WEBVTT 00:00:00.000 --> 00:00:01.800 Well, Anshul, it's been a while since you and 00:00:01.800 --> 00:00:05.500 I've done a 5G podcast, our standalone edition. 00:00:06.280 --> 00:00:09.140 And I'm excited. We've got Shahid Ahmed, leads 00:00:09.140 --> 00:00:12.740 NTT Private 5G. But Shahid, I love your title, 00:00:12.919 --> 00:00:16.600 New Ventures and Innovation. Before we get started, 00:00:16.699 --> 00:00:18.219 can you talk a little bit about that? Because 00:00:18.219 --> 00:00:20.140 that sounds like it's a lot more expansive than 00:00:20.140 --> 00:00:23.660 just private 5G, right? That is true. That is 00:00:23.660 --> 00:00:26.660 true. Thanks. And great to see both. Yeah, look, 00:00:26.920 --> 00:00:31.559 the world is shifting as well, even with AI in 00:00:31.559 --> 00:00:35.859 pretty much everything we do. Much like that 00:00:35.859 --> 00:00:40.920 in corporate setting, we have to have flexible 00:00:40.920 --> 00:00:45.079 titling. One of the advantages of working in 00:00:45.079 --> 00:00:47.500 a large company is to be able to get one of those. 00:00:47.780 --> 00:00:51.469 But joking aside, I think it's... This is mostly 00:00:51.469 --> 00:00:54.810 because I'm responsible for coming up with new 00:00:54.810 --> 00:00:58.189 ideas, new offerings that we can take to the 00:00:58.189 --> 00:01:01.890 market. As markets are shifting regularly on 00:01:01.890 --> 00:01:06.310 a daily, hourly basis, and we got to keep up. 00:01:06.430 --> 00:01:09.469 And market fit is a very important part of our 00:01:09.469 --> 00:01:12.870 overall roadmap, and we got to continuously... 00:01:13.549 --> 00:01:17.069 Build new things, new offerings that manage and 00:01:17.069 --> 00:01:20.250 fit to our customers' requirements and needs. 00:01:20.549 --> 00:01:22.890 Yeah, I've gotten to know you over the years. 00:01:22.930 --> 00:01:24.950 I've gotten to know NTT as a company as well. 00:01:25.010 --> 00:01:27.909 I've spent time at R &D Forum in Tokyo. I've 00:01:27.909 --> 00:01:30.849 been to a few upgrades in San Francisco. And 00:01:30.849 --> 00:01:32.709 I'm really impressed with what the company is 00:01:32.709 --> 00:01:35.010 doing, not only from a data center perspective, 00:01:35.170 --> 00:01:37.890 but with private 5G. And I really, when I look 00:01:37.890 --> 00:01:40.090 at 2025, and this is a subject that Anshul and 00:01:40.090 --> 00:01:42.420 I have talked about. on our podcast and like 00:01:42.420 --> 00:01:44.980 we're about to celebrate our fifth year and it's 00:01:44.980 --> 00:01:46.599 crazy to think that he and i have been doing 00:01:46.599 --> 00:01:49.239 this for five years now but really from my perspective 00:01:49.239 --> 00:01:52.620 2025 is really becoming the year of private cellular 00:01:52.620 --> 00:01:55.920 and private 5g that hockey stick is finally taking 00:01:55.920 --> 00:01:58.650 off and He and I have talked about, and I think 00:01:58.650 --> 00:02:00.769 you and I have talked about as well, Shahed, 00:02:00.829 --> 00:02:03.469 some of the challenges, different paths to get 00:02:03.469 --> 00:02:06.010 to 5G and that sort of thing. But I'd love it 00:02:06.010 --> 00:02:07.849 if you could spend a little bit of time and discuss 00:02:07.849 --> 00:02:09.849 from your perspective, what's really driving 00:02:09.849 --> 00:02:13.330 the momentum behind private 5G in 2025 and beyond. 00:02:13.449 --> 00:02:15.750 Yeah, you're right. We're also seeing that little 00:02:15.750 --> 00:02:19.530 uptick. And you and I had spoken about our private 00:02:19.530 --> 00:02:22.310 5G offering when we first launched, like almost 00:02:22.310 --> 00:02:25.150 three and a half, almost four years ago. Yeah. 00:02:25.250 --> 00:02:29.990 Things were very... rough and fresh and new to 00:02:29.990 --> 00:02:32.789 a lot of constituents, whether it's the suppliers 00:02:32.789 --> 00:02:36.590 or the customers or the other ecosystem players 00:02:36.590 --> 00:02:40.409 like device manufacturers. It was new and we 00:02:40.409 --> 00:02:44.550 kept at it. And now we're beginning to see and 00:02:44.550 --> 00:02:47.969 reap the investments we made earlier on in private 00:02:47.969 --> 00:02:52.409 5G. So yeah, absolutely beginning to see that 00:02:52.409 --> 00:02:55.509 uptick effect. I think that's been going on for 00:02:55.509 --> 00:02:59.270 the last six months. And so why is that? And 00:02:59.270 --> 00:03:01.430 we've looked into that quite deeply ourselves 00:03:01.430 --> 00:03:04.870 because we're sold out and we're trying to figure 00:03:04.870 --> 00:03:09.270 out where to shift our resources so we can manage 00:03:09.270 --> 00:03:13.129 this uptick in demand. And we see three major 00:03:13.129 --> 00:03:16.370 forces that are driving this new demand, if you 00:03:16.370 --> 00:03:20.430 will. One, I think devices are finally beginning 00:03:20.430 --> 00:03:25.250 to make it. to the overall ecosystem. In the 00:03:25.250 --> 00:03:28.250 beginning, we didn't see natively supported 5G 00:03:28.250 --> 00:03:31.370 devices for enterprises. And it was a major problem, 00:03:31.509 --> 00:03:34.810 to be frank with you. And that kind of impeded 00:03:34.810 --> 00:03:37.370 a lot of the progress. Many of our customers 00:03:37.370 --> 00:03:40.949 wanted 5G. There simply weren't enough devices 00:03:40.949 --> 00:03:44.289 or the variety of devices we needed to support 00:03:44.289 --> 00:03:46.650 those use cases. And so it was like always a 00:03:46.650 --> 00:03:50.389 chicken and egg situation. And that really slowed 00:03:50.389 --> 00:03:54.250 down a lot of deployments. In fact, to the extent 00:03:54.250 --> 00:03:57.949 even customers that had the budgets, had the 00:03:57.949 --> 00:04:01.650 SOW signed with us, wanted to delay it just because 00:04:01.650 --> 00:04:03.469 they're like, we don't want to start with the 00:04:03.469 --> 00:04:07.250 wrong foot forward. And let's wait till the devices 00:04:07.250 --> 00:04:11.389 make it to the market and then we'll get going. 00:04:11.610 --> 00:04:14.870 So I think that was one. One area that we thought 00:04:14.870 --> 00:04:17.550 has really matured and we're beginning to see 00:04:17.550 --> 00:04:20.610 a plethora of devices out there today. So that 00:04:20.610 --> 00:04:24.889 was a big tailwind behind our backs. The second 00:04:24.889 --> 00:04:28.629 one, I think, was just the availability of spectrum. 00:04:29.490 --> 00:04:33.350 I think that is something not just U .S., but 00:04:33.350 --> 00:04:37.170 outside of U .S., we were challenged with many 00:04:37.170 --> 00:04:40.730 regulators. not really fully getting on board 00:04:40.730 --> 00:04:44.250 and allocating Spectrum to enterprises directly. 00:04:44.810 --> 00:04:49.329 And in fact, even as recent as six, eight months 00:04:49.329 --> 00:04:52.490 ago, we had to work with Ofcom to see if they 00:04:52.490 --> 00:04:56.850 could figure out a way to allocate Spectrum directly 00:04:56.850 --> 00:04:59.189 to an enterprise. And there was a lot of process, 00:04:59.389 --> 00:05:02.790 a lot of red tape. But I think... By and large, 00:05:02.910 --> 00:05:04.990 we're beginning to see a lot of these regulators 00:05:04.990 --> 00:05:08.529 now get on board. I think that has been very 00:05:08.529 --> 00:05:12.050 helpful as well. Because without spectrum, you 00:05:12.050 --> 00:05:15.310 can't run a cellular network. And that is a fundamental 00:05:15.310 --> 00:05:18.529 requirement. Yeah. And CBRS has been an important 00:05:18.529 --> 00:05:21.970 part of democratizing access in the U .S. But 00:05:21.970 --> 00:05:24.980 I think to your point. It varies by geography, 00:05:25.220 --> 00:05:27.560 right? Europe is a whole different ball of wax, 00:05:27.560 --> 00:05:30.100 European Commission, and how individual countries 00:05:30.100 --> 00:05:32.860 manage spectrum as well. So you got to democratize 00:05:32.860 --> 00:05:35.300 that spectrum because that's how this all works, 00:05:35.360 --> 00:05:38.079 right? Wi -Fi is unlicensed spectrum. So, yeah. 00:05:38.220 --> 00:05:40.360 And we should talk about what's happening with 00:05:40.360 --> 00:05:45.079 CBRs in the US. There's some political headwinds 00:05:45.079 --> 00:05:48.420 around that. Sure. We're recently coming across. 