Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Is 5G killing 6G?
31 Plays12 days ago
What does it really take to move from 5G to 6G?
In this episode, I am joined by Sebastian Barros, Managing Director at Circles to unpack how the realities of 5G deployment—its wins, its gaps, and its growing pains—are directly shaping the development of 6G. Will 6G be a clean generational leap? or should it be emerging as a response to the lessons learned from 5G from densification challenges to energy efficiency, edge computing, and the economics of network investment?
We explore the cultural transformation telcos must embrace to make 6G viable—not just faster networks, but a platform addressing the big market gap of "monetization". AI takes center stage in our conversation—not only as a driver of new traffic patterns and services, but as a foundational element of 6G architecture itself.
Finally, we look beyond 6G: Is this the end of cellular technology long developing cycles?
Recommended
Transcript
A new guest to chat about 5G, 6G, and what is coming next
00:00:01
Pablo Tomasi
Hello everyone and welcome back to the Telco Drift. I'm Pablo, your host. And in today's episode, I'm joined by Sebastian Barros, Managing the Director at Circles. And we're going to be talking about an amazing topic because we're going to discuss whether 5G has been killing 6G or has been killing the opportunities for 6G to be an innovative technology. But ah before deep diving into that, Sebastian, welcome to the Telco Drift.
00:00:30
Sebastian Barros
Hey, Pablo, thank you very much for this invitation. I'm really happy to be here. So I think it's a super hot topic, especially as we're getting closer to Mobile World Congress. So I think it's a very important talk discussion to have.
00:00:46
Pablo Tomasi
So before again, we you know we start talking about 5G, 6G, and I'm one on the on the skeptical side of 6G. Probably people that have been following me know that, and you know who hopefully you're going to be little bit more optimistic.
00:00:59
Pablo Tomasi
But can just provide a very short intro about yourself, ah you know a little bit of what you've been doing so far and why you are the right person so for this topic?
00:01:04
Sebastian Barros
Sure, sure.
00:01:07
Sebastian Barros
Sure. So for the ones who don't know me, I've been almost 23 years in the telco industry. I've been working with several companies. I was with Liberty, so I was on the telco side for a while.
00:01:23
Sebastian Barros
Then I moved to Ericsson, and then I worked with Ericsson in Latin America, Europe, and Asia. Then I moved to Google Cloud. so they went at At that time, they were starting their Teleco Verticals, so I worked with them, specifically especially in Asia.
00:01:40
Sebastian Barros
and Then I moved back to Latin America. My home country is Chile, so I was closer. My base now is Mexico. I was with Ericsson for five years. I went back Ericsson to help them launch 5G in the region.
Passion and Publications in AI and Telecom
00:01:55
Sebastian Barros
And now I'm in a new job since last year with the Circles. Circles is an enabler of digital telco infrastructure. We provide all the infrastructure you need as a telco, but as a service.
00:02:06
Pablo Tomasi
Mm-hmm.
00:02:11
Sebastian Barros
And it's still based in Mexico. So that's a little bit about me. And well, I write a lot. I have launched now my third a book on AI.
00:02:23
Sebastian Barros
I also have a lot of research papers, I spend a lot of time writing. For the ones who know me, also link it in, I usually publish one one post a day and also have my Substack So a lot of writing and researching. So I think, yeah, and this is something that really get me very, very excited to talk about the future of our industry, which I think is in our hands to make it better.
00:02:51
Pablo Tomasi
Fantastic. And you know for for those that you know haven't had the chance to connect with you on LinkedIn, definitely recommend. you know You're posting lots of interesting articles and articles are always you know ah creating quite quite a bit of conversation back and forth, you know lots of comments.
00:03:06
Pablo Tomasi
So very good to see that the industry is ah is alive and well, in a sense. Every now and then it's good to hear that. But if we go back to to the sort of 5G vs.
00:03:12
Sebastian Barros
Thank you.
00:03:14
Pablo Tomasi
6G, ah call it and I call it 5G versus 6G, probably I shouldn't, but
5G Disappointment in Monetization
00:03:20
Pablo Tomasi
anyways. so So my starting point is this, 5G can potentially be seen as a bit of a hit and miss type of scenario for many telcos. And what I want to try to understand from you is that, first of all, do you think that is the case? And if that's the case, how is that going to affect or how is that affecting the development of 6G, are the processes that we have in place correct?
00:03:42
Sebastian Barros
Thank you.
00:03:45
Pablo Tomasi
ah Is going through the same process exactly as we did for 5G the right solution for this new technology? Or do we need to sort of radically change the way that we want the new technology to be developed?
00:03:58
Sebastian Barros
oh it's it's it's a good It's a good topic. I think, of course, in hindsight, it's always easy to criticize. i think 5G probably is a bit of a strange animal, right? It was launched in 2019.
