E2E Networks Ltd Q3 FY2024-25 Earnings Conference Call

ladies and gentlemen good day and welcome to e2e networks limited Q3 fy2 earnings conference call hosted by go India advisor lpp as a reminder all participant lines will be in the lesson only mode and there will be an opportunity for you to ask questions after the presentation concludes should you need assistance during the conference call please signal an operator by pressing the star then zero on your Touchstone phone I now hand the conference over to M somia from go India advisers thank you and over to you thank you Steve uh good evening everyone we welcome

you to e2e networks limited Q3 fi25 results con call we have with us on the call today Mr tarun Dua managing director Miss Mega Raha the whole time director and Chief Financial Officer and Mr Ron Gaba the company secretary I must remind you that the discussion on today's call may include certain forward-looking statements and must be viewed in conjunction with the risk that the company May face I now request Mr T daa to take us to the company's business and financial highlights subsequence to which we'll open the floor for Q&A thank you and over to

you sir uh thank you Sam uh good afternoon everyone thank you for joining us for the E network Q3 fy2 earnings form I hope you are all doing very well uh we are happy to share our journey and provide an update on the steps that we are taking to sustain our growth and strengthen our position in the cloud GPU and AIML workload space uh as you are all aware uh last quarter uh we achieved a strong operational momentum driven by key advancements uh deployed in our Cloud infrastructure our cumulative deployment have now grown to nearly

700 h00 gpus 256 h200 gpus and around 700 uh non h00 SL non h200 gpus and we also successfully raised another round of more than thousand crores from L&T for expanding our accelerated Cloud infrastructure and focusing on the Next Generation cloud gpus and GPU clusters uh this quarter one of the most important developments for us has been uh deepening of our strategic partnership with L&T a collaboration that marked the sign ific milestone in our journey towards revolutionizing Ai and Cloud infrastructure in India with lmt's deep expertise in data center management and our Advanced AI compute

infrastructure and Cloud experience both the companies can jointly offer robust and scalable solutions to the Enterprise government and organizations this collaboration will open new revenue streams Drive growth in high demand AI services and we will also improve our operational efficiency leading to uh enhance profitability and revenue uh together we will leverage each other's strength to capture a larger market share and create a more powerful presence in the AI and Cloud sector each quarter we continue to stay ahead of the curve by embracing latest advancements in cutting a GPU technology uh in line with our growth

plan we have also we are also expanding our Data Center capacity uh from 4.2 megawatt to uh nearly 10.2 megawatts uh which is a move which is supported by our recent fundraise we continue to innovate at a rid SP to expand our B tested scalable high performance Cloud infrastructure for Education research Enterprises government and startups the sege in demand for cloud and AI Services has helped us achieve strong uh growth and strengthened we have strengthened our position in our in the market by leveraging our Advanced AI cloud platform we are committed to driving India's digital

transformation empowering government and Enterprise initiatives and positioning India as a global leader in Ai and Cloud Innovation uh so our platform supports a wide range of cloud native services today including CPU and GPU environments virtual machines uh Native containers servicess uh architecture mode uh we offer flexible High perform performance Storage Solutions across uh object storage across uh block storage parall file systems and we also offer like Advanced networking load balancing uh firewalls and uh relational database Services no seq database Services Vector DB Services uh we have continued to develop our AI ml platform called te

which is designed for data scientists and developers to streamline their AIML workload uh simil like training insurence and model endpoint deployment with an early M advantage in the AI space since 2020 uh we continue to offer a superior price to Performance ratio helping customers scale without any long-term commitments to our Cloud as the AI Market especially generative AI rapidly expands we are positioned to help businesses capitalize on ai's transformative potential filling the demand Supply Gap in India and supporting countryes digital Evolution now I would like to hand over our call to CFO Mega uh who

will briefly touch upon the financial and operational highlights of the quarter under review uh over to you m thank you tarun and good afternoon everyone let me first start by giving you some of the key financial highlights I will summarize the performance of Q3 F by 2425 for Q3 f25 the total revenues to De irr 416 million which witness a substantial growth of 73.7% on year on-year basis Aida for the quarter is 246 million INR which further provides a growth of around 119% on year- on-year basis ailda margin for the current quarter is 59% which