00:05:48.720 --> 00:05:52.000 But I think spectrum, that was... Again, a big 00:05:52.000 --> 00:05:54.879 impediment, and I think that's by and large solved 00:05:54.879 --> 00:05:58.019 in most geographies. There's still a lot of geographies, 00:05:58.040 --> 00:06:02.300 especially in particular in Asia, in some of 00:06:02.300 --> 00:06:05.920 the countries, Southeast Asia, they still haven't 00:06:05.920 --> 00:06:08.459 really specifically allocated spectrum to the 00:06:08.459 --> 00:06:12.019 enterprises. But I think they're slowly moving 00:06:12.019 --> 00:06:15.500 forward with that strategy. The third one we're 00:06:15.500 --> 00:06:20.779 beginning to see is just the idea of... economics 00:06:20.779 --> 00:06:23.860 behind private 5g you and i have talked about 00:06:23.860 --> 00:06:26.240 that quite a bit in the past right yeah yeah 00:06:26.240 --> 00:06:30.699 and we always said that private 5g has to have 00:06:30.699 --> 00:06:35.540 wi -fi economics and what i mean by that is look 00:06:35.540 --> 00:06:40.180 it it cannot be telco -centric architecture that 00:06:40.180 --> 00:06:44.759 naturally drives a very hefty capex and opex 00:06:44.759 --> 00:06:48.600 bill of material for you, right? And so it has 00:06:48.600 --> 00:06:51.800 to come down to a point where it's not even an 00:06:51.800 --> 00:06:56.339 economic decision whether to pick Wi -Fi or 5G. 00:06:56.500 --> 00:07:01.420 It's just so easy in terms of making a decision 00:07:01.420 --> 00:07:05.009 that, yes, we'll do. We'll do 5G in the warehouse. 00:07:05.189 --> 00:07:08.990 We'll do 5G in the parking lot where thousands 00:07:08.990 --> 00:07:12.829 of cars are waiting to get shipped out to the 00:07:12.829 --> 00:07:15.329 rest of the world for an auto manufacturing companies. 00:07:15.569 --> 00:07:19.310 The economics, I think it became much more clearer 00:07:19.310 --> 00:07:24.509 in terms of what is the ROI. And it wasn't really 00:07:24.509 --> 00:07:26.709 the price points that we were worried about. 00:07:26.889 --> 00:07:29.970 Yes, indeed, it did drop over the last three 00:07:29.970 --> 00:07:34.199 years. And the way we sold it became much more 00:07:34.199 --> 00:07:38.800 articulate in terms of what are we selling as 00:07:38.800 --> 00:07:42.079 opposed to what is the bill of material in terms 00:07:42.079 --> 00:07:45.959 of CapEx and OpEx. And what happened was a lot 00:07:45.959 --> 00:07:50.980 of these pilots start to show the clear and clear 00:07:50.980 --> 00:07:54.939 ROI for private 5G systems. And I'll give you 00:07:54.939 --> 00:07:59.199 an example. There was an auto manufacturer. that 00:07:59.199 --> 00:08:01.860 we worked with and we did a pilot for almost 00:08:01.860 --> 00:08:04.379 a year and a half with them and what happened 00:08:04.379 --> 00:08:08.480 was they said look we're going to try to do wi 00:08:08.480 --> 00:08:12.379 -fi at the same time just to see if the two things 00:08:12.379 --> 00:08:16.899 add up and compare with one another the result 00:08:16.899 --> 00:08:21.720 was netnet was that the wiring and cabling needed 00:08:21.720 --> 00:08:25.300 to do all those wi -fi access points and i'm 00:08:25.300 --> 00:08:30.040 talking about For north of a 1 million square 00:08:30.040 --> 00:08:35.500 feet of area, you need to put a lot of Wi -Fi 00:08:35.500 --> 00:08:37.639 access points. I don't care if you have Wi -Fi 00:08:37.639 --> 00:08:43.059 7, 8, 9, 10, you pick any of those. It's a lot, 00:08:43.240 --> 00:08:48.379 okay? And one of the big cost items was just 00:08:48.379 --> 00:08:53.220 the blue CAT 7 cables going to these access points. 