00:04:13
Sebastian Barros
By many more political reasons, the standard was not really launched as a full standard. It was actually divided into, you could say, chapters. The first one being non-standalone and then standalone, which required, of course, changes in in your architecture.
00:04:34
Sebastian Barros
ah and And the problem, I think, was expectations, right? I mean, it's always about expectations. When vendors and the industry started pushing 5G around 2016 until 2019, there was a lot of conversation about 5G being the savior of monetization.
00:04:54
Sebastian Barros
So at that time, and still is the case, telco were struggling. They were growing at 1%, 2%. below the their inflation rates, and 5G was positioned as, hey, this is going to be your monetization tool.
00:05:10
Sebastian Barros
And when we launch start launching 5G, we realized that 5G NSA was really LTE with steroids. It could actually increase the capacity per site massively. It could also drive more download link speeds.
00:05:27
Sebastian Barros
In fact, if you look at India, right, the difference between LTE and is 19 times faster for the later.
00:05:31
Pablo Tomasi
Mm-hmm.
00:05:34
Sebastian Barros
So definitely there was an improvement, right? But all these amazing cases that we talked in the past about autonomous drivers, cars, remote surgeries and many others, x XR, didn't happen, right?
00:05:51
Sebastian Barros
And then around 2022, we said, no, no, look, the problem was that what we launched was not the real 5G. Now the real is going to happen with 5GSA, get your core ready, and all operators did that. And by now, I think probably around 100 telcos have launched commercially 5GSA, some of them are already nationwide.
00:06:13
Sebastian Barros
and revenues are not coming. So that's because I say, look, new between 2019 until now, I have spent around $2 trillion USD so in capex and spectrum, and revenues are flat.
00:06:30
Sebastian Barros
And of course, that created super, super big clash of
Technical Success vs. Monetization Challenges
00:06:34
Sebastian Barros
expectations. Now, On the performance side, if I look at other part of the equation, the cost side, the performance, 5G is really bringing a lot of performance enhancements, right? coming Starting from the NR, which is ah um the the radio spectrum signaling and processing, which is like 10 to 15 times more efficient, to energy savings in every single generation of radios, to ways to apply ai like sleeping cells when you are not using so much traffic, and many, many, many many other things.
00:06:55
Pablo Tomasi
Mm-hmm.
00:07:06
Sebastian Barros
Battery from the devices that can be actually reduced. So there is a lot of improvement. Now, that is not really solving the one and biggest elephant in the room for telcos, which is monetization.
00:07:20
Sebastian Barros
So then now the question for telcos is like, hey, Ericsson, Nokia, Huawei, I'm going to need to go into another big spending round from 2030 onwards, and I want to understand what is going to be the business case.
00:07:36
Sebastian Barros
And I don't think we have a real answer today. We have good ideas of what 6G can do, but I don't see in any of those good answers
00:07:41
Pablo Tomasi
Mm-hmm.
00:07:51
Sebastian Barros
a good solution for the monetization problem we have today.
00:07:54
Pablo Tomasi
So is is it fair to say, i mean, so technically this the the solution 5G delivered, right? You mentioned improvement, you mentioned, you know, energy and cost savings, but sort of ah the the more qualitative like aspect or like, okay, I have something beautiful, but i need to sell it and I need to make money.
00:08:12
Pablo Tomasi
That hasn't really been been successful.
00:08:15
Sebastian Barros
Yeah.
00:08:15
Pablo Tomasi
and and And I'm wondering, So from a telco perspective, right? i imagine I'm the you know big CEO of a telco. I probably got burned. Like I'm spending lots of money. I don't see enough returns, but do I have another choice?
00:08:31
Pablo Tomasi
I mean.
00:08:31
Sebastian Barros
Wow, that's a super interesting question. The problem is like this, and and I think this is a little bit about our industry, is full of vested
Telcos' Challenges with Vendors and Innovation
00:08:43
Sebastian Barros
interests. I mean, someone wants to sell you spectrum.
00:08:47
Sebastian Barros
and I think they will continue trying to sell you spectrum. Someone wants to sell you more radios for that spectrum and more fiber to support those radio sites that use more spectrum and bigger radios and so forth, right? So the telco in a way sits in the middle. Telcos doesn't doesn't have any R&D capabilities.
00:09:08
Sebastian Barros
In the telco world, they have outsourced this to the Huawei, the Ericsson, the Samsung, the Nokia of the world. a telco even large telco like a tier one telco might have one or two percent of their sales into r&d while techno company will have 20 percent every single last 15 probably i think is above 20 20 so in a way telcos are prisoners of what the vendors decide do.
00:09:29
Pablo Tomasi
Yeah.
00:09:42
Sebastian Barros
Of course, they have a big say, but but in a way they are influenced. So I think that's that's the biggest problem, Pablo, that I don't think telcos have a way out because what's going to happen, and I think I was relating in one of my articles. So by 2030, you will have three times more traffic coming in your networks.
