AI Infrastructure Will Accelerate: Zac Smith’s on Tech Evolution, Networking Bottlenecks

Diane King Hall coming to you live from the of the New York Stock Exchange. The AI infrastructure boom is entering a new phase. So says Zach Smith, the CEO and co-founder of datum, joining us here at the New York Stock Exchange. So we' in a new phase.one says we're in different innings. What are we in the first, second, thirdig are we in? Well, I can't I'm a big baseball fan. All right. But I think we're still in the first i You know, over the past couple of years, we've justeen a huge investment in the foundational parts of this new technology shift. So I'm sure investors or companies are plowing int picks and shovels or as I like to say, thekhoes and bulldozers of this build out. And, you know, I think we're just starting to see benefits of that. We haven't really seendoption in the ense. And the things that go along with thise technology shift. It feels like one day we're super positive about AI story and the build out, right? excited about Nvidia. We're excited about core. We've until we're not, you know, like core has been under a lot of pressure this week. And there's worry about NIO cloud companies.re do you see the where are you most optimistic about the AI iructure build out? Well, having been in the industryround data centers, networks and cloud computing for 25 y what I know that we're investing in infrastructure that has been, you know, really foundational to our economy as we've justuseI see that happening here now, and I feel very optim that the players in in this iy are taking it seriously. You don't go build a very large data center or a power plant, or putlions of dollars of chips into a facility without thinking throu who's going to use it, how are we going to pay for it? Who's the custoo I think overall, we have a relatively although it's robust and a very big build and of the industry and supplyparts chain, most decisions are being made with care around specific customer investment thesis.ou say 25 years in AI for the average person, myself inc John and Jane Doe. Wk about it in terms of just a handful of of AI. And I guess that's because for us, what we'lly seeing is the what some call AGI, right? Generative AI. It's happening, or maybe we call it superintelligence. Now, you also s infrastructure is movingeyond the GPU. So where are we going next? Yeah, I meana foundational move in, in model generation and what we've experienced through chatbots or maybe automations. And now we're start see this move into the enterprise change, real business processes, and that's going to affect things security and how people move their data or different ways to do analytics and, and just kind of moving software deeper throughout these, these industries. But I think you're going to also seehysical world. We haven't really started to see anything aroun robotics. And as people start to think of these applications, not only as LLMs and, you know, models built around text code or software, we're going to start to see more intelligence happening wi machines around us. Well, I see the bots delivering. You canee my wheels turning as I'm listening to you. Like you say, you say see morerobotig food. I mean, could have some food delivered right now on a robybe not to the floor, because I don't know if they can get through security. Right. Maybe we still need people for that. But, you know, I guess it starts to bring u like just the existential questions, you know, how much of this will be robotics and how much human, especially when you about Agentic your thoughts on that? I mean, I think we have big parts of our economy that have a shortage of labor where we need more skills and more capabilities. I think humans are pretty good at finding their best use case., you know, I'm hopeful that we can use some of this technology in companies. We'll find, you know, benefit beneficial use cases toward that,work with society around things like robotics. And, you know, I look at that in a city New York we're constantly repairing our streets or our, you know, need to be constantly repairing them. That's questionable. Maybe we could use robotics there, but go ahead, be able to use more robotics. We have huge am of like big, big challenges in our society around infrastructure outside of AI, that we could use more And I'm e about seeing some of the applications around robotics these areas. cI feel like the roads here are,u y weren't as bad as they they have been. This is just, youknoy networking could become one of AI's biggest infrastruct bottlenecks. Tell us a little bit more about that viewpoint. We've been focused a lot on chips, GPUs, data centers, theks there. But as you start to put these applications around inference using this technology in businesses or part of our lives, connectivity is going to be super impor, not only how wh these things and move the data to them, but also the security around that. And I think we've seen some of those, you know, red that AI is very powerful, but it can be used for positive and negative things. And so we h make sure that we're securing this data. We'reble to maintain these kinds of services without r f cybersecurity incidents. So when you think about the cybersecurity incidences does that So it makes me think about you m a point about software t determine how efficiently AI infrastructure operates. So of course, that makes me thinkke the CrowdStrike, Palo Alto networks of the world and demand for increasing demand for those in need, for those, because of the prolifn of AI. Do you do you see that an increasing needy simp in my, you know, little infrastructure build mind, whic that we're going from a where we had maybe X amount of software to X times ten. The more software in the world, there's more digital, there's moreata, it's being used me're going to need to increase our investments around sec around protecting tha around governance. And one more question for you. The economics of infrastructure shifting from capacity to effy. When you think about that, you know, it makes me think about thent that you made earlier saying humans have, and I may be paraphrasing this, have we have a use case. I hate hearing it like that as a human. Is that what that's about? Yeah. I mean, I think like this businesses, companies, you know, are going to find and justify, you know, business use cases around this technology. I think we're in this fir phase where we're r building the make sense. We're going to have to opt it. We're going to have to play the pI think you're going to see a lot of that happen over the coming year or two, as enterprises really figure out wher the juice can be squeezed. All right. Thank you, Zach. eciate t.