00:08:53.720 --> 00:08:57.690 Very expensive. Not only just getting the power 00:08:57.690 --> 00:09:00.070 there, but just the cabling there. And you can 00:09:00.070 --> 00:09:05.169 get away with power by putting PoE systems, but 00:09:05.169 --> 00:09:10.610 it is very expensive. Very expensive. Now, you 00:09:10.610 --> 00:09:14.769 compare that to a 5G system, same area, same 00:09:14.769 --> 00:09:18.370 outdoor coverage, putting in five, six, maybe 00:09:18.370 --> 00:09:24.659 seven at top at most versus thousands. of Wi 00:09:24.659 --> 00:09:28.980 -Fi access points. And so not only from a cost 00:09:28.980 --> 00:09:31.759 perspective, you've got security ingress issues 00:09:31.759 --> 00:09:34.600 where you've got so many ingress points coming 00:09:34.600 --> 00:09:39.620 in versus five ingress points. So it became abundantly 00:09:39.620 --> 00:09:42.440 clear for many of our customers that, look, this 00:09:42.440 --> 00:09:45.899 makes sense. And we're talking use cases like 00:09:45.899 --> 00:09:50.059 we want to auto flash all cars. all at the same 00:09:50.059 --> 00:09:53.960 time. I need bandwidth. I need coverage. I need 00:09:53.960 --> 00:09:57.000 reliability. You get all of those attributes 00:09:57.000 --> 00:10:00.659 come in into a 5G network. It made so much sense 00:10:00.659 --> 00:10:03.480 for them to do this. Yeah. Yeah. The economics 00:10:03.480 --> 00:10:05.399 are compelling because you get to a point, right, 00:10:05.539 --> 00:10:08.240 when you were also dealing with outdoor, indoor, 00:10:08.460 --> 00:10:12.059 massive spaces where the economics just make 00:10:12.059 --> 00:10:14.830 more sense to your point. to have five or six 00:10:14.830 --> 00:10:16.830 radios and you've got your core infrastructure 00:10:16.830 --> 00:10:19.649 in place. And then the security, like I didn't 00:10:19.649 --> 00:10:21.330 really, you brought this point up and it's something 00:10:21.330 --> 00:10:23.309 that I didn't really think about, but securing 00:10:23.309 --> 00:10:26.190 that as well is a heck of a lot easier when you've 00:10:26.190 --> 00:10:29.210 got five or six radios versus thousands of APs. 00:10:29.230 --> 00:10:31.789 But I know Anshul's dying to ask you a question 00:10:31.789 --> 00:10:35.210 about, about edge, because there's a lot of focus 00:10:35.210 --> 00:10:38.629 on modern AI going hybrid and moving to network 00:10:38.629 --> 00:10:42.639 edges. He and I have speculated that you need 00:10:42.639 --> 00:10:46.240 the 5G low latency and high throughput to make 00:10:46.240 --> 00:10:48.399 that all a reality. But Anshul, I'll let you 00:10:48.399 --> 00:10:52.519 fire away. Yeah, I wanted to ask you about Edge 00:10:52.519 --> 00:10:55.460 AI, how compute is moving closer to the edge, 00:10:55.580 --> 00:11:00.019 but also how agentic AI paired with Edge AI is 00:11:00.019 --> 00:11:02.460 driving 5G networks and private 5G applications 00:11:02.460 --> 00:11:06.080 and what you're seeing. Yeah, Anshul, so we launched 00:11:06.080 --> 00:11:09.879 Edge AI just almost eight months ago now. The 00:11:09.879 --> 00:11:12.100 whole idea, let's take a step back. What is Edge 00:11:12.100 --> 00:11:16.059 AI? And so we're all familiar with ChatGPT. You've 00:11:16.059 --> 00:11:19.940 got OpenAI. You've got Gemini. You've got all 00:11:19.940 --> 00:11:24.279 from France. You've got a bunch of very large 00:11:24.279 --> 00:11:28.279 language models out there today that are, not 00:11:28.279 --> 00:11:31.720 to get too technical, but that are huge in terms 00:11:31.720 --> 00:11:35.220 of parameters, its supports, and its neural networks. 00:11:35.279 --> 00:11:40.029 So to give you some sense, ChatGPT 4 .0. uses 00:11:40.029 --> 00:11:43.389 north of 100 billion parameters in its neural 00:11:43.389 --> 00:11:46.110 network. So that's all the tokens you need to 00:11:46.110 --> 00:11:48.370 parse through a decision tree, right? So essentially, 00:11:48.610 --> 00:11:52.309 that's what a neural network is. If then else, 00:11:52.370 --> 00:11:55.850 most of the time, it's a probabilistic heuristic 00:11:55.850 --> 00:11:59.529 model where it literally guesses what, based 00:11:59.529 --> 00:12:02.090 on probabilities, what the next word might be. 00:12:02.330 --> 00:12:08.279 You don't need that for a edge. type use case 00:12:08.279 --> 00:12:11.220 in a manufacturing setting. So if you think about 00:12:11.220 --> 00:12:15.379 a robotic arm that uses a lot of kinematic decision 00:12:15.379 --> 00:12:18.940 trees, so it has to decide several different 00:12:18.940 --> 00:12:22.039 permutations of moving an arm between different 00:12:22.039 --> 00:12:25.539 parts of the conveyor belt. It doesn't require 00:12:25.539 --> 00:12:30.799 100 billion parameters. It requires, you know, 00:12:30.820 --> 00:12:34.980 maybe a hundred thousand parameters. and tokens 00:12:34.980 --> 00:12:38.580 to build a neural network to be able to support 00:12:38.580 --> 00:12:42.500 the use cases for that robotic arm. And more 00:12:42.500 --> 00:12:48.860 importantly, latency has to be almost zero. Very 00:12:48.860 --> 00:12:53.220 ultra low latency is the major requirement, right? 