00:10:04
Sebastian Barros
a we have seen a decrease, but still is. in Nominal size is huge. I mean, you are in 180 exabytes per month.
The Need for 6G Amidst Unclear Monetization
00:10:13
Sebastian Barros
In 2030, you need to manage 500 exabytes.
00:10:13
Pablo Tomasi
Mm-hmm. Mm-hmm.
00:10:17
Sebastian Barros
That's a very conservative scenario. Now, there is no other way than adding... new layers of spectrum bands which probably is 6 gigahertz and if you have 6 gigahertz already you will need to buy your radios And that, if you multiply for 10,000 sites or even a large telco with 100,000 sites, that's a lot of money.
00:10:44
Sebastian Barros
So in a way, telcos will need to go into that 6G cycle. ah What I think it might change will be the mindset, because maybe telcos this time will know, look,
00:10:56
Pablo Tomasi
Mm-hmm. Mm-hmm.
00:10:58
Sebastian Barros
I'm not going to do this for monetization. That revenue side will not be impacted massively. But I need to do this in order to have better energy savings, more capacity per side, all other things that will control my costs.
00:11:15
Sebastian Barros
that's That's the reality. So I think 6G will happen anyway. Now the question for me is, how we can apply some of the learnings from 5G into that process that is actually starting as we speak.
00:11:34
Pablo Tomasi
So for me, and and one of the notes that I prepared for this is sort of a tech versus culture debate in a sense, because it seems from from what you're saying that effectively the telcos are forced to have a larger pipe. And I know no one likes the comparison of the pipes, but does does bring the point home in in a way.
00:11:54
Pablo Tomasi
um And that's going to be sort of a more efficient pipe. rather than trying to make, ah you know, using 6G for a more direct monetization is sort of just bringing the cost down. But is is that good enough? I mean...
00:12:08
Sebastian Barros
No, no, it's not good enough. But I think this is my point, huh? I don't think it's good. i I mean, you need to do it. Of course, you need to try to make sure that your factory, your big factory works with the latest technology.
00:12:17
Pablo Tomasi
Mm-hmm.
00:12:22
Sebastian Barros
That that makes sense. And that might imply new radios, new spectrum, new capabilities, new features, new architecture. I think that's something that needs to happen anyway. And there's one big thing that actually will drive the 6G development, which is that the traffic composition is going to be shifting radically.
00:12:43
Sebastian Barros
from humans to actually intelligence, artificial intelligence inferencing intelligence moving from one side to another. So that actually requires different architectures of
Outdated Monetization Models and Cultural Challenges
00:12:54
Sebastian Barros
thinking. Now, ah the problem is that we are trying to look for the solution in the wrong place.
00:13:04
Sebastian Barros
technology or the G's will not fix our monetization problem. and I think this is the part of the industry, it's super stubborn to learn. I think there is a cultural aspect and think it's great that you mentioned um because i mean culture beats technology every time. So so the problem we have in is this industry is culture.
00:13:26
Sebastian Barros
And I give you an example. We're still attached, very attached. to 20 years old way of looking at monetization. At that time, technical feature like an SMS was the product itself, right?
00:13:41
Sebastian Barros
I mean, you don't need to think too much. You just say, if you want to communicate, send an SMS.
00:13:44
Pablo Tomasi
Yeah.
00:13:46
Sebastian Barros
You pay me some few cents, done. Billions of US dollars and in SMS. Always was the same. Hey, the the technical feature of connecting you with someone else, you pay me per minute,
00:13:59
Sebastian Barros
80% of the revenues 20 years ago was a purely voice. And then data, right? Okay, I have a pipe. You want to do whatever you want. Cats videos, YouTube, Netflix. I don't care.
00:14:09
Sebastian Barros
i sell you this pipe. You pay me per gb Done. So in a way, the technical feature was the product itself. but There was no much thinking.
00:14:21
Sebastian Barros
But the world has changed. The world now...
Learning from NVIDIA's Platform Success
00:14:25
Sebastian Barros
doesn't monetize straight out of the box features. right You need to have at least two components. And I will give you one example.
00:14:33
Sebastian Barros
One is you need to have some sort of a platform or orchestration or a way to expose those things for more complex cases. The second thing is you need to have an army of developers or enterprise who can consume those capabilities.
00:14:50
Sebastian Barros
and And I give you i always give this example of NVIDIA. People wrongly think that NVIDIA makes their money from GPUs. And that's not the case.
00:15:03
Sebastian Barros
GPUs are just... the physical layer, right? But the real monetization for NVIDIA comes from CUDA. CUDA is their operating system.
00:15:15
Sebastian Barros
9% of all the libraries that you use for GPUs are CUDA.
00:15:15
Pablo Tomasi
Mm-hmm. Mm-hmm.