demonstrate a growth of 1 2 to three bips year on year bat is reported at6 million which demonstrate growth of 108% on year on-year basis Pat margin for December quarter is 27.8% and diluted EPS is 7.03 for the quarter which is around 86.5% on year on year increase now we can do a quick comparison with last quarter that is September 24th we have witnessed total revenue decline of 12.6% on quarter quarter basis from 476 million to 46 million in the current quarter Aida for the current quarter is 246 million as against 314 million on the

previous quarter as a result of this pad for the current quarter is 116 million exhibiting quarter on quarter decline of 4.3% from Q2 f25 while our Outlook overall outlook on AI compute infrastructure in all forms remains very strong there has been some dip in the Rue numbers as compared to previous quter so the decline in revenue and growth in the current quarter is mainly due to churn and downscaling of training deployment which are uh as you know busty nature and the revenue fact of De provisioning has tended to be concentrated during this quarter due to

concentration and due to a smaller scale we anticipate the effect of buy training workloads would eventually become muted at a larger scale and then we intend to build Cloud GPU Computing infrastructure we anticipate the Resurgence in demand for advanced AI solution over medium term uh now important update as well during 2024 we had raised a total of 14849 billion through preferen issue of equity shares out of this we have utilized in 1508 158 million till q325 and we have balance of 13349 million as in December 31st and that concludes the update for the quarter and

now we can open the floor for question answer session thank you very much we will now begin the question and answer session anyone who wishes to ask a question question may press star and one on your Touchstone telephone if you wish to remove yourself from the question queue you may press star and two participants are requested to use hand while asking a question in order to ensure that the management is able to answer questions from all participants please limit your questions to two per participants if you have any further questions you may please fall back

in the question Que ladies and gentlemen we will wait for a moment for the question Q assembles the first question is from the line of Asa from PK advisers please go ahead hi am I audible yes ma'am yeah so ma'am uh the first question is that what was a kex deployed in Q3 fi25 and how much kex we planned to do ahead in q1 Q2 Q3 of f26 so we have deployed 207 million de milon that is 21es H and for the next year ma'am so we we are uh we are kind of like like

very agile in terms of like how we deploy the kex uh so we are anticipating that like uh majorly based on the pipeline we will evaluate and based on that like we will uh kind of like uh do just in time deployment as much as possible over next several quarters and uh uh based on that like the Outlook is not based on a plan but like uh practically by looking at at the end of the quarter that like what was the kex that was done so it would be based on the demand Outlook uh that

we uh kind of like come up with okay okay so my second question is I understand the answer was given earlier but I missed it I wanted to understand what was the reason for a drop in quarter on quarter Revenue uh sure sure sure see we currently have a fair small scale and compared to that like some of our larger customers like uh their deployments were like quite large now our goal obviously is to kind of like build a much larger infrastructure which is spread across like much larger footprint of larger customers uh now training

workloads by their very nature are bursty so which means that like typically although you don't expect them like okay at what time does like a training get over or like a particular data science group decides that okay uh the training needs are over for now uh so unfortunately all of these got concentrated during the Q3 and uh we have also seen somewhat muted demand uh uh for uh the end of this quarter December typically is a slow month so where we have not been able to kind of like uh uh recover back uh that unutilized

uh infrastructure to a higher utilization rate so that is that is the current status today okay got it so obviously in the medium term we expect that like as the footprint of customers grows larger uh then the effect of like one or two large customers uh declining their training needs like should not like uh uh impact us so that is that is subject to growth of our overall scale understood sir okay that's it thank you you thank you the next question is from the line of Amar Mora from Lucky Investments please go ahead yeah am