00:12:53.320 --> 00:12:56.539 So you don't have the luxury of going to the 00:12:56.539 --> 00:12:59.899 cloud where the massive large language model 00:12:59.899 --> 00:13:03.960 is and be able to decide on a particular decision. 00:13:04.279 --> 00:13:09.159 It has to be very quick. It uses linear programming 00:13:09.159 --> 00:13:11.980 models that are mathematical -based as opposed 00:13:11.980 --> 00:13:17.779 to language -based. So it's solving for maximizing 00:13:17.779 --> 00:13:21.539 a set of parameters and equations as opposed 00:13:21.539 --> 00:13:24.059 to figuring out what the next word might be. 00:13:24.340 --> 00:13:28.409 So it's fundamentally very different. And so 00:13:28.409 --> 00:13:34.269 that's the idea behind edge AI versus a, say, 00:13:34.289 --> 00:13:37.509 large language model. Now, in comes agentic AI. 00:13:38.649 --> 00:13:43.789 Agentic AI is mainly related to the workflows. 00:13:44.389 --> 00:13:47.009 And so what we're talking about, you may have 00:13:47.009 --> 00:13:50.250 a workflow around procurement, right, where you 00:13:50.250 --> 00:13:54.470 have to approve a set of POs or invoices, right? 00:13:54.549 --> 00:14:00.690 And so an agent. could essentially learn all 00:14:00.690 --> 00:14:04.730 of those particular business processes. And based 00:14:04.730 --> 00:14:08.230 on some rules, it will approve that PO or invoice 00:14:08.230 --> 00:14:12.250 to get through. So it's a very IT -centric model 00:14:12.250 --> 00:14:15.429 as opposed to edge AI. Again, think of edge AI 00:14:15.429 --> 00:14:18.970 as something that might be useful for a drone. 00:14:19.250 --> 00:14:23.110 And I was in Japan recently, looked at our chip 00:14:23.110 --> 00:14:26.750 set that we just recently came out. That is AI 00:14:26.750 --> 00:14:30.169 chipset, but that's not a GPU. It's a regular 00:14:30.169 --> 00:14:33.750 old CPU. You can run some of our edge AI models 00:14:33.750 --> 00:14:37.769 actually in a Raspberry Pi. Yeah. That's how 00:14:37.769 --> 00:14:40.169 small we're talking about. Yeah, you can inference 00:14:40.169 --> 00:14:42.970 on an x86. Intel's all about that, right? Yeah. 00:14:43.600 --> 00:14:46.740 100%. You can inference a lot of these things 00:14:46.740 --> 00:14:49.740 as long as they're small and mathematical -based 00:14:49.740 --> 00:14:52.299 as opposed to language -based because that requires 00:14:52.299 --> 00:14:55.779 massive processing. Right. And think of a drone, 00:14:55.860 --> 00:15:01.440 like an autonomous drone going inside a building 00:15:01.440 --> 00:15:04.639 where inferno burning, right? It's burning. It 00:15:04.639 --> 00:15:07.039 has to go from point A to point B and come back 00:15:07.039 --> 00:15:10.799 to point A. It has to know how to navigate through 00:15:10.799 --> 00:15:15.350 and learn. on the fly, no pun intended, as it's 00:15:15.350 --> 00:15:18.230 going through and being able to tackle smoke, 00:15:18.470 --> 00:15:22.269 fire, other debris, and be able to come back 00:15:22.269 --> 00:15:25.570 and report in real time what's happening inside. 00:15:25.870 --> 00:15:31.009 And so that requires an AI model. It's got to 00:15:31.009 --> 00:15:35.129 fly the plane, right? That's learning and flying 00:15:35.129 --> 00:15:41.129 at the same time. That's the mindset or the overall. 00:15:42.579 --> 00:15:44.799 architecture we're talking about when we talk 00:15:44.799 --> 00:15:48.000 about edge AI. And I think to me, that is still 00:15:48.000 --> 00:15:53.320 a market and a technology yet to be completely 00:15:53.320 --> 00:15:58.080 understood. I think there's some progress being 00:15:58.080 --> 00:16:01.200 made. We're making some good progress with lighting 00:16:01.200 --> 00:16:04.759 these up for some of our same P5G customers. 