Trusted Connections_ Ethical AI in Telecom & 6G Networks

Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. being organized on the sidelines of India AI Impact Summit 2026. Today, we are gathered to discuss the new elements of AI and telecommunication. This event is organized by Telecom Regulatory Authority of India, TRI, in collaboration with India AI under Ministry of Electronics and IT. Today, interestingly, on 20th February, TRI marks 29 years of its journey in shaping India’s telecommunication landscape. Representatives from telecom operators, technology OEMs, policymakers, government, academia and media are present here. Aap sabhi ka hardik abhinandan hai. Kalpna kijiye ek aisa telecom network jo khud ko heal kar sakhe that can detect faults even before we know them and deliver seamless connectivity to billions without interruption.

This is not a science fiction. This is the power of AI in telecommunication. Today, AI is transforming industries. And as we look ahead, AI is all set to become even more transformative. From predictive network management to intelligent customer experiences, the possibilities are humongous. Now, it’s my proud privilege to invite Shri Anil Kumar Lahoti ji, Honorable Chairman TRAI, whose dynamic leadership continues to provide direction and strength to the telecom regulatory ecosystem of our country. Chairman sir needs no introduction. His vision has been instrumental in steering TRAI through a rapidly evolving digital landscape. I respectfully request Chairman sir to kindly deliver his inaugural address.

Distinguished leaders from the technology companies, from telecom service providers and industry associations, representatives from government, my colleague members from TRI and other colleagues from TRI, ladies and gentlemen. Good afternoon to all of you. It’s my privilege to welcome all of you to this session on Responsible AI in Telecom. This is a session by the side of India AI Impact Summit. During last few days, we have been listening to the world leaders from governments, technology companies, academia and civil society. AI is now and here. In this context, the very composition of this gathering reflects a shared responsibility of recognition. that artificial intelligence is no longer an emerging add -on to telecommunication. It’s a foundational capability shaping how networks are designed, operated and experienced by users.

Artificial intelligence and telecommunications complement each other to form the backbone for the intelligence era. Telecom networks are emerging as the primary carriers of AI, while AI itself is becoming the intelligence layer of telecom. In the upcoming 6G technology, AI will no longer be an application layer. It will be intrinsic. The telecom networks will be AI native. In this sense, telecom networks are no longer mere data carriers, but these are central pillar of India’s AI infrastructure. Our nationwide fiber backbones and mobile broadband networks constitute one of the most widely distributed digital infrastructures in the world, operating within mature operational and regulatory frameworks. India’s scale gives special significance to this convergence. With over 1 .3 billion telecom subscribers and over 1 billion data users, India operates telecom networks at a scale, where AI -driven automation is no longer optional.

It is indispensable. AI is already being deployed to optimize network performance, predict faults, improve energy efficiency, enhance customer experience, and combat fraud and spam communications. These deployments demonstrate how AI can improve service quality, resilience, and consumer safety when applied responsibly at the network level. India is already witnessing clear gains from the responsible use of AI in the telecom sector. Operators are stating significant energy saving with use of AI. Due to the effectiveness of AI and blockchain -based filtering operators are now flagging or blocking nearly 400 million suspected spam calls or messages each day. Enhanced enforcement and improved oversight of service providers has already led to the disconnection of about 2 .1 million spam numbers. The authority is also advancing the rollout of a digital consent acquisition framework following successful pilot runs with the banks to ensure consumers have digital control over consent for commercial communications.

At the same time, the scale at which AI systems operate is also increasing. The impact of AI in telecom also amplifies their impact. automated decisions taken by algorithms can affect millions of users simultaneously this makes trust the central pillar of AI adoption in telecommunication efficiency gains cannot come at the cost of transparency, accountability or consumer rights as telecom is an essential service public confidence must remain at the core of AI enabled transformation the government of India has been proactive in addressing this balance the India AI mission and the recently articulated AI governance guidelines emphasize a huge role in the development of AI and the recent articulated AI governance guidelines emphasize a huge role in the development of AI and the recent articulated AI governance guidelines emphasize a huge role in the development of AI emphasize a huge role in the development of AI and the recent articulated AI governance guidelines emphasize a huge role in the development of AI emphasize a huge role in the development of AI emphasize a huge role in the development of AI and the recent articulated AI governance guidelines one that encourages innovation while embedding safeguards by design.