I audible to hello yes yes Amar we are able to hear you hi hi hi sir couple of questions from my side firstly if you can help us understand this LMT Tire uh I mean how it is going to help us in Long Run in terms of gaining some large Enterprise clients that is number one number two is in terms of our current I think we have something around 4,200 Mega so what is the peak Revenue potential of uh this capacity and by when we will be reaching that kind of peak utilization level and in

that how much of the percentage would be Enterprise versus the uh you know the retail these are two questions uh these are three questions actually versus retail okay sure sure so your first question is like about the lnt ti up how it would help us in the long run uh see obviously uh a small company when they tie up with a big company like uh you get the advantage of the size and scale and connections that a bigger company brings in and also so their own capabilities in terms of Technology uh as well as infrastructure

so over here like we obviously intend to leverage uh all of those things it's quite early days I think it's been like practically like I think like uh uh less than less than a uh few weeks so basically where we have started uh interlocking started working together and uh jointly exploring uh opportunities with the customer so in the long run I believe that like uh uh the uh Interlock in terms of like uh jointly exploring opportunities would uh obviously help both the businesses who bring their own strength to the table and uh will obviously be

pursuing a lot more uh what we internally call as like uh long cycle business opportunities so compared to working withes or smaller companies the long cycle business opportunity are obviously long-term in nature and more rewarding in terms of uh lifetime value and uh smaller companies obviously make the decisions fast that has been the main stay of our businesses our business till now uh I think that natural change will come over like a medium term where we start working with larger customers uh in conjunction with our partners including lnt and uh uh I think that is

one of the major changes that we are going to see where we are pursuing uh more Enterprise business uh apart from what we continue to pursue in the and startup and uh other organizations that we have continued to do uh now in terms of megawatt capacity like that's the measure of the data center capacity like how much it load you can put in uh typically these are are fairly well-known numbers in the industry which are point in time numbers like that one megawatt of what it pertains to today uh like would vary between like say

$25 million to $50 million uh over a spectrum of say uh a smaller player uh and like a large hyperscaler so each megawatt would represent like different Revenue levels from for those so it could vary anywhere between let say $10 million to 50 million and it is also dependent on the kind of workloads now regarding percentage of Enterprise versus retail uh I think like it would be very hard to predict today but uh my assumption uh is that like mostly uh wherever in the industry there are like uh large revenues large profit pools obviously they

are the ones who invest the maximum amount of money into compute so what I believe is that like would be a shift from smaller companies to bigger companies in terms of overall concentration of our Revenue over a over the medium and long term now would be very hard to say like how the percentages would vary but the kind of capabilities we are bring building today are F very focused on meeting the requirements of the larger Enterprises for sir thank you sir I'll come back in sure sure the next question is from the line of Ashwin

Kia from Alchemy Capital Management please go ahead me thank you for your time um just one question was uh how has your Revenue percentage changed from training to influence quarter on quarter and how do you expect that to change moving forward for the next year um see the major concentration of the revenue obviously uh has been training will continue to be training I think training will continue to play play a very major role uh I think like for the foreseeable future uh that being said like of course the impact of uh large wory training workloads

is what we have seen in this quarter and uh but as we scale up like I think like uh that that effect would be less and less pronounced so I think it's a function of us being in our very early days of our journey uh as of today uh which is what we are seeing uh I think that is probably the question you're trying to ask right so like uh the other part being that like I believe that ultimately it's going to be somewhere the ratio eventually like over the medium and long term I think

it should become like more like uh 50/50 or maybe 6040 in favor of like training and what is your inference Revenue today percentage of your total revenue we don't really like uh track it to today very separately because there is like it's a very nebulous line between uh building training infrastructure for uh foundational models being built from the ground up versus foundational models being uh fine tuned at what level uh so like the graduation from training to inference is like uh kind of like there's a whole range so you can't really classify uh what constitutes

training versus inance uh so insurance and production uh like would continue to remain smaller like even in the in the global context itself like so ultimately it's the training uh so today we are like all the AI is like in its early phase so for significant future like training is going to constitute of majority of revenue for not just for us but overall in the cloud ecosystem like training would constitute the majority of the revenue okay thank you okay thank you the next question is from the line of garit coil from invest analytics advisory please

go ahead am yes yes brother can you help me understand our geographical concentration particularly in Data Center business like North America is the biggest market for the data center and if interest rate remain elevated hypers scalers like meta Microsoft May scale down the CeX which in return will impact our business so can you put some color on it like is that going to happen or otherwise what your take on it uh so I uh can't really uh comment on like what uh large Global players are doing like from the you uh from whatever we are

