Telemetry News Now

Telemetry News Now.

Welcome to another episode of Telemetry News Now. Today is, March twelve that we're recording, so you're likely listening on the thirteenth, maybe the fourteenth, and spring has sprung here in upstate New York. Actually, I I say that a little bit too soon. Spring sprang yesterday. It was seventy two degrees here in the, capital region of New York, and it is forty two degrees today. At least that's what the expected high is. Justin, I don't know what it's like by you, but, you know, we have about another week till spring starts, which means where I am, we have about another six weeks of winter.

Yeah. We're in that period of the year where you're wearing a sweatshirt in the morning and you're stripping it off because you're sweating, by the afternoon. Right? The pollen is in the air for sure.

So Oh, we're not quite there yet, but, we will be soon, and I'm looking forward to it.

So let's dive into today's headlines starting off with something out of CNN from a couple weeks ago, but I did find it interesting. Skype is shutting down after two decades.

So the news here is that, maybe it's no surprise to some of our listeners, but Microsoft is shutting down Skype, which I'm sure you remember was one of the most dominant Internet based phone and video services. And if you you remember that iconic ringtone, Justin?

Oh, yes. Of course. Yeah.

Yeah. Absolutely. So Skype will no longer be available starting this May twenty twenty five, and current users will be able to log in to Microsoft Teams, their free tier, using their Skype credentials.

Now, Microsoft acquired Skype back in twenty eleven for eight point five billion dollars. Now at the time, that was its biggest acquisition, so there were a lot of high hopes for Skype, obviously. Right? And then, what they did was they integrated Skype into several of their products like Office and then, the now defunct Windows Phone. Right? But it really never took off, except for, during the pandemic five years ago, there was a resurgence when they saw an uptick in usage.

I personally remember Skype, and I remember how they kinda led the industry in that, you know, the beginnings of that collaboration technology over the Internet. Mhmm. And then and when they came out, international calling was still very expensive. You know, that was still a thing that people had to deal with. So Skype made sense at the time, but I think those days are gone. And, honestly, I don't see how they were ever gonna compete with, other solutions like Zoom or my kids use WhatsApp and FaceTime, and they're almost seamless integrated into their devices. So, I mean, how they were gonna compete without changing the product significantly at least, you know, I don't know.

Yeah. Definitely, an end of an era. I remember I was in a long distance relationship at one point in my life, and Skype was how we communicated and kept in touch. Right?

Because back in the days when cell phone minutes, you paid per minute. Right? Skype, you didn't pay per minute. So even if it wasn't an international call, it made it really easy to keep in contact with somebody at a at a really low cost.

So, you know, sad to see it go personally, but I get the business decision. There's all kinds of other collaboration tools that compete with it, and so it is kinda hard for it to keep up. You know, it does have significant overlap with Teams and, you know, they mentioned if you have a free tier Teams login, you can continue to use that, which is basically the the Skype functionality. So pour one out for Skype.

Next out of the TSMC press center, that's the Taiwan Semiconductor Manufacturing Company. TSMC has announced an additional one hundred billion dollar investment in advanced semiconductor manufacturing in the United States, bringing, their total US investment to a hundred and sixty five billion dollars Now the expansion that we're talking about here builds on its existing sixty five billion dollar investment in Phoenix, Arizona. And so now this includes plans for three new fabs or fabrication plants, right, two advanced packaging facilities and a major r and d center. So this is gonna make it one of the largest single foreign direct investments in US history.

This is a big deal. The entire initiative is expected to generate hundreds of billions of dollars in semiconductor value for AI, of course, and presumably other technologies. Right? And according to TSMC, it will support about forty thousand construction jobs over the next four years for this initiative and create tens of thousands of high paying, high-tech jobs.

Now as far as direct economic impact, they're projecting to drive over two hundred billion dollars of direct benefit to the economy across the US over the next decade.