These principles are particularly relevant for telecom, where AI systems interact continuously with the citizen, enterprises and public institutions. TRI has been aligned with this approach. In July 2023, TRI issued recommendations on leveraging artificial intelligence and big data in the telecommunications sector, proposing a risk -based regulatory framework for AI in telecom. This approach recognizes that not all AI use cases are safe, but that all AI use cases carry the same level of risk. While low -risk applications may be guided through self -regulation, high -risk use cases, especially those directly affecting consumers, require stronger obligations around transparency, explainability, and human oversight. In April 2024, TRI further facilitated this approach through its recommendations on the regulatory sandbox, enabling live network testing of AI -enabled solutions, including those relevant for 5G and future 6G networks within defined safeguards.

This reflects our regulatory philosophy of enabling innovation while ensuring that public interest remains protected. The MANOV vision announced yesterday by the Honorable PM of India emphasizes a human -centric framework for ethical, accountable, inclusive AI governance. The principles of the Mana vision are equally fundamental to AI governance in telecommunication. Coming back to the agenda of today’s program, the two plenary sessions, we have planned to capture this responsibility very well. The first session featuring technology developers will focus on preparing telecom networks for the AI era and examine how networks must evolve to become more intelligent, autonomous and resilient, while remaining secure and sustainable. The second session with representatives from telecom service providers and GSMA will address building customer trust through AI -driven operations, highlighting governance, ethics, accountability, and customer protection in an environment where AI -based decisions increasingly shape everyday connectivity.

As AI -driven telecom operations scale across borders, issues of interoperability, standards, and ethical alignment become global concerns. India’s experience of deploying AI in telecom at population scale offers valuable lessons, while international cooperation remains essential. To address shared challenges, let me conclude with this thought. AI will undoubtedly shape the future of telecommunications. But it is the way we design, govern and deploy AI that will determine whether this future is trusted, inclusive and resilient. TRI remains committed to working with all stakeholders, industry, policymakers and international partners to ensure that AI in telecom serves both innovation and public good. I wish this session fruitful deliberations and look forward to the insights that will emerge from today’s discussions. Thank you.

Thank you very much, Chairman, sir, for your inspiring address. You have illuminated how regulatory frameworks and policies are evolving AI -driven telecom. Sir, your words make us believe that this transformation is moving forward in a positive way. We are delighted to hear your perspective, sir. Heartful gratitude to all the esteemed speakers and guests. The inaugural session has set a vibrant context for our upcoming discussions. Now our first plenary session will begin. Our first session is on preparing telecom networks for AI era. In this session, our experts will discuss AI adoption in telecom, transparency, security, safety, sustainable AI networks, and embedding responsibility by design. To moderate this insightful discussion, we are honored to invite Sridhar Ranjan Mittal, member TRAI, an eminent telecom policy expert with over three decades of experience in telecom networks, global standards, and spectrum policy.

We are honored to have distinguished panel of industry leaders. I welcome on dies. Our first panelist, Mr. Magnus Eberberg, Chief Technology Officer for Asia Pacific at Ericsson, a global telecom innovator who had played a key role in developing region’s long -term technology vision from 5G deployment to 6G readiness. Joining us next is Dr. Vinesh Sukumar, Vice President, Product Management from Qualcomm, a seasoned product leader with over 20 years of experience in large -scale AI, deep learning, and mobile technologies across global telecom ecosystem. We also welcome Mr. Parsi Tovnen of Nokia, who leads strategic engagement with governments and industry on AI and connectivity initiatives, driving large -scale ecosystem collaboration in cloud. AI and AI RAN. Our next distinguished speaker is Mr.

Shanti Gram Jagannath from Tejas Networks, a technology strategist leading wireless products, network management system, and AI -driven innovations. I request all the panelists to join for a quick photograph session on the demand of the organizers. You all may stand for a moment. Thank you, sirs. We look forward to a thoughtful exchange on how AI -driven innovations are being used in the future. We are redefining telecom capabilities from our panelists. now I hand over the stage to our moderator Sri Rutheranjan Mithrasar to start the session

Thank you Madam Sri Lahoti Chairperson TRAI my colleague member Dr. Tangirala doyens of the industry OEMs your staff accompanying you my colleagues from TRAI industry associations young officers representatives of start -ups so it’s a very important session on how the telecom networks will actually evolve to AI we saw a lot of use cases in the last 2 -3 days related to farming education healthcare but today the focus is telecom and the first session, this session is on the network, the next session as you’ve been informed is the subscriber so as a telecom network we are introduced to term access network, so that is one thing we would like to understand how your access network is evolving with the AI, we all know that in the 6G, AI and communication is one of the important use cases one of the 6 cases similarly going to core what kind of changes do you envisage when you implement AI, especially with respect to core, chairperson spoke about the benefits that are already accruing for the network management when the network management is getting into the AI is getting into the network management management.

