WEBVTT 00:00:00.000 --> 00:00:03.279 It is episode 220 of the GTO on 5G. It's the 00:00:03.279 --> 00:00:06.719 latest inside scoop on everything 5G and sometimes 00:00:06.719 --> 00:00:09.960 6G. We cover six topics in about 20 minutes and 00:00:09.960 --> 00:00:11.439 it's brought to you by More Insights and Strategy. 00:00:12.240 --> 00:00:14.660 I'm Will Townsend and as you can tell by my background, 00:00:14.759 --> 00:00:19.399 I am not in Bastrop. I'm on the road. I'm in 00:00:19.399 --> 00:00:22.500 Manhattan. We've got a few other team members 00:00:22.500 --> 00:00:25.980 here as well for an IBM infrastructure day, but 00:00:25.980 --> 00:00:28.260 I'm here to spend some time with Palo Alto Networks 00:00:28.260 --> 00:00:31.079 tomorrow. But joining me again this week is Federal 00:00:31.079 --> 00:00:33.479 Analyst Anshul Sag. Before we get started, Anshul, 00:00:33.520 --> 00:00:37.740 are you traveling this week at all? No, I'm traveling 00:00:37.740 --> 00:00:42.179 next week. I'll be at Intel Vision Global in 00:00:42.179 --> 00:00:46.039 Vegas. So we will be hearing from Intel's new 00:00:46.039 --> 00:00:49.960 CEO, amongst other things. Awesome. I can't wait 00:00:49.960 --> 00:00:52.840 to hear your insights from that event, my friend. 00:00:52.880 --> 00:00:55.399 But let's jump into my first topic. I want to 00:00:55.399 --> 00:00:57.859 talk about Verizon. And you and I have spoken 00:00:57.859 --> 00:01:01.320 about them in the past on other podcasts. I felt 00:01:01.320 --> 00:01:03.439 like they were a little late to the AI game relative 00:01:03.439 --> 00:01:06.760 to their competitors. News broke this week, and 00:01:06.760 --> 00:01:08.799 this is actually an article that I found on RCR 00:01:08.799 --> 00:01:11.599 Wireless News, that they've actually launched 00:01:11.599 --> 00:01:15.659 an AI assistant for small businesses. And this 00:01:15.659 --> 00:01:17.379 is really interesting. I don't know if this is 00:01:17.379 --> 00:01:19.280 something that they're necessarily monetizing. 00:01:19.439 --> 00:01:22.299 It may be something that they're doing. to provide 00:01:22.299 --> 00:01:25.040 some stickiness with their small business customers. 00:01:25.180 --> 00:01:28.840 But at a very high level, what the Verizon Business 00:01:28.840 --> 00:01:32.799 Assistant is intended to do is support text messaging 00:01:32.799 --> 00:01:36.560 and allowing small businesses to automate customer 00:01:36.560 --> 00:01:39.780 interactions and improve engagement with this. 00:01:40.000 --> 00:01:44.040 It's basically a text chatbot. But yeah, within 00:01:44.040 --> 00:01:47.329 the press release, it speaks to an example. of 00:01:47.329 --> 00:01:50.310 a customer texting a bakery about gluten -free 00:01:50.310 --> 00:01:53.189 options and how the AI assistant can provide 00:01:53.189 --> 00:01:55.670 an instant answer with context and that sort 00:01:55.670 --> 00:01:59.189 of thing. And from Verizon's perspective, Iris 00:01:59.189 --> 00:02:01.810 Meyer, who's the chief product and marketing 00:02:01.810 --> 00:02:05.010 officer for Verizon Business stated that typically 00:02:05.010 --> 00:02:07.069 small business owners are constantly juggling 00:02:07.069 --> 00:02:09.509 multiple responsibilities. And this is another 00:02:09.509 --> 00:02:13.509 opportunity to provide support for them. And 00:02:13.509 --> 00:02:16.969 a lot of small businesses take advantage of SaaS 00:02:16.969 --> 00:02:19.189 -based security solutions and those sort of things 00:02:19.189 --> 00:02:21.750 because they don't have security operations teams. 00:02:21.849 --> 00:02:24.169 They don't have network teams. So they tend to 00:02:24.169 --> 00:02:28.169 go for services that provide a turnkey type of 00:02:28.169 --> 00:02:31.169 solution. So I think it's interesting. Nothing 00:02:31.169 --> 00:02:33.909 from my perspective, super innovative about it, 00:02:33.949 --> 00:02:36.469 but what are your thoughts, my friend? Yeah, 00:02:36.509 --> 00:02:38.689 I think it's really interesting because small 00:02:38.689 --> 00:02:40.770 businesses are actually some of the best applications 00:02:40.770 --> 00:02:43.650 for these kinds of AI assistants because they're 00:02:43.650 --> 00:02:47.000 force multipliers. And I see that there are definitely 00:02:47.000 --> 00:02:50.319 applications where if I want to call