And, the TSMC chairman and CEO doctor C. C. Wei credited the expansion to strong customer partnerships, government support, and the success of TSMC's first Arizona fab, which began volume production late last year. So, obviously, this strengthens the US semiconductor ecosystem. Right? It completes this whole agenda for, you know, strengthening the domestic AI supply chain, you know, in spite of TSMC not being an American company and now with, you know, TSMC's first advanced packaging investments here in the US itself.

Yeah. And I mean, I think when the Trump administration announced the two hundred billion they were going to invest in Stargate or I guess was it five hundred billion they were gonna invest in in Stargate?

Five hundred. Yeah.

Yeah. I expected to see more articles like this. Right? Foreign companies coming in wanting to build manufacturing plants, in this case, fabrication of Philip chips, on US soil to benefit from some of that money.

So, yeah, it's interesting to see how this being done. Looks like they're gonna be building plants in Phoenix, Arizona, Thomas, Washington. I'm not familiar with that. I'm presuming that's somewhere outside of the greater Seattle area.

And then, Austin, Texas and San Jose, California. Right? Silicon Valley, of course.

Mhmm. Mhmm. Now the only thing that I have here, I don't have incredible insight into this one other than to say that, my experience with fabs and chip fabs here where I live, and there are several, is that it's gonna take a decade or more to build the thing anyway. So the direct impact is in construction jobs, which they called out, the forty thousand construction jobs.

With the long term economic impact, we're not gonna see that, and it's not even gonna begin to really take any kind of meaningful effect for a decade. I mean, we're talking about fabs here and not, you know, building a local retail, you know, store, whatever, you know, on the side of the road. So the greatest impact that we're gonna see in the in half a generation, if a generation is twenty five years, is just the building of these things and then all the, economic impact that it has to the local areas there as a result. And that's great.

That's fine. But we're not gonna see this huge influx of benefit to the AI space and the tech space and just this to the Philip space in general for some time.

So as a follow-up, from Reuters just this morning, that's March twelve, TSMC also proposed a joint venture with major US chip designers, and that's including NVIDIA, AMD, Broadcom, and Qualcomm.

Why? To operate Intel's foundry division, though it wouldn't own more than a fifty percent stake when all is said and done. Now this whole thing is in the early stages of discussion, and it follows a request from the Trump administration for revive Intel. And I I don't know, Justin, if you've seen in the in the news recently, Intel has been struggling with financial losses and, yeah, and and a declining stock value.

So according to Reuters, Intel's board supports negotiations with TSMC, but some of Intel's executives oppose the deal. And and the reasons that they're citing are, the company's differing manufacturing processes and also the protection of trade secrets. And I think that's understandable. I understand that. Now as far as timing, this joint venture pitch was made before TSMC announced its hundred billion dollar US expansion that we spoke of, you know, just a couple minutes ago. And so, you know, while this is in discussion, NVIDIA Broadcom and AMD, are actually testing Intel's eighteen a chip manufacturing technology, which, Intel claims is superior to TSMC's two nanometer process.

Yeah. So I think it's important to remember that most of the companies involved in this joint venture design the chips. So NVIDIA, AMD, Broadcom, what they do, they're what they call a fabless company. So they will actually design the chips.

They have engineers that figure out what the specs need to look like, design those chips, and then they'll contract with a company like TSMC, who's the biggest company that I'm aware of, in the world that does this, to actually fabricate the chips, to actually manufacture the chips. Right? Intel has taken a, historically, taken a different strategy with their business where they do both. They have engineers that design the chips, and then they have engineers that actually build them.

They have their own foundries where they build their own chips. So the theory here is that TSMC and these other design companies want to build a joint venture to help Intel migrate from their current business strategy to one similar to the other companies where they only do their design and they outsource their manufacturing Mhmm. To TSMC.

The theory here, whether it's true or not, theory here is that that will help Intel with some of their struggles, right? That simplifies their business model. They don't have to do both sides of that equation. They don't have to do both the design and the manufacturing because like we just said in the last article, it's time consuming to build a fabrication plant.