Another thing I would also request you to dwell on is that ultimately the AI is going to come on the handsets. So once AI is going to come in on the handsets, what kind of a challenge it will throw to the network? Another important aspect is also raised by chairperson is the security. Are are we going to be challenged by the AI being used for attacking the networks? And what kind of steps we intend to take? Another one thing with the AI is it is compute intensive. So the sustainability as also is listed is going to be important. So it will be important to know what kind of steps the OEMs are envisaging to take care of the sustainability part of it.

You will be all are kind of signatories to the or the countries signatories to the sustainable development goals of the UN so this aspect also is very important now without taking too much time I will, everybody wants to listen from the experts here, I will first like to invite Mr. Magnus Everbring from Ericsson to kindly share

Thank you very much, it’s a great pleasure to be here now, bottom of my heart I’m very impressed by this event and the messages we’ve heard and I think we just heard at the onset of this very much central message to leverage AI fully for ordinary people, consumers, for industries, for enterprise and government functions. It needs to use compute resources and that we often attribute to the data centers and indeed they will be there. But also we need to connect them. And here I think India comes out being very much in the pole position having a well over 90 % population coverage I’ve heard numbers even up to 99 % population coverage for 5G today. And not only that, but truly well performing networks.

That is a fundamental platform to drive innovation on and to drive innovation on together with AI. And looking then for the next five years to come, that indeed is what we will see. And the nations that in a few years time end of this decade are at the leading edge with the best of what 5G can do connecting the data centers with the AI applications in devices will be at the cutting edge and will have an advantage. And they are the ones having the least step into 6G. And again I think India is just in a supreme pole position to be there. The networks already today use AI. They use that to a large degree but still I argue it’s only the beginning.

We use it as the systems are being configured becoming more and more autonomous. The goal in the industry is still a good challenge but reachable is to reach what we call level 4 in TM Forum. By 2028 many mobile operators aspire to be there. and that’s beautiful, that’s very good but it is a big undertaking AI is being part to reach that level taking the next step I argue is what we really do with 6G, that’s to reach level 5 and be fully autonomous to cut the rope, cut the ties a bit and let it run and that we have in mind now as we gradually set the standard for 6G 6G shall be AI native and that will then be a baseline for bringing on the knowledge we’ve gained in 5G and to take the next step with 6G lastly then, networks for AI AI on the application side and here I argue it’s important to go society wide.

India is building its digital stack in an impressive way. Leverage AI in that. Industries come in, develop applications on top of it. We’ll drive efficiency locally and also will be export possibilities. So leverage the 5G systems with AI and cloud and then drive the way into 6G. Thank you.

Thank you. I got We at Qualcomm have been trying to really democratize AI and then try to see if we can translate AI to be resident on devices. These devices could be personal devices like phones, could be a laptop, could be your smart watches, your smart glasses, anything that you can think of. But doing that kind of AI inference on the edge is not easy, and especially when you want to go towards more personalized inference. I always joke with my colleagues that AI historically lacks common sense. How do you really translate that to something that’s meaningful to the user needs a lot of investment. In India, we’re seeing this significantly change. We’re seeing a lot of players putting a lot of focus, and how do we get it attached to the user and make a more important connection that drives a lot of these experiences.

At the same time, it’s also very critical of how do we look at coexistence between what runs on the cloud and what runs on the edge. It’s what we call a concept of hybridization. The concept of hybridization is the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built around the idea of a system that is built Hybrid AI, you know, working with our network operators is also not an easy concept.

It’s always been a challenge to understand, you know, which of these experiences would be transitioned towards the cloud, how do you make that decisions, and what runs on the edge. I think, you know, there’s a lot of research activity happening in this space, and India is definitely leading in this front. I totally expect in the next couple of, you know, months, we would see a strong transition where there’s a, you know, fundamental element of hybrid where you can have both edge and cloud coexist. And last but not

Thank you, Mr. Sukumar. I would like to invite Mr. Pasi from Nokia now.

