Service over self: How an ex-submarine captain is tackling America's Digital Divide

On this episode of Connected Nation, we talk with Tony Grayson – a man who has held leadership roles at such tech giants as Amazon, Facebook, and Oracle – roles that are perhaps eclipsed by his time as a submarine captain.

We’ll explore his new approach to closing the Digital Divide in America; look at the  challenges of providing resources for growing needs of AI; and share how the skills of our nation’s veterans can be better understood within and applied to high-tech jobs in the civilian workplace.

Related links:
Tony Grayson’s LinkedIn - NOTE: Veterans, active military, and transitioning military are invited to connect with Tony directly.
Compass Datacenters

Chapters

• 2:06 Tony Grayson background
• 6:00 Life in a submarine
• 13:25 The trajectory of Tony's career
• 16:38 Transitioning into the civilian world
• 19:02 Tony's opinion on leadership
• 27:38 Compass Data Centers does versus Quantum's
• 35:57 The future of AI
• 43:00 The benefits of rural broadband expansion

Transcript

Jessica Denson (00:05):
This is Connected Nation, an award-winning podcast focused on all things broadband from closing the digital divide to improving your internet speeds. We talk technology topics that impact all of us, our families, and our neighborhoods. On today's podcast, I talk with Tony Grayson, a man who has held leadership roles at such tech giants as Amazon, Facebook, and Oracle. Roles that I would argue are eclipsed by his time as a submarine captain. We'll explore his new approach to closing the digital divide in America. Look at the challenges of providing resources for growing needs of ai, and share how the skills of our nation's veterans can be better understood within and applied to high tech jobs in the civilian workplace. I'm Jessica Denson and this is Connected Nation. I am Jessica Denson, and today my guest is Tony Grayson, general manager with Encompass Data Centers overseeing Quantum, a subsidiary of the company. Welcome Tony.

Tony Grayson (01:08):
Hey, thank you for having me, Jessica.

Jessica Denson (01:10):
I'm excited to talk to you. My CEO actually jumped in before we got started to say a little hello. He's a big fan of yours. I appreciate that.

Tony Grayson (01:20):
No, definitely a big fan of what Tom is doing and kind of bringing broadband to the rural areas to really to fix the digital vibe. It's something we all should be looking to try to do.

Jessica Denson (01:32):
Yeah, he really called you, he said that you were really shaking things up and making things interesting. So I'm excited to talk to you today. I really appreciate you joining us.

Tony Grayson (01:41):
No, no, thank you for having me.

Jessica Denson (01:43):
Okay, let's start. I really like to give our audience an idea of who we're talking about and who we're talking to, a little bit of your background, that kind of thing. So let's start, I know you have a deep history in the military and you attended the prestigious US Naval Academy. Talk about what led you to serve and your time within the Navy.

Tony Grayson (02:06):
Yeah, sure. So I grew up here in stores of World War II from my grandfather who was a Catalina pilot, flew with a black cat Raiders and these were basically scout planes that flew a miles an hour, kind of the jee and went after Japanese shipping, but he kind of taught me the whole thing about service above self. And he was actually shot down and spent three weeks in a raft in the middle of the Pacific and washed up on an island. So, oh wow. He definitely had a very unique story and if you ever look at some interesting books about the black hat Raiders, but they basically painted 'em all black, but he also was in Pearl Harbor and lost everything and actually had to take his plan out there. So I basically heard growing up like that and I know it instilled in me that service above self and I kind of really wanted to do that. Now, granted I expected only serve five years in the military, ended up spent roughly 21 years, but I'm very, very happy I did. It made a good foundation for tech and I feel like I was able to give back.

Jessica Denson (03:17):
As I mentioned. Just briefly actually, I want to go back. Hearing those kinds of stories from your grandfather, did it kind of give some mythology to the idea of serving too? Just the idea of being on a raft and then ending up on an island and being part of some of our historic battles and that type of thing that just had to add some sort of mythology to that?

Tony Grayson (03:43):
Yeah, I think you join that history of wanting to do it, whether it's the Marine Corps or the Air Force or the Army. I think everyone has their unique history. I think what drove me to submarines or to the Navy really was the opportunity to go on submarines, and I really wanted to do that because it's kind of like the last vestige of real command that's out there. And by that I mean by no means saying that aircraft wing commander's not in charge of his wing or a squadron is not in charge of their squadron or a ship cap is not in charge of their ship, but when you're on a submarine, you submerge and that's it. You can't call for help. You can't call for backup. In today's kind of connected world, your boss is always there and can tell you what to do or provide you feedback. Even in a shooting war on submarine, it's not like that. And if you get caught your front page news across the world. So that's kind of really what I wanted to join the Navy for and really what I wanted to join the Naval Academy for where really to get out and do that.