or text 00:02:50.319 --> 00:02:52.860 the business off hours, there should be like 00:02:52.860 --> 00:02:55.419 an AI assistant that can manage that and then 00:02:55.419 --> 00:02:57.400 provide a summary to the business owner when 00:02:57.400 --> 00:02:59.479 they wake up in the morning of what calls and 00:02:59.479 --> 00:03:01.840 texts they missed, if there were any important 00:03:01.840 --> 00:03:05.219 ones. And even maybe if it's really complicated 00:03:05.219 --> 00:03:08.699 or really sophisticated, what business leads 00:03:08.699 --> 00:03:11.060 or what business it generated. I think there's 00:03:11.060 --> 00:03:12.800 a lot of opportunity there for small businesses, 00:03:12.879 --> 00:03:16.939 especially. My most curious thing is, is this 00:03:16.939 --> 00:03:19.360 something that they're tacking on to like phone 00:03:19.360 --> 00:03:21.919 service? So that way, like if you're a Verizon 00:03:21.919 --> 00:03:24.699 mobile customer as a business, like I'm trying 00:03:24.699 --> 00:03:27.439 to understand how they attach it. I assume it's 00:03:27.439 --> 00:03:29.340 got to be with a cellular connection because 00:03:29.340 --> 00:03:32.060 they don't have that many small businesses running 00:03:32.060 --> 00:03:34.800 internet connections. But yeah, I could see this 00:03:34.800 --> 00:03:36.300 being something that they offer for broadband 00:03:36.300 --> 00:03:38.879 customers, for mobile customers. But yeah, like 00:03:38.879 --> 00:03:42.680 in my mind, it's okay. I'm going home at 5 p 00:03:42.680 --> 00:03:45.580 .m. as a lot of small business owners you end 00:03:45.580 --> 00:03:48.659 up having to answer hours after hours called 00:03:48.659 --> 00:03:51.000 but it'd be great if you could maybe use ai to 00:03:51.000 --> 00:03:54.080 handle that and so when people call they don't 00:03:54.080 --> 00:03:55.860 get a voicemail box they get an ai assistant 00:03:55.860 --> 00:03:59.300 that can guide them and answer questions for 00:03:59.300 --> 00:04:01.680 you so you can spend more time with your family 00:04:01.680 --> 00:04:04.270 because i grew up in a household of two small 00:04:04.270 --> 00:04:07.090 business owners i know that yeah i'm very familiar 00:04:07.090 --> 00:04:10.490 with the overflow of small business ownership 00:04:10.490 --> 00:04:14.889 into weekends and evenings and i used to go home 00:04:14.889 --> 00:04:17.769 with my mom in her salon we'd go from school 00:04:17.769 --> 00:04:20.250 to my mom's salon and then go home at eight o 00:04:20.250 --> 00:04:22.569 'clock when she was done so or sometimes when 00:04:22.569 --> 00:04:25.069 she got to get off at five but it's small business 00:04:25.069 --> 00:04:28.740 owners only have 24 hours in a day so I think 00:04:28.740 --> 00:04:30.899 AI tools like this are actually really useful. 00:04:31.060 --> 00:04:33.740 And maybe this is a great way for AI to get access 00:04:33.740 --> 00:04:37.600 to people more with less friction. So I think 00:04:37.600 --> 00:04:40.220 this is a good idea. I wonder if we'll see others 00:04:40.220 --> 00:04:43.000 follow suit. Yeah, no, I like your perspective 00:04:43.000 --> 00:04:44.860 on it because I know you and I have talked about 00:04:44.860 --> 00:04:47.180 you come from a family of entrepreneurs. And 00:04:47.180 --> 00:04:51.060 so maybe I'm being a little too hard on. on Verizon. 00:04:51.360 --> 00:04:54.000 There's definitely value here. To try to answer 00:04:54.000 --> 00:04:55.759 your point around how they're bundling this, 00:04:55.899 --> 00:05:00.120 not abundantly clear to me. And again, not abundantly 00:05:00.120 --> 00:05:02.180 clear to me that this is going to necessarily 00:05:02.180 --> 00:05:05.500 be a big kind of monetization opportunity for 00:05:05.500 --> 00:05:08.120 Verizon, but I see it as sort of a retention 00:05:08.120 --> 00:05:11.259 mechanism, something that will keep small business 00:05:11.259 --> 00:05:15.480 customers sticky, potentially with Verizon. And 00:05:15.480 --> 00:05:17.639 yeah, and what's also interesting when you look 00:05:17.639 --> 00:05:20.860 at a lot of what The mobile network operators 00:05:20.860 --> 00:05:22.959 are dealing with the AI. Initially, it's been 00:05:22.959 --> 00:05:26.319 around customer support. So a very similar use 00:05:26.319 --> 00:05:29.300 case just coming from the MNO. We've