Namaste. My sincere thanks for the opportunity to be here. It’s exciting. It’s exciting for the sake of this AI Summit, but it’s also very personal for me. I have been coming to the beautiful country of India since 1999, and it has been fascinating to see this country evolving all these years, over these years, in different locations. We were talking earlier today already in how many places I have been experiencing this evolution and development. It’s fascinating. So thanks for the opportunity to be here. It’s very important. I don’t know what to add after such a wonderful keynote by the chairman of the other opening talks. I think it is obvious that AI will also impact telecommunication networks, the network evolution, the business models, the innovation, everything around it.

So instead of going deeper in the technology discussion, because they were very well covered by the previous panelists, industry leaders, I would focus this couple of minutes for how. Not what. But how. How to capture the best of AI era. And for me, it comes around who is able to build the most value -rich, most welcoming, most compelling ecosystems. Who is able to gather like -minded technology players, regulators, government agencies to think together the opportunity, the risks, the challenges, that whole 360 of AI. I don’t believe that any player, even though they would be the smartest people on the planet, are able to… fully capture the 360 by playing alone. Ecosystem it is. How you are able to proactively define the overall value of this AI evolution.

And then transparently and proactively… agree how that value is distributed, how that value is maximized within that ecosystem. It might sound ideological. It might sound a little bit naive, but I’m a firm believer that it is key for success. It’s a key to capture and deliver all the opportunities within this AI era. It is the only way to address the security risks, which are now going to be different than any time earlier. I mean, we are going to have different access network. We are going to have richer number of applications, applications which are transferring more data, which increased amount of data is contributing for the security risks. We have to address that topic together, end to end.

And let’s do it. India can show the direction forward. For whole world. There is a tradition for great. collaboration, great innovation, so let’s do it. Thank you.

Thank you so much, Mr. Pasi. I would now like to invite Mr. Shantaram from Tejas.

Yes. Hi, good afternoon, everyone. Am I audible? Not audible? Now it’s okay. So my colleagues and my boss warned me that, you know, if you’re coming number four, then you may not have much to say, so think of something different. So there are a few things that I would like to say. A few things, you know, being from here, born, brought up, everything in India, making stuff for India, so I’ll try to, I’ll try to anchor my… few comments in terms of what we want to do here. I think the fundamental problem that, I don’t know if you happen to read this book by Sir C .K. Prahlad, where he talked about the economy at the bottom of the pyramid.

So we carry with us a responsibility of solving problems for the bottom of the pyramid, primarily. And Indian telecom especially has that additional responsibility of making sure that access is provided to, and at a very low cost, it is provided to pretty much everybody in the country. Now, in this context, if you want to accelerate the adoption of AI, I leave you with a framework of thought. This is what I use to basically figure out what it takes. So you can either be looking at cost, or you can be looking at revenue. Or both. So when telecom operators have to figure out what are they going to invest in, then these are the sort of the two guidelines or markers that they can use.

When it comes to cost, it is optimization of either the CAPEX or optimization of OPEX in a simplistic sense. OPEX would be operational efficiency. There’s a lot of literature which is available. And my friends from the global OEMs are far ahead in implementation of some of these. We in India are chasing it as well. In terms of the hardware, it leaves you with, again, two choices for architecture. Do you find a way to completely do an AI -first, AI -native architecture? Or should we find a path that has a bolt -on capability? Because, you know, you have equipment that’s already in the field, and there are a lot of investments that we are making, you know, very fresh.

unlike some of the more mature networks in the West where the capital cycle has already gone through you know 4g 5g and it’s it’s paid off over 5 10 years already but in India we have made fresh investments even as early as last year now that kind of equipment if it has to be leveraged for the next 10 years how do you deliver AI that is one challenge that we’ll have to solve so there will be a choice you know do you do AI first or do you do a bolt -on if you are looking at the revenue side of the equation again in simplistic sense there are two sort of big buckets there is a product enhancement where the telecom network itself is enhanced in terms of efficiency there’s a lot of work happening in 3gpp, bouth at 6G alliance, A lot of thought is going in in terms of how do you make, how do you bring in efficiency into the product, essentially optimizing the cost per bit, so to speak, and doing it in a way that is a lot smarter than what we did before.

