
You also don't want to confuse people with segments which are not at scale. I think you're right.
The business is fast acquiring scale, and if we keep on expanding it at the pace it is, it can become an independent number sometime in the near future.
Yeah. Thank you.
Sir, just wanted to also understand, as you said, we are investing significantly and if we are at the tail end of it, how much have we invested till now, if that number could be handy, and what's the future plan, as a percentage of revenue? What's our R&D going to be like, sir?
Ashish Rai
It's very hard to tell, right? I would love to have the crystal ball which tells me where the world is moving all the time.
I think the way we looked at it, typically we used to be in the 8% to 10% range in terms of R&D. We had to meaningfully step it up last year, and like I was very transparent in the last couple of quarters as well, there is no other right reaction if we are the long-term sort of what is best for the Aurionpro owner and shareholder over the long run.
We've never seen this scale of opportunity before, and if we say, yeah, let's save a few dollars and not invest against it, I don't think that's the right thing to do as well. You don't know where the world will be.
I think we've got a very specific point of view. Within all the variables that you don't know, I think we like to stick around things that we do know and can invest against.
I think we have a very very specific and very clear point of view on the segments we are in, on what will really help us make revenue, right? If you look at the whole AI sort of investment going on, and obviously the world has never seen a cycle like this before, there is a lot of conviction that a lot of investors have and capital allocators have in terms of where they're putting capital.
Typically, if you just split the world into three three three layers, you say there is a model layer, there is something above the model, there's something below the model. The model layer honestly, of course, with Arya and with our specialized models and the research around tabular foundation models will continue to invest there to some extent, because we ourselves are consumers of it, and that we can predict.
Beyond that, it is very hard for us to predict where sort of the return comes in on that layer. What we have is a very specific point of view on above the model layer, where the world is going.
We feel very strongly that the banking software 2.0 thesis that I had more than a year back where we said software needs to change because the agents need the rule rails, they need the tools to really perform in the banking world. My statement to that was, you can't run a bullet train on steam engine rails, you can't really make the AI move in the enterprise unless, on back of 30-year-old sort of back offices and 30-year-old technologies, you need to upgrade the architecture.
We are seeing a very strong sort of evidence of that in the market. We invested ahead.
Now it's time for us to sort of monetize that, and continue to invest and continue to grow on where the software puck will move. Similarly, below the model layer, we had a very clear sort of view on what the world needs, what India needs, in terms of data center capacity, and the fact that we are so woefully short of the right kind of players existing in the space.
In some cases, we're obviously very strong on the design side, very strong on the integrated engineering side. Some problems you're trying to solve are fairly basic problems, it's like sending Sachin Tendulkar to play, I don't know what, street cricket, but the advantage of that is we are able to really understand that and convert that into a productized sort of offering, which will help us drive margins, right?
I think both above and below the model layer is where we feel the monetization opportunities are, and it allows us to take advantage of this massive investment cycle that the world is going through right now. We will, of course, run it.
We've traditionally been a very capital-light business, and we'll continue to be a capital-light business. The question is, look at where the capital is going, and see how you actually meaningfully allocate the limited capital we have to make most of it, right?
I think the way we are executing is with a very clear thought on where we think the opportunity is for the business to grow in each of these segments. Above the layer, of course, the transit side is doing its own thing, the banking and fintech side have their own bets, and below, you've got the data center side with its own bets, right?
That's how the whole thing is sort of playing out, or at least that's how we intend to play it as we go.
That's really helpful. Just two last questions from my end.
One is in terms of just the threat of cybersecurity, like recently due to AI attacks and everything. Are we looking for something in that space?
I think we're using significant AI in generation, right, in our own captive consumption. Is there some token cost that we see that going ahead, because if these guys have to reach profitability, they want to IPO, they might increase the cost.
Is that something that's going to be a significant cost challenge that you see coming forward, or how do you look at
Ashish Rai
it? Cybersecurity, I think we see that as obviously it's a threat to the world in general, but it's a fairly strong opportunity for us to play.