talked about 00:05:29.300 --> 00:05:32.740 T -Mobile and its Intent CX platform, which intends 00:05:32.740 --> 00:05:35.079 to take modern AI, generative and eventually 00:05:35.079 --> 00:05:38.920 agentic, and really transform. their customer 00:05:38.920 --> 00:05:42.000 support operational staff yeah it'll be super 00:05:42.000 --> 00:05:44.540 interesting to see into your point will others 00:05:44.540 --> 00:05:48.259 at t and vodafone and you name it will they come 00:05:48.259 --> 00:05:51.120 out with similar offerings as well but with that 00:05:51.120 --> 00:05:53.120 let's go to your first topic you want to talk 00:05:53.120 --> 00:05:57.009 about satellite in a recent funding i guess it's 00:05:57.009 --> 00:05:59.029 a grant from the spanish government and actually 00:05:59.029 --> 00:06:00.610 i didn't i don't know if i mentioned this on 00:06:00.610 --> 00:06:03.009 our mobile world congress wrap -up but i got 00:06:03.009 --> 00:06:05.910 to meet with the ceo and i got to see the size 00:06:05.910 --> 00:06:08.750 of the actual leo satellites that they're putting 00:06:08.750 --> 00:06:11.829 in space and dude they're super tiny they're 00:06:11.829 --> 00:06:13.689 like a little bit bigger than a shoebox but anyway 00:06:13.689 --> 00:06:15.990 what's this news about the spanish government 00:06:15.990 --> 00:06:18.339 and the 15 million yeah so I like to call these 00:06:18.339 --> 00:06:20.959 briefcase satellites. Okay. And they remind me 00:06:20.959 --> 00:06:24.339 of the satellites Sony sent up for space photography. 00:06:24.660 --> 00:06:27.519 But these, yeah, this was a $15 million round 00:06:27.519 --> 00:06:31.740 of funding from the Spanish Society for Technological 00:06:31.740 --> 00:06:37.639 Transformation, also known as SET. And this follows 00:06:37.639 --> 00:06:40.319 a $30 million loan they got from the European 00:06:40.319 --> 00:06:43.839 Union's investment arm. And it's basically just 00:06:43.839 --> 00:06:45.939 more funding to allow them to continue towards 00:06:45.939 --> 00:06:50.449 their goal of operating 250 satellites. And as 00:06:50.449 --> 00:06:54.290 we know, this is a primarily IOT, very small 00:06:54.290 --> 00:06:58.410 data application. And I think when you consider 00:06:58.410 --> 00:07:01.230 what's going on in the market right now, satellite 00:07:01.230 --> 00:07:05.269 tends to be the best way to get to monetization 00:07:05.269 --> 00:07:08.990 quickly. And if you consider that satellite has 00:07:08.990 --> 00:07:11.779 been around for a while, they're already They 00:07:11.779 --> 00:07:14.779 already have contracts representing 270 million 00:07:14.779 --> 00:07:18.439 euros in annual recurring revenue across 400 00:07:18.439 --> 00:07:21.399 customers in 50 countries. They're operational. 00:07:21.699 --> 00:07:24.920 Let's say it that way. And I actually was invited 00:07:24.920 --> 00:07:27.420 to one of their events. I think it was last year 00:07:27.420 --> 00:07:28.699 or the year before. And unfortunately, I had 00:07:28.699 --> 00:07:31.579 to turn them down. But I definitely would love 00:07:31.579 --> 00:07:34.100 to meet with them. I'm glad you were able to. 00:07:34.300 --> 00:07:37.459 Is there anything that you learned at MWC that 00:07:37.459 --> 00:07:41.050 might shed light on? this investment, right? 00:07:41.149 --> 00:07:44.209 Yeah, I did. And what's really interesting, like 00:07:44.209 --> 00:07:47.589 number one, again, I was amazed at the size of 00:07:47.589 --> 00:07:49.610 the satellites, right? AST SpaceMobile and the 00:07:49.610 --> 00:07:51.509 Bluebirds, they keep growing bigger and bigger, 00:07:51.649 --> 00:07:54.350 right? And so it's about the antenna arrays with 00:07:54.350 --> 00:07:57.610 what Abel and his team are doing in Midland. 