Jessica Denson (04:49):
The thought of getting on a Marine makes my claustrophobia it, I get chills all over, just even thinking about it.

Tony Grayson (04:57):
Are really bigger than people think. Not real big, but they're definitely bigger than people think. But you keep pretty busy. And I love the fact it's an all volunteer force and it's just a very, very tight knit team. Almost it'd be like 120, 130 people and you live on top of each other day to day really. It makes a unit much different. That's a little bit different than what else you have out there in the service. You just can't escape each other for better or worse, for better or worse.

Jessica Denson (05:28):
You said it's an all volunteer. So if you're on a submarine, you've volunteered for that type of service.

Tony Grayson (05:37):
You have to be a volunteer to go on submarines. If you don't want to go on submarines, then you don't have to stay on submarines. You can go do a service ship, you can go do aviation, you could do a bunch of other things. But on submarines it's all voluntary. You have to be.

Jessica Denson (05:49):
Oh, wow, I had no idea. That's interesting. There really isn't anything else in the services. The armed services like that is there where you have to volunteer to be a part of it.

Tony Grayson (06:00):
You have to kind of be okay with submerging and going under the water and it's one of those things that hey, it's 130 people and if you don't come up, you all don't come up. And so it's really, everyone has won that big group and you all survive or succeed as one. And so the worst thing you could do is have someone that gets claustrophobic and you have to surface and get them off the boat or someone has a medical problem, you have to get 'em off the boat. That's kind of why they're volunteers because it is a very unique environment. In fact, some of these submarines don't even have enough beds for people. So you have to be okay with basically getting in a rack behind someone else. They call it hot racking and dealing with their stinkiness. As you kind of go over there, we know we try to limit that stuff, but there is a good 15, 20 people could be that are doing that hot racking.

Jessica Denson (06:57):
Wow. I bet you have stories. I could just talk to you about that forever.

Tony Grayson (07:02):
Definitely a lot of stories and I wish I could, I mean you would be amazed at what we're doing out there. I mean it's all stuff. I don't know if you've ever read the book Blind Man's Bluff. It kind of goes into the secret life of submarines and all the sketchy stuff that we do, but we're still doing that stuff. It's something that you just never know that happens. There's not blind man's bluff.

Jessica Denson (07:23):
I'll make sure and put a link to that so people can look into it. I'm going to have to get it because it's fascinating. I didn't even know that it was a volunteer. I didn't know. So it must really just be fascinating work. As I mentioned, the Naval Academy is a prestigious institution and you attended that. What was it like to be a part of that and do you have fond memories of your time there or was it just really tough?

Tony Grayson (07:50):
Well, during my first year when you're a plebe and you get nothing and you get yelled at all the time and you do pushups and stuff like that, I regretted my choice of not going to MIT or UVA or something like that. But I think it's one of those things that it's one places that it's a really good place to be from and it kind of gave me the centering that I probably needed in college. My dad was, you've seen the movie like Animal House. He was that person who had a 0.0 and I was afraid I might have been that same way, but you really didn't have that much choice in the Naval Academy. You couldn't really do that. But it kind of gave me my centering. It kind of made me mature a lot quicker. So when I actually graduated, I think I was probably definitely ahead of what I would've been doing some rothi job at some regular college and just kind of got me into it and I knew what to expect more. You kind of lived and breathed the Navy, and a lot of people think the Naval Academy is like this. It's not a real college. It's one of the top engineering in the country right now. And they have everything from a nuclear reactor on campus to really good engineering portions. They also have really good humanities portions too. And to be honest, it's really not a bad place. If you look at West Point in the army, middle of nowhere, air Force, middle of nowhere, Naval Academy, downtown Annapolis, it is

(09:15)
Not granted. We didn't have air conditioned then, so we actually turned on the heat. If you could imagine during Annapolis summer when it's a hundred percent humidity, so now they have air conditioned, but we didn't have all that when I was there. So it was kind of like this fifties, sixties place and buildings that you kind of got to live with. But it had this history that you could really trace back and look at pictures and see who came before you. And they have, if you ever get a chance to go in there and they have old battle, the battle flags from the US S Constitution, you could see all the scorch marks from the candidates. It's just that history. It's really tough to compete with that.

Jessica Denson (10:00):
Yeah, I can't imagine going to a place that has such a storied history. And then is that traditionally where a lot of naval officers come from is from the Naval Academy or are there different paths to that direction?

Tony Grayson (10:16):
There's definitely different paths. I think it's a small number. I mean, I think we graduated seven or 800 per year that are going into the service. I mean there's other things like basically you graduate from college and you want go and become an officer, you go to Officer candidate school or you are supply and you go to called Fork and Knife School where they teach you to march and where the dentists are and stuff like that. And then they have Roxy, which is you kind of wear uniforms at a regular college and do some drill. And then they had this thing that is probably the greatest thing in the military, which is called och. And it's specifically for people going in Navy nuclear where they pay for your college starting your sophomore year. You're also given the pay as an E six or an E seven and accounts to your retired time. So it's like three years of actually living pretty large at a college and not doing anything military wise and then coming out and doing what I did the post time. So it really is the best thing out there that kind of, there's one thing I might've done, different Naval Academy. It's och ak.