It's also time consuming to retool it. Like when you come out with a new chip design and each chip design has different nanometer technology, it's more efficient as far as the number of transistors that are on a chip and so forth. You have to retool your entire factory to be able to manufacture that new chip. Right?

So that's expensive and time consuming for somebody like Intel to do internally, whereas the other companies, NVIDIA, AMD, Qualcomm, Broadcom, they can all just come up with a new design and all of the retooling cost and struggle is all TSMC's problem. Presumably, that's the strategy here is to help Intel be able to only have to do the design and outsource a lot of that expense and struggle to another company.

Yeah. You know what? Completely separate. But what boggles my mind is not that long ago, like, twelve years ago, I ran network engineering for research organization where I am and, which was tightly integrated with some of the local fabs and and some of the global, initiatives.

You know, they were talking about four hundred and fifty nanometer and then three hundred nanometer. And when I left, there was, you know, work towards, I don't know. It it just boggles my mind that we're talking about smaller sizes like, you know, three and two nanometer technologies here. It's crazy.

Moore's law. Alright?

Guess so. Yeah. You gotta keep up. You know, what's also interesting to me is the it's not interesting. I think it goes without saying at this point considering where we are in the industry, where we are in technology and globally, is how these companies all have to work together in order to advance their own interests and, and advance the state of technology, which leads right into our next, article right out of the Cisco newsroom.

GeoPlatforms Limited will be collaborating with AMD, Cisco, and Nokia to develop an open telecom AI platform, which they unveiled at Mobile World Congress twenty twenty five. So the idea here is to integrate AI driven solutions into telecom operations. And according to the announcement, therefore, enhance things like efficiency, security, revenue opportunities, and they're specifically talking about service providers. Interesting stuff.

So this would create a type of, like, a central intelligence layer that incorporates AI and automation across all network domains, including things like the RAN, the radio access network, routing security data centers, things like that. And when I said AI, the first thing you thought of maybe out there is, large language models and the large language model revolution that we've been going through. So the platform will be LLM agnostic and, of course, leveraging open APIs. Not open API, the company.

Open API, separate. Right? Agentic AI, LLMs, small language models as well, and also other machine learning techniques, which we do need to include in a discussion about AI if we're going to be honest, because LLMs is one corner of it. So all of that with the goal of enabling real time intelligence and autonomous network management.

And I think we're gonna see more and more of this in our space, Justin, in the networking world, whether it's on the telecom side, on the enterprise side. Folks have been using these technologies like predictive analysis and and classical, AI and the application of LL models and all these things, and now large language models, to mine information out of data, to understand patterns, to do all that stuff for years and years in health care, in aerospace. And that's kinda normal for many industries now. And and I feel like it's only in the last, you know, chunk of years, five years, eight years that we're really talking seriously about applying it to IT operations and and to the network?

Yeah. First of all, Mobile World Congress, if folks aren't familiar with, is an annual conference that takes place in Barcelona. It's one of the largest IT events that takes place every year, and, all the mobile operators will release new technology around the mobile network as well as new cell phones, new technology around cell phones, and so forth. So just took place last week in Barcelona. A lot of interesting articles that came out of that. But, yeah, I mean, it's interesting that I think historically most companies has seen how they apply AI as a little bit of their strategic advantage, right, if they can figure out a method that others haven't figured out, that gives them a little bit of a leg up against their competitors.

And we're starting to see a change in that way of thinking, it looks like, to where companies are wanting to collaborate and realizing, like, if we can all solve some of these problems in the same way and build a little bit of an alliance and share some of the research, share some of the knowledge and how we solve some of these problems, it's gonna be better for the industry as a whole instead of trying to keep it private to ourselves and try and view it as a as a strategic advantage to our organization. So, yeah, I think it'll be interesting to see how much more quickly that helps with, you know, development of the models or even just how to take those models and apply them to various different use cases that apply to different areas of of networking, which, you know, kind of leads us into the the next article from ARM Newsroom, another article coming out of, Mobile World Congress last week talking about AI RAN, which is basically similar to the, OpenTelecom AI platform that we talked about the last article.