00:07:57.870 --> 00:08:01.449 And then you look at Starlink and they're not 00:08:01.449 --> 00:08:05.560 shoeboxes, but they're not huge as well. So what 00:08:05.560 --> 00:08:07.319 I found interesting when I talked to the CEO, 00:08:07.600 --> 00:08:10.939 they're only going to need like less than a dozen 00:08:10.939 --> 00:08:13.879 of these very small satellites to completely 00:08:13.879 --> 00:08:17.560 support their global connectivity of their narrow 00:08:17.560 --> 00:08:21.060 band IoT offering. I think that's amazing. And 00:08:21.060 --> 00:08:24.079 I haven't researched the intellectual property 00:08:24.079 --> 00:08:27.240 that resides underneath what he and his team 00:08:27.240 --> 00:08:30.199 have done in Barcelona there. But I think that's 00:08:30.199 --> 00:08:32.059 pretty mind blowing, right? You know, Abel with 00:08:32.059 --> 00:08:33.580 AST Space Mobile, I think they're going to need 00:08:33.580 --> 00:08:37.860 45. or 50 of those larger lights for in and what 00:08:37.860 --> 00:08:40.820 they intend to do let's be honest is broadband 00:08:40.820 --> 00:08:43.440 connectivity and i talked about this in my forbes 00:08:43.440 --> 00:08:46.379 article what what starlink is doing with these 00:08:46.379 --> 00:08:48.179 smaller satellites but it's going to require 00:08:48.179 --> 00:08:52.879 thousands of those is really it's text messaging 00:08:53.909 --> 00:08:56.769 It'll be interesting to see how they can extend 00:08:56.769 --> 00:08:59.950 the broadband up and down. But it's primarily 00:08:59.950 --> 00:09:02.889 text, emergency text messaging. And then you've 00:09:02.889 --> 00:09:05.370 got this, with Satellite, you've got a full complement 00:09:05.370 --> 00:09:08.230 with narrowband IoT. So it's really interesting 00:09:08.230 --> 00:09:10.669 to see each of these companies, these three companies 00:09:10.669 --> 00:09:13.149 we're talking about right now, and what they're 00:09:13.149 --> 00:09:17.269 doing with the physical size of the satellites. 00:09:18.360 --> 00:09:20.559 Totally different approaches, totally different 00:09:20.559 --> 00:09:23.399 modernization approaches as well. And it's super 00:09:23.399 --> 00:09:25.779 fascinating. But I think there's room for all 00:09:25.779 --> 00:09:27.639 of these folks to play, carve out a niche for 00:09:27.639 --> 00:09:29.399 themselves. That's my take on it, my sense of 00:09:29.399 --> 00:09:31.700 it. Yeah, I was also going to add that from what 00:09:31.700 --> 00:09:34.600 I saw. it seemed like this 15 million dollars 00:09:34.600 --> 00:09:38.139 or 13 million 13 .8 million euros was mostly 00:09:38.139 --> 00:09:42.139 being used to help them accelerate to 100 satellites 00:09:42.139 --> 00:09:45.559 but also they still have a ways to go i think 00:09:45.559 --> 00:09:47.700 they're also looking for other funding opportunities 00:09:47.700 --> 00:09:50.429 to continue that growth Yeah. And the number 00:09:50.429 --> 00:09:52.049 that I'm quoting there are less than a dozen 00:09:52.049 --> 00:09:54.509 of that. That might be based on one particular 00:09:54.509 --> 00:09:58.389 narrow band use case globally. But that was my 00:09:58.389 --> 00:10:00.490 understanding is that they didn't need hundreds 00:10:00.490 --> 00:10:03.110 of birds to accomplish what they're doing. But 00:10:03.110 --> 00:10:05.789 anyway, really fascinating stuff. But hey, let 00:10:05.789 --> 00:10:08.610 me go to my second topic. And I know you were 00:10:08.610 --> 00:10:11.009 at GTC. You were on the ground. You were live 00:10:11.009 --> 00:10:13.129 tweeting. You were with Patrick Moorhead, our 00:10:13.129 --> 00:10:15.710 fearless leader. You were with Matt Kimball. 00:10:15.950 --> 00:10:19.240 I was literally probably across the street or 00:10:19.240 --> 00:10:22.580 across the highway with Synopsys in Santa Clara 00:10:22.580 --> 00:10:24.279 at their executive forum. It was a great event. 00:10:24.460 --> 00:10:26.960 I got to do some panel moderation there. But 00:10:26.960 --> 00:10:29.980 there was a lot of news, obviously, out of GTC. 00:10:30.259 --> 00:10:33.379 I recorded the Data Center podcast with Kimball 00:10:33.379 --> 00:10:36.919 and with Paul earlier this week. And I talked 00:10:36.919 --> 00:10:39.279 a little bit about the co -package optics and 00:10:39.279 --> 00:10:42.820 the silicon photonic switches and what NVIDIA 00:10:42.820 --> 00:10:46.080 is claiming it can drive from a performance and 00:10:46.080 --> 00:10:49.980 a power efficiency perspective. to provide connectivity 00:10:49.980 --> 00:10:53.980 for what's expected to be thousands, if not millions, 00:10:54.139 --> 00:10:58.059 of AI factories down the road. But this is our 00:10:58.059 --> 00:11:01.720 5G podcast, our slash 6G podcast. So I want to 00:11:01.720 --> 00:11:05.500 talk about AI RAN. And NVIDIA continues to push 00:11:05.500 --> 00:11:08.639 this very heavily. At GTC, there was an announcement 00:11:08.639 --> 00:11:12.019 with T -Mobile and with Cisco on how they're 00:11:12.019 --> 00:11:16.710 fostering an AI -native approach. to set the 00:11:16.710 --> 00:11:21.149 stage for what 6G could be. What's really interesting, 00:11:21.350 --> 00:11:26.529 Anshul, about this is that 3GPP drives the standards, 00:11:26.730 --> 00:11:30.289 but NVIDIA, they are driving the bus on modern 00:11:30.289 --> 00:11:33.509 AI, right? It appears to me that they want to 00:11:33.509 --> 00:11:37.029 drive the bus on 6G. And I do think everyone 00:11:37.029 --> 00:11:40.350 agrees that there is going to be an AI integration. 