All right!
Let's go ahead and get started. Welcome, everyone, to Cisco's Technology Strategy Investor Briefing event. Really glad that all of you came to our event today, and of course, we have a lot of folks on virtually. Just so you know, we will be making a replay of this event, and our content and all of our decks will be made available post the conclusion of this event today. I don't think I got my slides. Well, I'll do it without the slides. I'm gonna give my favorite part of the event. Oops, Oh, my favorite part of the event, the forward-looking statement. We will be making forward-looking statements, so I refer you to our Forms 10-K and 10-Q, which you can find on our investor relations website. Let's get to the agenda.
I'm really excited about the agenda that we have today. We've got all of our key players on the executive leadership team. We've got the amazing investor relations team here as well, and this event was really designed to really hit about your key care abouts. One, deep dive into our technology strategy across the portfolio. Second, the massive growth opportunities that we've got playing in our portfolio, including cloud, security, AI, plus other growth opportunities. Then we're gonna talk about the innovation we're driving across our portfolio and some of the announcements you heard today at Cisco Live. Kicking us off will be Scott Herren, our CFO. He'll provide some opening remarks. Then we'll be joined by Liz Centoni, our Chief Strategy Officer.
She'll give you a deep dive into our strategy, why we're so well positioned to go after those massive opportunities of cloud, security, AI, plus some other areas like full-stack observability, as well as others. We'll be joined by Jonathan Davidson. He heads up our Cisco networking business. There is a ton of opportunity. He's driving a ton of innovation within that space. He's also gonna talk about web scale. As you all know, we're accelerating our opportunity, our footprint and gaining share. You'll hear more about that. From there, we'll have Jeetu Patel. As you all know, he heads up our security and our collaboration business. He's gonna bring up a key member of his team, Tom Gillis. They're gonna talk to you about our security vision and why we're gonna win in the security market.
We'll take a quick 10-minute break, then we're gonna have Rakesh Chopra. He's gonna join us virtually. He's gonna talk to you about our Silicon One advantage and why that's an unmatched differentiator in the market that no one can beat, as well as how it ties to our sustainability efforts and our customer priorities. Then I'll come up and wrap it up. Just so you know, for each of the presentation, you'll have plenty of time to address Q&A to each of the speakers, then I'll come up and wrap it up. With that, call up Scott.
Thank you.
Yep.
Thanks. I don't actually need that. Yeah. My team is the only people in the room applauding. I just wanna start by, again, thanking you. I've had a chance to speak to many of you, and I know this may not be the most convenient place in the world for you to be. I really appreciate you making the effort to come here and see this face-to-face. I hope you had a chance to listen to the keynotes or watch the keynotes this morning. We have another keynote tomorrow morning that's gonna go deep on some different topics, so if you're staying tonight, I'd love to see you stick around long enough to go to the keynote tomorrow. The only other thing that I'd like to add.
By the way, thank you to those of you online as well. I know everyone's busy. Like, everyone's got a lot to do, and I appreciate you dedicating some time to this. I think the message that I wanna deliver is many of you have been very generous with your feedback, on things that we could do better, right? One of them is like: "You're so complicated. Can you help simplify the message for me, for my portfolio managers, for, you know, the people that I'm dealing with on the buy side? Help simplify.
You seem like you're doing so many things." Liz is, you know, who leads our corporate strategy, has pulled together, I think, in a way, and you saw some of this in the keynote this morning, how you can let the unifying themes over that, so you can get a better sense of trying to simplify the story. The second is there's more that we can likely do on the reporting front. We're looking at that as well. I'm not gonna change mid-quarter, but to, again, to try to simplify and make it easier for you to digest what we're doing. The second is, one of the things that I've heard from several of you is: "I'm worried about market share. I'm worried about the innovation. I'm worried about the long term.
You know, what does this mean?" This is one of the phrases that what someone used with me, "Is this a melting iceberg?" It's not, but again, part of that is we need to now demonstrate to you what is it we're doing. What is the innovation that we're driving organically, and what's the innovation that we're driving inorganically? We've carefully curated the topics here to talk about the strategy overall, to very specifically focus on Cisco Networking, and Jonathan Davidson went through some of that this morning. We'll get to a click deeper today. Security, in particular, is critically important to us, and you heard Jeetu Patel on stage this morning. He'll be here with Tom Gillis, who leads the security space underneath Jeetu Patel, to talk through where we're going there and some of the innovation in that space.
Then AI is the, the current, you know, the current hot topic for good reason, right, with everyone. So we'll talk a bit about. We talked on stage this morning about how AI will get built into our products and make our products better and make it easier for you, as a user, to use those products. We also have a significant opportunity to sell, you know, if you think of the gold rush, to sell picks and shovels, right, into the AI build-out, both on the enterprise side and on the public cloud side. So Rakesh Chopra will come and talk about what we're doing specifically within Silicon One to meet some of that demand. Jonathan will touch on that as well, but in the, in the large public clouds, as well as on the enterprise side.
Trying to address, again, we listen to your comments. I really appreciate that you provide those to us, not just in writing, but in the, in the dialogue that we have. Try to listen to that. That's what has informed the way we pulled this together today. Hopefully, you'll enjoy it. With that, I'll bring up our first presenter, Liz Centoni.
I thought I was done with stage this morning, but happy to be here. Clicker, forgot about that one. Thank you, Marilyn.
Yes, you're welcome.
All right, as Scott talked about, you know, we've spent a bunch of time around simplifying our... I know Marilyn covered this already, around simplifying our corporate strategy. If you attended the keynote this morning, Chuck talked about it and said, "Simply put, we securely connect everything to make anything possible." Underneath that, what we hear from customers is, what we do is complex, right? We're talking about our customers' networks in how they secure it, how they observe it. Our focus is really around how do we cut through those layers of complexity and simplify technology in a way that anyone can consume it everywhere, and do it at scale as well.
As Chuck covered this, but I'd like to reinforce this. All of this is driven by our purpose to power an inclusive future for all, and we live it every single day. It's part of the reason I've been here for 23 years, and while corporate strategy has been my role for the last three years, the bulk of my time, I think all my time in Cisco, has been in engineering. I lead our full-stack observability and applications business as well. I truly feel, right, when I look back over these last 23 years, is it's a time where truly our scale, our breadth, our technology can help us shape the future along with our customers.
Speaking to our customers in our global customer advisory board about, probably about a month ago, these are our customers, one of our top customers, who can also be, give us some, I was gonna say cynical, but more around, give us some valuable feedback. Maybe that's the way to look at it. It truly resonates with them when we talk about our strategy and purpose, and how together, we truly believe that, you know, we can help them scale to enable their business outcomes as well. Chuck touched on these this morning. I won't go into detail on these, but a couple of things. The first four are the ones that we've had, in our strategy for the longest period of time, I would say, over the last three years.
Those have remained unchanged, whether it's around how applications are a combination of traditional hybrid cloud native, and they live anywhere from on-prem, and that's why I focus on on-prem as well, as well as in the cloud, and how hybrid work. You know, this is about how do you provide those best experiences, no matter where the person is, and no matter what they're accessing as well, and all these together, in terms of distributed workforces, distributed applications, are also driving different traffic patterns. Then add automation, then add the need for more intelligence and simplicity as well. You really have to modernize your infrastructure. Then security underpins everything. It's built in ground up for us from day one.
we should probably be talking more about how we even drive security as part of our development life cycle as well. The new set of things coming up is more around, you know, AI-driven security attacks are getting so much more sophisticated, right? The prince out there who's sending that email sounds so eloquent nowadays. These are the new things that are coming up, and no matter the number of security point solutions out there, and a number of teams, the number of security attacks and the frequency of those attacks are getting increasingly high. Security top topic. Sustainability. Over the last year, we've added this as one of the top priorities that our customers continue to talk to us about.
As they've said that, "Hey, sustainability is a big part of how we operate and run our businesses," right? What we build, whether it's in Jonathan's business, the software business, in security and observability, absolutely is around how do we enable our customers' sustainability goals, including with Silicon One, more efficiency, higher bandwidth, but also we have our own sustainability goals of getting to net zero by 2040, as you know, Scope 2 and Scope 3 get a lot more challenging than Scope 1. When we look at what our customers tell us as their top needs, whether it's around, "Hey, the world's hyper-connected.
My life is really around hybrid multi-cloud, is a de facto way I operate because I have assets that sit across all of these, and I need to be able to provide that seamless experience to my users." We see as a tailwind for us because our portfolio comes together beautifully for that, and we'll talk about that as part of our platforms. Security, I touched on that. Hybrid work, bringing together our portfolio on collaboration, networking, and security. In applications, why do we care about observability? Most of the time we're looking at it and going, "Hey, it's the app dev persona." Most often than not, Cisco doesn't touch on it. It's more than the app dev persona. The IT ops, the net ops, the sec ops, they're all important in making that applications environment much more simpler.
In addition to what our customers tell us, that they see us helping them enable, we look at it as a huge tailwind for us, but we're working with them also around how do they consume our technology in a much more simpler way. How do we bring multiple parts of our portfolio together to enable business outcomes for them? How do we deliver capabilities through a platform where it's much more easier to actually consume those capabilities as well? A lot of our capabilities, especially in the software, is delivered as a service.
If customers want it on-prem, we make that choice available as well, because more and more what we're hearing from our customers is, "As soon as you develop it, we want to deploy it in the environment." Long gone are the days where it took 12 to 18 months to test it, so you'll see a big push in terms of as a service, not just on the software side, but on the hardware side as well. Scott talks to you about this, Scott and Chuck do, every quarter, with the exception of probably Edge at the bottom. These are our four pillars that we report out as well. What I do want to highlight is that when we look at these pillars, these are key areas of investment, where we invest more than $6-plus billion every year.
These are cross-portfolio in terms of what it takes to deliver the customer outcomes that they're looking for, whether it's around hybrid work, where, you know, our portfolio comes together through collaboration, security, and observability, or it's at the edge, where it goes across our portfolio in networking, security, and observability as well. When I think about, you know, or what I would advocate, like, when you think about our portfolio today, it's comprised of four core businesses: networking, security, collaboration, and observability. Wrapped around it is our customer experience piece, everything from service to the support we provide to our customers and the customer success motion as well, that is really about how do we help customers through their own life cycle journey?
The customers that we serve, and I know we don't talk about a lot of these on our, on our earnings call, has expanded from what you would traditionally look at Cisco and say, "Yeah, I know you talk to network engineers and IT ops and maybe NetSec and net, you know, NetSec ops," for example. What's not clear or what we haven't articulated very well is we talk a lot to DevOps and SREs, especially as part of our cloud portfolio, part of our full-stack observability portfolio. We're increasingly talking to app developers as well. When we talk KPIs to our customer, on one end, we're about reliability, redundancy, on the other end, we absolutely talk to them about how do we enable them to drive their revenue? How do we help them in terms of business risk as it relates to observability and security?
We span the gamut in terms of how we talk to the CIO and CXO organizations. I think that's pretty critical to see that the number of personas that we touch in our customer base, whether these are influence or these are budget owners, it has spanned from probably those who've been covering Cisco for a long time, have seen us, you know, kind of talk to as well. When we enable this in terms of how do we help our customers navigate in these environments, we also turn it around and see it as opportunities, like in connectivity, for example. Connectivity is far from over. That's why we talk about one of the key trends is hyperconnectivity. There's so many more devices and this need for more data, more intelligence.
When you look at IoT use cases, I used to run our IoT business about four years ago, when we started talking about IT and OT coming together. The IoT use cases, we've just touched the tip of, you know, kind of what customers see as IoT use cases that they can deploy across their industries as well. Everybody's asking for more immersive experiences. You know, there's autonomous everything, for that matter. There's digital twins. All of these are driving this need for that connectivity. Work from anywhere. Every customer of ours is trying to figure out, in addition to the technology, what does this mean for me? Spending a lot of, on the future of work, as you can look at that $1.2 trillion. $750 million new apps coming on by 2025.