You know, there's a lot of complexity in these RANs and the radio access networks where your cell phone is connecting to the tower. Everything from knowing if you've got a radio that's going bad, you're starting to have a lot of loss, customers cell phone signals are being dropped. Sure all of us can sympathize with all of a sudden your call just drops even though you have a good signal or you should have a good signal but you can't get one. You know, that's still a challenge that a lot of mobile subscribers deal with, and, you know, it could be really complex, especially at scale.

If you think of, you know, a huge mobile operator like, you know, we have here in the US that has millions of subscribers, finding one radio that's bad until customers start complaining can be a bit of a daunting challenge. But there's a lot of data that comes out of these RANs, but how do you you can't just have staff sitting there watching all those stats trying to find the signal from the noise in what's actually a problem. But that's what AI is really good at. Right?

Being able to correlate data, being able to sniff out from various different pieces of data what's actually a problem that someone, needs to get involved in and take a look at and fix things. So the idea here is to build an alliance across multiple mobile operators to be able to build those use cases that are repeatable across companies like AT and T and Verizon in the US and, of course, other mobile operators globally.

Yeah.

And the issues around power management too. When we're talking about individual devices, you know, there there's a lot of data that is necessary to do that analysis that you mentioned that is either on the device itself. It's in the provider network, it's an intermediary networks in between service providers. So I think that as we start to apply, you know, AI technologies to the telecom space, to enterprise network, all of it, you're going to see a need for more partnerships.

We're already seeing a lot of partnerships. We we just talked about Cisco and Nokia partnering. Like, that aren't don't they both, like, make a lot of the same stuff? You know?

So, you know, we're we're gonna see more and more of that because if you want to have an effective AI solution, you want your AI initiative to succeed and for the the results to be meaningful. So you're not doing AI for the sake of AI. Right? So you can get your next round of VC money, but you're actually trying to get a result.

Data is the lifeblood of that. And the more data, the cleaner the data, the more accurate the data, all of that, that's what's going to empower that, which is why I say that if your if your goal is to just have an AI initiative and you're going down that, like, you know, head headstrong and you haven't figured out your data situation yet, it's gonna fail. You know, when I'm looking at this and, yeah, you know, I I hear what you said, Justin, about the specifics regarding, wireless technology and and then how that interplays with individual devices that the providers don't own yet. They want their data, and they still have to manage power and all that kind of stuff.

There's a lot of client driven stuff there with regard to, with five g and things like that. It's gonna be all about those partnerships for sure. You know, we're gonna keep hearing, like, NVIDIA partnering with so and so and, in order to make it happen, and rightly so.

Well, keep in mind, in five g, presumably, you have a lot more radios, a lot more access points. They're doing what they call microcells, where you'll have, you know, a lot more radios that cover less devices per radio. Right? So this is gonna become even more important, with five g architectures to be able to find where you've got issues that are affecting your subscribers.

Mhmm. Yeah. For sure. And then also think about the kind of reverse CDN thing that people are talking about.

I talked to somebody. I think they were at Equinix, and they were talking about this thing that they're sort of seeing and preparing for is all of the data that wearables and your devices and stuff at your house and in your business, right, is sending back to the various SaaS platforms that we use, you know, even even on a consumer basis. So they can do, like, the cool AI stuff and then, you know, give us the meaningful results, blah blah blah, which presupposes that there is this reverse CDN. We're like, how do I you know, all the information from the streaming services, and then we put them into pops, and then we send them out blah blah blah.

Well, imagine that in reverse, and five g is gonna be a big part of that. And, and it's for the purposes of fueling a lot of this, you know, data analysis that happens somewhere else. Although we have been we haven't talked about edge computing and how some of that's even happening locally, regionally, and even at the device level. So we're gonna see more of that, as well.

So really interesting stuff.

