Chris Stansbury, Lumen | theCUBE + NYSE Wired: Business Transformation Edge

Palo Alto studio connecting Silicon Valley and Wall Street. >> I'm John Furrier with the Cube here with Dave Vellante, my co-host. Hello, I'm John Furrier with the Cube here at the Cube's NYSE studio. Of course, we have our Palo Alto studio connecting Silicon Valley to Wall Street bringing in deep tech coverage to Wall Street and Wall Street to Silicon Valley. This is our business transformation series. We talked to leaders who are leading the the era of modernization around AI infrastructure, AI intelligence as the transition to businesses who are transforming and driving revenue. Of course, taking the cost taken out using AI. Chris Denberry here, president and CFO of Lumen. Great to have you on. We all know Lumen Field in Seattle. We see that all the time. Of course, we've covered you guys on the Cube. Thanks for coming on. >> Great to be here. >> You know, networking is so hot right now. This is your DNA. Uh you guys just had your earnings. Uh you had an acquisition. >> Yeah. >> Alkira, a company we've covered on the Cube. You're starting to see the role of networking change. Um let's get into the news first. The acquisition and earnings. Take us through the earnings and then acquisition after. >> Yeah, to your point, the the the landscape is changing rapidly and one of the big uh pieces of data that we shared in earnings is that the strategic part of our portfolio, the growing part, the part that customers want today is now more than half the company. It's about 53% of revenue. And we see that continuing to grow in a really significant way as more customers are adopting NaaS and and to your point, we'll get into Alkira which just accelerates that. And there's a shift going on because networking historically has just been lazy dumb pipes. >> Yeah. >> That uh others then had to figure out and innovate around so they were secure and they worked. And that just doesn't work anymore. Uh there's too much sprawl, too much complexity. Especially with data now being geographically everywhere. We don't even know where our own data geographically sits. And so, you've got to have access. It's got to be on demand. You've got to have low latency. And it's a totally different networking environment today to meet those demands. >> Yeah, it's interesting. I'd love to get your thoughts and we'll get to the El Carro thing. I want to get some, you know, know that deal further. But that what's changed is the intelligence. Yeah, moving packet from point A to point B, been there, done that. The role of storage, networking, and computing, and even databases changed so much. If you look at how the best AI systems are being built, it's those three elements, but they've changed. The role of storage has changed. HBM memory, how important that is. The density of these large-scale rack-scale clusters. And you now have a distributed computing environment at the edge. >> Yeah. >> You got robotic, we see autonomous vehicles. Physical AI is certainly going to become super fast. It changes the role of networking. How do you guys see that? Because this You mentioned the dumb pipes. That was the old model. >> Yeah. >> Yeah, you put the intelligence at the edge. That's just Where does it go? Now you have to know that. If you look at the the rise of the importance of KB cache and Nvidia. >> Yeah. >> Everyone's looking at that saying, "Okay, tie to a software model, CUDA, networking, moving data around, tokens makes it really important." So, networking now is front and center. It's the operating layer. >> It is. >> Of AI. >> And and we like to refer to it as the nervous system for AI, right? If you think about the GPUs as the brain, the brain can only function with a nervous system that can alert it, that can uh that can help the body react. And in this case, AI's uh body react. So, we we see a world where the uh the network has intelligence built in to know, "Hey, wait a minute. Those GPU clusters are nearing capacity or there's an energy limitation in that geographical location. Let's just reroute the traffic to a different location where there is energy, where there is capacity." Uh but to do that, you have to have a programmable layer that lets you do that with little to no latency, on demand, yeah, from anywhere to anywhere. >> And that's what we're building. >> Yeah, I like that that approach and I'm seeing other people vertically integrating on the neo cloud side. N scale just bought and any scale for over a billion dollars privately held company. That points to this I got to know the resource. I got to know where to route things and cost is too important. I mean finance we heard from the leaders here in New York this week. Jensen came out say hey, here's a half a billion dollars. Here is KKR, Blackstone, Goldman Sachs CEO they're all saying hey, we're going to invest in AI infrastructure. >> Yes. >> And that is huge signal to the market that this is not a bubble. There is massive demand. >> Right. I totally agree and