00:16:05.059 --> 00:16:10.820 But I think the likes of the big guys like... 00:16:10.990 --> 00:16:14.769 OpenAI, Microsoft, I think once they get into 00:16:14.769 --> 00:16:17.450 this will blow up, much like Agentech AI has 00:16:17.450 --> 00:16:19.429 over the last few months. It's just going to 00:16:19.429 --> 00:16:21.870 take time and proof of concepts have to get out 00:16:21.870 --> 00:16:24.509 into the wild. And you've got to drive awareness 00:16:24.509 --> 00:16:27.450 for that. And I know, Shai, that NTT is focused 00:16:27.450 --> 00:16:30.370 on this concept of factory of the future. where 00:16:30.370 --> 00:16:32.210 you're taking what you're doing with Edge AI, 00:16:32.490 --> 00:16:35.029 you're utilizing private 5G, right? And you're 00:16:35.029 --> 00:16:36.750 working with some of the largest companies on 00:16:36.750 --> 00:16:38.830 the planet. I've spent time with you and the 00:16:38.830 --> 00:16:41.509 team at a Schneider facility, Schneider Electric 00:16:41.509 --> 00:16:45.090 facility in Barcelona a few years ago at Mobile 00:16:45.090 --> 00:16:48.909 World Congress. Lionel Basil, which is, I believe 00:16:48.909 --> 00:16:51.490 they manufacture trains, right? Can you speak 00:16:51.490 --> 00:16:54.649 to maybe Schneider and what you're doing there 00:16:54.649 --> 00:16:57.549 to enable this whole notion of factory of the 00:16:57.549 --> 00:17:01.070 future? Yeah, yeah. No, what is a factory of 00:17:01.070 --> 00:17:03.190 the future? And we've been hearing that quite 00:17:03.190 --> 00:17:06.230 a bit lately, especially here in the US where 00:17:06.230 --> 00:17:08.910 the administration is trying to get all the manufacturing 00:17:08.910 --> 00:17:12.309 back in the US. And yet we don't have the labor 00:17:12.309 --> 00:17:15.750 workforce to work on it. And so if you think 00:17:15.750 --> 00:17:19.990 about that, what does that factory look like? 00:17:20.109 --> 00:17:23.650 I think these are mostly dark factories where 00:17:23.650 --> 00:17:28.230 you'll have most of the work. done will be done 00:17:28.230 --> 00:17:31.589 through robots. And automation is going to be 00:17:31.589 --> 00:17:34.690 critical. And actually, if you want to take a 00:17:34.690 --> 00:17:38.690 peek of what that world will look like, you just 00:17:38.690 --> 00:17:41.849 have to go to China. And there's a lot of movement 00:17:41.849 --> 00:17:44.829 and progress being made in terms of what that 00:17:44.829 --> 00:17:47.730 factory of the future looks like. A lot of dark 00:17:47.730 --> 00:17:50.910 factories, a lot of robotic arms, a lot of automation. 00:17:51.589 --> 00:17:54.640 It's amazing. If you have an opportunity, well, 00:17:54.720 --> 00:17:57.339 check it out. Actually, you can even go to Taiwan 00:17:57.339 --> 00:18:00.039 or Philippines, and they also have been doing 00:18:00.039 --> 00:18:02.839 that. If China is not an easy place to get to 00:18:02.839 --> 00:18:06.859 these days. If you think about what that world 00:18:06.859 --> 00:18:09.619 looks like, connectivity becomes a paramount 00:18:09.619 --> 00:18:14.140 capability. And I don't mean this like... just 00:18:14.140 --> 00:18:18.319 a private 5G network, but it's the whole fabric 00:18:18.319 --> 00:18:22.099 there from the wired infrastructure to the wireless 00:18:22.099 --> 00:18:26.940 infrastructure. It is so important that without 00:18:26.940 --> 00:18:30.599 it, nothing happens. You could potentially have 00:18:30.599 --> 00:18:33.240 the whole factory shut down just because of a 00:18:33.240 --> 00:18:36.660 routing table that wasn't properly configured. 00:18:36.900 --> 00:18:41.210 And that's something... we are very focused on. 00:18:41.309 --> 00:18:46.109 We think as manufacturing moves to the US or 00:18:46.109 --> 00:18:49.230 other parts of the world, where they double down 00:18:49.230 --> 00:18:54.930 on manufacturing capabilities, where labor workforce 00:18:54.930 --> 00:18:59.250 is becoming, they're aging out and re -skilling 00:18:59.250 --> 00:19:01.369 them becomes a very challenging proposition. 