00:11:41.419 --> 00:11:44.320 within the 6G standard, but we just don't know 00:11:44.320 --> 00:11:45.960 what that's going to look like until we get to 00:11:45.960 --> 00:11:48.639 3GPP. But hey, man, you were on the ground. What 00:11:48.639 --> 00:11:50.980 was your perspective tied to this news? I'd love 00:11:50.980 --> 00:11:54.559 to hear that. It was very clear that this is 00:11:54.559 --> 00:11:59.159 a strong intention from these three partners 00:11:59.159 --> 00:12:05.230 to establish an early... group of companies that 00:12:05.230 --> 00:12:08.690 are trying to push the standard towards ai i 00:12:08.690 --> 00:12:12.190 think cisco's become increasingly tied to nvidia 00:12:12.190 --> 00:12:15.250 especially with the new networking stuff and 00:12:15.250 --> 00:12:19.049 t -mobile is pretty pretty closely tied to nvidia 00:12:19.049 --> 00:12:22.950 as well t -mobile is also a very leading carrier 00:12:22.950 --> 00:12:26.009 when it comes to implementing 3gpp standards 00:12:26.009 --> 00:12:29.250 yeah i think that them being part of this and 00:12:29.250 --> 00:12:31.899 i think This will go up to Deutsche Telekom as 00:12:31.899 --> 00:12:34.120 well, not just T -Mobile US. There's going to 00:12:34.120 --> 00:12:37.100 be a lot of influence on how the standards are 00:12:37.100 --> 00:12:40.019 set and what features get implemented, especially 00:12:40.019 --> 00:12:44.299 if NVIDIA wants to be that primary vendor and 00:12:44.299 --> 00:12:47.960 supplier for equipment. And I think this is one 00:12:47.960 --> 00:12:50.700 of those things where a vendor sets themselves 00:12:50.700 --> 00:12:53.659 up for the next G five, 10 years ahead of it. 00:12:53.820 --> 00:12:56.519 And this seems like this is NVIDIA trying to 00:12:56.519 --> 00:13:01.529 set up themselves as the default. for 6G, I actually 00:13:01.529 --> 00:13:05.330 posed a question to the NVIDIA telecoms team 00:13:05.330 --> 00:13:09.230 during our session about 6G specifically, because 00:13:09.230 --> 00:13:12.409 we haven't set the standard yet. And I was trying 00:13:12.409 --> 00:13:15.950 to get an answer on what are you guys doing to 00:13:15.950 --> 00:13:19.289 influence the standards and get these things 00:13:19.289 --> 00:13:22.909 done? And I got a nebulous answer, to be honest 00:13:22.909 --> 00:13:26.590 with you. But truthfully, I think what NVIDIA 00:13:26.590 --> 00:13:30.460 is doing is leading on research. working with 00:13:30.460 --> 00:13:33.500 different vendors like Ericsson, like T -Mobile, 00:13:33.720 --> 00:13:37.639 like Cisco. And I think a lot of it truthfully 00:13:37.639 --> 00:13:40.059 is going to end up on how well do they deploy 00:13:40.059 --> 00:13:44.179 5G and how well do they play with the existing 00:13:44.179 --> 00:13:46.879 networks? Because true, the 6G networks will 00:13:46.879 --> 00:13:49.559 be standalone, but there will still need to be 00:13:49.559 --> 00:13:55.080 a lot of interoperability and acceleration and 00:13:55.080 --> 00:13:59.360 open software. I think NVIDIA's... wants to set 00:13:59.360 --> 00:14:02.200 themselves up for success down the road. I think 00:14:02.200 --> 00:14:05.539 their current roadmap is pretty good, but they 00:14:05.539 --> 00:14:07.899 did say we're already doing a lot of the things 00:14:07.899 --> 00:14:11.480 in 5G that we want to do in 6G. I think the T 00:14:11.480 --> 00:14:14.639 -Mobile and Verizon partnerships with NVIDIA 00:14:14.639 --> 00:14:16.720 are the first steps in that direction. Yeah, 00:14:16.740 --> 00:14:19.279 no, I totally agree. And I'm not faulting NVIDIA. 