Jessica Denson (11:26):
So what does PAC stand for?

Tony Grayson (11:28):
That's a great question.

Jessica Denson (11:31):
Always, we can look it up, but I'll put it there for you.

Tony Grayson (11:33):
It's those people whenever you talk to PAC and Lucky Bastard. Oh, weird. It's Nuclear Propulsion Officer Candidate program.

Jessica Denson (11:44):
Okay, gotcha. And when you say stuff like E six, E seven, that's a level that someone's in?

Tony Grayson (11:50):
Yeah, yeah. There's like O, which is Officer E, which is enlisted E six is like a first class and an E seven is a chief.

Jessica Denson (11:58):
So as though they're already doing time or working within the military and building that time up even though they're got it.

Tony Grayson (12:05):
So where the Naval Academy, they like to tell you Canada towards your retirement. It doesn't but PAC does.

Jessica Denson (12:12):
Gotcha. And so while you were at the Naval Academy, you got a control systems engineering degree, correct. Bachelor of Science. Yep. What made you love that? Why did you choose engineering to go that direction?

Tony Grayson (12:26):
I guess I'm a big nerd and what I liked about control systems, I wasn't sure if I wanted to do MEKE or double E and control systems is really the blend of MEKE and double E and you can go and go into any field once you did that. And I think it prepared me well for nuclear power school afterwards where you kind of get basically a master's in nuclear engineering and you jam it in for a year and a half. I think it prepared me for that, but it also gave me my lifelong love of computers and hardware, which it serves me well right now at my current job.

Jessica Denson (13:03):
So before we go into what you're doing now, I still want to ask a little more on your background. Just it's so fascinating from the Naval Academy, that's where your naval career took off, correct? Yes. You were a naval officer. We talked a little bit about being a submarine captain. What was the trajectory of your career? You said you thought five years and it was 21. 

Tony Grayson (13:25):
I thought five years. And so I did my junior officer tour out here in Seattle where I'm at right now on USS Nevada Epi Ballistic Missile submarine. And I had talked to my wife and they give you a bonus to keep you in for that junior officer, which is your first year short tour. So we said, Hey, let's just stay in for that. And I was an aide for Marine Corps one star down in Miami during that time and I worked with that person. His name was General Huck, but he taught me a lot and it really kind of got me involved in day to day and kind of what the military was doing. So I remember looking out towards the end of that and I was like, Hey, how about we just give this department head thing a try and if we don't like it then we'll get out then.

(14:13)
And so it just never dropped from there. So from there I went up to Connecticut and was Webs on USS Connecticut. And then I was fortunate enough to be a detailer in Millington, Tennessee, which is where the Navy personnel command is. And I did my EXO tour back out here on a gn and then I was lucky enough to go out to Bahrain and was executive assistant to Fifth Fleet and then came back from my CO tour out of Groton on USS Providence. So very fortunate tour. I met some great people and some great friends along the way that I really still keep up with.

Jessica Denson (14:49):
So during that time, technology has been something that wove through all of that because submarines now are not like what you'd think you'd see in Hunt for at October or other Hollywood movies. They're much more high tech, right? Yeah.

Tony Grayson (15:07):
And so my J tour, I felt like looking back, it felt like the boat was like a pong boat, this whole analog green screen thing. And then you look at a Virginia class, which was one of the last boats I took, the Periscope deck, it looks like the Starship Enterprise, I mean, oh cool. We're using a lot of computer processing, they're using a lot of containers to do that processing. I think there's like 14, 15 megawatts of it load on it. It is a very, very different ship than it was when I first went on. And that's what people don't really realize that you kind of on a submarine if you get into the networking and that the computers, it really is kind of like a cloud. You understand all the layers of how cloud works, everything from the base infrastructure to the platform to the individual programs and the software layer that runs on top of it. So if you really get into that and you nerd out like I tended to do, then you get into that and you kind of really understand how cloud works. And that's kind of what got me into Facebook. I was originally supposed to go do facilities in Dublin because my wife and I have been trying to go to Europe forever, but I ended up going to San Francisco and was doing more towards the networking platform portion. 

Jessica Denson (16:26):
Yeah, your resume is really a who's who in the tech space, Amazon, Facebook, Oracle. Tell me what it's like making that step into the civilian corridor. How did that feel?