00:19:01.730 --> 00:19:04.910 And so retooling is where we think industrial 00:19:04.910 --> 00:19:07.970 companies, manufacturing companies will look 00:19:07.970 --> 00:19:11.789 at. Yeah, I think Anshul's got a final question 00:19:11.789 --> 00:19:14.250 for you around just access to all this technology, 00:19:14.529 --> 00:19:18.309 right? Because it's a high bar. And to your point, 00:19:18.509 --> 00:19:22.140 a lot of what's driving. private cellular and 00:19:22.140 --> 00:19:26.839 private 5G is moving manufacturing offshore back 00:19:26.839 --> 00:19:29.440 into the U .S., right? There are huge investments 00:19:29.440 --> 00:19:32.460 in semiconductor production now. Intel's opening 00:19:32.460 --> 00:19:35.420 a big fab. They're launching a foundry business, 00:19:35.720 --> 00:19:39.019 right? TSMC is investing in the United States 00:19:39.019 --> 00:19:42.059 as well and in places like Arizona and others. 00:19:42.279 --> 00:19:44.539 But it comes down to democratization. Yeah, this 00:19:44.539 --> 00:19:47.660 is a twofer. I wanted to see if you could maybe... 00:19:47.960 --> 00:19:50.220 Tell us a little bit about the convergence of 00:19:50.220 --> 00:19:51.940 a lot of these wireless technologies and how 00:19:51.940 --> 00:19:54.720 that's driving future applications, but also 00:19:54.720 --> 00:19:57.660 real -time computing that takes advantage of 00:19:57.660 --> 00:20:00.380 the convergence of these for secure communications. 00:20:01.799 --> 00:20:04.180 Yeah, that's a great question, Anshul. Actually, 00:20:04.200 --> 00:20:06.299 one of the most interesting things we're seeing 00:20:06.299 --> 00:20:11.420 is where 5G and Wi -Fi coexist on the same radio. 00:20:11.619 --> 00:20:15.160 And I don't mean just two different radios, but 00:20:15.160 --> 00:20:18.460 they're actually integrated. To me, that convergence 00:20:18.460 --> 00:20:23.539 of 5G, private 5G cellular and Wi -Fi, think 00:20:23.539 --> 00:20:27.220 about that in a warehouse setting or factory 00:20:27.220 --> 00:20:31.319 setting or even a hospital setting. You can roam, 00:20:32.819 --> 00:20:35.819 if you will, or ping pong between the two networks, 00:20:36.039 --> 00:20:40.400 depending on how you have configured it. I think 00:20:40.400 --> 00:20:42.400 to me, that's going to be very powerful. And 00:20:42.400 --> 00:20:44.539 I'm really looking forward to that convergence. 00:20:44.880 --> 00:20:48.220 And we're beginning to see some of the OEMs now 00:20:48.220 --> 00:20:52.339 really start to launch those kinds of capabilities 00:20:52.339 --> 00:20:55.880 and products. Stay tuned for that. And more broadly 00:20:55.880 --> 00:20:58.819 speaking, I think from a security standpoint, 00:20:59.339 --> 00:21:03.460 we are working with many of our customers. 5G 00:21:03.460 --> 00:21:06.940 for us, private 5G for us is a tip of the spear. 00:21:07.160 --> 00:21:10.839 As we go to our large, very large customers, 00:21:11.160 --> 00:21:13.700 it opens up all kinds of other opportunities 00:21:13.700 --> 00:21:17.700 for us. So when we walk in there, typically we 00:21:17.700 --> 00:21:21.500 go with private 5G, we end up with transforming 00:21:21.500 --> 00:21:24.569 their whole network architecture. And a lot of 00:21:24.569 --> 00:21:28.390 it has to do with security. But because as you 00:21:28.390 --> 00:21:32.150 start to increase your automation, there's a 00:21:32.150 --> 00:21:35.950 lot more security challenges associated with 00:21:35.950 --> 00:21:39.029 that, right? You open up OT devices that have 00:21:39.029 --> 00:21:42.089 never been opened up or haven't been opened up 00:21:42.089 --> 00:21:46.009 in decades. Now you're opening that up too. And 00:21:46.009 --> 00:21:49.069 so, yeah, I think where security is going to 00:21:49.069 --> 00:21:52.019 be a big part of this, a lot of our A lot of 00:21:52.019 --> 00:21:55.140 our network architecture work that we do starts 00:21:55.140 --> 00:21:58.640 with that. And it's just, I don't think anybody 00:21:58.640 --> 00:22:03.660 has the magic wand to just fix that in one shot. 00:22:03.700 --> 00:22:06.539 I think this is an ongoing challenge that's always 00:22:06.539 --> 00:22:10.140 going to be around. Yeah, I upgrade this year. 00:22:10.299 --> 00:22:13.440 I learned that NTT is the third largest provider 00:22:13.440 --> 00:22:16.859 of data center services. So your point around 00:22:16.859 --> 00:22:20.880 P5G being the tip of the spear. And I have to 00:22:20.880 --> 00:22:23.420 educate a lot of people that NTT is not just 00:22:23.420 --> 00:22:26.480 an integrator as well. You've got this bevy of 00:22:26.480 --> 00:22:29.279 solutions. You're very deep. You're very broad. 