00:14:19.440 --> 00:14:21.500 They're being aggressive. They should be, because 00:14:21.500 --> 00:14:24.379 like I said, everyone agrees that they're driving 00:14:24.379 --> 00:14:27.580 the bus with respect to modern AI. And Silicon 00:14:27.580 --> 00:14:30.440 and software and all of that. And hey, look at 00:14:30.440 --> 00:14:34.580 Qualcomm, right? With early and NTN, they made 00:14:34.580 --> 00:14:36.919 some bets. They tried some similar things that 00:14:36.919 --> 00:14:41.029 didn't quite pan out. Remember? And hey. Qualcomm 00:14:41.029 --> 00:14:43.129 continues to be a successful company. NVIDIA 00:14:43.129 --> 00:14:45.370 continues to be successful. And sometimes you 00:14:45.370 --> 00:14:47.789 got to be bold, right? And you've got to be willing 00:14:47.789 --> 00:14:49.909 to get out there and lead and see where it goes. 00:14:50.049 --> 00:14:52.309 So I think it's interesting. But let's hit your 00:14:52.309 --> 00:14:54.350 second topic. And we want to continue to talk 00:14:54.350 --> 00:14:57.210 about T -Mobile here. And I saw this. They're 00:14:57.210 --> 00:14:59.870 pretty close to some agreements with their competitors 00:14:59.870 --> 00:15:03.799 on swapping spectrum. Yeah, so this is a... 800 00:15:03.799 --> 00:15:06.799 megahertz spectrum deal. It's actually with Grain 00:15:06.799 --> 00:15:10.080 Management LLC, which is actually a private investment 00:15:10.080 --> 00:15:13.279 firm specifically focused on digital infrastructure 00:15:13.279 --> 00:15:16.500 and solutions and broadband. So basically what's 00:15:16.500 --> 00:15:21.600 happening is they are going to sell their 800 00:15:21.600 --> 00:15:24.820 megahertz spectrum to this firm with a combination 00:15:24.820 --> 00:15:29.500 of cash and 600 megahertz spectrum. So Grain 00:15:29.500 --> 00:15:32.919 has 600 megahertz spectrum. T -Mobile. has cash, 00:15:33.320 --> 00:15:36.340 T -Mobile is giving them this 800 megahertz spectrum 00:15:36.340 --> 00:15:39.200 and getting cash in 600 megahertz spectrum that 00:15:39.200 --> 00:15:41.960 they were already leasing. It's like a swap. 00:15:42.080 --> 00:15:45.480 It's a swap. We don't know how much cash is being 00:15:45.480 --> 00:15:48.360 changed hands, but basically what this does is 00:15:48.360 --> 00:15:51.519 it bolsters T -Mobile 600 megahertz coverage 00:15:51.519 --> 00:15:54.840 and spectrum, which they already had, and it 00:15:54.840 --> 00:15:57.220 helps them with coverage. And then that gives 00:15:57.220 --> 00:16:00.120 away the 800 megahertz spectrum. that someone 00:16:00.120 --> 00:16:04.120 did not use. I wonder who that was. And yeah, 00:16:04.360 --> 00:16:06.919 basically what I'm hearing is that this will 00:16:06.919 --> 00:16:11.120 most likely be used for utility companies because 00:16:11.120 --> 00:16:13.259 800 megahertz is still a very low band spectrum 00:16:13.259 --> 00:16:17.139 and it could be used for IoT applications that 00:16:17.139 --> 00:16:20.539 would be nationwide. And yeah, it just seems 00:16:20.539 --> 00:16:22.919 like it's a good deal because T -Mobile gets 00:16:22.919 --> 00:16:25.399 spectrum that they needed. They don't need the 00:16:25.399 --> 00:16:28.980 800 because they have the 600. And 600 has better 00:16:28.980 --> 00:16:33.379 propagation anyways. And the 800 megahertz blocks 00:16:33.379 --> 00:16:35.960 are going to be about the same as 600 megahertz 00:16:35.960 --> 00:16:38.539 blocks. So getting more 600 megahertz where they 00:16:38.539 --> 00:16:40.360 already have a lot is the best approach for them 00:16:40.360 --> 00:16:44.360 anyways. Yeah. I agree. No, it makes sense. They've 00:16:44.360 --> 00:16:47.059 invested in 600 megahertz and to be consistent 00:16:47.059 --> 00:16:48.879 with that, I think is important. And you're right. 00:16:48.960 --> 00:16:53.519 It propagates much better. This all goes back 00:16:53.519 --> 00:16:57.519 to T -Mobile's initial objective, right? With 00:16:57.519 --> 00:17:00.159 5G, which was they wanted to get the most coverage 00:17:00.159 --> 00:17:02.500 out there. And they were very aggressive with 00:17:02.500 --> 00:17:05.259 that. They've invested in rural. You can argue 00:17:05.259 --> 00:17:08.779 that AT &T is invested in rural as well, especially 00:17:08.779 --> 00:17:12.039 when you look at fiber buildouts and places like 00:17:12.039 --> 00:17:16.299 Evansville that didn't have the same proliferation 00:17:16.299 --> 00:17:19.750 of fiber as more densely populated areas. No, 00:17:19.769 --> 00:17:22.210 I think it's a smart move and T -Mobile continues 00:17:22.210 --> 00:17:24.970 to refine things and sharpen the checkbook. And 00:17:24.970 --> 00:17:27.049 it's interesting to watch, man. I still remember 00:17:27.049 --> 00:17:30.630 when the whole Sprint deal was teetering, you 00:17:30.630 --> 00:17:33.009 and I, we had spent time with executives. We 00:17:33.009 --> 00:17:35.630 understood the fundamentals. You and I put our 00:17:35.630 --> 00:17:37.369 reputations on the line and we said, this deal 00:17:37.369 --> 00:17:40.190 is going to get done and watch what happens with 00:17:40.190 --> 00:17:43.049 T -Mobile. I got to tell you, man, let's take 00:17:43.049 --> 00:17:45.069 that victory lap again. And I'm not throwing 00:17:45.069 --> 00:17:47.450 shade on Verizon. They've had their challenges. 00:17:47.509 --> 00:17:50.710 I'm not throwing shade on AT &T, but it was really 00:17:50.710 --> 00:17:53.450 interesting to see basically the third place 00:17:53.450 --> 00:17:58.130 player emerge with what's arguably the strongest 00:17:58.130 --> 00:18:01.150 5G network on the planet right now. So anyway, 00:18:01.309 --> 00:18:03.029 I thought I'd just bring that up so we could 00:18:03.029 --> 00:18:05.549 take that victory lap. I'll just add before we 00:18:05.549 --> 00:18:09.369 start talking about SoftBank, that merger was 00:18:09.369 --> 00:18:13.240 the perfect example. of an actual merger that 00:18:13.240 --> 00:18:17.440 benefited the entire industry and consumers. 00:18:17.920 --> 00:18:21.559 T -Mobile and Sprint were both kind of capped 00:18:21.559 --> 00:18:25.359 in growth and not really being successful together 00:18:25.359 --> 00:18:28.500 or separately. And the combination of the two 00:18:28.500 --> 00:18:32.400 just, it boosted competitiveness from AT &T and 00:18:32.400 --> 00:18:35.400 Verizon. It catapulted them to first place. And 00:18:35.400 --> 00:18:37.640 I think it accelerated 5G deployments in the 00:18:37.640 --> 00:18:40.599 country. Yeah, no, I agree. Competition breeds 00:18:40.599 --> 00:18:42.559 innovation. I'm going to say it for the millionth 00:18:42.559 --> 00:18:45.480 time again. So totally agree with you. We're 00:18:45.480 --> 00:18:48.799 going to close out the podcast this week and 00:18:48.799 --> 00:18:52.299 with a SoftBank super topic. So SoftBank has 00:18:52.299 --> 00:18:55.019 been on a tear. It's funny, man. You and I have 00:18:55.019 --> 00:18:57.599 talked about them, the ebb and flow over the 00:18:57.599 --> 00:19:00.839 years. But lately, it's clear to me that they 00:19:00.839 --> 00:19:04.859 are doubling down when it comes to just AI in 00:19:04.859 --> 00:19:07.809 general. You're going to talk about Ampere. big 00:19:07.809 --> 00:19:11.230 acquisition there actually matt kimball weighed 00:19:11.230 --> 00:19:13.750 in on that in a forbes article and i would encourage 00:19:13.750 --> 00:19:16.269 our viewers and listeners to check that out but 00:19:16.269 --> 00:19:19.849 before you you get to ampere i want to talk about 00:19:19.849 --> 00:19:22.349 their announcement of a telco large language 00:19:22.349 --> 00:19:24.869 model that their positioning is foundational 00:19:24.869 --> 00:19:28.730 for telecom during the ai application so What 00:19:28.730 --> 00:19:32.369 they're calling their LLM is the large telecom 00:19:32.369 --> 00:19:35.529 model. I like that, the LTM. And it's built on 00:19:35.529 --> 00:19:37.789 SoftBank's own network and operational data. 00:19:38.009 --> 00:19:41.769 So remember, SoftBank, it does operate a mobile 00:19:41.769 --> 00:19:46.069 network in Japan, but they also have this vision 00:19:46.069 --> 00:19:48.730 fund. They also invest in a lot of technology 00:19:48.730 --> 00:19:51.470 and they acquire companies and that sort of thing. 00:19:52.460 --> 00:19:55.400 So they've designed their LTM, their large telecom 00:19:55.400 --> 00:19:57.799 model, based on their own network and operational 00:19:57.799 --> 00:20:01.660 data. They've been supposedly fine -tuning this. 00:20:02.220 --> 00:20:04.480 And in particular, they cite a recent example 00:20:04.480 --> 00:20:07.079 how it was configured for a base station optimization. 