Even if you don't if you say, "You know what? I'm gonna put my own, you know, kind of little judgment on top of that and take a fraction of it," and say it's $100 million, $200 million in terms of applications, that's still pretty significant. That's where our portfolio in networking, security, and observability comes together. Expanding attack surface, I think all of you know that pretty well, and Jeetu's gonna cover that a little bit. Surging AI workloads. You know, customers are talking to us about the need for being able to have infrastructure that's a lot more efficient than the infrastructure that they can get today, especially when you talk more Generative AI workloads. Because we've been enabling, they've been building their, or running their workloads on our AI-driven infrastructure already.
It's just that when you take it to the next level and talk about Gen AI, it brings a whole new dynamic in mind, especially as they wanna run multiple workloads, not just individual workloads like high-performance computing does. Good point. Time to talk about AI. This is one where somebody asked a question in the press and analyst meeting, going: "When is Cisco really going to market more than just do stuff?" I think now is our time. If you think about. When I think about AI, it is pervasive across our entire portfolio. I mean, we use AI today in automation. We use it in vulnerability assessment. We use it in baselining, both in security as well as in full-stack observability. We use it in noise reduction.
If you look at the Webex team, they have reduced over 100 billion minutes of background noise. That doesn't happen without AI, right? It's this massive scale of data that we have, that I think there are very few peers in the industry. Most often than not, we don't talk about competition, but I'm like, "Hey, we should talk about it a lot more in terms of how we're vastly differentiated versus the others out there." Jonathan, I know you're gonna cover this in a little bit, and Rakesh is as well. When I think about the opportunity that we have on enabling, or our aim for building an AI network at scale, and I think about our Silicon One optics and our data center networking portfolio, I don't think there's any parallel.
Or I should say it this way, there's few parallels in the industry that would actually touch that. I know Rakesh is gonna talk about Silicon One. One of the things that our customers trust us is about giving them choice. Over the last few years, we've solidified that with our customers. Instead of locking them down into this, here's a full stack for AI, and here's a very narrow deployment model, our view is, like, the Silicon One that we're developing for this is around helping them with, hey, your architectures could be anything from Ethernet, enhanced Ethernet, and fully scheduled fabric. We will always drive a differentiation, but we're giving them that choice and flexibility, and having them make more data-driven decisions, rather than being locked down to something just because they wanna put, you know, AI out there.
There's a number of product enhancements. Jeetu will talk about it on security, around how we're using AI to enhance our cybersecurity solutions, how his team is working on making AI systems themselves to be secure. Then he's gonna talk also around, on the collaboration side, how we're including more I think he's gonna showcase that tomorrow on main stage for the day two of the keynote, around how we're making, you know, more immersive and more realistic experiences in, yeah, in hybrid work. Jonathan, I know you're gonna touch a little bit on AI in your portfolio, so I'll leave that.
In observability, the team's ready, my team's ready in a couple of three months here to be able to get to market, how we're helping observe large language models and APIs as well, and that's just the beginning of the roadmap that we have in terms of coming out. Better user experience, all of us are driving our products towards a better user experience. More towards, you know, gone are the days of CLI, but not just UI, but moving more towards, like, prompt interfaces because we want our customers to be able to consume our products faster and get the full benefit of them. Our customer support team has been using, you know, AI for a long time already.
They're looking at how they can do this much more customized, and give customers more of a control about being able to address their or troubleshoot issues as well. Internally, within the company, we already have been using AI around how we can optimize our own development teams in terms of better productivity and software quality as well. I call it automated inspiration, because with AI, we always believe it's a combination of AI plus deemed human insights as well. How do we get and think more creatively versus thinking about some of the tasks that could be automated? Just our own IT team, which probably runs a bazillion bots today, is also looking at Gen AI around: How do you reduce the friction so that our own users and developers don't go around IT and, you know, start building their own tools, for that matter?
Just an overview around how we're harnessing the power of AI. I'd be remiss if I didn't touch on why we're so bullish in networking around AI, around AI networking opportunity, 'cause we believe that we're really well-positioned. While today, the majority of the market is InfiniBand, we believe that. Something that Siri didn't like that I said or wanted more explanation. We believe that this market's gonna shift, with the majority of it being Ethernet. Why? If you look at Ethernet in terms of the install base, the economies of scale that you can give this massive ecosystem, that's not gonna be just a narrow InfiniBand itself.
That's gonna be a move towards Ethernet. We believe that Silicon One, with the flexibility, and I can't wait for you to hear from Jonathan and from Rakesh as well, will give us an opportunity to shine above the others as well. I would love to touch on our differentiation. I did this at our global customer advisory board as well, and they said, "You know what? It's about time. It's about time you guys talk about some of your differentiation. Try not to be as humble and brag a little bit." I did put some brag points on there for us, 'cause it's true, when I look at, you know, what do we hear from our customers?
Sure, they give us a lot of feedback around, you know, what are the things we could do better, and we try to drive that into our roadmap. Innovation, it's really around: How do we deliver customer outcomes by bringing our portfolio together and do it at massive scale, right? We're a market leader in secure networking. We're bringing together networking security and observability and collaboration, and many of these solutions, from full-stack observability, which is a $36 billion market growing at double digits, and We're hugely under-penetrated, and we've just gotten started. There's hybrid work. There's SSE.
When I think about, you know, our massive portfolio, we also have access to data across all of this in terms of different data sets that we collect through, you know, via telemetry, as well as, you know, metrics, events, logs, and traces that we collect. We have a pretty unmatched set of data that's out there that we use and will continue to use to help our customers with their AI-driven insights as well. I mean, that, when I think about just even in security, right, you look at email, endpoint, and network, and you look at the scale of telemetry that we have access to, that's the 400 billion security events that we look at every day.
When I think about observability, 630 billion observability metrics to enable the KPIs that I was talking about, the user experience, the revenue impact, the business risk, for example. Trust, by the way, part of the reason for in some ways, previewing this with our global customers is for them to also provide us feedback. It's like, "Yes, we agree with you. No, we don't agree with you on this." They completely agreed with us. They do trust us as a trusted brand. They know that for years, data protection and security has been number 1 for us. Even through very challenging pandemic situation, we've been continued to be rated as a top-rated in supply chain. We like I said, we have our own goals, and we're enabling
you know, kind of, carbon-free or net zero goals as well. When you look at our global reach in terms of our customer base, a million-plus partners and customers. Outside of China, 99% of the world's largest companies and 82,000 plus in terms of government organizations. Said a different way, I think it's pretty fair to say that there are very few parallels out there, that can compare themselves to Cisco today. Business transformation. Our customers have continued to talk to us about this fact that, "Hey, I want you to bring the solutions together to deliver real value. I want you to simplify the buying and enablement.
I want to be able to get the choice if I want to consume it in the cloud, if I want to consume it on-prem, as a service if it's hardware, I want that flexibility, as soon as you develop it, I want to be able to consume it. You'll see, we've talked about this on multiple calls, how we've accelerated a big chunk of our business to SaaS and as-a-service as well, and providing those, you know, flexible consumption capabilities for our customers. We're also have not just enabled today, but already enabled new routes to market. Outside of MSP and MSSP, we've also enabled cloud marketplaces as an example, where a number of our products today are sold via the cloud marketplaces as well.
I touched this on main stage when we think about the four tech trends that will shape our collective digital future, it's gonna continue to be around hyperconnectivity, and so hybrid multi-cloud becomes absolutely important. How we deliver our capabilities from a cloud platform absolutely becomes important. full-stack observability plays a big role in there. Our goal towards enabling sustainability and in around the secure part, outside of, you know, what we're delivering already here around Security Cloud, just even making things like access easier. Again, it's around abstracting complexity away from users. Regardless of where they're accessing it from or, you know, what they're accessing, it should be simple.
Jeetu's gonna talk about this a little bit more, and how we're doubling down on AI even more so, especially Gen AI, around making things more secure, making AI systems themselves more secure, delivering those hybrid experiences that are more immersive, and enabling our customers' needs for that, building those AI networks at scale as well. Before I leave this and go and just briefly touch on quantum, one of the things that we've always talked about, which now everybody seems to talk about a lot more, is around responsible AI. I know I've stood up on that main stage over the last three Cisco Lives and talked about, you know, what we do in responsible AI. Now I hear more and more of the industry talk about it, which is awesome. There needs to be more.
For us, responsible AI is not negotiable. As much as we look at Gen AI as this huge opportunity, we do it with a responsible, you know, kind of framework in mind that looks at everything from bias, to unanticipated content, to fake content, to IP infringement issues. We don't want to stand in the way of innovation for our developers, but on the other hand, it's not a free-for-all either. We've got a team that includes myself, our chief legal officer, and our engineering and other teams, our security and trust organization, that come together very regularly around how we should build and deploy AI models within Cisco, and how do we enable that for our customers as well.
At our global customer advisory board, in fact, a few customers approached me after the fact and said: "You know, given our trust in Cisco, maybe you should be thinking about providing this as a service in terms of, you know, I don't want to go out there and buy a foundational model and then train and deploy it. Maybe Cisco should be thinking about that, 'cause I trust you with my data, 'cause you've had vast amounts of data, but you don't turn around and try to monetize it." I thought that was a pretty darn huge compliment.
Then in quantum, while quantum use cases are still further out there, we have a quantum lab and a team that continues to work on quantum, including things like quantum key distribution, because our fear in this space is generally there could be a rogue actor or a bad nation out there just harvesting a ton of data, At some point in time, they can decrypt it using quantum. For us, when you think about classical networks and quantum networks, we believe We are the company that should take our customers to that future, where you can deploy You know, you can have classical packets and payloads on the same kind of ports that are deployed around the world.
You have quantum key distribution that keeps those communication channels secure in such a way that the keys and secrets are safe. They don't, in the future, have to have a separate path for classical and quantum networking. That's our aspirational goal as we look at into the future, and that's one that we aim to enable. You know, having been here for 23 years and done Cisco Live for the last few years, I don't think we've had this much of an expansive set of capabilities that we've released across the board and a significant number of them through organic innovation. I mean, Given my CSO role, I also am part of our M&A team, working with Scott's team, and while that will always be part of our strategy-
I'm pretty proud to say that the engineering teams, and tomorrow we'll learn more on the, on the customer experience side, have delivered some pretty fantastic capabilities this week. I know I talked a lot out there. Happy to take questions. A deep breath, take some questions if you have any.
Make sure you have a mic.
Yeah. Hi. If you don't mind, I'll start with two AI questions for you. Leaving aside Jonathan's portfolio, where we understand the infrastructure sort of upside when it comes to AI infrastructure, how are you thinking about the rest of the portfolio when it comes to AI in terms of monetization versus creating a differentiation, like, in terms of some of the security that you sort of talked about at the keynote, or Jeetu talked about? There's clearly differentiation, how do you think about monetization? Secondly, maybe this is more in sort of Jonathan's wheelhouse in terms of AI, sort of when you think about, as you said, like, as a service model, how are you thinking about how much of that increases the relevance of on-prem versus sort of going to the public cloud?
What does that do to the relevance of on-prem?
I'm happy to touch the first one. Nathan, feel free to add more, and Jonathan, absolutely. When I think about AI, there's two things that we look at: How does it enhance the current feature and functionality that we can deliver to our customers in a differentiated way, and what are the new use cases that we can do, right? When you think about in hybrid work, for example, and Jeetu and team are gonna talk tomorrow about AI assistance, what you can do about much more immersive experience. Because in hybrid work, every company is still trying to figure out what works for them. Hybrid work will be around for quite some time. That's our future. Not everybody's gonna go back to the offices.
There's more and more that's needed to make those experience a lot more sticky, so when you're not in, you know, a campus or wherever it is, you still feel like as much of a realistic experience. You'll find us focus more on that immersive experience, and customers are willing to pay more for that. In observability, for example, some of the top challenges that our customers have is, I need to be able to not just observe, you know, what my application is dependent on, but also what my application is feeding or what my teams are feeding into these large language models. Is that confidential data?
Monitoring things like, you know, APIs, OpenAI's API, for example, and thinking about a way that we can show, "Hey, you know, this is confidential data that's being fed into those, you know, into those LLMs, that, by the way, you don't own. It's not in a private instance," that folks are just looking at it and going: "You know, it's automated inspiration. I'm just looking for some help on it." These are things that our customers are willing to pay more for because they are using AI more around even how do they enhance kind of their outcomes at the end of the day, but they also want to do it in a way that doesn't put them at risk.