and look it, we don't know who the winners and the losers are going to be. What we do know is that whoever wins in the space is going to need a programmable network layer. And that's been our vision now for a number of years as we've you know, turned the company around and now we're really transforming it because ultimately the only way AI is going to work is if the network that supports a very dynamic environment is dynamic itself. It can't the network cannot remain static to support a very dynamic AI world. >> Yeah, we've been I've been covering Lumen for a long time. Love the networking approach. You guys have a great business. Talk about the asset and the strategic assets that you guys have because we're predicting that after the AI factory era when all the scarcity is still going to continue to be scarce with HBM memory, but you're going to have the monster centers the AI factories you see in the big places in Texas. All this big build out that's going to continue to go. Okay, if you believe that to be true which everyone does, the distributed computing nature of AI means that there's going to be other nodes. >> That's right. >> The edge, metro areas. When I hear people talk about metro pops or metro factories, it kind of has that same trajectory of networking. >> That's it. >> You guys have a huge asset base. Share just some general stats on Lumen's position because if the distributed computing comes full throttle, that's going to make AI factories everywhere. >> Yeah. >> That means they're just nodes on a distributed computing network. Talk about that and the role of networking. >> Yeah, and we we think about it really in a in a couple of different ways. And it to us it's a three-layered cake. And at the base, the foundation, is that physical layer of fiber, the most modern fiber in the world. And and we are on track to deliver about 58 million miles by 20 31. It when when Kate joined the company just 4 years ago, we were at about 12 million miles. So, a massive investment has taken place. And we've done that very efficiently because as we built network to support LLM development for the hyperscalers, we've been able to deploy additional network to support broader AI for the inference frame. >> have been digging the trenches, paving the roads. >> For the >> Really? Most of it has been AI. >> Yeah, but the but most of it has actually been the utilization of underutilized conduit that was put in the ground 25 years ago. >> Yeah, so you're not digging new fiber. >> digging. We're deploying. And but that's not the unlock. The unlock is the programmable layer. So, think about all those pipes and the way we traditionally do it, frankly, is a lot of north-south traffic. It's it's prem to cloud. And because of our proximity with the hyperscalers and because of our now multi-cloud gateways that sit physically in our network, we can get direct cloud access with super low latency, super high speed, much lower cost than going through uh the normal internet. >> on that? What's What does that mean for the customer? >> Well, what it means for the customer is lower cost, uh much faster ability to connect into the cloud. What it means to the hyperscalers is is faster time to revenue in terms of that cloud uptake. And what Alkira brings us when you connect that in is the programmable layer for east-west traffic where you're now moving data cloud to cloud, cloud to DC, to the edge with it never having to touch the prem. And that layer is actually agnostic to carrier and cloud. It touches every cloud. It'll ride on any network. So, as others are building trenches, we can control that traffic that ultimately rides on those lines. >> north-south, east-west. And what that means is you can go anywhere >> anywhere >> based upon what workloads, what the infrastructure is doing. Um talk about the demand because um you know, I've heard I love when people talk about the bubble because uh you know, I've been through bubbles before and especially the dot-com bubble. I lived that you know, I got a lot of scratches from that one. But at that time, there was not a lot of demand for the fiber. >> That's it. >> And and so, people compare a lot of the AI bubble to the day of the internet. >> Yeah. >> And I think that's a false premise because if you look at the online population at that time, it was just ramping. Now, the demand curve for AI is off the charts. The demand for intelligence is off the charts. So, share your thoughts on the demand curve for intelligence and the why you're investing so much in the fiber. >> Absolutely. So, when you think about the the dot-com bubble, it was a solution looking for a problem, right? And and ultimately uh industries grew and found a way to you know, consume that and utilize it. This is totally the opposite. This is a situation where there's demand uh today. Uh obviously, you need fiber. People are building fiber, deploying fiber, but you need a way to program it and consume it easily. And so, for too long, networking has been about deploying the