like obviously all of us are hearing in the industry news is that like the investment into AI is definitely going up now regarding data centers themselves like we rely on data centers only in India so we have one facility in the North in the Delhi NCR region and we are in the process of establishing the second facility uh uh near in the Chennai greater Chennai area uh uh and uh uh like so geographically from a infrastructure point of view we are concentrated in India from a customer point of view primarily the customers we tend to

serve uh are in India now uh over the medium and long term obviously we will uh continue to look for opportunities to expand out of India into other markets as well U so and this may necessarily not be on the uh public Cloud but these opportunities could be in the form of software and services and support as well okay underst thank you the next question is from the line of Ken kapasi from taus Investments please go ahead uh hello yes you AUD please go ahead yeah sir what will be the impact of the company of

this new GPU GPU rule proposed by the US government in the short term uh I don't think there would be any immediate impact so one is of course there is like about like a more than a period of a quarter before these regulations come into effect uh secondly based on uh uh what India is doing today in terms of overall volume see currently we are not really hitting the limits uh that are being placed now uh and there are like certain exceptions which we have not fully studied but like there are exceptions around uh certain

end users who are consuming like up to 1,700 gpus then not being counted in the overall country limit uh there would be companies based in the US itself when they bring their infrastructure to India potentially they not being counted so overall like I think like that impact uh assessment would be more uh like I our belief is that like that should be assessed like in about like twoe time frame rather than uh on a immediate basis thanks uh does that answer your question K yes yes yeah the next question is from the line of kha

from n please go ahead yeah thanks for the opportunity I hope I'm Audible hello yes please go ahead yeah so as I can see that your non h100s and h200 has seen an increase of 100 gpus so is it like that the demand for the h20s and h100s are currently not much so that is why you have deployed the demand environment continues to remain strong so the the pipeline is there obviously for both h100s as well as h200 uh what we are potentially uh kind of like missing is like uh uh like immediate and closures

that we were seeing in the past so I think like the sales Cycles have grown uh I don't think the demand has gone away like I think the sales Cycles have become longer I think that is what we are seeing okay so are like the s100s and s200 are they completely utilized or they are underutilized as of now no no they are under of today so how much that would be the percentage if you have that number uh it's a a fluctuating number because like a lot of uh workloads in the cloud it's very very

hard to measure like a point in time very very easily so but we have reasonable capacity on both sides both being consumed as well as being uh made available to the Des Okay and like while we are like sitting on a like pile of cash so we see like the utilization has only been 150 crores so how much time you think that the funds going to take time to utilize or like how the Apex CLE if you can comment on that like see very very hard to uh commit to kind of like uh spend that

money in a market where the uh infrastructure infrastructure as like the high-tech equipment like GP and CPUs uh they get like upgraded versions so it is very very important not to kind of like build a lot of inventory of uh what would potentially uh kind of like will get superseded by a newer version although these are all long Cycle products they will sell for seven eight years uh but it is also important to have dry order for the latest versions so that has always been our strategy that like always be able to invest for every

version so so that way we will continue to kind of like evaluate the demand and continue to kind of like uh try and keep the utilization High uh but not have so little capacity idle that like we are not able to service new demand so it's a balance that we have to draw out I think like that balance will get established and uh I think those uh future kex Cycles like they will obviously continue to get established and when we look back we will be able to say with certainty that okay uh this was the

KX that was spent as opposed to trying to predict today so we we are reactive not predictive so that is what we'll continue to do like overall outlook for both medium-term and longterm we continue to uh see that the AI uh AI transformator potential and its adoption is definitely picking up okay got it so my last question is Nvidia has recently launched bits I guess their new product uh if I'm Rec correcting the name correct so is that a threat to the you know Cloud GPU since the the device is really smallet in the compute