Yeah. That'll be fascinating to keep an eye on because a lot of those networks were designed for one direction, right, from the content creator and content owner out towards the people consuming that content on their mobile device at their homes or whatever. Right? So they the networks weren't really designed for large volumes of traffic going in the opposite direction. Right? Telemetry data coming from the edge of the network back towards, you know, the content provider. So it's reverse CDN ideas, you know, really interesting to keep an eye on.

Alright. Moving right along, the next article is from Infosys. This one's a little bit dated from the end of January, but it was a really fascinating kind of to what we've been talking about about how AI can transform telecommunications. A couple things I took away from this one that we haven't already talked about is being able to collect all the data about subscribers subscriber behavior. This is a little bit frightening from a from a big brother perspective, but being able to analyze a lot of that data, pull out customer trends, and and do better targeted marketing, Being able to find patterns in custom preferences, customer behaviors that allow companies to be more effective in their marketing and sales strategies, improve their customer acquisition, their customer retention.

I Like, we all know you go and you browse for something in your browser, there's cookies that are tracking your behavior, and you're getting targeted ads. Right? Like, that's been a thing for quite a while. But being able to, like, pull back your mobile behavior and where you roam, where you go and be able to use that for targeted ads.

I don't know, Phil, how you feel about that. That starts to get a little creepy to me. Yeah.

It's creepy, but this is stuff that's been going on for a long time specifically. And I'm not talking about targeted ads. We already know about that. And, you know, the behavior analysis thing that social media does and that kind of thing.

But I mean with regard to the application of, like, ingesting data so that you can build a model and then apply it to new data, make some sort of prediction about churn or about whether a customer or prospect will buy or cancel their subscription, you know, regarding marketing stuff, right, or business decisions. That's that's common, and it's been going on. And it's difficult. It's not difficult in theory.

It's simple in theory, but it's difficult to do because of problems like misclassification.

And I think that a lot of folks this happens in politics a lot, like, where you assign a candidate, you know, some, like, data scientist or quants. You know what I mean? They'll assign a candidate, like, seventy percent likelihood of victory, and everybody's like, oh, well, they're gonna win. And I'm like, no.

That's a seven out of ten chance. A three out of ten chance for the other person is still a very significant probability.

And so you're gonna see a lot of this misclassification where people would look at it binary. You know what I mean? Like, oh, you know, as far as marketing is concerned, like, oh, they're gonna buy, they're not gonna buy. No. It's just a little bit more likely.

There's something in in machine learning called Lyft as well, right, where you have well, you know, I'm not even gonna go there because it gets into the technology. But what I wanted to talk about was this idea of having an AI first strategy. I read that right in this article. I don't know how I feel about that because it's like, if that's what your business is all about, unless your, like, business is making building models, like academia, research and development, you know, building a foundational model.

Right? So it's AI first because that's your actual product. But if it's not, and for most people, it isn't, what in the world is an AI first strategy? Shouldn't your strategy be like solving problems first strategy or generating revenue strategy?

And then if AI is the right tool to do a thing, then you use AI? There are a lot of times when we have this this idea, like, I'm gonna apply this model. I'm gonna do this thing when it could have been solved with a simple some relatively simple software development. Or do you understand what I'm trying to say?

It's like Oh, yeah.

I mean, this has been a trend in our industry for a long time where people hear about a trend.

Right? A new technology that sounds really cool.

They know on the surface level that it can solve problems for them, but they're not sure what the problems are. So they're like, oh, we're gonna be a cloud first strategy. We're gonna be an AI first strategy. It's like, well, if you don't identify what the business problem is that you're trying to solve, doing it in the cloud or doing it with AI is probably set up for failure. Right? What you really got to do is take a step back and be like, what are the what are the business problems that we're trying to solve and, you know, then find a solution that solves that problem. Right?

Yeah. And these are legitimate projects, and ML has broad industry wide multiple industry, like, across industry application because we are talking about math.

My concern is misrepresenting the ability of predictive performance. Mhmm.

You know, I think that's gonna be a problem that, you know, what can a model truly achieve?