And on the other side, you have a possibility of generating revenue by providing AI through the telecom network, which Pasi also kind of referred to. So that I see as a possibility of a two -way business model, which is an opportunity that’s available, where on one side you have all the users, you have communities, you have people in our villages, the farmers, etc., who are all on one side of it. And on the other side, you have startups, companies that are building models, companies that are building specific agentic applications. and in between you have the telecom player. And there’s a possibility if we do the frameworks right, and what do I mean by frameworks? The very top thing in terms of framework is trust.

There has to be trust, there has to be some amount of regulation, there has to be some amount of safety that comes with regulation. And allow people to dynamically upload, and it’s like an app store model. So the telecom network essentially becomes a platform where simple models, easy models can be actually uploaded and made accessible to all the users of the telecom network. It doesn’t have to be only the bottom of the pyramid, but of course the bottom and all the layers above. So I’ll leave you with this thought,

So thank you so much Mr. Shantaram for sharing your thoughts. Now for the next few minutes I will first have a round of questions with the panel here and then we will throw the session open for questions from the floor. So Mr. Magnus the first question is that we are talking of optimization with the AI but we always used to say that 4G, 5G networks are self -sorn. The concept of son was already there, self -optimization networks. So what does AI fundamentally change now?

Okay, thank you. Well, it’s always a journey. We are… For every move we make, we learn, we get more insights and then we take that and move further on. So in that sense, although we’re done… more and more autonomous parts of the systems in the past. It doesn’t say we can do more in the future. And we should do more, of course. Now, with AI, we get a very powerful tool to analyze a lot of data and to apply that knowledge onto a set of, in the system, some algorithms. And we’ve had some fantastic observations there. One was just the, in a very part that’s been optimized for decades on how we do the link adaptation, how we control the communication between the system and the device, we’ve managed to optimize the capacity by 10%.

So imagine a water… a loading spectrum of 100 megahertz. Then you had the equivalent of 110 megahertz. by applying this optimized algorithm. And that’s the cutting edge of what we can do today. I’m sure we can do improvements tomorrow also on that. How much? I don’t dare to state. In energy efficiency, as has been discussed there also, we, in another part of the system, applied AI analysis and make that part of the ongoing processing. And energy efficiency was optimized by 33%. For that part. And that’s an enormous savings applying that on a pan -India network, of course. So, again, it is about to, wherever you are, how far you’ve been, continue to do research to understand the future potential, and then step -by -step apply that with the knowledge that you have derived.

And you will continue to take steps and climb the ladder. Thank you.

so next question to you Mr. Sukumar so in the telecom hardware and software space which decisions should be pushed to the edge and what decisions should be centralized

it’s a great question by the way I think it’s really going to come down to elements of key performance indicators I would suppose if you’re looking at areas where you want to focus more on data privacy user privacy better responsiveness, better data management a predictable end to end performance to a large extent that’s going to be happening on the edge and this could look at experiences on data plane loops L1, L2 user kernel space operations anything to do with PII information or user privacy privacy, all that would be edge resident. And as you go more towards cloud, I would say the emphasis is going to be a lot more around fleet management, anything to do with AI, ML training as such.

There’s a concept which historically on the ML land was called MLOps from, you know, from training data all the way to inferencing and monitoring. We have a new model which looks at if there is areas there where things are breaking because of drift, how do we go fix that? That’s where I think cloud definitely helps. Now at the same time, it’s not very binary equation, which is one is edge, one is cloud. There’s also concepts of hybrid, where it’s going to be coexistence. As I was mentioning before in my talk, is that you have to find ways how edge can complement the cloud. And it could be elements of personalization, where you want to drive cloud.

What’s the edge? and you want to go towards more complex scenarios, then you position towards the cloud. A hybrid is in the very early stage. To a large extent, these days routers are very static in nature, meaning those workloads and experiences are predefined, saying X, Y, Z runs on the edge, you know, and A, B, C runs on the cloud. But the most difficult challenge has been is how do these routers can be intelligent enough when you happen to have a multi -turn conversation at some point of time positioned towards the cloud. I think that is something of a huge research topic, and I’m hoping in the next couple of months we’ll see some interesting results.

Thank you.

So the next question for you, Mr. Pasi. So I’m sure a lot of development is already taking with the telecom OEMs. Now, what are the decisions which are taken off -net? What kind of decisions do you expect will be taken off -net? Thank you. and what are the decisions during the operation of the network while using AI you think you would be taking?