00:15:20
Sebastian Barros
And then on top of that, you have around three, four million developers using e CUDA. So even if you take out from Jensen 1 the GPUs,
00:15:34
Sebastian Barros
you cannot take the whole platform and ecosystem you have developed on top. And that's the part we don't have as Telco. We don't have an army of developers that are ready to go and get into our operating system, build things. And we don't have a joint,
00:15:52
Sebastian Barros
operating system or platform that we can expose easily all these capabilities so so a new G will fix performance, will fix costs but will not fix revenue because that's not what it's supposed to do in a way
00:16:11
Pablo Tomasi
it's it's It's a little bit sad that it's not supposed to fix revenues, which is like the the big challenge for that.
00:16:16
Sebastian Barros
yeah
00:16:17
Pablo Tomasi
But when you're talking about sort of that that idea of platform that that you can expose, it it sounds to me a little bit like the whole network API conversation. and the idea of that programmable network.
00:16:28
Pablo Tomasi
So are you a supporter that do you see legs into it? And and I'm asking because, ah i mean, some some people spent a lot of money acquiring companies for the network APIs and it didn't really turn out so far, at least as they wanted without naming names, but billions were spent there.
00:16:37
Sebastian Barros
do
00:16:44
Sebastian Barros
yeah No, look, I think it's the right thinking, oh but we need to develop
00:16:50
Pablo Tomasi
Mm-hmm.
00:16:55
Sebastian Barros
further more into that thinking so i think if i take for instance ericsson i think ericsson did a pretty interesting move there they bought this company api cpas provider of course they were unlucky because they bought it at the hype of cpas so they pay overpaid and then they need to write off assets and then they say look what we're going to do with this and then I think ericsson was smart enough to say look the The business for Ericsson is not to sell APIs, but we need to create a flying wheel.
00:17:28
Sebastian Barros
What does it mean? That if we help operators to monetize the hundreds of features that we produce in Access and Core, we will get the money back. So they created this Aduna company.
00:17:41
Sebastian Barros
where they are not going to make any any any money. It's more like they're inviting, I don't know, they have 20 shareholders because everyone gets a piece. So you have all the way from America Mobile to Telecom.
00:17:54
Pablo Tomasi
Yeah, I have quite a few telcos because they're on.
00:17:56
Sebastian Barros
Exactly. And the idea is is really not to make money out of APIs, but to say, we need to build this platform where we can expose capabilities. I think the strategy is on the right direction.
00:18:12
Sebastian Barros
um They are, think, too challenged ahead. And I think, of course, they're starting with very basic APIs. I mean, verification.
00:18:22
Pablo Tomasi
Yeah.
00:18:22
Sebastian Barros
Google can do verification.
00:18:23
Pablo Tomasi
yeah
00:18:23
Sebastian Barros
You don't need telcos to do that, to be honest. I mean, of course, it can enhance. But I expect, and if you see the Camara API, it's becoming a little bit more advanced with MCP.
00:18:36
Sebastian Barros
and and few other things of course i think it's an early stage but but it's the right thing that a something is missing here we have the network the network has ton of capabilities but we don't have any any way to expose it securely to developers so that's the right place now there are two things that i think they're missing or three things one is developers uh you can have an amazing platform but the problem is developers and developers are would say a very
Need for Cooperation in Competitive Telco World
00:19:11
Sebastian Barros
different human species i mean i i used to deal with developers in google and google they're amazing we managing these communities but they are very special people right they are not that your typical telco attendee of mobile congress they are very special
00:19:29
Sebastian Barros
And you need millions of those. I don't think anyone doing that job today. I've been advocating that maybe GSMA should take again that role. Because if you don't build that community, you can have the best exp exposure exposing platform.
00:19:43
Pablo Tomasi
Mm-hmm.
00:19:44
Sebastian Barros
Nobody will use it, right? So that's the first thing. The second thing is that telcos have seen themselves always competing against each other. And I think now they need to think a little bit different.
00:19:57
Sebastian Barros
There are certain areas where they need to compete and there certain areas where they need to cooperate. And I was part of the initial setup of of of of this API, CAMARA API, and I think My conversations with Telco in many cases was, oh, no, but I don't want to cooperate with, I don't know, Telefonica or these guys. I want to do it myself.
00:20:19
Sebastian Barros
And then you explain, yeah, but the developer doesn't want to come to Mexico and only work with you and then talk.
00:20:22
Pablo Tomasi
Mm-hmm.
00:20:24
Sebastian Barros
He wants a global platform. I don't know, but that's not the way I work. That mindset is still there, like a silo mindset. And then the third thing is that I think the CAMARA API guys need to get this a little bit more exciting.
00:20:39
Sebastian Barros
Okay, fine. We were looking at the low-hanging fruit of identity, but okay, what else we can do with these networks, right? I mean, can we think a little bit bigger? Like I give an example.
00:20:51
Sebastian Barros
of there's gonna around 16 000 robots by twenty twenty seven
00:21:00
Pablo Tomasi
Mm-hmm. Mm-hmm.