Every single one of our customers is also looking around the room and saying: "How do I use AI?" That's part of us enabling them. There are new use cases that we're looking at, that I think I would love for us to share when we're ready. There are brand-new use cases that wouldn't happen if Gen AI was not here. The other point I'd make is, you know, when you think about products, we want our products to be adopted, and the time to value for our customers being a whole lot, maybe in minutes, in a few minutes. Things like prompt interfaces, I would say, are a great way for us to be able to drive that time to value even better and be able to monetize.
The more our customers get it adopted faster, the more you can. You've landed it, and then you can expand and sell them the next use case. Jonathan, do you want to cover the on-prem more now? On the as a service, maybe wait for Jonathan.
Yeah. You talked a lot about simplification, and that, you know, certainly makes sense. I guess when you have 1 million customers who probably have 1 million implementations, and, you know, you're still talking about, well, if you want cloud or if you want on-premise, we can kind of do any of those configurations. What is that process like from taking it from a simplified strategy to a simplified sales process and getting your customers to kind of translate it to the end customer?
I know Jonathan touched this, on this in our... He did it very eloquently in the session that we had with press and analysts. It's really around what we are training and the way we've enabled our sales team is to really pivot to talk use cases and business outcomes. While we'll build best-of-breed products, right? It's the best widget that we'll make in terms of speeds and feeds and features, the way we're looking at is: How do I enable that outcome? I can talk full-stack observability for that matter. You know, customer digital experience monitoring is a big use case for our customers, but when you think about it, you have to go across a lot of complex areas. It's the user device, it's the endpoint, it's the network, it's internet, it's the application, it's SaaS services.
When we talk about the outcome and how the capability that we're delivering allows them to bring this together as a solution where they don't have to stitch together each and every one of these pieces. Even when I'm talking to a large financial that has 30,000 developers, I said, "Hey, are you planning on stitching this together?" He said: "No, I expect somebody like Cisco to do it. I want my developers focused more around what new services do I. Because I am increasingly, you know, the competition is folks who probably are, you know, newer into this space, so they want to focus on where they derive the most value. They're looking for us to kind of stitch this together. This is what our sales teams are pivoting and talking to, is more around how do we enable those outcomes?
Depending, Mira, on who you talk to, sure, customers, they're, you know, you go and talk to IT ops or net ops, they want, like, "Hey, what's the latest and greatest in terms of speeds and feeds and features?" We can have that conversation as well. And also, as budgets get more constrained, customers are still spending on security, and they're still spending on AI. They're still spending on more in terms of automation. Customers themselves are coming and asking us this as well. I know we're out of time. Thank you. Thanks for being here. With that, I'll hand it to Jonathan.
Almost lost my first step there. That was dangerous. Thanks, everybody, for coming and spending time with us. Thanks to everybody who's watching online or listening. Let's talk about your favorite topic in the whole world, networking. I always like to do one little example. If everybody could just take their phones out, put it in airplane mode, put it away, and then leave it that way for the rest of the day, see how you feel. That kind of explains the power of the network pretty quickly. I'll skip through that part. Fundamentally, we believe that from a networking perspective, there is going to be more TAM in the future than there is now.
I'll go through some of the reasons for that and why we believe it, but you just got a few questions and answers from Liz around AI, and we see that as a huge opportunity. I'll talk more about some of those tailwinds. Fundamentally, I started at Cisco in 1995, and back then, we really dreamed of a day when IP networks would be the de facto way of communicating around the globe. We hit that a long time ago, actually. We still, in my opinion, are at the beginning of connectivity and what that can mean for the overall planet and for our businesses and users, and we'll talk a bit more about that.
We also see that the hybrid world is here to stay, so some of us are gonna work from circular tables inside of the MGM. Some of us are going to work from our home offices, some of us are gonna work from our corporate headquarters, or all three. We realize that this is gonna be important, and as Liz was calling out, we need to have the key technology in order to enable this transition. In fact, we believe that over the last seven years, we have gotten to the place where we have fundamentally the best technology in every area in which we operate, and I'll talk more about that. We also see a whole host of market transitions happening at the same time. Cloud is obvious.
The multi-cloud is also obvious. You'll hear this more from Jeetu. Customers are not gonna put everything in their private cloud. They don't want to put everything in one public cloud either. They want to have a multi-cloud strategy. They know they cannot get locked into any one public cloud. There's an opportunity to provide connectivity services, to provide security services, observability services, that's consistent across each and every single one of those clouds, and that's another opportunity for us. We see the internet and how networks are built as fundamentally changing. I'll talk more about that. We see that our ability to go and meet our customers where they want us to be is driving that new opportunity for us. There's also how applications are actually written.
You know, if you go back 15 years, we used to be able to point at a server and say, "My application runs there." We virtualized, and you're like, "Okay, it's in that general vicinity of those servers over there." We went to the public cloud, and we go, "I don't know where it is." We went to containerization in the public cloud, and we're like, "Well, now we not only know where it is, we don't really even know which pieces and parts are everywhere." The complexity that is there requires someone to simplify, and there's no one better in the planet to help our customers understand these things than for Cisco. Then, of course, there's IoT, and I'll talk more about where that is going.
The challenge that we're seeing, that our customers are telling us about, and why you're seeing a very consistent theme throughout this entire Cisco Live event, is that it's getting harder by the day. Complexity is continuing to increase, and we actually see that that complexity will continue to increase over time. We're going to go from billions to trillions of things that will be connected, and our customers need help in simplifying all that connectivity, simplifying that security, and then overlaying observability on top of that. The way to do that is by delivering these unified experiences, and I'll talk a bit more about what that means and what that looks like. The challenge that our customers have given us is very, very clear. How are we addressing it? Well, first of all, we are going to continue.
the critical, seminal technologies in order to solve the challenges to which our customers have posed to us. This means that we have to be the absolute best in silicon. We have to be the absolute best in optics. We have to be the absolute best in software. We have to be the absolute best in simplifying and providing a common, unified experience. We have to continue to innovate at a rapid pace. I often will tell new college grads who come into us through internships, and they go: "What's it like?" I say, "Well, you all know what a treadmill is?" They say, "Yes." I say, "Okay, well, get on the treadmill." That is what innovation feels like. Like you just have to continue to go and go and go, and it never stops, 'cause the need to innovate never stops.
We also realize that we need to be able to generate outcomes for our customers. It's not about a specific feature anymore. It's not about a specific application. They are looking to run their business, and they want the network to be as simple as walking over to a water faucet, turning it on, and having water come out. You know, they want things to be that level of simplicity. We have realized that the way to do this is to simplify our portfolio, and you'll hear more and more about how we're bringing our portfolios together. Integrations between the networking portfolio and our full-stack observability portfolio, and the networking portfolio and our collaboration portfolio, and of course, between networking and security.
You're going to see more and more of that to drive those unified experiences for our customers, so they can focus on their business outcomes instead of worrying about the infrastructure that's underneath it. The way we do that is all through a whole set of choices, which I'll talk a bit more about as well. Okay, this is what I would call a cartoon-level view of operational complexity. The real view is too complex to put on a specific slide. Think about you are a user, and you could be sitting here, you could be sitting at your home, and you want to go and utilize an application. This is basically what you're gonna be going through. Of course, you've got a Wi-Fi access network inside of this room. You're gonna be going and connecting to some service provider.
From there, you have to go and get some set of network services that you're receiving right now, even though you may not know it. You then have to go to some cloud. Let's say if you're using Salesforce or some other application that happens to be on one of the public clouds. What happens when something doesn't work? Is it the Wi-Fi network? Is it your laptop or your tablet? Is it a problem with the connection to the service provider? Is one of the regions at one of the public providers having a bad day? Is my application that I'm trying to get to having a database issue, and it's causing me to slow things down? This is the challenge that IT professionals have today. When something goes wrong, they don't know where it went wrong, and how do they go and solve it?
If it's under their control or do they need to call AWS and open a ticket, or do they need to call their service provider and open a ticket, or do you just kind of wait it out? We have the opportunity, and we are today giving our customers the tools and the ability to go and simplify this problem, so that when an issue happens, they can resolve it in minutes or seconds, and not days or weeks. That, from my perspective, is extremely powerful, and no one else on the planet has the data set that we have in this environment. Okay, I'll go a little bit deeper now. The way that we're going about doing this is by simplifying the networking portfolio. I told you I started in 95, and at peak, Cisco had about 52 different business units.
I know this because I was running a solutions engineering team that was trying to put all of these products together, and I did that job at Cisco for six years. I can tell you, it was challenging. It is fundamentally different now because we build our products, starting with what the outcome that our customers are looking to achieve, and then we work in, instead of thinking about the feature and how do we push it out. We start by understanding what the customer is looking to do, and then building that unified customer experience at the beginning. Not something you have to bolt on later on. We are going to continue to drive that consistency, and we are doing that.
Vision that we're driving is what we're calling the Cisco Networking Cloud, which ties back to a Jeetu launch last year, which is the Cisco Security Cloud. The Cisco Networking Cloud is a set of unified experiences that enables both on-premise management of your platforms and operations, as well as cloud-based management for your operations. That is one way that we're gonna drive those unified experiences. We also are pushing our Silicon One technology, because, one, it's so innovative, and two, because there's significant power savings, and three, it is feature-rich across our portfolio. We started in the carrier routing space, hyperscalers, and now we have it not only in our Catalyst switches, but we also have it in our Nexus data center portfolio. To give you a little bit of.
We launched our first piece of silicon three and a half years ago, December, three and a half years ago. In general, the industry generates a new chip every 18-24 months in a given product family. Between then, three and a half years ago, and today, we now have 14 different pieces of Silicon One cutting across our portfolio. The pace of innovation is dramatic, we're not done yet. We are going to continue to innovate at an extremely rapid pace with Silicon One, you're gonna see it become more pervasive, not only inside of our portfolio, but you're gonna see it with our customers buying just the silicon as well, that's, it's been happening, it's going to continue to happen.
The third thing that really has resonated, this is where I'll get to the network as a service question, is flexible consumption models. Three and a half years ago, we also went out and said: "Hey, you wanna buy silicon? You can buy it. You wanna buy systems with no software from us? You can do that. If you wanna buy just the software with no systems, put it on white boxes, you can do that as well." In the similar vein, we have had some customers say: "Hey, look, instead of buying something from you on net 30 terms, how about I buy it from you, hardware, software, everything as a hardware subscription?" To me, that is just another example of how we can enable our customers to procure things differently. Everyone has different metrics that they're tied to.
Our technology and how our customers want to buy, however that happens to be, we will meet them there. We announced something called Cisco Plus a few years ago, which is our ability of how we go and enable not only our customers, but probably even more importantly, our partners, to go and offer network-as-a-service type technology. We're gonna continue down that path as well. That ties back to that flexible consumption. Okay, talking about meeting our customers where they are. We have customers that, for regulatory reasons, they will always have on-premise-based management tools, and we will always build the best on-premise management tools, full stop. We also believe that while our cloud management platform has been growing extremely fast, we believe that there is an opportunity to grow it even faster.
In the keynote today, I really highly suggested everyone go and take a look at our cloud management technologies because they are market leading. We do have a large amount of customers that are in the middle or they're on that journey from on-premise management to cloud management, and we want to make sure that all of those customers come with us along that journey. This is where we've added a new technology to enable our cloud management platform, which is Meraki, to go and manage our largest portfolio of switches, the Campus Catalyst platform, so you can go and now see them as part of the Meraki dashboard, whether it's a Meraki switch or whether it's a Catalyst switch.
You can go and do troubleshooting, you can do packet capturing, you can even log in and help troubleshoot into the Catalyst devices through the Meraki dashboard, which was, believe it or not, one of the biggest feature requests that we've gotten since we first announced that 12 months ago. It's GA, it's available, and we see a lot of exciting things continue to happen. We will continue to add more and more of our traditional on-prem managed products. We will continue to integrate them more and more into our overall cloud portfolio. Stay tuned for what's going to happen there. All right. Now, one of the biggest pieces that we have is I'll share with you that, and you all know this, but the platform effect is real.
