Starlink's Wireless Ambitions, the Data Center Boom, and Sampath to Zayo

Full Transcript

Don Kellogg 0m10s

Hello, and welcome to three hundred and eighth episode of the week with Roger, a conversation between analysts about all things telecom, media, and technology by Recon Analytics. I'm Don Kellogg, and with me

Roger Entner 0m19s

as always is Roger Entner. Roger, how are you doing?

Guest Speaker 0m22s

I'm great. How are you?

Roger Entner 0m23s

I'm good. So we got some news about a actually, our most popular.

Guest Speaker 0m28s

Yeah. It's not news anymore. We're a little bit late.

Roger Entner 0m31s

Little bit late. Yeah. So friend of the podcast, Sam Path, who I believe has been on the podcast three or four times with us

Guest Speaker 0m36s

Three times.

Roger Entner 0m37s

Yeah. Has landed at Zayo. Do you wanna tell us about it?

Guest Speaker 0m41s

Yeah. So Sam Perth is the incoming CEO at Zayo. Zayo is a connectivity provider, especially on the back end side. So it's not a consumer brand, so most people will not have heard of it. But they provide a lot of the fiber, both lit and dark, to enterprises and data centers.

Guest Speaker 1m3s

Should be very exciting to see what Sample does there. You know, it's also a nice segue because as the industry expands, our offerings and what we do in terms of research also expands. You've been working very hard behind the scenes of putting together a newsletter with pod of what's happening in the data center world and who builds what, when, and a whole database around it. Maybe you wanna you know better than what you did.

Roger Entner 1m32s

It's more than newsletter. It's one of the things that we noticed, you know, as we were studying everything going on with AI and telecom. We've talked about it on the podcast before, this huge movement towards building larger and larger and larger data centers. And the amount of compute that AI takes is really monumental.

Guest Speaker 1m49s

And the spend. Yeah. And the

Roger Entner 1m50s

spend is monumental, and finding a place to even just put all the GPUs can be really difficult. So there's this entire submarket that we are really diving deep into around, you know, what does the deal flow look like? What does the financing look like? What are the power constraints? How big is the data center?

Roger Entner 2m7s

You know, there's a lot of pushback that's happening at local levels in some areas. So, you know, for example, there was a moratorium in New York on data center build outs that just passed.

Guest Speaker 2m17s

And Maine. Right?

Roger Entner 2m18s

And Maine. Yep. And so there's just so much going on. And one of the things that we've gotten really good at is reaching out and tapping into all these different sources of information that by themselves would be impossible for anybody to follow and using some of the tricks in our tool bag to make this really digestible. My personal reflection on this is, like, you think you know how big this effort to build out data centers is.

Roger Entner 2m43s

When you start to look at the numbers, are just absolutely mind blowing in terms of 10 gigawatt data centers being built in Texas. And, you know, that's huge.

Guest Speaker 2m53s

Yeah. Well, we can put one number in there. So the telecom industry for many, many years, decades, has been the largest infrastructure investor in this country. $50.70, $8,090,000,000,000 dollars. This year, just like the big stock market listed companies will invest $780,000,000,000 into data centers, 11 times what, like, the wireless industry did at its high, which is just, like, unbelievably crazy.

Roger Entner 3m28s

Crazy. Yeah.

Guest Speaker 3m30s

And so stay tuned. We're gonna come up I'm gonna come out with a really attractive offer for people to tap into our data. Somewhere, I think we will have a free newsletter, like a full rounded service offer. Talking about people investing a ton of money into data centers, talk about Starlink. He, like, sent stock prices and telecom providers tumbling again with a five minute little side comment in his earnings call.

Guest Speaker 4m0s

Right?

Roger Entner 4m0s

Right. And it's definitely, um, got the peanut gallery going too. And I know you have a really definite perspective on this. So why don't you tell us about what you think about Starlink getting into wireless?

Guest Speaker 4m10s

So I have always been of the opinion that he will build. There's a report on our website that says the s one says wholesale, the pattern says MNO. And people came to me and said, yeah, Roger, you're nuts. He's not gonna build a wireless carrier. On this week's earnings call, both Gwen Shotwell and Elon Musk said, we will build.

Guest Speaker 4m36s

People should treat Elon Musk the way I treat Donald Trump. I don't take him literally. I take him seriously. Because Elon and Gwen were like riffing on how they will build this with femto cells and this and that. And people went off the deep end and, you know, were like, this is not gonna work, blah blah blah blah.

Guest Speaker 4m56s

It's not gonna work like that. So when we look, what does he have? He has Starlink with hundreds of thousands, if not million customers in this country, with a phenomenal router. In our data, we show fiber like quality. Well, actually not fiber like quality.

Guest Speaker 5m16s

And what consumers report, yes. But I'm running every 15 a speed test on my Starlink connection, and I see exactly what's happening.

Roger Entner 5m24s

Well, let me ask you this though, because you're in a suburban area. Right? How many other Starlink dishes are near you?