Market you have to understand that like uh there would be multiple approaches to solve various problems so obviously there they would have Nvidia would have seen gaps in terms of what can be serviced from the cloud uh for like uh micro unit see the edge devices have always existed so uh like if you look at Edge compute like that has always existed now they have built a slightly more powerful Edge machine it does not take away the need for data center in any way I think that this could be like one more thing that would

add to the ecosystem I don't think like it's a zero some game that like whatever you are able to do on the edge there would be still further more workload from the cloud okay so technically like what is the difference if we could explain it a bit um see like like there's a whole range of devices that uh kind of like do a bit of AI and uh this is a slightly more powerful Edge so uh like when you talk about like what you can do in the cloud like it's a lot more number of

things uh compared to what you can do on the Edge see it's not just about like being able to do an inference or a training in isolation you also need access to Vector DBS you also need access to relational databases you also need access to a lot more storage uh you also need rcy you also need reliability you also need security so there is certain things that you can do on your laptop certain things you can do on your mobile certain things that you can do on workstations certain things that you will be able to

do on a edge device which is uh very powerful uh required for one particular organization or one particular business unit and there would be a lot more things that you will continue to do on the cloud so it's not like really uh comparative uh in terms of like a straightway movement of workloads from here to there okay got it that answers my question thank you so much sure sure sure sure the next question is from the line of ache from C integrated services please go ahead hello sir congrats on the good set of numbers sir

my first question is what is what is our sourcing strategy for the gpus like uh uh do we Source GPU from the uh Nvidia and then we do add some value like we make our AR arure or we do Source it from the companies like HP or Network Technologies okay so that's a good question so gpus do not work in isolation obviously so see I think like before the hopper architecture during the uh EMP architecture the a00 and a30s and a40s Etc like uh there was a very large prevalence of gpus being deployed in the

PCI form factor where you typically bought the gpus and deployed them in uh compatible servers which were uh typically uh on the hardware compatibility list of major vendors like Nvidia or AMD or Intel so that used to be the approach uh today it's a lot more integrated approach where a vendor like a d or HP or a netweb or a super micro I think there are like uh I think 10 to 15 such vendors uh who build GPU servers so now today's approach of like uh kind of like buy a complete integrated Hardware uh which

is a design which is done by a principal like an Nvidia or an AMD or an Intel and uh the major bill of material also comes from an Nvidia or Intel or AMD uh currently of course like we are very majorly focused on Nvidia that does not preclude us uh to work with other players in the future uh and we buy like these integrated servers uh typically that is called the hgx line that's the major one uh some of the other gpus like l4s l4s they continue to follow the older model of being able to

kind of like deploy PCI Cards into uh boxes which typically come without gpus so we do that as well uh but these are less powerful gpus typically the more powerful gpus come in their pre-built integrated boxes from the major uh GPU server vendors okay sir okay very well understood and sir my second question is uh what is the competitive intensity in our kind of business like do we face intense competition from the uh companies like multinational companies like Amazon web services and Google cloud or as as AI has become very very mainstream since uh 201920

since we have been working on AI the competition has become far more intense today and competition in the compute and Technology world has always been intense so that has definitely increased so uh typically there are a lot more players today who are doing the Cloud GPU infrastructure as the market has expanded there is space for even more number of players I think that competitive intensity and the market both will continue to expand okay sir so what is our competitive like if any player if any company would have to choose from the other company like ours

then what is our Competitive Edge and why they will choose us see like everyone will build Niche Solutions which are targeting particular solutions for data science teams particular industry Solutions and uh everyone will eventually find their nich to succeed in so again like I said these are early days we are we continue to learn from our customers we continue to learn from the ecosystem and we have an R&D team which has the ability to move very fast and build and we have built like a lot of uh intellectual property that uh is integrated into our

cloud and uh this will continue to play a big role in terms of differentiation okay sir and sir my last question is on the financial side so uh Mr aay can you please follow back in the queue for further question yeah sure thank you sir thank you thank you so much the next question is from the line of abishek shakar from incret Capital please go ahead hi thanks for the opportunity and congrats t for a good 3Q um my first question is um did 3Q play out as you had anticipated or were there any positive

negative surprises um sorry I didn't understand the question uh so did uh the 3Q uh in terms of you know both Revenue margins in terms of ramp up did it play out as we anticipated or there was a negative surprise and if yes it was towards the end of the quarter towards the midle of the quarter any color in terms of you know how the revenue PA out for us uh in in in terms of like when you are a cloud operator like uh all these things uh basically like you uh like we have maintained

like I think like over last many many calls that like we react to every situation we don't try to predict any situation so basically like nothing surprises us basically so like we operate on the principle that like we have the ability to react to anything understood understood and regarding uh you know you are mentioned about these bursty workloads uh was it was it part of um you know the change was among our top large customers yeah yeah yeah yeah so obviously these are the larger customers that's why the impact is visible if it was smaller

customers then we wouldn't have seen the impact now uh our whole philosopy over here is that eventually as we scale up our infrastructure as we scale up our customer base with larger customers uh larger deals So eventually each single large customer will not impact us that much and that is that is definitely the path that we continue to follow understood just last one data point in terms of depreciation we saw a significant jump in the depreciation how should I read this in the context of uh you know the hardware uh especially the GPU numbers that