00:20:07.940 --> 00:20:10.579 And they found that it reduced task times from 00:20:10.579 --> 00:20:13.579 days to minutes. And this is huge. And when you 00:20:13.579 --> 00:20:16.160 talk about what NVIDIA wants to do with AI in 00:20:16.160 --> 00:20:18.440 the RAN, a lot of that comes down to base station 00:20:18.440 --> 00:20:21.440 optimization, right? So this is precursor to 00:20:21.440 --> 00:20:24.099 potentially what NVIDIA wants to drive forward 00:20:24.099 --> 00:20:27.099 with what it's doing with silicon and curating 00:20:27.099 --> 00:20:29.259 large language models and small language models 00:20:29.259 --> 00:20:32.079 and that sort of thing. I think it's interesting, 00:20:32.160 --> 00:20:35.339 and I think the timing with it as well, the fact 00:20:35.339 --> 00:20:37.900 that SoftBank is a founding member of the Global 00:20:37.900 --> 00:20:43.000 Teleco AI Alliance, but also the timing around 00:20:43.000 --> 00:20:47.619 its announced intent to acquire Ampere. And I'm 00:20:47.619 --> 00:20:49.880 going to let you take it from there because there 00:20:49.880 --> 00:20:53.759 is a correlation between these two. Yeah, so 00:20:53.759 --> 00:20:56.279 Ampere is an interesting company because they're 00:20:56.279 --> 00:20:58.940 one of the few... arm server vendors that are 00:20:58.940 --> 00:21:01.440 out there arguably one of the most high performance 00:21:01.440 --> 00:21:05.480 arms from server vendors out there and it's interesting 00:21:05.480 --> 00:21:09.059 because softbank through arm already owned eight 00:21:09.059 --> 00:21:14.079 percent of ampere so this was like them buying 00:21:14.079 --> 00:21:17.519 themselves out yeah or just increasing their 00:21:17.519 --> 00:21:22.079 share of the company but yeah ampere has succeeded 00:21:22.079 --> 00:21:25.339 where others have failed in terms of arm server 00:21:26.089 --> 00:21:29.329 availability and products. I wouldn't say that 00:21:29.329 --> 00:21:32.170 they've necessarily been gangbusters, but they 00:21:32.170 --> 00:21:36.710 were founded by a former Intel exec, Rene James, 00:21:36.890 --> 00:21:39.730 who was pretty familiar with the space. And I 00:21:39.730 --> 00:21:42.369 think that this was very much SoftBank leaning 00:21:42.369 --> 00:21:48.650 into their AI ambitions, also their telecom infrastructure 00:21:48.650 --> 00:21:52.730 plans, and also making sure that ARM servers 00:21:52.730 --> 00:21:55.990 continue to grow in relevance. whether that's 00:21:55.990 --> 00:21:59.650 in telecom or in the data center or for ai yeah 00:21:59.650 --> 00:22:03.269 while this wasn't necessarily the full eight 00:22:03.269 --> 00:22:05.230 billion dollar valuation that this company had 00:22:05.230 --> 00:22:07.990 two or three years ago it is still 6 .5 billion 00:22:07.990 --> 00:22:10.970 so quite a bit of money and i think that this 00:22:10.970 --> 00:22:15.869 will also enable softbank to make arm chips without 00:22:15.869 --> 00:22:18.750 having arm be the company who makes them yeah 00:22:18.750 --> 00:22:23.130 and then what this obviously sets up is SoftBank 00:22:23.130 --> 00:22:27.650 wants to make a run at modern AI. And now they've 00:22:27.650 --> 00:22:30.430 got large language models that are foundational. 00:22:30.849 --> 00:22:34.049 They've got purpose -built silicon. They're putting 00:22:34.049 --> 00:22:36.609 all the pieces together. At the end of the day, 00:22:36.750 --> 00:22:39.609 there's really, I would say that AMD is a close 00:22:39.609 --> 00:22:43.490 second to really competing with NVIDIA, but NVIDIA 00:22:43.490 --> 00:22:46.869 clearly has a, they're out in front of all of 00:22:46.869 --> 00:22:49.900 these competitors quite a bit. But I see this 00:22:49.900 --> 00:22:52.700 as an opportunity for SoftBank. So they're doubling 00:22:52.700 --> 00:22:56.299 the ring on AI. Yeah. All right, my friend, it's 00:22:56.299 --> 00:22:58.059 been another great podcast. Why don't you take 00:22:58.059 --> 00:23:00.599 us home? Absolutely. We hope our viewers and 00:23:00.599 --> 00:23:02.319 listeners found this week's topics interesting. 00:23:02.539 --> 00:23:04.779 If anyone out there would like to provide insights 00:23:04.779 --> 00:23:07.759 on a specific 5G topic for future podcasts, please 00:23:07.759 --> 00:23:10.799 reach out to us on social media. Will is at WillTownTech 00:23:10.799 --> 00:23:14.400 and I'm at Anshul Saag. We hope you have a great 00:23:14.400 --> 00:23:17.079 week and don't forget to rate and subscribe.