The platform effect is real across the industry, and it's actually true inside of security, observability, and networking. We have seen the ability for cross-sell and upsell to work in this environment, but it only works if you have a platform. If you've got a bunch of platforms, it's very difficult to cross-sell and upsell because they don't work with each other. So we are on a journey, which we have publicly announced today, of simplifying this platform-based approach over the coming period of time. We are already well on this journey. In fact, we didn't put names on here, but we've already deprecated two of the platforms on the far left-hand side, and we are going to aggressively move from that to fewer platforms.
Now, the key is this: the goal is not to get all the way to the right-hand side, barring nothing. Like, we only want to get to the far right-hand side as long as all of our customers come with us. We are going to make sure that we make it very easy and very simple for our customers to move with us during this simplification process. We've had meetings with our global customer advisory board, we've had meetings with hundreds of customers, and the resounding feedback has been extremely positive. "This is great. We love it. Why didn't you do it earlier?
Tell me exactly when this is gonna happen, because I want to align to your timelines, 'cause I want to get to this endpoint with you. That has been a phenomenal piece of feedback, and we are gonna continue down that path. All right. More about what is the Cisco Networking Cloud? It is that integrated platform for both on-premise and cloud-based offerings. It is going to have a unified experience. As we go down the process of moving to fewer platforms, the platforms that we have right now, we started a year ago to make sure that all of the UI is the same across these platforms so that it's seamless for you. Look at AWS today.
You go and log in, there are lots of different teams building lots of different applications, but it looks similar, and you know what you need to do when you're working in that environment. We want the exact same thing happening in our environment as well. We also have been told that there are specific technology areas that would dramatically simplify the life of our customers. A few of them. One, how can you help me understand what's happening across all of these different domains, my data center, my campus, my wireless, my public cloud, my private cloud, and everything in between? We today announced the first phase of that, where we took ThousandEyes, and we expanded it from simply looking at the internet to now looking at everything and providing that end-to-end assurance.
The second is, we want to be able to offer a full topology view of all of your assets. You wouldn't believe how many customers actually don't, not only know what assets they have, but what their topology is. This sounds really simple, but this is a very, very big deal. The third one is, if I put a security policy into my security domain, how do I make sure that that policy is consistent across all of my networking domains? Jitu's team and my team have been working very closely on making sure that you just have to put policy in one place for your users, for your applications, for your things, and it will be pervasively pushed against, across all of the various elements. Then we talked about unified design as well. All right, web scale.
Let me talk a bit about that. I've told you about Silicon One. I told you about where we're going there. We are continuing to heavily invest in optics, and we are also continuing to heavily invest in optical. Part of optical, we actually include, from a architectural design standpoint, we think about Acacia as being part of optical, because it's a coherent transponder and a pluggable port, just so you're aware. We know that the choice and flexibility resonates with the web scale and hyperscale. I would say six-plus years ago, we didn't have a whole lot of business with them. Now that business has grown substantially, and Chuck and Scott share that data on our calls.
That said, the predictor of, I think, any business, at least in the Cisco space, is how much time are customers giving to you to help us understand what their deepest challenges are? I can tell you, we have more integration and more weekly conversations now than we have ever had in the hyperscale space, and we continue to be very excited about where we can continue to go with this. We also continue to be very focused on use cases and roles, but not only for our own software and systems, but for our silicon standalone. Also, we have been heavily investing in driving SONiC as an operating system, and that's being broadly adopted by one hyperscaler, and we see others moving more and more in that direction as well, and we think that that makes a big differentiation.
The second thing is around how is the internet evolving? This comes back to what you all have probably heard me talk about before, which is routed optical networks. There's really two key enablers to this. First of all, if you go back even just 10 years ago, IP routers were considered the most expensive elements inside of the internet infrastructure. By when I say internet, I mean all of kind of infrastructure, not just the internet proper. If you look at the innovation that has happened in silicon over that period of time, it is no longer the most precious resource. We built networks going back 20 years as though it was the most precious resource. It is no longer that way, which means that we can fundamentally re-architect the internet. That's number one.
Number 2, we can actually collapse layers of the internet. You do not need a separate DWDM layer. You don't need a separate OTN layer. You don't need a separate IP layer. Those can all collapse into 1 layer, giving greater efficiency, significant lower power consumption, and dramatically enhanced manageability. We see this resonating first with the hyperscalers, which has been growing extremely fast, and with the service providers as well around the globe. I had people that two years ago were like, "Yeah, we're not sure." Now we have carriers saying, "This is the future," and these are very large carriers. It is a matter of time before this becomes the predominant mode of deployment of the internet proper and most wide-area networks. Had some AI questions, I will talk about that. We talked about the hyperscalers.
We need to be able to provide them with the technology they need. There's a tremendous amount of work going on, how we need to change our Silicon One to enable that, our software, so you can build very large networks that don't drop any traffic. That's kind of a key requirement, and our extremely latency. Rakesh Chopra will talk more about that. We wanna take all of that knowledge that we've gained, and we want to make it very simple for enterprises to deploy their own AI clouds as well, on their own premise. We announced that capability today, and that is available. We also see the potential to continue to advance our portfolio.
We announced a year ago, the ability to use all of our ML and AI insights of the dramatic amount of data that we're consuming, and to give you predictability around what might happen in a network and how can you modify your network before it happens. That's been extremely well received in the portfolio. The last one is, of course, using it inside of our portfolio to make our products easier to use, easier to consume, and also make our, quite frankly, our own engineers even more productive than they are already. I'll let Jeetu talk more about what's happening in both of those different areas. I just talked about this. This is the data center blueprint for networking. It's out and available. What have we seen of all the things that we've talked about?
The power of the platform is real, and so there's a whole set of use cases, and you'll hear more details about each of many of these, I should say, tomorrow, during our main stage presentation. I'll give you a little bit of a, of a teaser. Honeywell was able to take ThousandEyes and fundamentally transform how they look at their 6,000 different sites around the world and provide a whole new level of transparency and simplicity to know where the issues are. We also have already announced integration with AWS.
One of the challenges that our customers have seen is that you can see what's happening across the global internet, and then when you got to the AWS boundary, you knew there was a problem somewhere in AWS, but you didn't know where, you didn't know why, how. Now we're actually sharing data between AWS and ThousandEyes, so we can now go and understand what's happening inside of AWS. Yet more visibility and more capability. AT&T, I will just say stay tuned. We have a very exciting announcement that is coming out with them tomorrow. Thinking about taking this simplicity-based approach across not just areas in my domain, but also in Jeetu's domain, and how we can simplify a joint customer experience to rapidly go and expand that.
We can kind of go on and on around where things are headed. With that, I am extremely excited about what's happening in the Cisco networking space, all the integrations that we have going across the various portfolios. Our plan and goal is to continue to be the connectivity and security technology provider for the entire world, and I think we're in good shape to be able to do that. With that, I'll pause, then maybe Marty or... Marilyn's over there. Marty's here. Okay.
Great. Simon Leopold, Raymond James. I wanna ask in terms of sort of the AI opportunities, how are you differentiating products? Because sort of the benefit of Ethernet is it's an industry standard.
Yeah.
how does your industry?
Well, I can give you the 30-second answer, then Rakesh Chopra will be here after the break, who has been working on this for quite some time, and he's in our silicon team, and he's been a fantastic partner for the six and a half, almost years I've been back. Basically, it is a combination of hardware and software. If you think about the way that big chassis systems used to be built, you had a line card. It had ports on it for Ethernet. You had a backplane that actually had another chip on it that would enable you to connect those various line cards to each other. Inside of the fabric, inside of that chassis, the fabric would be scheduled.
If a packet came in on this side, you don't really want to send it to the other line card unless you knew that that traffic could get out the other line card. That is a scheduled fabric. The challenge, this is one of the key innovations that we have because we're using Silicon One, that chip that's on the line card and that chip that's on the fabric card is the same chip. If you go to anyone else's technology stack, those are two different chips, two totally different silicon architectures completely. What we can do is a lot of these large AI networks, they don't want to use chassis for lots of different reasons. They want to use smaller boxes.
We have the ability to say, "All right, go and deploy a bunch of these smaller boxes in a, in a fabric," and then we can give them the software where all of those boxes act as one unified fabric. This is a technology that came in through the Lieba acquisition, and we have sustained, maintained, and grown since that beginning of time. I'll let Rakesh talk more about it, but that's the. Okay, that's maybe, like, a 75-second version.
Jonathan, last question.
Well, I'm gonna disappoint you 'cause I'm gonna build off of what Simon Leopold just brought up. This scheduled fabric thing that you're talking about... Sorry, this is Aaron Rakers at Wells Fargo. I guess the first question on this is that in the market today? I guess when I look at the forecast that Liz Centoni had presented.
Yeah.
It's a pretty robust amount of growth for AI fabric network build-outs for Ethernet. I'm just curious of, how do you foresee the inflection? What drives this change of these network topologies? Is it next year? Is it 2025? I'm just trying to understand the progression of this evolution.
I'll give it to you this way. I don't know that I can answer your question directly, but I will say that, you know, I've been looking to try to replace InfiniBand networks for over a decade. People have been before me trying to replace, but there was no fundamental impetus to doing so because it's like: Oh, they're not that big. It's not that big of a problem anyway. We could, but it's a lot of work. The conversations that I have now with hyperscalers are: I will do anything to be able to have alternative choice in this space, and whoever can show me that it works at the scale that I want it to work, wins.
That's the fundamental change, because the market itself is growing so dramatically, no one, no one who's spending this amount of money wants to be locked into any one given architecture. It is really that simple. All right, I literally just got the yank, so thank you all very much. With that, I will hand it over to Jeetu, my buddy.
Hello, everyone. I'm Jeetu Patel. I run the security and collab business. Just by show of hands over here in the room, and then I know that there's people online as well, but how many came to the keynote today? Most of you did. All right, awesome. What I thought we'd do is we'd walk you through a little bit about, you know, what the thinking is behind building a platform, how are we simplifying the platform. Of course, I'm sure you're gonna have some questions around, you know, growth and all of those pieces, and we'd like to leave some time at the end for that. Let me actually start out by just talking about the potential in the market. My headline on the market and the TAM analysis is we're not demand constrained.
There's plenty of demand, highly fractured market, many, many players in the market. No one owns dominant share, we should actually be able to more than get our fair share of it. If you think of the scale that we are at, security is a game of scale at this point. If you think about the scale of telemetry that we have and just the amount of traffic that's flowing through our network and what we're doing, it's actually pretty phenomenal. 80% of the internet's traffic flows through our network. We've got 300,000 customers. We've got 400 billion security events observed daily. There's a fair amount of volume of data, and in the world of security, volume really matters because we have to deal with this at machine scale.
You can't deal with it at human scale. The question that many of you have had is, "Hey, the market's growing at a certain rate. Why are you not growing as fast as the market?" We've said that we've had opportunity to go out and make sure that we can shore up our portfolio, and actually, today was hopefully very illustrative of that to you on the amount of work that's been done. We've probably done more innovation in the past 12 to 18 months than we've done the 10 years prior to that. It's pretty exciting to see what's happened, but it happened because there was a few structural things that we changed. The first one was we actually have completely upgraded the team quite a bit.
We have some great people from Cisco that got put into very influential positions. We also got people from the outside. If you think about Tom Gillis, who should be joining me over here shortly, both of us were doing a press event, and we're trying to play between 3 locations and try to get him into a car and to come to a certain place. Tom should be here shortly. He ran networking and security at VMware and is running our security charter on my leadership team here. The team underneath him is actually pretty phenomenal as well. Raj came from Netskope, built Netskope. Shayla came from McAfee, who is our Head of Engineering. Raj is our Chief Product Officer.
Our head of design came from ServiceNow, Compass, whose picture is not here, he works for Raj. You know, Greg Connors, who runs Advanced Systems and all of the new projects that are going on right now on the innovation side, he came from VMware and works directly for Tom. We've got the ops team, we've also got our go-to-market leader, came from Palo Alto, Emma Carpenter. Her sidekick, who's the head of SEs, also came from Palo Alto. Her head of Americas, came from Palo Alto. Ambika came from VMware. What you're starting to see is. We've got, you know, people who run the product teams that we made announcements for today.