capacity. It's now about enabling easy consumption. And so, you've got to have the capacity, yes, but you've got to have a pathway to easy consumption. And that's where we really see the growth. And and a couple of data points, if we look at our own um growth trajectory, again, over half the company's now growing. If you think about our NAS product, which is really north-south lanes, it's growing high 20s, low 30s quarter on quarter in a market where the dollar TAM growth is under 1% per annum. >> Yeah. >> So, you've got an explosion of data, huge data consumption, but very little dollar growth because that north-south traffic is commoditized. When we move to east-west, we're talking about growing our TAM from, you know, 20-ish billion to over 70. >> Yeah. >> And and in a market that's growing at 13% in real dollar growth because the need now with data centers being everywhere, and quite frankly, we don't even know where the next one's going to be. >> Yeah. And there's huge backlog, by the way. There's orders waiting to be fulfilled. >> That's exactly right. And and so, that'll all sort itself out in and when you watch the markets and it's like, oh, is it a memory thing? Is it a GPU thing? Is it an energy thing? Those Those are long lead items. That will sort itself out. But what is very clear is that we're moving to a much more distributed data environment, and we have to have an access layer that allows you on demand to move traffic anywhere. I mean, think about >> Yeah. >> um a regulated industry, uh financial, health care, where you've got data sitting in a in a private data center in an area that's about to be hit with a hurricane. Well, if you need to rely on traditional networking to go build a pipe to the new data center location, it's going to be too late. But if you can access that in a very program- programmatic way on demand, and move that through software, >> are networking fundamentals. It's not new, but the the the techniques are new. >> That's it. >> The the programmability and routing and policy, that's been around since networking was around. >> That's it. And and quite simply, what what are we doing? We're we're doing what other industries have done. We're cloudifying networking. >> Yeah. I mean, I think that demand thing is huge. And one thing I want to get your thoughts on cuz I think you you you brought this up and I think it's worth unpacking. The demand on the east-west and north-south are continuing to thrive because of the cloud. Now, if you go back two years, the number one conversation in the pundit circles was repatriation. Moving workloads off the cloud back to on-prem. But what's happening now is that's the cloud still growing. >> [laughter] >> And so, you have cloud growth, you have on-prem growth. Now, what happens? You have a demand for intelligence. First was like marketing, search, we saw the first wave. Now you got coding, agents are coming in, physical AI right behind it. The on-premises activity has nothing to do with mutually exclusive solutions on the cloud because cloud's growing. So, to me that's distributed computing. So, share your thoughts on this because a lot of people trying to squint through data center growth and on-prem is because the cloud's slowing down. That's not true. >> not true. >> The cloud is still growing, but it's just still connected. So, north-south you meant hyperscales. There's more hyperscale demand because people have workloads on there. >> People have workloads on there. And by the way, what what does on-prem mean anymore? Because on-prem doesn't mean that it's physically in the building that you're sitting in. It may be in your footprint, in your corporate footprint. Um but it's still geographically dispersed. And so, you still have to have that access. So, you're seeing growth in both. And what does that mean? It means that the environment is getting more complex. It means that there's more network sprawl using traditional methods to try to get all that connected. And so, for us, complexity is our friend. >> Well, well, the demand for data the demand for data growth on that's private data. I mean, the frontier models are public data. Let's face it. >> Yeah. >> On-premise data center, enterprise data is locked. So, the unlock there is how do I leverage that? Well, I want to have either VPCs in the cloud and/or a controlled managed service on a workload end-to-end. That's the demand that we're seeing. >> Exactly. And And look, if we go way back to when uh cloud first originated, there was the the doomsayers who said, "Oh, it's going to wipe out all on prem." And the reality is, where have we landed? We've landed with a hybrid model. >> Yeah. >> And a hybrid model, private cloud, public cloud, on prem. What it means is more points of connectivity that have to be secure, that have to be fast, that have to be accessible on demand. And that's where we fit in. >> you're hitting on the perfect storm of innovation because if you look at Alkira's acquisition. Now, let's talk about that because you're essentially