And if you do a little bit better than flipping a coin, that's from a technical perspective, good. Like, if you're an ML ops person, right, like, that's that's a great success because you beat out flipping a coin. Right? You're you're technically more effective.

Yeah. But how much money did you spend to beat out flipping a coin? It's like you and I could flip a coin.

Right?

Well, that's what I'm doing.

We don't spend billions of dollars.

What's the business value? So remember, whenever you start, I'm speaking to myself, I'm speaking to you, I'm speaking to our audience. Whenever you're thinking about AI initiatives or hearing people talk about it, you do need to tie it back to that. Like, what is the actual goal here, and is AI necessary to solve it?

And if so, what is the cost to get there, whether it's figuring out your data pipeline or in our industry or telemetry pipelines? And then, you know, is the result of a predictive model significant enough that it's worth the investment? Accurate enough. And then accuracy means something different in, like, the technical world than it does in, like, the business world.

So making sure we're all on the same page.

At the same time, though, this is so interesting to me and very, very cool to see tapping into AI's actual potential, right, like this article said. But Mhmm. You know, we're we're talking about early stages in the IT world and, in the telecom world, and so we're gonna start fleshing some of these things out. We've been talking about partnerships this entire podcast today, so we're gonna see some of those flesh out how we solve some of these data problems.

That's what a lot of these partnerships are, by the way. I mean, it's like, how do we get the data? Well, let's partner with them because they have data. Well, we need that data.

Let's partner with them. Well, we need these chips to run this in this cloud, so let's partner with them. And all of a sudden, you have a a solution through partnerships.

So really interesting stuff. I'm really excited personally because this has been, a big part of my life for the past few years.

Yeah. And I mean, some snark and cynicism aside, I do think as an industry, once we figure out some of these tangible use cases and tangible applications for AI, we're gonna solve problems that have been plaguing us as an industry for a long time.

Yeah.

And so I think I've probably said on the podcast before, like, I firmly believe that AI is going to be like the next Internet in the way that it changes our society and how we do things. Now, that being said, there's going to be a lot of, like, hype. There's going to be a lot of things that people don't fully understand the technology or getting people excited about things that AI can't actually do either because the data is not there, the data is not clean, like you're you're highlighting the infill. But that doesn't mean that it's not real, that it doesn't exist, that there aren't real, you know, tangible solutions to it. We just have to, as engineers, think through, like, what's real and what's Mhmm.

Hype. Right? Yeah. Absolutely. I I absolutely agree with you. This is very real technology. We can point to specific, you know, algorithms.

There's real technology.

But in order to do this effectively over the long term and solve these problems, we are talking about upskilling entire staff, you know, bringing on ML ops engineers or data scientists where you never had a data scientist before or a team of data scientists. Basically skilled AI talent to do this into perpetuity because you are talking about having to maintain these data pipelines and and these architectures. And we're only we're only now talking about the potential cybersecurity risks and the the governance frameworks, whether it's regulated or not regulated. We're only starting to talk about that. So the applications are awesome, but we are such at the early stages, and I I think it's fantastic. I don't know about you, Justin, but I was getting bored talking about BGP.

Right? You know? So I'm not sure.

About BGP all day long, but I I pick your point. Yeah. Yeah. Yeah.

I, you know, I think I was thinking as you were going through your statement about, like, political and stats and which candidate's gonna win, I think a lot of that confusion comes from people who, you know, maybe never studied statistics or the basics and don't understand sample size and whether or not the sample size is representative of the whole.

And so to your point, having at least a basic working knowledge of statistics and how some of these algorithms work and which algorithms are a good application to certain problems is gonna be really key to the successful outcome Yeah.

Of a lot of these initiatives.

Right?

Yeah. I agree, but not for most people. I think most even technical leaders can get away with a foundational understanding, what to ask, what's important, and then high level architecture stuff. Kind of like a a really good VP of engineering doesn't necessarily need to know the syntax of how to configure VTEPs for their VXLAN environment.