00:21:00
Sebastian Barros
now Those robots will need to have some sort of identity and actually certification and proof of humanoid.
00:21:12
Sebastian Barros
That for me will be something that Telco wants to do, right? You don't want any of these humanoid robots become rough and, don't know, kill someone. That is something you might want to do with leveraging on technologies that we have, like eSIM which is amazing technology.
00:21:27
Sebastian Barros
So I think those are the three things that need to happen for making this like a, okay, like the CUDA of Telecom.
00:21:34
Pablo Tomasi
yeah it's It's interesting because as when when you're saying that telcos don't get along well, I think ah that that's still the case 100%. I mean, most of the conversation that that I have with telcos, whether they say directly or not, but there is always that sort of underlying mistrust of their fellow telcos as like you know it's never going to be a sort of one plus one is always going to be um i am one versus the others and and i don't trust those guys it's it's a little bit again mind shift that i think needs to happen
00:21:51
Sebastian Barros
Hmm.
00:22:07
Sebastian Barros
No, and look, I think they they're they're competing with the wrong people. I don't think their competition or their market is competing against another telco.
00:22:17
Sebastian Barros
I mean, I think they they need to think better where the value chain is moving, right?
Redefining Telcos' Role in the AI Economy
00:22:23
Sebastian Barros
And and how they can participate in that value chain.
00:22:23
Pablo Tomasi
Mm-hmm.
00:22:26
Sebastian Barros
One example I love, and I'm really in love with the what they're doing, is Deutsche Telekom. I think they they are getting...
00:22:38
Sebastian Barros
the right idea that a telcos r needs to move ah away from just selling SIM cards. We need to make investments. We need to make partnerships. We need to make ah targeted investment in new infrastructure.
00:22:55
Sebastian Barros
They're trying to find a way that we need to participate in every single, I think they participate in every single initiative, all the way from AI, RAN, ORAN, you name it.
00:23:08
Sebastian Barros
I think that's the way you you you need to think on a new type of Maybe rethinking your role, right?
00:23:14
Pablo Tomasi
So try and try different things and see what what works.
00:23:21
Sebastian Barros
i mean Think about this. I think 20 years ago, our identity identity as a telco was very obvious. We needed to connect people.
00:23:33
Sebastian Barros
Nokia probably s one of the best campaigns. obviously not only in telco but in the world, ad campaigns connecting people. I mean, so straightforward, right? That's what you were doing. Now, after 20 years, we have 9 billion subscribers connected. Still some way to go, but pretty much every everyone in the world can connect to each other or to the world.
00:23:55
Sebastian Barros
So we have lost a little bit about what is our next identity. and And I think that identity comes from the fact that we are moving into a world or an AI economy.
00:24:08
Sebastian Barros
And that AI economy will require not moving bits, ah but actually moving intelligence. And that intelligence cannot only come from global data centers. will need to also be distributed locally. And I think connecting intelligence, moving intelligence is where telcos should start aiming.
00:24:33
Pablo Tomasi
So so let let me throw this out there for you. AI is not a bubble then.
00:24:36
Sebastian Barros
Mm-hmm.
00:24:38
Sebastian Barros
No, no, I don't think so. I think it's very concrete. ah There's always hype. I mean, there are two things, right? There's always hype. so So you need to try to kill some of the hype.
00:24:50
Sebastian Barros
and But there are very concrete things about AI. First of all, that, and this is our problem one of our problems in in industry, we need to understand that AI has been with us for 60 years. Actually, telecommunications...
00:25:06
Sebastian Barros
if you look back Bell Labs and then at&t was actually the first users of ai machine learning and actually even Bell Labs was probably the first company who commercially used a neural networks in the 60s for signaling processing was of course one layer a you know neural network so it's nothing compared to today but but we've been using for 60 years or so we know this thing now
00:25:08
Pablo Tomasi
Yeah.
00:25:28
Pablo Tomasi
yeah
00:25:31
Sebastian Barros
What happened now is, of course, 2017, Transformers comes and say, hey, attention, it's all you need. And then all of a sudden, you have LLMs, right? Now, then on top of LLMs, you have this Gen-AI trend.
00:25:44
Sebastian Barros
And I think that's where we need to be a bit careful. Gen-AI is just a small piece of massive, massive practice, which is usually intelligence
00:25:53
Pablo Tomasi
Mm-hmm.
00:25:54
Sebastian Barros
And I think the overall artificial intelligence will bring value to telcos. ah Probably the Gen-AI part is a little bit overhyped.
00:26:06
Sebastian Barros
ah But one thing I've noticed this time is that telcos CEOs have been extremely extremely fast to adapt.
00:26:17
Sebastian Barros
I mean, usually if if AI would have, I mean, if this AI frenzy would have come 20 years ago, I think CEOs would have been very reluctant.