Tony Grayson (16:38):
I tell you it's not at all what I thought it was and this is why I try to do as much as I can to give back. I mean we're always told that the Navy or the military in general prepares you well for the civilian community and it really doesn't. I mean, to be honest, leadership's different finances are different. Everything is just very different. And I think the people that have the most problems transferring from the military to the civilian world are people that treat it like the military. You can't in the end, you can always tell someone what to do, you can yell at 'em and you could throw stuff at 'em and nothing's going to happen. You do that in the civilian world and you're going to be written up by hr. It just takes a lot more to motivate people. In the Navy, you definitely have people from different backgrounds, but everyone tends to think the same and you're all there for the similar reason.

(17:29)
It's not like that in the civilian world. And I think I was very, very fortunate where I had an incredible boss at Facebook who really at the time says, you can't do anything for three to four months. And I was pissed at the time. I was like, oh, I can do this kind of stuff. But it really was, it taught me to really sit back and understand why people do it that way, what they think, how you learn. And I consider myself very fortunate in that. So that's kind of one of the things I try to pass on to people is you cannot think that you read some leadership book or some maybe book and think that applies to outside. It really, really does it. And the people who do that or fall back on that, you will have some issues and you just have to realize that you are learning something new. And we pretty much get no street cred probably past about 12 years in the Navy because that's kind of what the equivalency is after that. It's just very, very different. And I think people really need to understand that when they get out.

Jessica Denson (18:32):
So let's talk a little bit about leadership on LinkedIn. When I was doing your research named one of the top leaders on top leadership voice is what it was said, LinkedIn top 1%. Do you do a lot where you help others make this transition? How things are different and the same, that type of thing? What is it that, how did you earn that moniker?

Tony Grayson (19:02):
I think to be honest, I think you can't read a leadership book at all and learn leadership. I think you have to be honest. I think you have to fail and I think you have to be hit that ditch where you can get up and walk away from that ditch and learn something. If you hit that pole, then that's not so good to get to find that balance of learning continually. But I have a lot of scars both in my navy time, in my post Navy time that I don't want people to learn from my mistakes and continue to grow and from that. So I tend to write a lot about that and try to be very open about where I messed up and what I learned from it and how to lead truly lead without ego in putting your team first. I think a lot of it's just saying it.

(19:50)
I think we hear the term servant leadership, I think very view practice that or even really know what it means because I hate to say it, we all tend to put self first and I think the military gets that out of you, but it doesn't necessarily get out of you that much. And I think in times of crisis or times of pressure, it's just human nature to defend yourself first. And I think that it's very dangerous to do that. So I try to really get up there and make sure people really understand it. And just because you read some book about someone who succeeded doesn't mean you're going to be successful either. And it might work for that person, but everyone's work environment is very different and you can't just read a book, write some notes and expect it to do it like that. You really have to develop your leadership for the company you're in, the culture that you're in and your team that you have.

(20:42)
And it's going to be the same. It's going to be different day to day, month to month, year to year. Your teams are always changing, the culture is somewhat changing as you grow, but definitely the goals of what you're trying to do change. So it's a moving target and leadership, you have to treat it like that. And yeah, I just have a huge problem with, and I have no problem with leadership coaches and tell 'em that kind of stuff, but they need to recognize that that's going to give you found the tools to be successful, but you're going to have to implement a lot of tools to be successful and what made you successful at Facebook would not make you successful at AWS because they're very, very different companies and how they worked was very different.

Jessica Denson (21:26):
That's interesting that you can apply that and I think that's a sign of good leadership being able to recognize that mistakes are part of learning, don't you think?

Tony Grayson (21:35):
Yeah. But I had to make a lot of mistakes to get to that. Not proud of everything I've done and how I got there, but I think it makes me, at least at this point in time, a much better person that I feel like I can finally give back a lot to the military and they can take it or leave it on what I try to say. But even if I can just help that one person and help that transition or help someone go into the Navy doing a lot now with Navy recruiting where every quarter I'm taking tech leaders out to ships and submarines and showing them what these on a carrier, 18, 19-year-old kids are launching the aircraft and recovering aircraft. I think it's eyeopening for tech leaders to see that. And these are the people that we aren't hiring because their resume doesn't say X, Y, or Z when these people could be the best pre-build, hire, you really can teach academic stuff. You can't teach will, you can't teach loyalty and you can't teach that kind of desire to push. It's just something you have or you don't.

Jessica Denson (22:40):
Yeah, and I would think in the military you would have to have that self starter in you to work as a team too. Just the idea that living in this environment you have to live by certain rules, but you have to also, I mean I'm pontificating obviously I've never been in the military, but I would think that it takes someone who's very committed to an idea. And so translating that to a company can be very powerful I would think.