You' have shared between the last quarter and this quarter where the increas is only in the uh a00 v00 so how should I read that uh depreciation number uh like not not very easy to answer that question so how do we read depreciation number I guess like uh I can hand that question over to Mega to kind of like understand it better and answer it better um the idea to understand here is that did did any of our assets were built uh I mean were in you know are part of the depreciation but yet not

build um maybe one reason could be the bursty workloads but depreciation is charged on slm basis over a period of 6 years which is the life as per the company's act as well so once we do addition in a particular quarter then depreciation will continue over a straight line method over period of six years understood got it thank you thanks for taking my question thank you ladies and Gentlemen please limit your questions to one per participant as there are several people waiting for their turn the next question is from the line of pankit sha from

the narrow wealth please go ahead yeah hi t uh so actually I wanted to understand uh on the platform side like what are we doing on the platform side which will say differentiate us or uh which will make our customer acquisition uh Journey more smoother uh something on the cloud side like yeah yeah P so there would not be any one single thing which will have the Major Impact but having an integrated platform that works seamlessly across multiple functions uh and uh having a good user experience for the data scientist for devop for developers and

uh incorporating like a lot of abilities that uh are being derived from AI into the platform uh I think that is the key to success and uh uh essentially it's like any any software adoption uh cycle where kind of like you keep taking the product feedback from the customers from uh the industry from the market and uh continue to build at a rapid pace so I think that creates the sustainable long-term Advantage with the platform working on this uh like an end to end as you're saying uh so is it like currently very limited few

players are there or how should we look at this see like one one key Advantage for our platform obviously it has been in continuous operation for past 10 years now so that's one of the so decade long experience of running some very critical Services uh has made the entire platform like very battle tested so that uh core platform continues to be very robust and secure and uh we build on the we continue to build on the same principles and uh continue to focus both on the reliability scalability and uh the features that are required for

our customers okay actually I was trying to understand on the integration side the software capabilities where where we can like move from more like a kex Le business to more like a software software Le business which will be more recuring in nature over a longer term so uh like I didn't really understand what is the question over here like see Cloud consists of like all these things basically so cloud is a uh catch all term for like basically like uh uh what data center provides to you what uh physical Hardware in terms of server switching

infrastructure storage infrastructure that gets deployed all of that gets integrated with the software and uh like all these abilities obviously there are multiple uh delivery approaches towards these abilities like so it could be in the form of a public Cloud it could be in the form of a private Cloud it could be in the form of on premise infrastructure or eventually uh it could be in the form of like some services so like all of them constitut like various parts of the cloud infrastructure like uh you can say that like various ways of uh looking

at the uh like the same thing and software obviously constitutes like a key piece in all of this thank you sir the next question is from the line of harik Gandhi from HP mg shares and securities please go ahead hello sir am I audible yes yes yes hi sir so just uh two questions from M the first question I just wanted two two timelines first uh you mentioned that we are doing a data center expansion so I just wanted to know by when are we planning on executing that and the second uh second part of

the same question is that we've applied for a tender in government AI project right so any update or any time where we can expect an answer on that we we are trying to kind of like uh build a second new location as quickly as possible uh in the South uh near Chennai uh and uh second like basically like whatever is happening in India AI I think like that is uh highly visible and public information so as we kind of like see uh any uh impact due to that like we'll obviously inform all the stakeholders So

currently uh I think what has happened is that like there has been a technical evaluation uh where we have qualified a technical evaluation now uh the second part is the U Financial bid uh will uh have been opened but like they have not been uh uh the l1s have not been declared as yet so that is the current status that India so like once the indiai team declares like okay these are the El we have received I think at that point of time they would ask for Elvin to be matched by players who want to