They just need to know these things exist and, like Sure. How they work together. So I think, you know, business leaders, even technical leaders, but business leaders, understanding how to spot AI initiatives where it's applicable, what the considerations are in in getting there. I think that's gonna be, like, the main skill sets and not necessarily, like, which algorithm do I apply here?

Which model do we use? Do we use, you know, cosine similarity or do we clustering? Like, whatever. You know, we have engineers that can do that.

So I think that's gonna be where we see the real meaningful benefits and results of applying AI to to our industry. Yeah. And and I think that's sorely lacking, by the way.

Yeah.

And you and me, I I'm speaking for you. I'm sorry. You and I included tend to be more nerdy and focus on a lot of the technical stuff, which I love. But if you got into the weeds, statistical analysis is just scratching the surface. You need to get deep into, like, linear algebra and all these things where it's, you know, it's it's much more advanced. So and good luck with that.

Alright. Moving on. Final article for today from the Associated Press on March eleventh, Musk's Starlink and Bahati Airtel, who's a big mobile carrier in India, signed a deal to explore bringing satellite Internet to India. So we've covered a number of similar articles like this on the podcast before, but, Starlink is going out and signing a lot of contracts with a lot of the carriers around the globe to allow customers to roam from terrestrial five g mobile networks to Starlink's satellite Internet. This article didn't highlight it, but some of the previous ones we've covered, they're usually starting with, text messaging. I think a lot of the people who are based in the US saw the NFL Super Bowl commercial where the T Mobile partnership was announced and the use case there. And the commercial was someone who's out hiking in the woods and needs to be able to get a text out to a loved one to let them know they're okay or share pictures of the beautiful scenery.

So presumably, this is something similar to that where, subscribers can send text messages through Starlink when they don't have a cell phone signal through a normal, radio tower. But, yeah, just continuing to see more evolution of this, more partnerships announced signing Starlink with the various mobile carriers around the globe. And, of course Yep. This is a huge one being in the how how big of a country India is and how many mobile subscribers are in India.

I don't know if, you mentioned it, but I read here that over forty percent of India's population, their one point four billion people population, does not have Internet access, especially in remote regions.

Right? Mountainous regions. You know? Because, it's a very large country, subcontinent.

So this is a big deal, and I'm surprised this had didn't happen sooner because that is such a important market. So, yeah, very, very interesting.

Yeah. I always my bias is I always think of India as, like, the cities. But, yeah, to your point, there's a lot of rural areas in India just like a lot of other countries. And, yeah, all of those people are are underserved with Internet or don't even have access to it in twenty twenty five, which is something we kinda take for granted in where we live, Phil.

Yeah. That's a good point. There are security concerns. There's government approval. So this is not a a quick win for Starlink or for India for that matter. So there's gonna be a process to go through. And, you know, there are gonna be local telecoms that oppose this as well because they're going to see potential loss of revenue and a subscriber base.

And so moving on to upcoming events, we have the, Philadelphia networking user group on March nineteenth. I don't know if that's in Philadelphia proper. I think it's outside Philip. Well, you could check the USNUA website, USNUA dot com.

Click on groups and events and look up the Philadelphia NUG there and see the town if it's near you. We have networking field day thirty seven, March nineteen and twenty out in Silicon Valley and also live stream. So make sure you check out the live stream. Some really interesting vendors are gonna be presenting.

I've been checking them out. NVIDIA GTC AI conference, March seventeen through twenty one. I don't have it in front of me, but I believe that's also in the Bay Area, San Jose. Justin, you aware?

I'm not sure. I have to go look.

Yeah. Yep. Yep.

I see it here.

DCD Connect. So that's data center focused. In New York City, March twenty four and twenty five, I will be at that one. In fact, Justin, I'm gonna be on a panel discussing or debating InfiniBand and Ethernet, our beloved technologies. Yeah. So that'll be an interesting one.

That'll be good.

And last but not least, we have the Connecticut NUG, Connecticut networking user group, March twenty seven. So make sure to check that out on the US anyway website if that's near you for time and exact location. So with that, those are the headlines for today. See you next time.