Some of them came from Zscaler, some of them are from the networking side of the business. We've actually built probably one of the best teams that we have ever built in security, and that's a great leading indicator for what you can start to see happen in product velocity, which is then a great leading indicator for what happens on the growth side. Those are all things that kind of go one after the other. If you think about where we shine and what we're gonna do is where security meets the network is where we'll actually meaningfully differentiate, right? You know, if you think about the networking side of the house, fast and reliable kind of connections to go out and build, and you saw this in Jonathan's deck, I'm assuming.
I missed the first part of it. They're pretty hard to build. It gets even harder when you have these fast, reliable connections that are pretty complex, where you have to then figure out breakpoints where security, vulnerabilities can exist. So what's happened in this market, like I mentioned about in the keynote, is this market's evolved as a result of patchwork. You've got way too many point solutions in the market. No one actually owns a dominant share, and each threat that emerged had a new point solution for it. There's frankly, only a couple credible end-to-end platform providers in this market.
Cisco will be one of them, Microsoft will be one of them, Palo Alto will be one of them. Then we have to just make sure that we think about, like, what specifically are gonna be our key differentiators, and we'll walk you through some of that today. As we think of our role as Cisco in security, what is our role gonna be? You know, we thought long and hard about this, and we said, if you think about a hybrid multi-cloud future, where everyone wants to make sure that they get the benefit of the public cloud economics, but don't get public cloud lock-in. Most companies at least have one is too few public cloud providers to bet on, and three might be too many.
Usually, most will have at least a couple public cloud providers that they'll actually bet on. The way that it's worked so far is, you know, the four major computers in the world, you think of Microsoft, Amazon, Google, and then you've got your private data center. Your networking and security stack tends to be very siloed in each one of these public clouds. A role that we can play very credibly, that actually would be completely aligned with the customer's interest of not getting locked in, is can we abstract networking and security above that, so you can acquire and steer any and all traffic to any of those providers and make sure that you can transcend policy across any of those providers, and the policy should persist? That would be a, you know, meaningful value add to the market.
It's something that we know how to do well. There's no business model conflict. You know, everyone, even the public cloud providers, are saying that they're gonna be that they will be multi-cloud. The reality is there's a structural disadvantage that they have, because there's no incentive for them to be multi-cloud. Whereas over here, for us, we have no dog in the hunt over there, so you actually have a much more credible story with the customer to be multi-cloud. We launched the Cisco Security Cloud last year. Many people asked us, saying, "When are you gonna start delivering this?" What you saw today was a lot of announcements that had a ton of momentum around building out this vision of the Cisco Security Cloud.
Then Jonathan, you know, launched the Cisco Networking Cloud, and those are gonna be just fused together. We will have many, many more point integrations between those two platforms, and they will look and feel like the same because we've standardized on the design language, so we've standardized on a common policy object model. All of those pieces are starting to get done so that we don't have. You know, we don't come across as a holding company. We wanna come across as a platform company, right? Even the acquisitions we make, we will make sure that we strip out the name of the acquisition. They'll be integrated into the platform. That's the way that we're gonna do it. We're not gonna go out and have each one of these product brands that'll actually stay on by themselves.
That's frankly, one of the areas where, you know, we have not done it that way in the past, that's something that we wanna do. The only brand that you'll actually see is the Cisco Security Cloud. Now, as you think about what we've done for the Cisco Security Cloud, there are 4 key areas, like I talked about earlier, that we wanted to focus on. How do you protect the user? How do you protect the cloud infrastructure? How do you protect applications? How do you protect breaches from? When the breach happens, how do you detect it, and then how do you respond and remediate that as fast as possible?
you know, like I said earlier, one thing that we wanna really be clear, be clear on is, what are we gonna be able to do that's uniquely different for Cisco, that the others can't do as well as us? That, in my mind, is a very clear, when security meets networking, that intersection point is where we have very, very unique set of capabilities. What we'll do over there is we'll make sure that we can provide a common design experience so that Meraki, and the user and the administrator of Meraki, doesn't feel like they went into this very jarring, different experience when they go in to set a policy with with the Cisco Secure Access. They can actually do it with a common design language. It's all gonna be you know, kind of fluidly integrated.
We will have common policy objects underneath them so that you can have, you know, more integration at a very. There'll be shared telemetry, and we'll have open APIs. That's, that's the goal of making sure that these platforms come together. When they come together, the personas we are serving, there are four personas. We are serving the end user, for which, you know, I joke around on this one, but what we wanna do with the end users is just deliver the most boring demo possible for the end users. They shouldn't have to do anything. They open up the computer, they connect, they get to work, you know? The second thing is IT. What do we do to make the IT job easier?
The third one is the SOC center, where you might be monitoring for breaches. How do we make sure that we have a great incident response database with Talos, and we can actually pull that data and not only go out and help the SOC analyst directly, but we can actually make sure that every product is instructed with that data as well, for the breaches. Lastly, the DevSecOps side of the house, right. The announcements we made that you folks saw, and I'm just gonna skip through this fast in the interest of time. You know, we gave this analogy of, if you wanna turn on a faucet of water, you know that you wanna. You don't ask which pipe you're gonna get the water from.
You just turn on the on the faucet, and you should be able to get water, but yet that's not the way that we do things when people have to go out and connect to their applications at work. If I have to connect to an application in my private data center, I have to log on to VPN. If I have to connect to my application, my Salesforce or Workday in the cloud, I need to make sure that I use ZTNA. If I have to connect to Netflix, I go out and do a direct internet access, and that just seems like it's way too much cognitive load. So what we wanna do is make sure that that does not end up being this kind of, you know, overly burdensome, like, a you know, burdensome task.
That's why we launched this kind of notion of Cisco Secure Access, which is our version of SSE, you know, which is this Secure Service Edge, which is what Gartner calls that category, so this is our solution for that. It's a single, you know, experience, regardless of what location you're logging in from, what network you're logging from, what device or what application. It's one experience. You have no idea. You just turn on your computer, and it works. What we'll also do is make sure that these connections are very, very kind of easy to use, and they will be focused on with high-performance throughput. We'll have, you know, we've got about 42 different pops right now.
Think about us having another 75 or so with Webex, that we'll be able to leverage that infrastructure. We'll use a partner ecosystem so that it's just a few milliseconds away from any pop at any point in time, and that's the goal as we move forward. Once we do that, we'll also tie that into ThousandEyes, so that if something goes wrong in the experience, you can actually pinpoint that down to the smallest degree and say, "This is what we need to do, to go out and solve that." That's the first big announcement. We believe this is a huge opportunity. This is gonna be GA in October, at the end of October. It's gonna be in early access in by the end of June. This is a very...
This is imminent right now in how it's happening, and we've actually got a lot of customers signed up for early access as well. Then, the second thing was around this cloud, you know, kinda cloud protection side, where this notion of having the private cloud and public cloud, one of the things we talk about is this notion of having a ground to cloud offering. You know, we can protect the private data center, we can protect the public data center. They speak different languages. One speaks IP addresses, one speaks services.
When an application in AWS wants to talk to an application within the private cloud, which is what 95% of the applications that sit in the public cloud need to have some access to some resource in the private cloud, we can give them that same zero trust affordability, so that you can have access to that application and only that application. If they wanna go out and connect to anything else in a different environment or a different application, we don't give them that access. Retail application, you can connect to your inventory application in the production database, but you can't connect to that same inventory application in staging a development, or you can't connect to a customer database in the production environment.
We've given you zero trust, least privileged access, only to the application that you're looking for, and if you port that workload from one cloud to the other, guess what? The policies transcend. Going back to my slides earlier, that layer above the cloud providers, we wanna acquire and steer any and all traffic to any of the providers, and persist the policy that you might have for security. And then, essentially, this is essentially what we have with Multicloud Defense, and that's available today. Well, once again, the pace of innovation and the velocity... By the way, this same management console, CDO, that manages our firewall, also manages this. It's, like, one single management console that does that.
In addition to that, we will continue to keep investing in our data centers, and we just launched a high-end version of our firewall, Firewall 4200. It's twice as fast, it's more reliant, it's more resilient, and we will have this notion of an Encrypted Visibility Engine capability, which is through machine learning. What we're able to do is, most traffic is end-to-end encrypted. There used to be this kinda technology in firewalls, which was deep packet inspection, where you would look and inspect the packet to know whether there's malware. The reality is, when it's end-to-end encrypted, you can't do deep packet inspection. What do you do?
Well, you can infer, based on the movement of the packet and interpret whether or not that actually looks like a breach, and that's what Encrypted Visibility Engine does. We just enhanced that capability in our latest version of the OS. One of the challenges that we have is, we have had a pretty low NPS score in Firewall. The reason we've had a low NPS score in Firewall, is people are actually in the older versions of Firewall. They haven't moved to the newer version. One of the things we're doing right now is, very heavily making sure that we actually migrate our customers from the older version to the newer version. You have an insertion point now with Multicloud Defense. Even if a customer is using Palo Alto firewalls, they don't have a multicloud, you know, kind of strategy.
We can go out and make sure that we can do Cisco Multicloud Defense, when it comes time for renewal, you can definitely give them an alternative to go out and move over from them to a single platform with a single management plane. That's basically what we've done with the firewall. It's basically the most amount of functionality packed into 1 rack unit kind of form factor. Then we talked a lot about. I'm gonna have Tom Gillis come up over here, because I think we're both playing relay and tag, so we're gonna go from 1 to the other.
The two things I wanna talk about over here is on what we're doing with Generative AI. Tom runs our security business, like I mentioned earlier. On what we're doing over here from a policy standpoint is, we are making it brain that simple for people to go out and set policy with natural language, rather than having. Like, Tom was giving me this example recently, where we were talking to a customer. They have 6 million policies, right? 6 million policies in the firewall.
Yeah. That's a lot of policies.
It's a lot of policies.
Yeah.
It just gets cumbersome to go out and manage these policies. Firewalls are hard to upgrade. What we wanna do is make sure that generative AI can be used over here. You just put in a natural language instruction, and the beauty about this is, the demo that you saw in the keynote today, for those that are online that didn't see it, you just put a natural language instruction.
The system goes back and forth with this policy assistant that says, "Do you want to set a policy for Tom, who's a new employee, to give him access to the clean room, as we are gonna go out and buy a company, VaultX?" You know, what it will do is, it'll ask you a couple different options, and then you pick the options you want, and it'll say, "Do you wanna deploy the policy?" That's a beautiful way to go out and set, you know, kind of automation of a policy. Over time, what you'll see is, we'll be able to abstract that layer, so that any other policies we have can all kind of fit underneath that layer as well, and it'll all be...
You know, think about most of our user experiences over the course of the next few years, will actually have a Generative AI component to it, where you'll have a prompt-based interface tied with a GUI-based interface. That'll actually make massively simple, you know, kind of, it'll simplify the way that people use these products, and the dexterity that's required is gonna be much lower than what you would have otherwise required to go out and set firewall policy. That's that. I'm gonna hand it to Tom at this point so that I can make a dash for the next meeting. Hopefully, this was useful. Thank you all for coming, and, you know, I'm sure that Tom will be able to answer questions at the end as well.
Thanks, Jeetu.
Okay, thanks. Thanks.
All right. Cool. All right. Hey, everybody, I'm Tom Gillis. I run the security products that work for G two. I can keep walking through. I think we covered most of the highlights here. I mostly was here to answer any questions you might have. A big push for us is building these products with APIs, and so the APIs allow us to start to move security into what we call the CI/CD pipeline, which is software developers, when they start thinking about how we build an app, can think about how does security interact with that application, and that all lives in software. At go time, you push a button, and all this stuff is automatically configured. Nobody opens a ticket. Nobody's updating firewall rules, right? It's done ahead of time, and that makes a more efficient DevOps-type deployment.
At the end of the day, the reason we do all this stuff is to stop attacks like ransomware. What we announced at the RSA show was Cisco's XDR. XDR is extended detection and remediation. This is a really big opportunity for Cisco. It's generally a market that we consider to be white space, right? Generally, customers haven't chosen an XDR solution. There's certainly lots of competition. If you walked around the RSA show floor, almost every vendor was talking a story about XDR. How are we different? The way that we distinguish ourselves is we are probably the only, maybe there's You could argue, one other company in the world that has telemetry from each one of these domains: email, web, endpoint, and network. If you think about an attack sequence, it usually starts with email.