00:26:24
Pablo Tomasi
Yeah.
00:26:27
Sebastian Barros
But I'm really impressed and i um I'm really glad to be honest that every single CEO recognize that this AI trend or wave is is different to to what we have seen in the past. So you can look at, well, Deutsche Telekom is a good example, SKT
00:26:49
Sebastian Barros
ey it is a Etisalat I think what Hatem is doing there is is like out of this world. um Vodafone doing partnerships that are very interesting Telstra with the partner with Accenture and Microsoft so everyone is like hey no no no this is different let's let's really put effort money CEO agenda topic so now This year is going to be the CFO year because, of course, now the CFO will say, show me the money, right? Okay, you have invested in all these amazing pilots, AI agents.
00:27:27
Sebastian Barros
Okay. Yeah,
00:27:29
Pablo Tomasi
Yeah, the show me the money part is always a bit tricky, isn't it? At this stage, at least for AI.
00:27:34
Sebastian Barros
yeah yeah, yeah, exactly. So, it and it's a lot of experimentation. There is a lot of fluff as well. I mean, for instance, in my case, I... I'm a bit more conservative on the AI agent perspective.
00:27:49
Sebastian Barros
because I believe that AI agents are not ready yet. They are based on large language models, which are very prone to hallucinations. So if you put an AI agent that is based on an LLM you will just create a much bigger level of hallucination. And if you look at labs and tests, you're talking about 20 to 30% hallucinations.
00:28:15
Sebastian Barros
And that level in an enterprise, I mean, it's like for every 10 customers, three customers, you say something stupid.
00:28:15
Pablo Tomasi
Oh, wow, that's that's a lot.
00:28:23
Sebastian Barros
so
00:28:24
Pablo Tomasi
Yeah.
00:28:24
Sebastian Barros
ah So so it I think it's not ready. Nevertheless, I think the the hype is real. I think there's something bigger. and And I think it's good to see Telcos, I mean, betting.
00:28:39
Pablo Tomasi
And that's a positive note. And then if sort if you circle back to what we're discussing six G and and stuff, you already mentioned that sort of AI will change a little bit, so to you know the type of traffic, you know the the use cases that are to be needed to be supported. But is is can you have 6G without? So does AI need to be embedded into 6G?
Integrating AI into 5G and 6G Networks
00:29:00
Sebastian Barros
Yes, yes.
00:29:00
Pablo Tomasi
or How do you see it?
00:29:01
Sebastian Barros
I think actually mean actually one thing that nobody likes to talk about is that in the 3GP releases, in 5G, we miss a little bit that window.
00:29:10
Pablo Tomasi
Thank you.
00:29:14
Sebastian Barros
Of course, now we we catch up a little bit on release 17 and so forth, but 5G was, I mean, AI as basic machine learning models was there, but was not the main driver. Now the industry quickly recognized, look, I mean, we need to make AI being a native part of of radio access, of core networks, of everything we do.
00:29:37
Sebastian Barros
And now if you look at release 18, 19, and now the upcoming 20, 20, which is going to be in 2027, which finished the 5G era, you could say, and give us the...
00:29:47
Pablo Tomasi
Mm-hmm.
00:29:49
Sebastian Barros
There is a massive amount of AI and machine learning natively attached to the infrastructure. And I think it's critical in in in this sense, right?
00:30:01
Sebastian Barros
And maybe ah a bit of a different perspective. So networks are becoming super complex, super complex. ah And compared to other type of infrastructure, they are distributed, so which makes your life even more painful, right?
00:30:15
Sebastian Barros
So if you're a telco here in Mexico, you probably have 30,000 macrosites in all kinds of places.
00:30:22
Pablo Tomasi
Mm-hmm.
00:30:24
Sebastian Barros
um You have maybe two vendors or three vendors that manage all this infrastructure. But you need to use 70% your sales of your sales to manage that.
00:30:36
Sebastian Barros
So it's it's massive, it's crazy. and And why? Because a lot of the things are manually driven. A lot of the incident management is still manually driven. Optimization has a lot of manual components, even design and planning as well.
00:30:52
Sebastian Barros
Even when you implement the network, it's also bit manual even when you want to do an upgrade although there have been a lot of improvements you need to do some some few man manual activities so for me the the biggest focus of AI my my view will be guys we cannot be spending 70% on OPEX
00:31:14
Pablo Tomasi
Mm-hmm.
00:31:14
Sebastian Barros
we to actually reduce this we need to make these networks much more automated cannot be so manual work and that's where embedding ai natively or could could really really help us a lot
00:31:32
Pablo Tomasi
But but will will that, so, you know, let's say you have like the 2G network, 3G network, 4G network, 5G network, then you put 6G on top with AI native. Will that be able to sort of also optimize your existing?
00:31:47
Sebastian Barros
Okay, good good point.