Tony Grayson (23:09):
No, it is, it is. And you got to remember the military doesn't pay that much. I mean they junior military that have families that are literally on food stamps that are defending our country and living in rough conditions to keep us free. I think we tend to forget about that these are the people we don't want to hire because their resume doesn't say X, Y, or Z, even though they could be exactly what the company wants. And that's what I'm trying to get people to understand is just because someone says they don't have that they've dealt with computers or design cloud or whatever, I can guarantee you could probably take any IT tech out of the military and they do a very good job for you and you have to read into the resume and what they probably grew up with. And so I think it's an education on both sides.

Jessica Denson (23:57):
So is that part of what you're doing with your role as veterans chair at Infrastructure Masons?

Tony Grayson (24:03):
Yeah, it's that and it's more IT is trying to get apprenticeships programs stood up for people who are enlisted and go to college and sign up for the reserves that they can get internships that are set up for the Navy. It's kind of how do you still keep our nations strong by keeping the best and brightest goal in the military, but how do you enable them to be successful when they transition out? This you can't treat it like you have to keep everyone in the military. People are going to get out no matter what you do. And so you have to make sure they're prepared for when they leave. And so that's what we're trying to do in granted, I don't know anything about financial services or stuff like that. I am specifically just talk in a tech field, but I do think the tech field is where we could do the best.

(24:48)
And to be honest, I think we get pigeonholed a lot into, we can only take orders March turn weds basically hook up electrical circuits, that's it. That we're great program managers, we're great ops, we're great engineering. We just don't know we can do that because we do it all in the military. So we are definitely not one discipline. We're taught a lot of disciplines and the most interesting thing about the Navy or the military in general that I think people forget about is we change job every two years. So if you can imagine you get dropped to in a completely different job with a completely different authority, with a completely different remit every two years and you have to pick it up really, really quick or you're going to not be successful. And so that's how we're able to really kind of get on the governor. We just have to give us maybe a little more time, maybe three to four months to like I would got at Facebook, then learn what's going on and then we can really can contribute. But unfortunately I don't think we're offered that much time. I think we're judged from day one.

Jessica Denson (25:49):
And that's a shame because obviously you're able to, if you're changing jobs every two years, you're able to adjust and make things work. And so it's really just helping the civilian population understand how those positions translate into the economy. Correct.

Tony Grayson (26:10):
Exactly. But it's also helping the people leaving. It doesn't matter if they're off.

Jessica Denson (26:14):
Understand as well.

Tony Grayson (26:15):
Yeah, understand as well where they fit in because it's going against the bias that some senior person thinks they can walk in and be some C-level exec at some Fortune 500 company. They're making the understanding that you might be a manager of in the military of 10,000 people, but it's okay to be just an individual contributor in the civilian world because in the military, as you get more up in the ranks, you lead more and more people. That's not equated you have in the civilian world. You can make some of the boast impact by not managing anyone and just really focused on the business and helping the business get ahead. And all of those are really, really tough things for the military to learn. What you've done for the last 20, 30 years is you're taught this promotion and how to be successful and it's just different on the outside.

Jessica Denson (27:11):
Yeah, that's an interesting and important thing to take on, a challenge to take on to really help on both sides, the understanding of that. Let's shift a little bit now to what you're doing currently and you're working under the umbrella of Compass Data Centers running Quantum. Correct. Can you talk a little bit about what Compass Data Centers does versus Quantum's role?

Tony Grayson (27:38):
Yeah, and so Compass Data Centers is really focused on building data centers for the bigs of the industry, building 500 megawatt plus campuses and enabling the clouds to be successful by giving them space when they need it. And so it's tough for the clouds to catch up with that because they're focused on a lot on the platform. But to build a data center takes a long time. It takes, you have to find the land, you have to buy the land, you have to get everything permitted and the clouds want 'em basically now, now, now, now. And so they, they're much better set up to choose from the companies that are out there and lease those data centers and the other company runs and they focus on what they're good at the facility side and the cloud focus on what they're good at, which is the IT side and up with the platform aspect to it. So that's what Compass has done is really to help those companies be successful and deploy anywhere.

Jessica Denson (28:39):
And Quantum, it's a smaller version of that, correct?

Tony Grayson (28:45):
Yeah, it is. It's really going after a different group, which is really enterprise and DOD with this premise that hybrid multi-cloud is the future where people are going to want on-prem and they're going to want the best of all the clouds. And this idea of distributed compute for inference for ai, meaning you're going to have to deploy computers everywhere. You're not going to have meaning, you're not going to have these big campuses, which you will, but you'll have compute you need on each corner to help with self-driving cars in smart cities and all that kind of stuff. And we're trying to make that successful. So think modular but very different than what it's done in the past. And this is where it's tough for me because I'm going it against years and years of people using modular where everyone has a horror story and I definitely have my own horror stories where I've had modular data centers and it's because they were really built out of steel.