Our incident response team said that 80% of the ransomware attacks we saw last year came from a phishing email. It's not those goofy, badly written emails from that friendly foreign prince who just needs to borrow your bank account. You know that guy, "Please, could I borrow your bank account for a few hours?" Now, these are emails that, you know, with tools like ChatGPT, they're gonna be coming from someone you know, referencing someone you did in perfectly written English, so much harder to figure out friend from foe. You need to look at email. The email will take you to a website. You need to look at every single click on that website. That website's gonna look indistinguishable from, say, a photo-sharing site, right?
You click on a link, and all of a sudden, code starts to run on your machine. When that code runs, there's a tool in Windows called PowerShell. If we see PowerShell spawning a new process we've never seen before, that there's legitimate reasons why that might happen, 80% of the ransomware attacks, just like with the phishing, 80% of the ransomware attacks we saw last year came from an unknown process that spawned out of PowerShell. If you could see these events side by side, weird-looking email, funny-looking website, PowerShell runs, strange process is now connecting to a customer database, that's clearly a ransomware attack. If you only saw one of those domains, you're missing more than half the picture.
This is really our unique advantage in the industry, and we have an awful lot of telemetry to back that up, right? 400 billion security events per day because we process so much of the world's email, so much of the world's web. Of course, nobody knows a network like Cisco, and, you know, a whole large, deep technical bench to understand these types of attacks. Just as Jeetu talked about earlier, how we're using AI to really streamline and dramatically improve the process of managing a firewall, we're doing the same thing for the security analysts. In the SOC, there are these analysts that process alerts, and they process tens of thousands of alerts every single day. They just come just flowing into the security operations center.
One of the unique advantages we have is because we run an incident response business, our IR team, not the investor relations, the incident response team, writes up a case study of, "Here's what happened." We have 10, 20 years of written case studies on how we responded to a particular attack. When we feed this stuff into a large language model, it creates, like, a smart security system. It can recognize patterns and be like, "Oh, you know what? We see this moving to here, and then it, you know, we see that PowerShell running, and then it's making a move to this customer database. Based on what I saw on the, you know, that 20 years of history, you should probably be setting up a honeynet.
Would you like to turn on packet capture?" Having the computer suggest to the analyst actions to take, we think, can create an order of magnitude increase in the efficiency of the SOC analyst. This AI stuff, I know there's a ton of vendor hype around it, but it's a pretty big deal for us. You know, it creates a whole new way of thinking about how our customer is interacting with our system, right? Imagine a firewall that you can actually talk to. That's basically what we're doing. Imagine a security analytics platform that can actually alert you, like, "Hey, this looks weird. You should pay attention." It's that level of interaction which we think is gonna be transformative. This is really our focus, is platforms. You're gonna hear us talk a lot about it.
You'll hear us talk about exciting news specific to this platform strategy. There's three components to the platform. There's all the stuff that we use to protect users. That's what secure access is, the email thing I talked about, all of that, think of that as one product. There's all the stuff we use to protect hybrid cloud infrastructure, so private and public clouds. That's the firewall and the Multicloud Defense all fit into that cloud protection suite. Then there's the stuff we use to tie those two together and look for a breach, should it get through these defenses. That's the pitch. Questions? Yes.
Hi, Ben Reitzes, CIS with Melius. I wanted to ask, the SOC assistant.
Yeah.
Makes a ton of sense to me.
Yeah.
When do you think people will be using it en masse? Does that create an ARPU lift for you? Do you charge more for that capability?
Two parts to that. This is always going to be an assistant, right? It will always be helping the administrator, we think, creating greater efficiency. The firewall capability, we are aiming to have the end of this calendar year. Like, we are actively working on that. The SOC assistant, we think we will have into next calendar year. Your question was en masse. You know, give it six months to really, like, get adoption. The other thing I think, you know, if I am being honest and candid, is that this is radically new technology, and we are going to have to adjust and tune it, right? Like, if a firewall policy administrator is not right all the time, people are going to learn to trust it, and how do I catch when it makes mistakes, right?
There's going to be a tuning element that we go through. It's a pretty big step.
If it takes off.
Yeah.
Do you charge more for it, or is it already in the product?
Well, it's definitely going to be.
Is it a work in progress, and you'll figure out as you go?
Yeah. It's, the honest answer is, it's a work in progress, and we'll figure out as we go. Just so you understand the context, it's a feature on a product.
Yeah.
I don't think this is something you buy separately.
It's a feature that you would agree makes it entirely more usable to many more organizations.
Entirely more usable. It's a very, very meaningful feature on the product, like a very clear ROI. Do we make that a different tier of our product offering? Maybe.
Uh-huh.
You know? One of the things we're trying to do is really simplify the way we take our products to market. We're really trying to get to the point where we think about a user suite as one thing, and there's kind of small, medium, large flavors. This would be in the, let's call it the large flavor or the high, medium, low flavor, right? We'll put it in the top end of our bundles. Bundles are the way that Cisco sells effectively.
Okay.
Yeah.
With regard to what you've announced cumulatively back to the RSA, you know, with XDR.
Yeah.
I guess, do you feel like these products start gaining momentum and impact revenue?
Yeah. Well, the RSA product, we announced it in April, and it was going into beta right at the time of announcement. We've got 50 customers running on it, which is we can't support anymore, so we're getting strong demand signals from that. It'll go into general availability in July, and we'll start booking revenue on that right away. You know, now at your level, when does it impact revenue is a different question. Look, we'll start to see meaningful results with it, you know, as early as the beginning of next fiscal year for us. Yeah. Same with the, with our SSE, the secure access thing. Like, that's right in our wheelhouse. This is something that we're selling to our AnyConnect install base.
Customers are lining up for this stuff. There's a little. If I'm being candid, customers' view is like: "Hey, Cisco, you're kind of late here, but we really want it from you." We're backlogged with demand for that and, you know, very excited about the number of customers that we have in early access. The GA date for the Secure Access is October 27th. It's a big summer for us. Multicloud Defense available now, the XDR available this summer, and then the Secure Access available the end of October.
The value proposition of networking and security.
Together.
Being at that intersection is?
Yeah.
You know, it makes a lot of sense. I think there's a lot of security companies trying to make that same kind of value proposition for why they should move more into networking.
Yes.
I think you guys.
Don't say a lot, but there's some.
Yeah.
Yes, yes.
But.
We don't like them.
Right.
Yeah.
You guys are more differentiated in networking, going into security and making that value proposition, but how do you better defend against people making the value proposition that security should move more into networking?
Yeah, it's, I don't want to sound flip, but there's one simple answer, is we need to be great at security, right? We haven't always been, but that's why I'm here. Like, I'm a product person. That's why Jeetu is here. We're product people, and we're focused on building great products first, and everything else comes from that. All the revenue forecast and market adoption and customer, it starts with, like, build something amazing. Our customers want that from Cisco. That's how we came to be as a company in the first place. So, you know, we have to have that level of commitment and passion to build amazingly good security products. Now, we can't be all things to all people. Security is a big industry. There's lots of categories I don't even want to touch.
The places where security meets the network, things like secure access, things like firewall, we can be the best in the world, meaningfully and measurably better than the competition, and we will.
Hey, maybe if I can just clarify. When you say the platform will be GA in end of October, does it mean all of your point solutions are at that point?
I'm sorry. What I said was GA in end of October is the Cisco Secure Access.
Right.
Secure Access is an advantage of significant product integration. In order to deliver that thing that we demoed at the keynote and have been talking about, the VPN meets zero trust, that involves web proxy, firewall, browser isolation, endpoint posture assessment, ThousandEyes. There's, like, They used to be disparate products that we brought into one integrated solution, one integrated data path, right? We scan the bits once, one management console, and that will be available October 27th.
Okay.
I should say end of October.
Yeah.
Yes.
When I think about the entire then sort of product solution portfolio moving onto one platform.
Yeah.
Is there a sort of timeline of how you think about sort of going through with that? Sort of a second part, when you think about sort of customers adopting it.
Yeah.
Let's say they say, "I'd like some parts of the platform still. I'm not completely on board with buying all parts of the platform.
Yes.
How does the sort of experience look like for the customer, particularly as multiple vendors come in.
Yes.
in different parts?
Let's be clear. The Secure Access is an example of a tightly integrated solution. If you remember the slide I showed up just a minute ago, where I said, let's talk about user suite, is a super set of what we have in Secure Access. There's even more. There's email, right? Which is a huge differentiator for us. There's EDR. Pulling all of these pieces together, they won't necessarily be managed by one interface. Like, it doesn't make sense to build one console for email and for, you know, sort of network. Those are just two different things. Integration where integration makes sense is what we're actively working on. Expect to see more from us in Q1 about how we package and bring that to market.
We're gonna increasingly talk about our product portfolio in terms of these three suites: a user suite, the hybrid cloud infrastructure suite, and then, you know, what we think of as the breach detection suite or the analytics piece. Does that make sense? The Secure Access is part of the user suite, but it's not the only part, right? There's a couple more pieces, and we're actively stitching these things together in a sensical way.
One last very quick question, so we can get to a break.
Eric Suppiger with JMP. Cisco has is overcoming some brand credibility on the security side. How are you going to demonstrate the effectiveness of this solution?
Yeah
To the prospective customers?
Yeah. I'll be very, very candid. Cisco has had some credibility issues with, you know, the quality of the security and also the efficacy of the security products we build. It's really. I'm a security person, you can tell, right? Like I came from VMware. I ran the security, business there. It's really encouraging to me how customers want it from us. Like, this Secure Access thing I talked about, like, we, I have way more customers lining up to get their hands on that than we can even accommodate. You know, the world needs this integration of network and security together, but they can't compromise on the quality of security. When we get it right and deliver great security products, our customers are super willing to believe, absolutely leaning in. We've got to deliver that.
Remember those dates I just told you? Multicloud Defense, available now. You can, like, turn it on and measure it. That's significantly ahead of the market. There's a couple startups that are out there doing that. We're ahead of them, and so a very differentiated solution that you can see right now. Our Secure Access is gonna be differentiated versus Zscaler and Palo Alto. Were we clear on how that works? Right? The advantage we have is that we take legacy apps that need a VPN, as well as modern apps that live in a zero trust framework, and we accommodate both of them with a seamless end user experience. None of our competitors do that today. I did a show of hands with customers who were in an event yesterday.
I said: How many are running some sort of a zero trust product today? 100% of the hands go up. How many are also running VPN? 100% of the hands go up. It's this kind of two-headed dog that everyone is, you know, managing in the market. We make that one thing. Wonderful differentiation, simple to understand, easy to get your head around, and, you know, it comes from Cisco. I think if we just keep executing with this drumbeat, there's a wind at our back here where the customers are looking for this stuff from Cisco. Their expectations are high, but they're looking for it from us, and we're ready to deliver.
All right. Great. Thanks, Tom. There'll be some time to ask Tom questions later.
Okay.
We're gonna take a quick 10-minute break, and then we're gonna reconvene.
I'll be here during the break. You can ask me questions.
Yep, yep.
Okay. Thanks, everyone.
We'll reconvene after the break, to go over Silicon One and that advantage it provides for AI and many other opportunities.
Cisco has made it its purpose to power an inclusive future for all. Where will we be in 50 years? Let's go see. Cisco, the bridge to possible. Humans and nature, we're in this together. Nature has given and given. It's our turn to do more. Cisco Smart Building Solutions and our partners' technology benefit both humans and nature. Catalyst switches connect securely, delivering Power over Ethernet, reducing costs and greenhouse emissions. Cisco Wireless and DNA Spaces use intelligent automation, creating efficiencies that help the workplace and the planet. Collaboration tools enable hybrid work, decreasing environmental impact. Sustainability is essential to powering an inclusive future for all. That's why Cisco is committed to achieving net zero emissions by 2040. Between meeting human needs and a sustainable future, there's a bridge.
Cisco, the bridge to possible. A cyber attack can grind everything to a halt. Cisco Security keeps your network and your company moving forward. Because if it's connected, it's protected. Cisco.