seeing cloud-native growth, rise of the cloud, we saw that generation, the rise of AI-native together is distributed computing that's hybrid. >> That's right. >> Okay, that's a fact. It's interesting, Alkira was on our supercloud program that we ran. At that time, we called it supercloud. We didn't like the term multicloud because we predicted that supercloud was essentially workloads running across clouds as a resource, not I'm running my teams and office on Azure. It was I want to use Azure for compute and now AI. So, you're starting to see that play out in a way the multicloud is just a substrate of resource and networking. >> Yep. >> So, that played out. So, it kind of supercloud name was just kind of didn't really take, but call it distributed computing. >> Yeah. >> Now, you have AI layering on top. You have the systems of intelligence. Okay? >> And that's going to be key. Uh term that we coined here in the Cube was underneath system of intelligence is systems of adaptability. >> Yes. >> And that's what you're kind of getting at on programmability. So, talk about that piece because I think Alkira fits into that adaptability piece because you got to have the data, it's got to be adaptable, real time, available, horizontal, harmonized, contextualized, >> Yeah. >> context graphs, but then that produces the intelligence, the superintelligence, specialized intelligence, general intelligence, that sits on top because I think this is a new layer of the infrastructure. >> It it is because again, historically the network has been slow to adapt. Again, very static. And so we're now in a dynamic environment where the hyperscalers want to control the movement within their own clouds. But what the customer wants to do, very few customers are in one public cloud. They've got data in multiple clouds and in multiple DCs and they want the ability to move from anywhere to anywhere in a dynamic way. And so that's where we see the growth. And by the way, let's expand that more broadly just in terms of business use cases. So think about the world of M&A where historically, oh, I've got to integrate these two compute and network environments. It's going to take me 18 months and 100 million bucks after I buy the company because their equipment doesn't match my equipment. Well, when that's all in the cloud, Koch Industries uses Alkira to integrate M&A because it happens in days because it's cloud native. Now think about real-time resilience. There was a situation, a large customer of Alkira's that I can't name in the Middle East where a data center was flattened by Iran and the customer never knew it happened because you had resiliency built into that cloud layer that let their business just keep on running. So when we think about the number of problems >> that's pieces out autonomous. >> Exactly. >> That's autonomous behavior. That's agentic-like adaptability and resilience. >> That's right. So so the more back to your point, the intelligence that can exist in that kind of a programmable layer is exactly what AI needs to function. And by the way, I mean you know this probably better than I do, but what's the rule of thumb on AI's intelligence level? Isn't it doubling like every 3 to 6 months? So with that is coming more business use cases, more data consumption and and this situation is only going to need that layer more and more. >> president and CFO, so scope the cost. You mentioned the acquisition on the M&A side. I like those numbers. Let's take that example of the data center in Bahrain or wherever it was. Um the cost scope of the old way versus having that adaptability and resilience. What would that have cost? I mean, just kind of order of magnitude in the solar system of cost. Like, how big of a savings does that yield? >> You When you think about it, it's it you almost can't compute it because there's reputational damage. There's the amount of time the business is going to be down. Um and and again, with that capability, it just didn't exist. So, you go from a situation of massive uh you know, recovery efforts, PR, everything else to oh, we didn't know that happened. >> just the risk management. It's huge. It pays It pays for itself. >> It's huge. And so now, um with with with a programmable layer that is cloud agnostic and carrier agnostic, well, what does that mean? It means it's also geographically agnostic. So, so we can create a global environment for a multinational corporation >> very handy. Talk about sovereignty because sovereign cloud was more about privacy and you know, regulation stuff. Now with AI sovereignty, you're talking about revenue in state geographies, regions, or countries. >> Yes. So, now you got people leaning in saying, "Whoa, whoa, whoa. I love the privacy for users and data that's maybe specifically has certain parameters. Now you're talking about revenue." Yeah. >> So, sovereignty will be very much programmable. >> Exactly. >> That becomes Talk about your vision and how you guys see the sovereign AI playing out. >> Yeah. So, so really when when