(29:40)
And the steel kind of depends on your welders and I'm used to the navy where you have non-destructive tests to them welds to make sure they're good. You don't have that fin weld. So you might have a good welder one day and a bad welder the next day in case you put five or 10 million worth of servers in this container that starts leaking immediately. You had a be welder, you automatically assume that all containers are bad or you have this old concrete one that's 64 tons that you deployed and it cracked in the same kind of situation. So what we're trying to do is a different approach. We are trying to use something modern technology. I'm a huge, not only a computer geek, but I love kind of material science and we're used in a composite that's new to really build these modules to give the benefits of the quality of a hyperscale but at a lower cost price point, but something that's more sustainable and stronger.

Jessica Denson (30:34):
And so explain to just for someone like myself the layman, what the modular pieces do. Do they amplify internet? Do they?

Tony Grayson (30:47):
I think it can do really anything. So imagine something that you take all the benefits, a mass customization, meaning you give the customer some choices on customization, but you're using a manufacturing line to build a data center. So I mean there's a reason why in the car industry you don't hand build everything. I mean you do with a Bugatti, but they're like a three or $4 million car.

(31:14):
It's built in assembly line. You have a couple different models and a couple different trims of those models. We're just doing the same thing but are applying it to data centers where you kind of have different densities or different coolings, but we're trying to keep it within this realm so we can build 'em quickly, look and build 'em to a high quality and we can build 'em at a lower TCO than what else is out there right now. And we're taking the best of what the hyperscalers do and we're making it as a service. So it keeps people, they can still maintain themselves cash rich, they can, we take care of the funding and the capital, we do the engineering, we deploy it and we run it for a customer, all our operating lease and they can lease it for five, 10 years. They can re-up that lease, they can buy it or they can pay for us to take it away and we can refurbish it and give it to another customer. So it's kind of the best of both in my opinion, where you're taking the business of the hyperscalers, which is this kind of providing this service to the cloud, but now extend that into DOD federal, state and enterprise, but also doing the best of what you can that it's tough for the hyperscalers, like the hyperscale data centers to do on this manufacturing approach to it.

Jessica Denson (32:27):
So it would serve, it's intended to serve a smaller regional area kind of thing.

Tony Grayson (32:32):
I mean I think it could be a regional area, but let's say you build a retail data center and now you want to deploy something that could be closer to the customer. You can't really move that retail data center, but I can deploy mine on roofs, I can deploy 'em in parking lots, I can deploy 'em in garages. And the fact that the office building have that whole real estate has kind of imploded. It's not tough for me to get two or three megawatts anywhere where you want because I might take up eight parking spots, but I can take the couple megawatts that that office building has and we sublease that area and they're more than welcome to give it to us as long as I can use some of those rack spaces for their own clients and their own building. So it gives you a lot more freedom on where you can deploy and opens all that up and you just deploy just what you need because it's more of a just in time delivery.

(33:24)
So imagine you can order your servers, you order your module at the same time and everything shows up at the same time and then you are not as worried as with the AI shift. But going from hopper to Blackwell is a big change. Going from 40 kilowatts to 132 kilowatts going from air cooling to direct to chip cooling, it's tough to retrofit a data center, but if you just build for what you need, it gives you a lot more flexibility. Then you could have a standard compute data center right next to a hopper data center right next to a Blackwell data center kind of all in the same area, but without having to do retrofits.

Jessica Denson (34:01):
Gotcha. What led to this form? Was it just something that Compass Data Centers saw a need or was it something that you really pushed for or how did Quantum come about under this?

Tony Grayson (34:17):
Yeah, so Criss Crosby, CEO of Compass really saw the value of Edge back in 2016 and he had bought two companies that were kind of doing stuff for Google, like the Google fiber huts for the fiber to the home. And one of 'em was a software based company, which is kind of B-M-S-D-C-I-M, which was now called Radux iot, and one of 'em was EdgePoint. And so I took over EdgePoint about three years ago and until then they were building kind of standard modules that were based out of concrete. And I saw really the value of it, but it was almost too early. If you remember, there's been a lot of these kind of edgy type, modular based business that have gone out. It's not that they were doing anything wrong, it was just wrong time. And now with the growth of hybrid multi-cloud, the growth of pushing broadband to rural areas into unique places, the growth of higher density racks and the rapid technology changes, we're going through it. The modular starts coming into the heyday and now you're with ai, you can be putting 36 million in a module. So you have to build 'em differently. You have to safeguard that hardware because a lot more expensive, it's a lot more of investment and they don't want 'em in the things of the past. They want something that's going to stand up to the test of time.

Jessica Denson (35:42):
So let's talk for a second about AI and other tech innovations. Is there anything that you're excited about? I know everybody's talking about ai, so it could be that, but is there that and other things that you're really excited about this space?