be impanel and uh then I guess like some of the players would get impan understood so but mention over there yeah so but you did not mention the timeline for the data center the Chennai 1 when would that be operational see we are trying to do it as soon as possible I guess like uh during the next quarter it would get operationalized whether we will be able to deploy all the services in the first quarter itself like remains to be seen uh our effort would be to kind of like uh get as many services as

we have in the current location to be also made available in the second location understood sir and I could see a huge amount of over other income uh is that a onetime income or we've uh can can you please explain on that I think majorly it's a it's the the treasury uh income for from the recent fundraises is my understanding understood and just one last bit last month we had this uh so last month I can see the arpu has reduced right for larger customers have some of the larger customers on the training side have

churned or downfill so that has resulted in the AR decline so yeah so when we say we are going to get back to normal that will normalize over a longer period of time right so how how long are we expecting are we are we expecting these numbers to remain in seven we can't put very sharp timelines over there like obviously like the uh AI industry the AI infrastructure comput all of it continues to expand we continue to build the capabilities that we are seeing are needed by our customers and uh we continue to uh convince

newer and bigger customers of our capabilities and uh uh I think like it would be hard to put a timeline on like when uh uh that starts to show up in the number okay so on the safer side we can take this as a conser conservative number and continue or do you expect much more churning going in the short term see like I said that like we are not really predicting anything but like obviously we are reactive rather than to build and expand so I think like we we see them as a part of the

journey like like these are like uh not unexpected events in terms of like training workloads being churn uh so the incre increase in the uh sales cycle I think like uh that is something that like we probably didn't anticipate but I think like over a period of time that increased sales cycle like gets mitigated by doing more effort on the sales by expanding in parallel the number of conversations that you are having understood so thank you so much for answering the questions have a good one thank you the next question is from the line of

amay from amet Capital please go ahead hi uh two questions quickly here you spoke about workloads they're more on the training side uh I want to understand uh are the workloads lower on the inference side due to a the demand from the customers or uh is it because of the capabilities of e2e that's the first question and the second question is uh obviously will try to increase your data center capacity with a lot of Colo capacity uh you know set to hit in the next 2 three years would you prefer utilizing a Colo capacity in

the future keeping the you know asset L model or would you prefer uh building out your own data center uh sure sure so like we obviously have always preferred not to build physical data centers so like we'll continue to rely on collocation uh that is one uh with regard to inference and training like I don't think like uh like see we are like a fully capable player uh in terms of like like anyone who has tested our platform to run like influence like they obviously have the ability to run influence on US influence by it's

very nature like the uh uh it grows slowly as like the adoption of AI grows in uh in terms of like the uh adoption by the end customers uh of Enterprises so the initial volumes required by influencing is always small and then that scales up over a period of time uh so and also inferencing obviously like it starts much smaller than typical training workloads so but over a period of time obviously we uh expect that like uh there would be a normalization between like training and entrance to be somewhere between 40 60 60 in favor

of training or 50/50 so just so that I get this right my understanding is your uh training happens during your development phases and inferences where you actually it's a continuous process although like training teams could take a break where they say that okay sometimes like they would be running like multiple trainings in parallel so trainings do tend to get downscaled for some period so for example uh typically like if you look at December period like when a lot of training teams uh would also be taking some vacation so you would probably not start like I'm

just guessing this is just all a conjecture so like the training teams could decide that okay it's like the uh year end so let's end the uh trainings that we were we were PR previously running and uh come back next year and then redo those trainings there could be of course other product related Reasons I'm just making a conjecture over here uh so ultimately like trainings would not be like a 24 into 7 activity or 365 day activity it would be more like maybe like uh 9 months out of 12 months kind of an activity

but yes training uh during development like it doesn't really go away because you're always building uh newer features for uh your customers or improving whatever you're doing today oh great thank you so much thank you the next question is from the line of par podar from vmpl please go ahead hello sir I have a small query so if we see the revenue part from last year I mean last year and last quarter so the de preciation part and other income part is very high in Revenue it's very high and in the again expenses is very