00:22:29.359 --> 00:22:32.029 You're working with companies like Cisco. and 00:22:32.029 --> 00:22:35.210 others. And it's been really interesting to see 00:22:35.210 --> 00:22:37.309 the evolution of the company. And there's been 00:22:37.309 --> 00:22:39.990 some consolidation as well. I think there was 00:22:39.990 --> 00:22:42.569 the recent news around the operating company 00:22:42.569 --> 00:22:45.809 and some consolidation there, which I think has 00:22:45.809 --> 00:22:48.410 driven a little bit of, I don't want to say confusion, 00:22:48.430 --> 00:22:52.529 but misunderstanding of what NTT does on a global 00:22:52.529 --> 00:22:55.329 basis as well. As we wrap things up, do you want 00:22:55.329 --> 00:22:58.700 to comment on that point? Yeah, yeah, absolutely. 00:22:59.099 --> 00:23:01.700 Let me first comment on Cisco. We're a longstanding 00:23:01.700 --> 00:23:04.660 partner. We're one of the largest partners we 00:23:04.660 --> 00:23:07.460 have. In fact, one of the most important things 00:23:07.460 --> 00:23:11.579 we did recently with them was embedded our eSIM, 00:23:11.720 --> 00:23:15.279 Transatel eSIM, into their Moroccan Catalyst 00:23:15.279 --> 00:23:18.700 boxes. So that if you're familiar with Transatel, 00:23:18.740 --> 00:23:22.400 it's one of the largest MBNOs in the world. We 00:23:22.400 --> 00:23:25.059 have connectivity in 200 different countries. 00:23:25.740 --> 00:23:29.619 Think about it. We put that eSIM in every catalyst, 00:23:29.940 --> 00:23:34.819 every Meraki box, enabling the enterprises to 00:23:34.819 --> 00:23:37.980 be able to provision, configure very quickly. 00:23:38.059 --> 00:23:42.440 You open up the Meraki router, right? And you 00:23:42.440 --> 00:23:45.440 go to Meraki .com. Immediately, you're connected 00:23:45.440 --> 00:23:49.460 without figuring out, hey, I need to have connectivity. 00:23:49.759 --> 00:23:53.240 So you can download the configuration and be 00:23:53.240 --> 00:23:56.380 done. in like less than five minutes. Powerful, 00:23:56.500 --> 00:23:59.619 yeah. Very powerful. And I think you think about 00:23:59.619 --> 00:24:02.539 a deployment where you have thousands of Meraki 00:24:02.539 --> 00:24:05.319 or Catalyst devices, very easy to provision. 00:24:05.500 --> 00:24:07.619 Once you're done with that, and we're giving 00:24:07.619 --> 00:24:11.119 away that for first 60 days, be able to connect 00:24:11.119 --> 00:24:14.099 right away as a bootstrap. And then after that, 00:24:14.220 --> 00:24:17.000 the customer can sign up for a longer term as 00:24:17.000 --> 00:24:20.900 a backup solution. Yeah, we... We do a lot of 00:24:20.900 --> 00:24:25.759 things. We've got a lot of stuff. And so, yeah, 00:24:25.859 --> 00:24:29.000 I think bringing all of our companies together 00:24:29.000 --> 00:24:32.200 as one, I think that made so much sense. Yeah. 00:24:32.420 --> 00:24:36.519 And I think hopefully we simplify for our customers 00:24:36.519 --> 00:24:41.059 all of our offerings all in one. Yeah. Ashad, 00:24:41.059 --> 00:24:43.380 you're a busy guy. Thank you for taking the time 00:24:43.380 --> 00:24:46.819 to speak with Anshul and I about what NTT is 00:24:46.819 --> 00:24:49.539 doing, the innovation that you're driving, not 00:24:49.539 --> 00:24:51.799 only with private 5G, but even beyond that, the 00:24:51.799 --> 00:24:54.279 data center and that sort of thing. And it's 00:24:54.279 --> 00:24:56.140 been a while since we've done a standalone edition 00:24:56.140 --> 00:24:58.599 of our podcast. So thanks again for being on. 00:24:58.660 --> 00:25:02.019 We really appreciate it. Yeah. Thank you. And 00:25:02.019 --> 00:25:04.240 thanks for having me today. Appreciate it. Thanks 00:25:04.240 --> 00:25:06.619 a lot. Yeah. Thanks, everybody. Take care.