Wow. And we have only this four minutes? Okay, okay. But on a serious note, I think jury is still out for this one. So how we are able to… And again, going back to my earlier comment, how we holistically address the topic, how we go through the overall value and that ecosystem and related functionality. I believe more will be done by the network itself. When we design it correctly, it’s able to do many of the security vulnerability assessments by itself. It’s… alerting upstream? Is it going to the edge or further? Let’s see. But I think it needs to be very intense dialogue between the network and edge. More decisions are traveling further to the regional data centers, more we are contributing to inefficiency and hence also the complexity.

So I would, my planning assumption is that I would push as much of the optimization automation to the network itself and then the limited cases to edge and less and less to the actual regional data center, to put it short.

Thank you for that. Shantaram, let me come to the ethical part of it. So let’s say we’re a base station serves urban area and also part of it serves the rural area. So how can we make sure that with the AI at the background the customer in a rural area is not deprived of the bandwidth vis -a -vis with respect to the urban area. So what kind of steps do you, I’m sure you will be looking at those things that the bandwidth is not constrained to an area or to a set of subscribers but what do you what are your thoughts? How can we check these things in a network?

Okay. So in just the easy answer is to buy more equipment. But But that’s a great question. It has always been, you know, this question has been live for many, you know, almost decades. I know we went through a case where, you know, we had net neutrality and those debates also happening. So it’s not different from those types of debates. I think while we look at access to AI, there is access to the central AI, right, which has to go through a backhaul capacity and so on. And obviously there one has to create different types of network slices for different types of use cases. And today’s technology, at least the way that we administer networks, it is quite possible to do that.

And it is possible to do that with single clicks. And with AI and us bringing in operational AI, it is, you know, it can even do it more efficiently. It can do it more efficiently without you having to think too much. you can just say that you know I need to create this kind of a bandwidth for this type of an application and the network management you know the assisted network management can actually go and do that for you. Now this coupled with having a lot of edge access for AI. So you know I’ll give I’ll share an example. In the US they recently launched an application where the telecom network can actually sense your voice metrics and identify you through that.

So imagine that kind of an application here in India where you know your identity etc. can actually be verified by the telecom operator not just by your number or by a digital information but by analog information that you are actually communicating and these types of applications can actually be launched. on the edge of the network so short answer step one try to have a lot of a GI and step to use lot more sophisticated network management capability to clearly separate out different types of

well thank you so much we can have a do the note that the times up but still we can have one or two quick questions yes mr.

good afternoon respected panelists it was a wonderful discussion but as I will try to keep myself short as I was hearing all that, we are having around 118 crores of mobile connections in India, built up over a huge network of say fiber and wireless and lead circuits and everything and once we are trying to introduce AI in that, as we know it’s not AI native, we have to build up in the form of external like apps and all and any single minute of disruption causes a huge resentment and the loss of the you know, all the time and resources, so how do we to actually progress from this 118 crores and we want to feed through the AI what is the vision in front of us that we can really carry forward from here, I will look for any of the panelists, thank you

Perhaps I can start thanks for it, it’s a wonderful question and sorry if I sound like a broken record but without thinking the network evolution end -to -end, you are not able to address it. So it comes to this ecosystem of players that you are able to model what the change is introducing to the network and optimize the network end -to -end. I think it is the only way. Otherwise, you are going to put patch fixes here and there based on certain application behaviors and you are not able to evolve the whole network.

Do you like to also substitute?

So I think essentially what you are saying is that it is a journey. And how do we sort of chart out that journey? There are two, three thoughts on this. I think one is that today we are telecom networks are mostly catering to human users. Of course there are enterprises etc. Three, four years, five years down the road I do expect AI users to be starting to dominate. So the business model and the regulation has to support that kind of an evolution. I think that is step one. We need to sort of think through what is, you know, how do we handle, so today we have 118 crores mobile phones. Five years from now we might actually have 500 crores of AI agents which are doing various things but they are still communicating either with each other or with central data repositories and so on.

And we need to basically figure out how do you charge for it? What is the, what is the economics of this? You know, how is it all going to work? Who is going to pay for all of that activity? So there is a lot of policy thought process that has to go in. On the physical side, we are building more and more and denser and denser fiber optics which are carrying 8 teras, 20 teras and so on and so forth. And in anticipation of something like this. So I don’t know if that answered your question but thank you so much.

Thank you. Thank you. Thank you.