If we're to build a bridge to an inclusive future, then getting health care to everyone, everywhere is critical. Take rural Europe, where local doctors leaving for big cities is creating a medical desert. For patients left behind, many lack the mobility or the flexibility to reach critical urban appointments. The remedy, it turns out, is as much a technological marvel as it is a medical one. Meet Medibus, a state-of-the-art clinic on four wheels. Designing such a wonder came with its own set of challenges, taking everything Cisco knows about mobility, connectivity, video conferencing, and security into account. Together with partner, Deutsche Bahn, dispatching it from the cloud to create a 21st century lifeline.
No area is too remote, no diagnosis or specialist unavailable, all because one company dared to wonder if the road to better health care could literally be the road that runs through town. That's the inclusive future. Cisco, the bridge to possible. Where will you be in five years? Where will we be in five years? In 25? In 50? Let's be here, and here, with her, and him, and they. Let's connect them. Let's connect everyone. Let's deliver technology that gives them access to power opportunity. Let's set a new standard for data security and personal privacy. Let's change the system, promote equality and fairness in the workplace. Let's tear down the barriers to social justice for a more inclusive world. Let's clean house, zero carbon, zero waste, because the health of our family is tied to the future of our home.
Let's gather resources and partners, steer toward our greatest challenges, and accelerate for the benefit for all. Cisco has made it its purpose to power an inclusive future for all. Where will we be in 50 years? Let's go see. Cisco, the bridge to possible. Humans and nature, we're in this together. Nature has given and given. It's our turn to do more. Cisco Smart Building Solutions and our partners' technology benefit both humans and nature. Catalyst switches connect securely, delivering Power over Ethernet, reducing costs and greenhouse emissions. Cisco Wireless and DNA Spaces use intelligent automation, creating efficiencies that help the workplace and the planet. Collaboration tools enable hybrid work, decreasing environmental impact. Sustainability is essential to powering an inclusive future for all. That's why Cisco is committed to achieving net zero emissions by 2040. Between meeting human needs and a sustainable future, there's a bridge.
Cisco, the bridge to possible.
A cyber attack can grind everything to a halt. Cisco Security keeps your network and your company moving forward. Because if it's connected, it's protected. Cisco.
If we're to build a bridge to an inclusive future, then getting healthcare to everyone, everywhere is critical. Take rural Europe, where local doctors leaving for big cities is creating a medical desert. For patients left behind, many lack the mobility or the flexibility to reach critical urban appointments. The remedy, it turns out, is as much a technological marvel as it is a medical one. Meet Medibus, a state-of-the-art clinic on four wheels. Designing such a wonder came with its own set of challenges. Taking everything Cisco knows about mobility, connectivity, video conferencing, and security into account, and together with partner Deutsche Bahn, dispatching it from the cloud to create a 21st-century lifeline. Now, no area is too remote, no diagnosis or specialist unavailable, all because one company dared to wonder if the road to better healthcare could literally be the road that runs through town.
That's the inclusive future. Cisco, the bridge to possible. Where will you be in five years? Where will we be in five years? In 25? In 50? Let's be here and here, with her and him, and they. Let's connect them. Let's connect everyone. Let's deliver technology that gives them access to power opportunity. Let's set a new standard for data security and personal privacy. Let's change the system, promote equality and fairness in the workplace. Let's tear down the barriers to social justice for a more inclusive world. Let's clean house, zero carbon, zero waste, because the health of our family is tied to the future of our home. Let's gather resources and partners, steer toward our greatest challenges, and accelerate for the benefit for all. Cisco has made it its purpose to power an inclusive future for all. Where will we be in 50 years? Let's go see.
Cisco, the bridge to possible. Humans and nature, we're in this together. Yet nature has given and given. It's our turn to do more. Cisco Smart Building Solutions and our partners' technology benefit both humans and nature. Catalyst switches connect securely, delivering Power over Ethernet, reducing costs and greenhouse emissions. Cisco Wireless and DNA Spaces use intelligent automation, creating efficiencies that help the workplace and the planet. Collaboration tools enable hybrid work, decreasing environmental impact. Sustainability is essential to powering an inclusive future for all. That's why Cisco is committed to achieving net zero emissions by 2040. Between meeting human needs and a sustainable future, there's a bridge. Cisco, the bridge to possible.
All right, we're gonna go ahead and get started. I'm gonna introduce our next speaker. He's gonna close us out. He's gonna talk to you about Silicon One. I'm gonna turn it over to Rakesh Chopra, who's joining us virtually. He's gonna talk about the advantage we have with Silicon One and what we're doing with AI, which I know is super top of mind for everybody. With that, Rakesh, over to you.
Thanks very much. I appreciate it. As we mentioned, my name is Rakesh Chopra, and I'm here to talk to you about Cisco Silicon One. Before we get into the details, I just sort of wanted to comment that, you know, frankly, I didn't use to have this gray hair. I've been in the industry for a long time, and I've been at Cisco for 26 years, and to be honest, I've never seen anything like Cisco Silicon One. Frankly, it's the reason I'm so excited about both the technology, but also what it means about the industry. With that, why don't we go ahead and jump into the slides?
Let's start off by talking about what it means to our customers and what they're going through. As new services like 5G, augmented reality, virtual reality, and AKB are sort of coming on board, we're starting to see this exponential growth in bandwidth, but it's happening at a time when there's a lot of other things going on. It's augmented by this notion of sort of remote work that we've all become used to with COVID. Frankly, the fact that I'm talking to you remotely is a good proof point that it's here to stay. We're also sort of seeing a move to the cloud that's sort of been going on for the last 10 years. All of this is hard to do if our customers are actually constrained.
You can imagine that that's even more difficult to do when you're trying to do this on top of a sort of complex and divergent network infrastructure. On top of that, we're living in a world today where there's environmental constraints, and so there's a new focus now on limiting our carbon and physical footprints. All of this stuff that I've talked about from networking is true for web scalers as well, but actually it's going through what I like to call a tectonic shift. The rise of AI is causing a massive infrastructure build-out for the web scalers that is unlike anything that we've seen in the past.
They have all the challenges that the general networking folks have, but on top of that, they've been using sole source technology that, frankly, isn't designed to scale to the types of new challenges that they're facing. Next slide, please. I think what's interesting here is that Cisco is in a very unique position in the industry because eight years ago, we realized that as an industry and as a company, we needed to evolve, and we therefore started what's now called Cisco Silicon One. Our goal here was to converge the chaotic and disjointed set of networking technologies that have driven complexity throughout the infrastructure and come up with a single architecture that can be deployed across the network, across routing and switching, across form factors, and across different customer types.
We realized that in order to really bring simplicity to the network, we needed to not just sort of converge the top of the cone, like Jonathan talked about earlier, but we also needed to converge the bottom of the cone with silicon. The combination of those two is really a huge simplification for our end customers. We realized that at the end of the day, convergence actually isn't enough. You also have to have best of breed. Silicon One has unmatched programmability, performance, efficiency, and flexibility. Now, we take this converged architecture and we build multiple purpose-built devices that are deployed across the network, addressing multiple different markets. As again, as Jonathan said, and Liz said, at the end of the day, we also have multiple business models.
You can buy our technology in a full system, or you can buy our technology as silicon only or anything in between. The culmination of all of these basically gives our customers, at the end of the day, one experience across the entire network without any sort of compromise. At the end of the day, that is the basic value proposition that we're bringing the market with Silicon One. Next slide, please. At the root of all of this, what we understood when we started Silicon One is that we've transitioned in the industry from a days where small technology charts that were limiting what we could build, to the fact that power is actually now the limiting factor in the industry.
If you look at what limits us in terms of what systems we can build, it's all about how much power we can deliver and how much heat we can sort of evacuate from the systems. That creates what I like to call a technology imperative for us to solve that problem. Interestingly enough, our customers were also out of power in their rack, in their room, and in their facilities, and that creates what I like to call a business imperative to solve the problem. As I think we're probably all intimately aware, our planet has only a certain amount of resources, and that is also creating a moral imperative to solve the problem. Taken together, we have these sort of three separate independent imperatives, all culminating at the same time for the first time in the industry.
This created a perfect catalyst for us to innovate in Silicon One. As we think about Silicon One and what we've done with Silicon One, we took this to heart into how we designed it, how we laid it out, and how we use it in our products. It's really driven by the belief that we can actually deliver sustainability through technology. Next slide, please. When we look around at what we were doing at Cisco and what all of our peers are still doing, we've found that we're all making the same mistakes and repeating history over and over and over again. It was clear that we frankly needed to do something completely different.
Said more simply, what we really realized was, if you approach a problem with the same organization and the same technology, at the end of the day, you're going to get the same outcome. What did we do? Next slide, please. We created a new organization that is focused on building one architecture across the entire network or, and across all the business models. Said more simply, we set our scope, as opposed to a narrow scope like everybody else in the industry, we set our scope very, very wide. The next thing we did was once we adjusted the scope, we realized that there's actually technology limits, which stopped us from achieving that goal.
We had to invent new technology, very low down in the Silicon layers, in order to be able to converge deep buffered routing and lean and mean switching. Finally, what we did was we took the large investment dollars of Cisco over eight years, and we poured that on top of the technology. The culmination of these three, I would contend, is actually nothing short than a fundamental change in the industry. What we've created, at the end of the day, is the industry's truly scalable networking Silicon architecture. I realize that that is a very, very bold claim, but as you begin to look at the details of Silicon One and begin to see how quickly we're moving, you'll start to realize that that is actually really a true statement. Next slide, please.
To better understand what we've managed to accomplish, I think it's good to look back at where we were as Cisco and where the industry still is, and then think about what we sort of evolved into. It used to be that requirements would drive unique architectures. We'd have one architecture for deep buffer routing. We'd have one architecture for switching. We'd have one architecture for HPC. On top of that, what we found was that based on the organizational structure that existed, there was a multiplication effect of these architectures. Different companies build products going after the same market segment. Even within a company, through acquisitions, you might have multiple product lines creating overlaps. Finally, whether you're a full system vendor or a merchant silicon provider, you're probably also building unique architectures.
Now, what happens here is that you have a bunch of different technologies and a bunch of different architectures, all who have reasonable success in the market, and you continue to invest in those. That's what creates what I like to call the sort of technical debt problem. It becomes too expensive to move away from these legacy architectures, and at the end of the day, it's really the classical innovator's dilemma. Do I focus on something for the future, or do I focus on what is making money today? At the end of the day, if you add this all up, I would contend that there's over 20 silicon architectures today in the industry, and that creates a massive level of confusion and complexity into the networking infrastructure. We at Cisco now, we have a fundamentally different value proposition.
Our goal here is not to be the 21st architecture. We want to be the 1 architecture, and the way we do that is requirements no longer drive unique architectures. They drive unique devices that we build out of Silicon One. We're erasing the boundaries between merchant silicon and custom silicon. You can now buy Silicon One as a full system, like we always have at Cisco. If you want to buy a white box from us and put your own software on top, that's okay as well. If you want to buy just our silicon and build your own hardware and software on top of it, that's okay as well. Finally, we did take the bold step eight years ago to wipe the slate clean and start completely over with a brand new architecture to solve for the future.
We invested a lot of money and came up with a lot of hardware innovations to solve that problem. At the end of the day, what this translates to is a faster pace of innovation and a better and easier to maintain network. Next slide, please. Now, what's a little bit unintuitive about the strategy that I just talked about is how this creates sort of a multiplication effect and compounding on itself. We end up creating differentiated products which add value to our customers, similar to what I just talked about. There's another benefit of convergence. The other benefit of convergence is that it allows us, inside of Cisco, to leverage the IP that we generate across many different devices and across many different products. That saves us cost, that improves the quality of our products, and it means that we have an accelerated development time.
It's not just the types of products and the IP that we're generating, it's how we build our silicon that is very different than it was before. Everyone here on the call probably realized that Cisco has been building silicon for a very, very long time, as do a lot of other people in the industry. The way that we used to build silicon and the way every system vendor besides Cisco builds silicon today, is what I would call the classic ASIC model. We worked with back-end providers to do a big portion of the development, and we would do part of it. Today, with Cisco Silicon One, we are a full standalone silicon company contained within the confines of Cisco, operating in what I call the COT or the customer-owned tooling model. That means we develop all of the IP.