We set up the cybersecurity practice now four quarters back. We went in for the cert in certification and stuff.
Basically, we will build out a fairly strong set of capabilities on the cybersecurity side. It's still a very fledgling practice, so, I think from a revenue contribution standpoint, well, I don't want to talk about it right now, but I think we feel that is a fairly strong capability to build out over the next four quarters, which we are already building out.
I think we initially had a cloud services business and that we slowly expanded into moving in on the cybersecurity side as well, right? You'll see more from us in that space.
I think it's also important to be mindful of the threat. I think, at least personally, if you look at it, I feel there is a tide waiting to happen.
I think many of us feel it in terms of the cybersecurity side. It's very important for us to be aware as an organization on where the threat is, and I think we're investing a lot in terms of that as well in where the sort of perception of threat is and what can we do to minimize the surface that's available to any attacker.
I think that sort of things will go on, and that's important for every enterprise, not just us, right? I think we will continue to sort of meaningfully invest on both sides.
By now, I've lost track of the second question. What was the second question?
I think we use significant amount of AI for our own generation of tools and
Ashish Rai
daily
internal use, right? If these guys want to IPO, they want to raise costs, what do you think?
Can that be an impact on us if you could just Oh,
Ashish Rai
yeah. Right.
Look, I think we do use some third-party tools across the enterprise, which are based on public closed-rate models, which do cost some money. Increasingly, first of all, I don't see too much of a cost inflation on that side.
We are very disciplined, very conscious of where do I get the sort of the most productivity gain out, and typically the productivity gain balances out the cost of the token to the extent that we need to spend on token. I don't see the cost going up.
In fact, and especially, I think you're referring to sort of the IPO talks and all for public model shops. I don't see coding as an area where there is a huge opportunity for sort of price expansion.
It's a fairly competitive space. Again, where the coding models stand today, you probably are very close to a level of kind of intelligence saturation on the regular coding task.
When that happens, I would expect the price to go down rather than go up. That's factor number one.
Because if you don't need the incremental sort of improvement in intelligence that a frontier model sort of gives you if you say my task is fulfilled at the saturation level. I think we're either already there or we are three months from there, but I think that's sort of point number one.
Point number two is, to the extent possible, what I see us doing, we built out our own coding agents, which we are using internally in the enterprise. I feel a scenario where in the next three, four months, we would probably be on completely in-house models.
We'll actually not really be using at least public closed-rates. Right?
We building out our own infrastructure, it gives us an opportunity also to build out agents which are overfitted to our own software and hence deliver a performance which is better than the frontier model anyways, right? Which means there is no token cost to it.
There's obviously an infrastructure cost to it, which we are, sort of progressively investing, and we will probably need to invest a little bit more over the next couple of quarters in terms of setting up our own infra to support the internal models, right? That's point number two.
Both of those together, I see the opportunity for sort of cost overall reducing rather than increasing on that side.
That does. Thank you so much, sir.
All the best for the
Ashish Rai
Thank you.
Operator
Thank you, Mr. Jhaveri.
We have the next question from the line of Mr. Vidit Shah.
I request Mr. Vidit Shah to accept the prompt on the screen, unmute your audio and video, and ask your question.
Morning, sir. Congratulations on a good set of numbers and the order book strengthening.
I have a couple of questions today. Sir, on the order book, want to understand whether the order book includes recurring revenues, which is your AMC, annual licenses, etc.?
If so, how and what is that proportion in the order book?
Ashish Rai
By and large, on the software side, we don't really count the AMCs and all as a part of the order book. I think that is more or less outside.
There may be, or there is some subscription-style revenues which would be in the order book, and there would be some transaction-style revenues which would be in the order book. We've honestly never been able to sort of size that properly in a way that's a meaningful piece of information to the investors, we kind of don't specify that overall, right?
I would say the way one should look at the order book that exists right now, it used to be 80%, it's probably now 70%, 75% of it is executable over the next four six six quarters, I would think, four to six quarters thereabouts, right? Then there will be some tail end, which is again not necessarily recurring, but just lasts much longer in the order book.