we think about Well, so think about a use case, financial sector, right? Big global uh banks. Um they're going to they're going to want one thing for sure. They need privacy. They need high-speed pipes and they they need not to be over public internet. And the reality is that can be stitched together with partners who build infrastructure. But what allows that to work is that programmable layer that makes those underlying pipes look like one network. And the ability to move in and out of cloud environment. So, to your point, you can program in the sovereignty to make sure that you're complying with whatever the regulations are locally, but you're given a global view and one pane of glass to manage that network and >> And the sovereignty is not country specific. It might be from a geography standpoint. >> Yeah. >> But because of the nature of the hyperscale, it's a global footprint. That's what you got you're saying, basically. >> Exactly. >> All right, let's talk about the transaction. You guys closed the deal. Take us through what was attractive on Alkira? You mentioned some of the multi-cloud domain specific advantages. How did it all come together? How did it go down? >> Yeah. >> How's it going to play into the future? >> Yeah. So, um we had our investor day in New York back in uh February. And we laid out a vision for what north-south and east west traffic could look like and the and the network that we uh were building and aspiring to build. And about the same time, Alkira came uh into our view and we said, "Hey, wait a minute. They've already built this east-west capability. And we can accelerate our transformation by two to three years uh and save uh the capital involved, but the bigger point was the ability to move immediately. And so, we engaged in that conversation, uh moved quickly, and uh the deal closed in July. And I would say importantly, in a lot of acquisitions, um you know, you've got uh competing views between the acquirer and the acquiree. >> Yeah. >> The uh the Alkira team is absolutely phenomenal. >> Yeah. >> We knew that they were bringing a skill set to Lumen that we didn't have. We've actually plugged Lumen's innovation into Alkira rather than the opposite. And the Alkira leadership team, the founders were thrilled that it was Lumen's network and Lumen's connectivity to the hyperscalers. Yeah. That really unlocked value. >> I mean they got the keys to the kingdom because I know interviewed the founders, very strong tech team, very excellent software layer. But you bring the network piece at scale. >> That's it. >> That's a huge deal for them because now they didn't have to do the biz dev on their side >> Yeah. >> to lock in relationships with the hyperscalers and also the edge is coming on strong for them. >> That's it. >> How they factor into the plans going forward? It sounds very strategic. You organized around them. >> Yeah. So we we obviously had to be careful with, you know, gun jumping and whatnot. We had to wait till the deal closed, but day one, as soon as that deal was closed, we had teams on the ground working together on how we uh first quickly deploy things that can be sold today, but then very quickly started to integrate road map so that there's an there's an ongoing selling motion here because this is really about scale. When it comes down to it, it's a new capability, but customers can choose how much they layer into that relationship. Uh and it can be through ecosystem partners that add additional uh value that we're working with. So it's about making sure that road map is unified and consistent. And and we're going as fast to the hoop and as hard as we can so that when we guide next year, the desire is for the first time in many, many years to actually guide revenue. Um because we think this pulls in our inflection back to revenue growth >> Yeah. >> uh more quickly. And so >> And by the way, their multi-cloud fabric and what they're doing to the hyperscalers works beautifully on edge. >> It does. >> Edge is going to be we we are very bullish on the intelligence of the edge. We think AI factors going to move the edge very quickly. That's going to change the game on what happens at the edge even from telecom providers becoming more service oriented. >> Absolutely. >> And you guys as a network, you have points of presence all over. >> We do. And we do have some edge environments all over. And so, the reality is is that again, complexity is our friend. Uh compute at the edge, I think is real to your point. And what we're going to have to do is make sure that for that to work, the data gets to where it needs to go when they need it to go there. And that's what we're enabling. >> Well, I'm super excited uh for this chat. One of the things I would like to chat with you briefly is the transformation journey you mentioned earlier. Our series on business transformation edge is about the competitive edge of the future. Most IT projects I've seen in my career over the past 30-plus years is IT, yeah, do the pilot out in the hinterlands