high we compar with the last year so I'm not able to digest I mean how how I mean are we are having planed for the revenues so that we can match the current like last quarter the share price reached to 5,000 rupees the current is trading as 3500 so are we comfortable or are we confident enough that we will reach the de price again no comments we never comment on the share price okay sir but it's still the revenue part we can leave the share price right the revenue that I mean we uh kind of

like don't predict the revenue we react to the revenue uh so in the sense that like we are obviously trying to build like a lot of scale of infrastructure and capabilities and uh over the medium-term longterm obviously AI is a very large Market with a large number of players and uh uh I think like that market will continue to grow and uh medium to long term obviously uh like we see that like will continue to grow okay thank you thank you the next question is from the line of Sumit jwal an individual investor please go

ahead good evening hello yeah hi hi hi recently I just went through your DT and I have been following the India indiai Mission the to procure the 10,000 so uh my question is that uh how looking the progress going ah and not not about this year I mean the next two three four years that the government is trying to make the democratized GPU and T masses yes see that's a medium to long-term Outlook like obviously that Outlook is very bright like so there is a obviously a government focus on expanding the role of AI in

the Indian economy and uh uh as a country we should not be left behind and I think like we are seeing that in Enterprises as well where uh everyone is trying to figure out like how to and there are implementations of AI that like a lot of Enterprises are already working on so like uh definitely uh that's the overall Outlook like so indiai Mission obviously it is a net positive for our entire AI industry in India where it would help in terms of expanding the overall Market uh regardless of whoever uh kind of like uh

uh becomes the biggest beneficiary of Indi Mission regardless of that the market would certainly expand because of uh the existence of AI Mission and the budgetary support from the government over there and uh overall like we continue to maintain a very positive outlook uh of for AI compute infrastructure and AI services in India thank you the next question is from the line of Ashwin Kia from Alchemy Capital Management please go ahead yeah welcome back hello can you hear me yes yes yes yeah uh I'm curious what is your planned capex for the next quarter or

next two quarters and have you all placed an order for the Blackwell gpus from Nvidia yet or what's the strategy with acquiring those gpus so like uh obviously the plan is to kind of like uh begin with like a uh immediate number of blackv gpus uh currently we have not placed the orders for blackv gpus although we have a lot of conversations going on for acquiring the blackf gpus and uh uh as we kind of like plan the initial capacity it will obviously uh kind of like keep all the stakeholders uh informed about that and

and uh uh yes absolutely like we do intend to build like a significant about of capacity on the blackw Range as and when it becomes available is there any planned caps before that outside of the black oil range on the hopper range in the meantime uh so like uh like there could be like reactive kit uh uh in the hopper range like so uh it would it would depend on uh kind of like the Outlook that is coming from the uh sales conversations that we are having thank you the next question is from the line

of hardik Satya and individual investor please go ahead good evening tun uh yeah hi yeah so on the India AI side uh there was a techical requirement of having th000 gpus made available within 6 months of time frame we we already have more than a th000 gpus so that requirement we practically made the should we assume that that is already part of your ecosystem and that is being consumed by others right now and whenever you get this tender through or so you will have it immediately ready or again we enough flexibility enough flexibility to be

able to kind of like significantly expand our capacity as in when needed so it can be procured in a shorter notice with all the absolutely absolutely ab and the second question on the um Chennai data center any particular reason why we are having that second location given that our current capacity is not fully we have one location one major location in Delhi NCR or rather two major locations in Delhi NCR which are uh kind of like joined together with like a big link and we have a smaller location in Mumbai now Chennai of course is

like uh kind of like uh gives us uh access to a very different seasonic Zone very different type of connectivity compared to what Mumbai and Delhi have so the landing stations uh in in the South uh would be significantly different from The Landing stations you get in Delhi and Mumbai so in that way like in all respects like Chennai or Bangalore was a good choice to have for the secondary or rather the second location not a secondary location both of these locations would be for the primary workload of uh every every possible variety thank you

sir ladies and gentlemen due to time constraint this was the last question for today's conference call I now hand the conference over to the management for the closing comments yeah thank you everyone for uh uh uh listening to our conference call and uh thank you for your questions uh we continue to uh look forward to uh working uh with all of you uh over the long term uh thank you everyone on behalf of cindia advisor lpp that concludes this conference thank you for joining us and you may now disconnect your lines