Right? I think that's how I would look at it.
AMCs and all, there's a lot of it which we don't count right now, because it does twist the number somewhat.
Right. That's over and above the order book?
Ashish Rai
Yeah.
Sir, I think you answered my question. My next question was how much of this 2,050 crore order book would be executed in H2 of '27 and how much in '28 and then whatever was incremental beyond that?
If you could get a sense.
Ashish Rai
I think it's hard to throw a very specific number there, but that's why I say a lot depends on project execution, right? I would say 70% at least is executable over the next four or five quarters thereabouts.
Okay. Couple more questions.
If you could get a number of what the unbilled revenue was as on 30th September? Wanted to get that number, if possible.
Ashish Rai
I think if you've not declared the number, I think it's hard for me to sort of mention that on a call, or mention that in anyway. Look, I think I would say the unbilled this is the way I look at it, probably that's helpful to the investor overall, right?
Under the hood, if you look at it, not just on unbilled, right, on unbilled, on AR, on cash, and probably also applies to profitability, but less so, banking has had a very strong H1, right? I think on those things, it's generally been fairly positive.
On the transit again side of the business as well as the small like some tail end of the government work and all, has had a very strong H1. In those parameters, it's actually the H1 is much better than last year's H1.
On the data center side, just given the very large orders we've picked up, and we are still kind of seating them, trying to get on top of them, and we've had some issues, contractual issues in terms of being able to bill and stuff, that is a meaningful negative on that sense in H1, right? The sum total of that is that you are more or less at the same level that you were last year.
Under the hood, the core business which used to exist last year actually had a much stronger H1, but the incremental new business we built out on the data center side kind of diminished that somewhat, right? I think that is how I would put it, but overall, it's more or less it's slightly higher than last year, but more or less at similar levels as last year as a percentage of the story, right?
I would say look at all these pieces together and then look at it from the standpoint of the business, and the big wild card we had, the big variable we had, was the large data center orders that we're delivering where we have, to be very transparent, a bit of struggle in terms of controlling the cash and unbilled side, but that is well under control in Q3, and we expect H2 to be very very strong on that side.
Sure, sir. Thank you for that.
How long does it take for this unbilled revenue to get billed, or what is the conversion cycle like?
Ashish Rai
Typically not long, it's just a regular part of the business,
right?
Ashish Rai
It's not just us, is the entire industry. It's a normal case of where you started, you have to commit capital, start work, commit people, and the contracts are taking care of themselves in the normal course.
Right, sir.
Ashish Rai
The billing. Yeah.
One final question on the deals. Ujjivan Bank, iCashpro deal and the metro line 5 order were significant deals in Q1.
Any significant deals in BFSI and transit that we've won in Q2?
Ashish Rai
We won a number of things. The thing is, the way we declare orders is we normally sort of try to keep a threshold of what we sort of send out a release for and not, right, and it has to be sort of a material thing for it to come.
I would say Q2 overall, we saw a very strong trend in terms of winning deals, and some of that announcement should be coming through as we complete those wins into contracts and all on I think all sides of the business. Transit, we did announce this entry into Nordics with the Finland win.
We are very close in a number of cases there, you'll see more wins coming in on that side. Banking, for sure, you'll see a lot more wins coming on all sides.
They have to be of a meaningful size, which we do see in Q3 and Q4 sort of coming through, including, I think most importantly, on the new AI-native stack, right? Because so far, I think what we've seen is sort of the result of the investments, the inputs.
I think what we'll start showing is the outputs, which is in terms of actual market wins with serious banks in serious markets, right? I think a lot of good stuff is going to come through on the banking side as well.
Right, sir. Thank you so much for this.
Wish you a great year ahead.
Ashish Rai
Thank you.
Operator
Thank you. We have the next question from the line of Hitendra Pradhan.
I request Mr. Pradhan to accept the prompt on the screen, unmute his audio and video, and ask the question.
Hi, sir. I hope I'm audible.