of the of the environment. Let's get it going. Months go by. Hands are in there. Multiple five firms are playing with it. And then it's like, okay, did it work? And there's a whole presentation. Now, it's much more rapid. >> It is. >> business-oriented. It's business model. How has your transformation journey gone? Take us through the transformation and how you as president and CFO look at that, directing capital. Does it change how you look at capital? Does it change how you look at the personnel? How do you see M&A's factoring in? Take us through your vision on the transformation. >> Sure. Uh you know, it's there's a lot of history here, so I'll keep it short, but I would say the most critical thing that happened was the change in leadership when Kate Johnson joined the company. And her number one objective was to change the culture. And this is an industry that um was uh was really defined by playing to not lose. And so, I actually met with the uh someone this week who's a huge Cleveland Cavaliers fan. And he talked about the NBA playoffs game where the Cavs had a huge lead over the Knicks. And then they just watched it drip away because they got overly defensive. And that's how this industry worked. And so, we had to change the mindset. But at the same time, we had to build a foundation that that didn't have cracks in it. And our capital structure was a mess. And so really um there was a coordination but a division of labor. Kate was very focused on driving the culture change, moving us to a more digital future. While I was focused on all right, we've got 20 billion in debt with half of it doing 2027. It's an existential threat. We had creditors that were upset and we had to deal with that. And so we did it in a very transparent public way. We sold non-strategic assets. We monetized underutilized conduit with these large hyperscaler deals, which allowed us to delever the company. It allowed us to fix the capital structure and put us in a position really importantly to have a balance sheet that could become a strategic asset >> got to clean up house a little bit. >> And that that's what allowed us to buy Alcuria. >> Yeah. >> And so so and and that's what puts us here now. So what in many ways the the turnaround was getting us to the starting line. We weren't even at the starting line. We had to get to the starting line. And now we've got a business that is last quarter grew 14% year-over-year that's now more than half the company that's really starting to pull the company forward. We'll inflict EBITDA this year and we we will inflict revenue sooner or later. >> I mean I I love that. If I if if you had to answer the question now that you have this strategic asset, things foundation set, cleaned house a little bit on the some of the loose ends. You're in position, you're going to take an offensive approach. >> Absolutely. >> What is the how are you going to leverage that strategic asset? What's the plan? What's the focus? >> It's a great question and and we've actually started to be public about the fact that look, um this company like every company needs to have a product life cycle approach. And so what we're doing with a with a pathway now to a much bigger growth company inside of Lumen that is our digital future is to de-emphasize and candidly um starve the uh the products that customers aren't buying anymore of resources that it takes to support them so we can plow that into the innovation engine and get even more aggressive on the digital side. So to your to your point on how we think about deploying capital, it really is about driving a return now. >> Mhm. >> It's it's less and less about fixed pipes. Again, 58 million miles is our target and it's more about that digital layer. So it's about redeploying resources from the past into the future and being super aggressive about it. >> I've been kind of very vocal on the whole AI factory distributed computing thing we're just kind of riffing on but I said the the secret sauce is what I say to Dave Vellante, the secret sauce is connecting the factories. That's it. So how do you view the AI infrastructure build out that's going on now because you know, we predict that the scarcity thing will sort of start probably next 12 months. >> Yeah. >> In that 12 months, enterprises and and and neo clouds are on a discovery phase of monetization. We think ages will come in quickly. >> Yep. >> Offer deterministic workloads, clear but I think, you know, in 2027, 2028 will be a massive um uh discovery of wow, this is how we monetize intelligence. What's your view of that? >> So it's so interesting you say that because this week when I've been in New York, I've met with customers, I've met with investors and I've had a couple of great conversations and that timeline is aligning with I think what that community is starting to see which is a lot of the speculation around shortages and limitations, that's going to resolve itself to your point in the next 12 to 24 months. What I think enterprise is really starting to discover is