00:31:47
Pablo Tomasi
Gs
Necessity of Shutting Down Older Networks
00:31:49
Sebastian Barros
I think if if if we move to year 2030, what will happen will be the following. So 5G, the 5G network ride will be already upgraded all the way to release 19, or actually to release 20. So now you will have feature comp compatibility, and then you will also have probably backward compatibility. So your 5G network should work well with 6G, and I think here's where vendors since 4G has been quite quite consistent, right? I mean that whatever software you are putting is backward compatibility.
00:32:27
Sebastian Barros
Now, by then I would expect that 2G and 3G networks are shut down. um We have seen a big trend in Europe and the US where they're shutting down 3G and 2G networks. So for me, one of the key requirements for 6G, if I were sitting in a CTO office, would be, look, by 2030, I need to shut down 2G and 3G, right?
00:32:50
Sebastian Barros
Wow. If not, I will probably maintain 2G. There are some reasons why you want to maintain 2G. Sometimes it's because of some voice coverage and maybe
00:32:57
Pablo Tomasi
Mm-hmm.
00:33:01
Sebastian Barros
so some some some even some IoT devices that are too expensive to change. But that one doesn't really use so much bandwidth. I mean, usually you can put it in one of the guard bands.
00:33:13
Sebastian Barros
it's It's very easy to manage. three G needs to go because also you want to use that spectrum. and then what should happen is that as 5g uh begins you need to in a way upgrade those sites you would probably have multi multi-technology radios so you will just change those 5g radios for 5g capable radios and then all the spectrum will move to succeed so that that should be the logic evolution uh but but and I think this is another important point for telcos
00:33:37
Pablo Tomasi
Mm-hmm.
00:33:49
Sebastian Barros
oh we also like to live with a lot of technical depth, meaning I work in the BSS space now. You have systems that have been there for 15 years, right? and And um um I understand, right? It's a right of assets.
00:34:08
Sebastian Barros
It's working. Why touch it?
00:34:10
Pablo Tomasi
Yeah.
00:34:11
Sebastian Barros
But that's also the reason why you cannot do things easily, right? And many times Telco have this be this big thinking saying, look, if it works, keep it there.
00:34:22
Sebastian Barros
Just build some logic on top and you connect it. but My view that by 2030, Telcos should have a clean slate, kill all their old systems.
00:34:33
Sebastian Barros
So we have at least a 5G6 major layer only to manage and to take care of it.
00:34:40
Pablo Tomasi
Okay, it seems quite a quite aggressive, i would say, during 2026, but it's good to good good to expect you know something positive from from the technical side. um but But I'm curious, I heard some people discussing whether or not do we still need to have the evolution cellular technology that is divided in Gs, sort of.
00:35:04
Pablo Tomasi
Do we really need to have that... categorization and the separation of tech or can we do something different? Is is there an ah an alternative as well or or is just people talking
Architectural Changes for 6G
00:35:16
Pablo Tomasi
because I don't know.
00:35:16
Sebastian Barros
No, I mean, it's it's it's ah it's a valid point. It's a valid point because you could also claim, okay, Sebastian, we have 5G until release 20. That's the last release of...
00:35:28
Sebastian Barros
5 g Can we keep doing releases, right? I mean, if we need a new feature, there's a release 21, release 22.
00:35:33
Pablo Tomasi
Yeah.
00:35:36
Sebastian Barros
Why do we need really this massive 6G thing? Now, that release, in my view, depends only of one thing.
00:35:47
Sebastian Barros
Do you really need to make a major architectural change right ah If not, you can go with upgrades like releases all the way until whenever you want. The only reason you need really to make a new G is when you need to jump to a new architecture. So for instance, if I take the example from 3G to 4G, the big architectural change you could say is we moved from circuit switch voice to all IP. That was like a big big shift.
00:36:21
Sebastian Barros
And of course, going all IP requires a completely different thinking. From 4G to 5G, I would say that the biggest change was that we moved to a service-based architecture, and that was core, right? So we copy-paste all the ideas that these hyperscalers have been bringing into cloud, and we say, well, guys, we cannot be designing monolithic networks.
00:36:42
Sebastian Barros
We need to design service-based architecture, microservices, and so forth, right? And that's what we did with 5G core in a way. um And of course, on top that, a lot of enhancements.
00:36:53
Sebastian Barros
Now, the question for me will be, you only need 6G if you need to make a big architectur ah ah architectural shift.
00:37:02
Pablo Tomasi
Hmm? Hmm?
00:37:02
Sebastian Barros
If not, yeah, I mean, you can you can have a grid. In my perspective, and I think this is debatable, I think we need ah a new architecture. Systems are too heavy, too complex.
00:37:18
Sebastian Barros
They require two-manage work. They need to have much more automation embedded. We need to eliminate, in my view, many of the functions. I think we have redundant functions and even functions that nobody uses.