Tony Grayson (35:57):
Yeah, I'm excited for the future because right now everyone is focused really on large language models and you don't make many money off large language models. They're a necessary thing to get to the moneymaking part, which is inference. A lot of the inference right now is based on chat GPT, and so it's some form of chat BGPT call an API call that you have your own program and it calls back to Jet GPT and they wrap it in a different format. But there are some companies who are starting to build their own language models like Tesla's building their own even video's, building their own. And so once you have those large language models, people will start deploying that for inference and that's where they'll make their money. Some good examples of that happening now are there's a certain chip manufacturer that has 16 high definition cameras on their manufacturing that uses an inference type ai.

(36:50)
It has to be less than a hundred microseconds to basically check quality T-Mobile is working with the city of Las Vegas where they're trying to prevent pedestrian deaths. And so they have cameras that are focused on the crosswalk and they're focused on cars coming down. If a person steps on the crosswalk, the cameras look down the road and see if it needs to turn that stoplight red and has it aside within five milliseconds. And once you start getting into this video kind of looking at video, you can't call back home. You have to deploy that compute close to where that's at because we just don't have the pipes or the latency need to kind of push that back to it like a centralized location. It has to be distributed and it has to be close to the customer and that's where the real money's going to be made.
(37:32)

It's inference. And you'll get to a point where you'll have some kind of radio access network antenna on a building and that will have a bunch of different waveforms and whether it's six G, 5G, some temping special for augmented reality and that will have its own edgy type data center at its base that will do local compute for the platforms that are on it and then send back what it needs to for the big hyperscale data center. So the hyperscale data centers are not going away, you'll just have more distributed compute that'll enable this inference.

Jessica Denson (38:04):
So when you say that, so I just wanted to clarify two things. One, when you're saying that you really need these other modular places to help with AI or inference, you're saying that because the time needed, you need to reduce that latency, correct? That they need to be closer?

Tony Grayson (38:26):
Yeah, it really depends on the platform, what you need. What's happening now is text-based and you just write, Hey, write me a paragraph on X. You don't really care about the latency. Now imagine you're in a self-driving car that has a bunch of video that pushing this video out via some radio access network to a local area. Do you really want that going back to Asper and Virginia to make a decision? No. You want it to make a decision because everything's based for cars is a human reaction time, which is 10 milliseconds. So you can't push that much data through a pipes no matter how big your pipes are. To go back, you're going to have to do something locally and then make the decision to push it back out. And people forget it's not just human interaction, it's machine to machine. That's really where you get all the data from it's machines talking to each other at a high rate, pushing a lot of bandwidth to each other. That's when you're really going to need this local compute to generate this inference. And we haven't really got into this yet. I mean we've kind of talked about it and that's why since the focus has really been on LLM or kind of Jet GPT inference, that's why some people think inference is latency sensitive. And I agree, text base it is. But once you start getting closer to the customer and send 'em data rich stuff, latency really does matter.

Jessica Denson (39:46):
And let's just define just even further, when you say inference, are you saying how the interpretation of the data is? Is that basically what that is?

Tony Grayson (39:53):
It'll take your input or it'll take some other input to go into a language model that it's been trained to do and a large language model that has to go through a training on how you want it to react and how certain parameters are. And then it responds back to that basically the model that it's sent back to. And so it's making decisions based on what it's learned. And it could be just based on you run it to write a paper, it could be based on things that it's learned on philosophies for cars or where people are at. So it doesn't necessarily have to be generative ai, it could just be where it comes up with something new. It could just be making decisions based on a bunch of different point sources of data.

Jessica Denson (40:38):
This is fascinating. We're in a strange new world.

Tony Grayson (40:40):
No, we definitely are. That's what I think is still interesting is we are, if you kind of look at standard compute really has not changed with Intel or A MD. I mean, yes, we've gotten a little bit faster, but we've been designing racks for really 14, 15 kilowatt peak and maybe an average of six to eight kilowatts for a long time. You can't do that with ai. I mean you're going from Hoppers that are roughly 41 kilowatts of rack, which are rear door air cooled to black walls into this year, which is 132 kilowatts of rack roughly. That's both air and direct to chip. And then you're going to follow that up in two years by black the Rubens, which are 250 kilowatts plus, I mean we are in a significant massive increase in capabilities but also in power consumption and performance needs. And it's just something we've not dealt with before. I mean we've been dealing with a pretty flat curve for standard compute for a while.

Jessica Denson (41:42):
And at Connect Nation we talk a lot about the digital divide. So the need is great, and the longer we wait to fill that gap, the more it hurts others as well.

Tony Grayson (41:56):
And I don't think people, I mean the US is spending, I know it's the bead, which is a lot of the stuff that connected nations working with, which is basically government funding to help fix the digital vibe by pushing kind of broadband rural areas. But there's also the Covid Relief Act. The Infrastructure Act is pushing 77 billion really to push this broadband out to rural areas. And people think that everyone has broadband, but there's a lot of people that don't going to libraries to get broadband and how can you really learn at this day and age if you're in a Dialup modem or DSL, you really need some kind of broadband. And I'm glad the US government has spend so much money and I'm glad companies like Connected Nation are doing that because it's the right thing to do. Until we fix that digital vibe, we're not going to be helping people. It's more you can give 'em a computer, but if they don't have a way to access the internet in a meaningful way, it makes no difference.