that you can't use AI to replace a really lousy process. You've got to actually have an AI first business process where you're re-wiring your company around AI. And and that's that's how you invest in AI in a responsible way. And that is really starting to ramp. We're seeing it in in financial sector, in health care, largely regulated industries first, but I think we're just going to see it explode from there. >> you know, I'm a kind of a a nerd on competitive advantage, competitive strategy. So I have to ask, you know, the transformation edge is about competitive edge. How do you view that because when you get I mean words like operating leverage, those are words that have been kicked around. We're talking about the ultimate dream scenario for operating leverage with AI. >> Yeah. >> So you have that coming into play. How is the competitive advantage of the future look like for companies that do take that approach? People who are watching are formulating plans for their business model transformation. And they all want to have a competitive They want to drive revenue, keep happy customers. And what is that leverage? What is the key secret sauce? >> I I if I were to summarize it down to one word, I would say scalability. And so the the network has been a huge limiting factor because it hasn't been scalable. Oh, I need another fixed pipe. Oh, I need, you know, another layer of security around all that. So to the extent that we can deliver that programmable layer that has that intelligence built into the network, you now as the customer, we now as the provider, have a network that's very scalable that isn't capital intensive. That allows you to add more, add more, add more in a very digital way. >> love the AI infrastructure build out. >> We love it. >> I mean more data centers, more connection points, more demand for >> more more demand for a programmable layer that allows you to consume that in a way that is on demand and digital. >> So Al- Alkira is a strategic layer not only to connect to the hyperscales today, but tomorrow connect to the other resource centers. >> Data centers, edge, all of it. It It It's the really the unifying layer that pulls all of that together to one pane of glass for the customer. I mean, it's really it's it's magic. >> Chris, what's interesting in the past 2 years, I probably said the word computer science operating system on the cube more than 17 years of doing the cube because what you just said is when you have that layer of control plane, it's essentially scheduling resource management and routing. >> That's it. >> Okay. That's an operating system. >> Yeah. >> You're building an operating system layer on the network side that kind of mirrors kind of what we're seeing on KV cache, the importance. That's the darling in Nvidia circles, KV cache, but it's inside the cluster. >> That's right. >> You're everywhere else. >> That's right. And And And again, historically, it was dumb pipes that other tech companies built around to enable those pipes to be secure and connect to each other, but it's inefficient. Uh it's created just a lot of complexity. Um and so, what we're doing is cleaning that up by saying, "No, no, no, that should be native in the network." And we go from there. >> And your customers are getting more demanding because they can have quantifiable reasons to see lag, latency. They can see when we all know we see dot dot dot thinking on our our AI prompts. That's just slow. No one likes slow. No one wants a slower product. >> And AI? I mean, think about the the early kind of AI engagements that we have as consumers. We're engaging with AI in a customer service agent, you know, at at any particular company. If that agent has to go and look for data to answer your question, and that's a 3-minute response time because you've got a slow network underneath it that's clunky and stitched together, that's not going to work. You've got to have that immediacy so that the AI works for you. >> the data's not available, you get the wrong reasoning engine. That's hallucinations can start, drift happens. A zillion things go wrong. >> A zillion things go wrong. So, it's imperative that that latency is kept to the >> Well, Chris, great to have you on. I'm glad we went extra time on the transformation piece and congratulations on the earnings and the closing the deal. We'll be keeping an eye on it. Thanks for coming in and sharing. >> Thank you. It was great conversation. >> Furrier. This is the business model transformation edge series. This is where also the AI factory kicks in. Distributed computing is coming into AI era. It kind of the same, but it's different mechanism. Low latency, the role of the network as an operating principle is becoming front and center. We're already seeing with Nvidia every day on the mainstream GPU and AI infrastructure side. But, as it moves to the edge, as it moves everywhere, the role of the network is critical. More of this is coming. We're doing our part. I'm John Furrier, host of the Cube. Thanks for watching.