Guest Speaker 5m29s

So I asked the installer, And the installer was telling me that they're installing in, like, their area, which is, like, South Of Boston, somewhere between five 10 a week.

Roger Entner 5m44s

I agree with you that we should take Elon seriously. I'm also a little bit more of a skeptic, think, on this front. Yeah. And for me, I think the issue is that you have to have some form of, like, terrestrial transport.

Guest Speaker 5m55s

Yeah. And he will do that.

Roger Entner 5m56s

Right. Right.

Guest Speaker 5m57s

Okay. Let me tell you. Let me tell you how he'll do it. So he's gonna use all the Starlink routers that he has, give it a second SSID, and provide that to everybody who has Starlink mobile. This is like what cable does.

Guest Speaker 6m14s

And this will reach like, what, 10 yards maybe outside the premise. But then, in my opinion, if you look at it, T Mobile just swapped spectrum with Grain. And the 800 megahertz spectrum that T Mobile hadn't couldn't get rid of, which they got from Dish well, no. They tried to sell it to Dish. Dish didn't want it.

Guest Speaker 6m37s

Grain swapped it for other spectrum that T Mobile wanted. Grain is not a terminal buyer. In my opinion, Starlink will buy the 800 megahertz from Grain. And then they will put this on macro sites, in the beginning on boomer sites, meaning really tall sites that blast out the signal. That spectrum, if I remember correctly, is seven by seven megahertz in most of The US and three and a half by three and a half at the border.

Guest Speaker 7m6s

This is not a capacity layer by any means. But that's where you put the pilot, and that's where he can, like, take a voice channel. It's owner, Vulti on that side. Right? Then where he needs more capacity, he can use the MSS and the AWS that he got from Starlink for a little bit more densification.

Roger Entner 7m30s

So how many megahertz are we talking total?

Guest Speaker 7m32s

Well, we're talking about 65 here from Starlink. The AWS four is 10, and the MSS is the other stuff. And then in the urban jungles or canyons where your phone cannot see the sky, I'm absolutely convinced he's gonna buy upper c band to densify the network even further. There's an ex parte where before the upper c band decision by the FCC, where they were talking about build out requirements. And they didn't get what they wanted because they argued that, oh, if we can do this with satellite, it should be counted as covered.

Guest Speaker 8m14s

The FCC rightfully said, no. You need to cover it terrestrially. That just means he's not gonna buy upper c band in rural Idaho because he's gonna cover that with satellite. And so in the earnings call, Gwen Shotwell also talked about that they will put the spectrum that they bought from EchoStar into each of the satellites. Absolutely, they will do this, but they will wait for that until the five gs NTN non terrestrial network standard is finalized, which is in 2027.

Guest Speaker 8m51s

Because right now, and if you listen to Elon, he said like, you know, Starlink works really well with WhatsApp and Signal with talking voice. Yes. This is very latency tolerant, but this is not a mainstream telecom offer. And yes, on t satellite, the number one data hog is WhatsApp and doing live video over Starlink in remote parts of the country. But, yes, it works with that, but that's not a telecom offer.

Guest Speaker 9m24s

For telecom offer, you need this to be a five g offer. Right now, Starlink is four g. He's gonna wait until it's five g NTN. He will have a lot more satellite with gen two and gen three, which right now, you can't do the faulty latency budget with Starlink because you don't have enough bandwidth and you don't have enough satellites. He has the permission to put more satellites up.

Guest Speaker 9m49s

The gen twos are 10 times more, and the gen three are a 100 times more than the current satellites. And that will take care of that problem. So we're talking about more like 2030, 2031, 2032, which is also where the build out requirements are hitting for the upper c band. So this all timing wise nicely comes together. And so, yes, he's gonna come.

Guest Speaker 10m15s

He's gonna probably do a very nice job in rural America. Will he be an AT&T, Verizon, T Mobile? No.

Roger Entner 10m25s

That's my issue. Right? Is the existing assets are really good at covering the parts where there aren't people. Right? And it requires a lot of spectrum buys and a lot of regulatory arbitrage as well as building out and spending, you know, potentially hundreds of billions of dollars to build out a terrestrial network basically everywhere where people exist in these sort of density.

Roger Entner 10m44s

Anybody that is willing to spend hundreds of billions of dollars to build out a net new

Guest Speaker 10m48s

But he doesn't need to.

Roger Entner 10m49s

And Dish just tried that, and it didn't work.

Guest Speaker 10m51s

Right? Here's the thing. Dish built a network for 70% of the country on a little bit over $10,000,000,000. He just made an announcement that he's gonna spend, what, a $160,000,000,000 on his TeraFab or something like that. Let me look up the actual number.

Guest Speaker 11m8s

16.8. The relative telecom numbers are for him not that big.

Roger Entner 11m14s

Right. But that's exactly my point too. From a capital allocation perspective. Right? Like, wouldn't you rather build huge data centers and build out a wireless network and then be valued like a wireless network?