Ashish Rai
I can hear you. There's an echo, but let's go.
Sir, my question is on the DC project and what is our working capital requirement on our DC projects? If you can give us some sense of how the revenue recognition is happening on the DC projects, that will be helpful.
Ashish Rai
Okay. Look, I think working capital requirement, of course, as you can see, is a little bit high in the H1, I think.
Obviously, there is a little bit of working capital requirement. I think relative to the size of the business and the size of the opportunity and the size of the profit you can make in it, it's not super significant.
Right now, there is a little bit of stress because we're starting off the projects, and we're trying to sort of iron out some kinks in terms of how you bill and when you bill and all that stuff. I think longer run, there will be some working capital requirement, I think related to the size of the profit opportunity not unmanageable, and we are getting into a lot more discipline in how that is managed.
Of course, those costs are factored into the profitability, right? We know what we're going to spend.
It's not either related to size of us or size of the business, overall, I think it's not going to be something we can't manage. Sorry.
Other than working capital, what was the other part of the question?
Also on, sir, the revenue recognition. Yeah.
Ashish Rai
Revenue recognition,
uh-huh.
Ashish Rai
I think that is a bit slow right now. I think it will again accelerate into H2.
Typically, I think the way to think about revenue recognition is these are projects with very project-style sort of revenue recognition, right? You need to progressively deliver and then progressively take that revenue, which is not different from what we are used to doing on the sort of, let's say, the services side or the IT side, or even the banking projects, although banking projects, you do tend to front-load a lot of stuff.
I think that's how you have to see, it's progressive recognition of revenue as the projects progress. Typically, what we are seeing is these are 12-, 14-month projects these days.
There is a increasing pressure on timelines that we see at least on that business where earlier clients were more prepared for a long gestation project, while now they're typically getting shrunk to 12 months or less, or 18 18 months or less.
That's helpful, sir. Thank you.
Ashish Rai
Thank you.
Operator
Thank you. The next question comes from the line of Neil Jhaveri.
I request Mr. Jhaveri to accept the prompt on the screen, unmute his audio and video, and ask your question.
Mr. Jhaveri, please unmute your audio and ask your question, please.
Thank you for the opportunity. Just one question from my end.
The equipment and product license revenue has increased materially, while software services revenue declined sequentially. What explains this shift and what can be expected as the baseline going forward?
Ashish Rai
Look, I would not read too much into sort of one quarter to the next quarter movement. I think largely it's a function of sort of the mix overall getting skewed towards the data center side a little bit.
Of course, we continue to deliver the large projects on the transit side and the banking and all that stuff as well. I would not see this as a trend.
I would say especially the license revenue on the banking side should be fairly strong going into H2. That will probably still add to that line a little bit.
Software and services again is a function of what you are implementing, right? I think that number will probably expand meaningfully in Q3 and Q4.
It's hard to really substantiate that mix. It's sort of one quarter to the next, it's just a question of the projects you're delivering that quarter.
I would not sort of draw a trend line from that.
That was helpful. Thank you.
Operator
Thank you. Due to the paucity of time, that was the last question for the day.
I now hand the conference over to Mr. Ashish for his closing comments.
Ashish Rai
Okay. Thank you very much.
Thank you for taking the time to join the call today. If I was to summarize where we are, I think it's been a humongous effort from the team to really get back into sort of solid growth as we get into Q2, is obviously a very very strong expansion from Q1 to Q2 in terms of what we could deliver.
I feel we are making the right investments right now to capitalize on the very significant growth opportunities that exist for the data center side, banking, transit, all of our businesses, payments as well. You will see a lot more sort of evidence of the sort of the efficacy of the investments or the right timing of the investments that we've made not just on the data center side and the banking side, but I think across the business overall going into the next half and the next few years.
I look forward to spending more time with you when the next quarter ends. Thank you for joining the call.
Thank you.
Operator
Thank you, sir. On behalf of Aurionpro Solutions Limited, that concludes this conference.
Thank you for joining us, and you may now exit the meeting. Thank you.
Already premium?