00:37:34
Sebastian Barros
So I think we need to make a simplification of the of the network. and And that will require a change in the architecture. The other part that is important as well is that, and tu also also this is debatable,
00:37:49
Sebastian Barros
The radio, you could say the fi well layer one, layer two of the radio also will have significant changes in the design, especially as we are bringing AI in RAN
00:38:03
Sebastian Barros
and And that also will require an architectural change. So in a way, i think we need that we need that change in 6G Now, if if if someone says, look, I don't think we need any big architectural change, yeah you could you could all go with release 21, 22, 23, until
00:38:12
Pablo Tomasi
No, no.
00:38:22
Sebastian Barros
till we we need something
00:38:25
Pablo Tomasi
So it it seems this time around is needed, but potentially we shouldn't be talking 7Gs. If 6G is done good you know well enough, you you would have that sort of programmable AI enabled let's call it platform, for the sake of a better word, that would then enable just you know more and more releases without like a big shift on 7G, unless something bigger than AI comes along, but that's a few years from now.
Agile Updates Over Decade-long Cycles
00:38:53
Sebastian Barros
Yes, I think, let me let me think. I mean, this is, what what you say is interesting. I think, I don't want to be, yeah, I don't want to be wrong, but actually Börje I think towards the last, end of last year, actually said that 6G will be the last G So so I think you're right, Boria. right,
00:39:21
Pablo Tomasi
Probably probably need to change the title of the podcast then, you know, reflecting that.
00:39:24
Sebastian Barros
yeah know i think here of I'm actually I'm um checking so I don't and don't like but I recall very clearly because for me was like wow that's a super bold statement coming from a company that leaps out of G's if you think about it and and this is the challenge also
00:39:34
Pablo Tomasi
and That's a bold statement.
00:39:41
Pablo Tomasi
Yeah.
00:39:45
Sebastian Barros
I see Ericsson, Nokia, Huawei not as a vendor. I think, and and this is also another problem that telco CEOs fail to see, oh is our R&D department.
00:39:57
Sebastian Barros
I mean, that's that's what Börje Hotard and Huawei CEOs are. i mean, they are they are in the R&D department of telecom. So...
00:40:06
Pablo Tomasi
Mm-hmm.
00:40:06
Sebastian Barros
um when you when When procurement is squeezed, they are not squeezing the price, are squeezing the R&D money that they're going to be spending in telecom. So for someone like Ericsson to say, look, 6G is going to be the last G, it's interesting because what happened after that? What happened with these networks?
00:40:26
Sebastian Barros
How they are thinking? my My perspective, probably, and this is only my speculation, is that... a The way that, I mean, and here's the part I also want to give thanks to all the researchers and um and developers of these standards. So imagine you, but me and Pablo, you and are standing in 2010. And our boss come and say, hey, Pablo Sebastian, you need to think how networks will be in 2030.
00:40:57
Sebastian Barros
Based on that, get together with everyone and and try to agree. And that is the standard. I mean, these guys are thinking 15 years in advance. So, I mean, what they do is is like amazing work.
00:41:07
Sebastian Barros
I mean, even in hindsight, it's easy to criticize. It's like science fiction, right? I mean, how in the world you will know where it's going to be?
00:41:16
Pablo Tomasi
Yeah.
00:41:17
Sebastian Barros
So it's pretty wild. Now... That's okay, but for a world that is changing so fast, I think this kind of long-year planning, multi-year planning, with a huge amount of capital spending and R&D spending, probably is not making too much sense, right? And I expect that from six years onwards, we will think more of... I don't want...
00:41:44
Sebastian Barros
I'm trying to think. It's more like a sprint. Like, like i mean, we need to change architecture.
00:41:47
Pablo Tomasi
Mm-hmm.
00:41:49
Sebastian Barros
Okay, we don't need to wait 10 years for 7G, right? I mean, let's look into a new architecture um roadmap. And that takes maybe two to three years, right? But I think these ways of working of 10 years, it's it's too heavy for everyone.
00:42:07
Sebastian Barros
and And it's is' is's just too crazy to... to to plan, right? that That's my view. so So we will have something, maybe they will not call it 7G, but it will be more like maybe jumps or small jumps and like a massive jump.
00:42:25
Pablo Tomasi
and and and And that's a very you know interesting point and and probably like a great point to end the podcast. I know we could be talking for ages about this, but need to wrap this up. And yet the idea of like sort of changing from that sort of very long 10-year type of roadmap in something you mentioned, Sprint, something that is a little bit faster, maybe can address problems in a little bit more of an efficient or effective way. does Does make sense for whatever 7G is not going to be. but Sebastian, thank you very much for joining me on the Drift really appreciate your insights.
00:42:59
Sebastian Barros
Amazing, amazing. Thank you very much for this invitation and this great conversation.
00:43:04
Pablo Tomasi
Perfect, thanks everyone for listening.
00:43:07
Sebastian Barros
Thank you.