Guest Speaker 11m24s

Here's the thing. He doesn't wanna be AT&T, T Mobile, whatever. By the way, he's gonna build everything himself.

Roger Entner 11m31s

Well, he's gonna have to because the carriers are gonna fight him tooth and nail.

Guest Speaker 11m34s

Nobody No. No. The argument is he would buy it from Ericsson or whoever.

Roger Entner 11m40s

Now he's also gonna design all of the radios and everything?

Guest Speaker 11m42s

Yes. Absolutely. That's what he did with his router, and that router is phenomenal. The Starlink router is the best router that you can get anywhere. It is a phenomenal quad band router, and they build it themselves.

Guest Speaker 11m58s

It's standard compliant, but better than standard. It's nuts. Why does he have Grok? Grok, his AI will basically crunch all the code for the software stack for that carrier. He already has billing.

Guest Speaker 12m11s

The buy flow on the website is pretty good. So if somebody can do digital transformation and digital selling, it might be him. We see it in our numbers that the stores, they are horrible, but it's mostly third party distribution. But he doesn't need to be huge to make an impact. And it's free money because the network is already there.

Guest Speaker 12m33s

And he's gonna add it. How is it

Roger Entner 12m35s

free money? Explain to me how this is free money.

Guest Speaker 12m37s

So the satellites are already in space.

Roger Entner 12m39s

Well, the satellites are already there. Yes. But, like, the rest of the network is not free money.

Guest Speaker 12m43s

Okay. The rest of the network is gonna spend $1,020,000,000,000 dollars on it. So for somebody like him, so what? And then look at this as a fully integrated I call it Musk stack, and there's also a Bezos stack. Because what Musk builds from the top down, from the sky down, Bezos is building from Amazon Prime up.

Guest Speaker 13m5s

There's a report coming, by the way. Look at Teslas. In every Tesla built by Tesla is a little data center that has the capability of an h 200 from NVIDIA. It's a distributed data center that sits on a battery with four wheels, electromotors on each wheel, that now and then moves around and transports a person.

Roger Entner 13m27s

Now we've fully passed into the reality distortion field.

Guest Speaker 13m30s

No. No. No. No. No.

Guest Speaker 13m31s

That's already there. That is the vision. And the Teslas will have or have already Starlink connectivity. It uses, right now, AT&T's IoT as a backhaul. This is all coming together.

Guest Speaker 13m46s

Look at the entire stack. He is building this, and the stuff is coming together.

Roger Entner 13m52s

I see wins and losses on the ledger for Elon in terms of making grand promises like this. Well Granted, SpaceX is a incredible achievement and, you know, the rocket's landing themselves. It's fantastic. He's managed to figure out some really fantastic things. Other things that he's promised for a long time, like full self driving, still have not arrived.

Guest Speaker 14m10s

Yes. But that's why I'm taking him seriously, not literally. Yes. The full self driving has not arrived yet. Yes.

Guest Speaker 14m16s

Some things are behind schedule. Do I think he will get there? Yes. Yeah. I absolutely think he will get there.

Roger Entner 14m23s

I think the point is more to drive the stock up, which, by the way, has not been performing well since the IPO than to build out the network. I can see that it is possible for him if the stars align and he spends lots and lots of money and everything else happens his way, that theoretically in five to ten years, it could happen. Right?

Guest Speaker 14m38s

Exactly. And we're talking about five to ten years. He has a long time horizon. And his real mission is to become an interplanetary species. So I get it that these are the stepping stones on the way.

Guest Speaker 14m50s

But the things that he has done and even failures like buying x are turning into an asset, for example, Grok. Grok four point five is actually not that bad.

Roger Entner 15m3s

It's not that bad. Yeah.

Guest Speaker 15m4s

It's not great, but it made huge jumps forward. And it's especially good for the integration with social media. And nothing else can look at Twitter feeds. And, yes, I'm calling it the dead name. Live with it.

Guest Speaker 15m17s

You can look at Twitter feeds like nobody else. And so for especially AI, social media centric applications, by far the best AI, and it's pretty good. As I mentioned, I created a Tableau automation tool with it where we can move dashboards. It's pretty nice. But yeah.

Guest Speaker 15m35s

No. He's coming. I don't think he's coming soon, but he's coming.

Roger Entner 15m40s

I think it's in the realm of possible. I'm not sure it's in the realm of probable. I agree that, like, it is theoretically possible if everything comes together. Right?

Guest Speaker 15m48s

You should take him seriously. The carriers I know are taking him seriously. Even the cable guys are taking him seriously. It's not another FWA event where people are like, oh, this is not gonna do anything.

Roger Entner 16m0s

Right. I mean, if if he's blowing a hole in the side of their stock price, right, like, they have to take him seriously. That's their job.

Guest Speaker 16m6s

You know, now Elon Musk is taking over from Charlie Ergen as our favorite guy to talk about. Yes. Yes. God help us. Alright.

Roger Entner 16m16s

Alright. We'll we'll talk to you next week.

Guest Speaker 16m19s

Talk to you next week.