Jessica Denson (42:52):
And then here's this new technology that's coming along and it's just expanding that unless we do things now, we do something about it now.

Tony Grayson (43:00):
Exactly. But the other thing that the benefit of what this network and pushing these network nodes in the rural areas, now you're actually pushing the means for inference to happen now because now you'll have more network nodes, but you'll have less traffic through those nodes and it'll be easier to route traffic. It's the great thing that Connected Nations taught me, all of Kansas routes to Atlanta, if you can imagine it doesn't matter because it doesn't have its own routing, but a company like Connected Nation will put down, let's say a hundred network nodes, which will now route your traffic. So qui will get less traffic through, but you in a rural area will have a lot better access and a lot less latency because a company like Connected Nation use government funding to build an internet exchange point right near you. And because they build an internet exchange point, you'll get like the Netflixes, you'll get the Hulus, in which case you'll not only have better access to the internet and academia, but you'll have better access to media too.

Jessica Denson (44:04):
I think the pandemic, before the pandemic, I would talk to journalists or just general people, they'd be like, you do what and why do people need that? And now I think people understand it's not a privilege or a luxury that if you want to be a part of the world in general government, whether accessing government services or watching a Netflix show that you need this. So Bravo, I'm glad to see you working with this.

Tony Grayson (44:29):
But I think there's a lot of people out there who still think that this kind of broadband access is everywhere that we live in a modern age and you have starlink and blah blah blah. That's only putting a bandaid on the huge problem that we have.

Jessica Denson (44:43):
It really is.

Tony Grayson (44:44):
It's only broadband, but it's affordable broadband. It doesn't matter if you can get broadband, if it costs a thousand dollars a month to get that broadband because that one rural customer is paying for the entire fiber line, that's cost prohibitive too and it doesn't make a darn bit of difference. So it's bringing affordable broadband to the masses and so they can basically build the business or build the industrial base or learn to help people to better themselves. It really is. It's a critical problem that we just don't understand. I think we live, a lot of us live at a privileged life myself, I bitch and know. And if I don't have a gigabit in my home right now, but in reality there are some people are dealing with nine, 600 bond modems and stuff. Maybe if you can imagine what that is and how frustrated it is when you download a video, how annoyed do you get if it takes you a minute to download something?

(45:39)
And there are people who can't even do that. So we would need to be better. And to be honest, because the edge has not happened because of two things, and we've been talking about it for the last seven or eight years, it has not had the network to be enabled successful, which is going to be fixed by bead and some other in Europe's doing the same thing too, of pushing broadband to rural areas. It's just through different regulatory or different EU sponsored initiatives, but you haven't had the platform, which has really inferences the platform and this other push for broadband to warrior areas is the network that'll make it happen.

Jessica Denson (46:16):
Well, Tony, I could talk to you all day. You are very interesting, fascinating, and your deep knowledge of this is very evident. I would like to close with one thing today, you mentioned above at the start of our podcast that you were in the Navy because you really believed in service above self. It seems like quantum, tell me if I'm wrong about this, but quantum, your role there is also somewhat thayy.

Tony Grayson (46:42):
Yeah, but I consider myself very fortunate where we're servicing definitely the enterprise and that's good, but we feel a huge need that the federal government, the state governments and the duty has for stuff like FEMA or forward deploying assets that enables our Navy to be more successful or Air Force be more successful. I mean, that kind of stuff, it's great. It's a great way to give back and as a business and know that we're given the military or federal or state the best technology they can have at the best price possible to enable them to be successful. And so it's just a different way to give back. But to me, since it's kind of morphed into this, I consider myself super, super fortunate. I tell you, there's nothing like going back into the Pentagon and talking to people and bringing the knowledge that we have of the industry now into what they do to help them be better.

Jessica Denson (47:42):
Well, Tony, I appreciate you and I appreciate all you're doing. Thank you so much for talking with us today. I would love to follow up down the road to see what you're doing in six months, a year from now, how things are going with that. So if you're open to that, I would love to talk again.

Tony Grayson (47:57):
No, definitely would love to do that too.

Jessica Denson (47:58):
Alright, I've been talking with Tony Grayson, who's the general manager with Encompass Data Centers. He oversees Quantum, a subsidiary of Compass include a link in the description of this podcast. I'm Jessica Denson. Thanks for listening to Connected Nation. If you like our show and want to know more about us, head to connected nation.org or look for the latest episodes on iTunes, iHeartRadio, Google Podcast, Pandora, or Spotify.