That means we do all of the physical design. That means we buy wafers directly, and it means that we run the full quality line. That means at the end of the day, there's a much smaller margin stack on top of the silicon that we develop than other people in the industry. One of the questions that I get quite often is: Why is Cisco selling silicon direct to our customers, and other system vendors aren't? There's really two answers to that. The first is, at the end of the day, their technology is not competitive, and so even if they tried to go out and sell it to the market, they would not win. The second is what I just talked about, which is their development model does not allow them to get to the cost points necessary to win in that market.
The nice thing about the cost points that we talked about here is not only does it help us win in terms of entering silicon-only markets, it also sort of helps our mainline system business in getting lower cost points and increasing our margin. We take these innovations, we go out to the market to win customers. We are no longer here to tell our customers how they want to consume our technology. We give them the flexibility. Again, if they want to buy full systems, white box or silicon only, those are all available from Cisco. Actually, interestingly enough, Cisco is the only company in the industry who offers all of the business models. Everybody else in the industry is operating on one or two of those business models only.
That flexibility is really, really important to win the web-scale customers that we talked about before and earlier in this call. What's important about sort of these web-scale customers is that not only do they drive large. They're actually now the ones who are driving the key technology transitions in the industry. Working directly with them allows us to ensure that we have the best silicon. Now, what that actually means, the consequence is that all the rest of our businesses that use Cisco Silicon One, pick up that level of innovation as well, and we end up with much better products as a consequence. You see that when you look at sort of market share gains with various different Cisco products as a consequence of this. Once we win all of these different customers, we're able to drive the volume up.
Again, because Cisco offers all of these business models, we're able to achieve a scale that nobody else in the industry really can. What that translates to is we're able to take our costs down and generate additional margin dollars. With that, we can take those margin dollars and reinvest them in our technology and create this sort of compounding effect, where we sort of accelerate the innovation. Next slide, please. How has this really been playing out at the end of the day? You heard Jonathan talk about this a little bit earlier as well. If you look at sort of the traditional norm of the industry, between 18 and 36 months is usually how fast a competitive piece of silicon comes out in the market.
Because they have a narrow scope, that piece of silicon is addressing one market segment. Cisco Silicon, within 24 months of launching, we announced 11 devices in 11 markets. Said another way, is we're now operating 8x faster than anybody else in the industry. Really, the way to think about this is, we started this journey eight years ago, and we're now able to sort of enjoy the benefits of all that work we've been doing over those past eight years. Next slide, please. Interestingly enough, I know everybody here is interested in hearing about AI/ML. What's interesting about Silicon One in many ways is everything that I just said in the past few slides actually fully applies to AI/ML as well.
Let's talk a little bit about what's going on in terms of the AI/ML market and why we believe we're well positioned for this market. It used to be that infrastructure was built for high-performance compute, or HPC. These were reasonably small clusters, built in a reasonably small number of places, built on sole source GPU, sole source switches, and a proprietary interconnect called InfiniBand. This worked great at the scale that they were operating at. As you think about what's happening with AI/ML, there are several trends which are really important. One is that the scale of the networks of AI/ML is unlike anything else we've ever seen as an industry. It is much, much, much larger than HPC and much larger than most other markets in general.
As that happens, what you're seeing is there's many different GPU vendors all coming to the market with their own solutions, including some web-scale customers themselves. A few weeks ago, Meta actually announced that they're building their own GPU. As that happens and the scale of AI/ML grows, this notion of sort of a proprietary interconnect like InfiniBand being the long-term interconnect, starts to become an impossible thing to realize. Ethernet is getting integrated into all of these GPUs, and we're starting to see the effect of this transition to Ethernet-based technologies as AI/ML grows. What's the value proposition that Cisco Silicon One has in the space? Clearly, based on all the previous slides I mentioned before, I would contend that Cisco Silicon One is the best Ethernet switch on the market. Full stop. The best Ethernet router on the market, full stop.
This flexibility that we have built into our infrastructure allows our customers to deploy a hardware infrastructure and select whether they want to run the interconnect as Ethernet, getting ultimate compatibility and interoperability with multiple vendors, or fully scheduled fabric, where you can get the true ultimate performance, or what we're calling here enhanced Ethernet, which is the middle ground between those two areas. Again, similar to what we said before, Silicon One allows convergence of the AI/ML infrastructure, and we're enabling our customers to have the tools to make trade-offs that are important for them, rather than us sort of trying to force on them the technologies that we have. This is something that no other vendor can actually offer in the industry. We can advance one more slide.
Now, remember what I said, though, before, is: at the end of the day, convergence is not enough. You must be the best in order to win the markets. Now, if what we were talking about was the performance gain for a fully scheduled fabric was small versus Ethernet, nobody would frankly really care. Before we go into the details, I just want to sort of normalize us on a few key points. First is that the traffic patterns for AI and ML are very unique compared to generic Ethernet infrastructures. It's made up of high bandwidth, long-lasting flows, and those characteristics exacerbate the traditional load balancing schemes that exist in Ethernet. The second thing is how the GPUs talk to each other. There's what's called an all to all collective, where every GPU talks to every other GPU in the job.
What they do is they send this data to everybody, and then they synchronize and wait for the very last piece of data to get there. What that means at the end of the day is that one bad load balancing decision affects everybody else and all of your GPUs. If I think about just an analogy to that case, imagine that Me and my friends were all gonna fly to Europe to go and watch a soccer game, and we all got on a plane, and we flew over to Europe, but what we decided to do was mail our tickets across a ship, and it took three weeks to get there. The fact that we got there in eight hours doesn't help us get into the game.
We're stuck waiting for that envelope, and that's exactly what happens with AI/ML infrastructure. Now, what you see in the graph is as you add more and more jobs to the infrastructure, which is a big change from high-performance compute, but exactly what happens in web-scale AI, and I love the structure, the performance of traditional Ethernet begins to strain. With fully scheduled fabric, it is always giving you the perfect load balancing decision, to the point where actually you get 1.9 x quicker job completion time, which translates to being able to put 2 x more GPUs on the same network infrastructure. Now, at the scale of AI/ML that we're talking about, this really is a huge game changer. Half the networking cost, half the networking space, half the networking power, increased GPU performance.
This, at the end of the day, is a truly sustainable network. It's part of the reason that we're so excited about it. If we could advance to the next slide, please. At the end of the day, how has Silicon One played out since we launched it at the end of 2019? I'm gonna break this into two different plot tracks. The first is talking about the traditional full system business that Cisco has always been doing since our early days. For full systems, it's now available in all of Cisco's flagship products. It's available in the Cisco 8000 with IOS XR, targeting the service provider and web-scale customers.
It's available in the Catalyst 9500X and 9600X, running IOS XE for our enterprise campus customers, and it's available in the Nexus 9232E and 9800 running NX-OS for our Nexus enterprise customers. We have, across our three major product lines, now all offer Silicon One. We really have converged the bottom of that cone that I talked about earlier. Earlier in the talk as well, I talked about this notion about delivering sustainability through technology. It's frankly, really easy to make that claim. It's harder to make a real difference. Luckily, we have customers who, like us, share that passion for sustainability, and they've run their own studies around what the benefits are for deploying Cisco Silicon One. One example is DT. As you probably all know, a major European service provider.
They've announced that by deploying the Cisco 8000 based on Silicon One technology, that they've been able to reduce their power bill by 92% and reduce their physical infrastructure from 8 racks per router to 1 rack per router. This is a real sustainability change based on technology innovation. If I think about the other pit of the story, One of the big announcements that we made in December 2019, is alternate business models. How has that been going? We now offer silicon-only. We offer white box with various different operating systems on top of it, we go after primarily the web-scale customers. As you probably all realize, that is the hardest set of customers to penetrate. They're the most demanding set of customers.
At this point, we've actually penetrated five of the six global tier one web-scale customers with these alternate business models. A very, very resounding endorsement behind what we're talking about doing. Next slide, please. At the end of the day, if you, if you ask me, "Why are we set up to win for Cisco?" At the end of the day, it's because I think we have the right technology, we have the right cost points, the right investments, the right scale, and the right business models. We're the only company in the industry with all of the technology across silicon, hardware, optics, and software, and that allows us to innovate across these siloed pieces of technology.
We've become mature enough and flexible enough, where we're now the only company which allows our customers to consume the technology across all of the business models, from silicon-only, white box or full systems. Said another way, we are empowering our customers with the choice and flexibility to deploy the technology in the ways that work for them. In conclusion, I'll just start off or just end with what I said before, is this all makes me, as an engineer, incredibly excited to be at Cisco and be involved in this technology, because it is truly a fundamental change in the industry. Next slide. With that, I'm happy to answer any questions.
Yeah, that was a great presentation. I really appreciate it. One of the presenters earlier today when we were talking about the AI fabric build-out, you know, talked about this idea of having silicon on both ends of the network, common silicon on both ends of the network. I can appreciate the switch layer and the east-west traffic layer. What's the other side of the network? Are we talking about data processing units or DPUs? I guess I'm a little bit confused of what that is.
Yeah, great question. I'll give you just a little bit of color here. Pretty much everybody else in the industry, besides Cisco, is running horses who can run a single race. You have people in the industry who are out running InfiniBand. You have people in the industry who are out running Ethernet switches. You have people in the industry who are coming out with line card devices for scheduled infrastructure. You have people running horses which have technology for the spine layer of a scheduled infrastructure. All of these are unique architectures, and the customers have to make a very tough decision of which one of these horses am I going to bet on?
I think what's really unique about Silicon One in this space is because we have this flexible infrastructure, and we can reprogram our device to take on unique personalities, we can run a single horse who can do all of those things and allow our customer to actually evolve their choices over time, into what they actually want to deploy.
Thanks. Simon Leopold from Raymond James. Maybe following on that question. I think I appreciate the aspect of being able to program Silicon One for different options. What I'm trying to get a better understanding of is other Ethernet switch providers are supporting the RDMA over Converged Ethernet or RoCE protocol. I'm not quite sure.
Yeah.
I follow how your scheduled fabric is different from the RoCEv 2 implementation?
Yeah. Great, great question. When people talk about deploying Ethernet-based switches on AI and ML infrastructure, really, they're all running sort of RoCEv 2, running RDMA protocols, running priority flow control to sort of ensure lossless delivery through the fabric. All of that infrastructure is basically trying to sort of distribute the traffic across the network without interference. That's actually a really, really hard thing to do, and people are attempting to improve the performance through various different enhancements with Ethernet. It's what I refer to as sort of the enhanced Ethernet middle ground. What you end up seeing is that in some cases, you're able to improve the performance, in other cases you aren't.
Every time you can't improve the performance, the performance of your AI ML cluster drops like a stone because every GPU is waiting for the very last packet to be delivered across your network. One bad decision is like you've made a million bad decisions. What you get with the scheduled fabric is 100% non-blocking, guaranteed performance through your network, regardless of what else is going on. It's always doing the right thing. It is always delivering your traffic in an optimal fashion, giving you the optimal throughput of the network. That graph that I showed you before in the slides with Ethernet versus fully scheduled fabric, those are sort of generic Ethernet and fully scheduled fabric. In that, we are showing that 2 x more GPUs can be hooked up to the same network as a consequence.
What people are trying to do is improve Ethernet performance with this sort of enhanced Ethernet cell topologies, and you're able to begin to close the gap, but you can never really converge all the way to what that fully scheduled fabric is.
Hi, it's Ben Reitzes, Melius. Two questions.
Sure.
One on the web scale, hyperscalers, five out of six.
Yeah.
Can you just give us a little more color there on the materiality, the ramp, and how that flows through, you know, to us seeing the results in the business? Then I had
Sure.
Follow-up after that.
Sure. I'm not sure I can disclose actual numbers. What I, what I'll say is that, you know, obviously the global web scalers are the ones who have sort of very significant volume sets behind them. They're also the most picky customers, in terms of trying to make sure that they're making the right technology choices. There's a long time from design win to sort of production ramp. We've been at this now for three years, so we do have a bunch of stuff that is in full production ramp and other stuff that's sort of coming along.
Okay. I mean, others in the room may know this, but I guess qualitatively, how big a driver is it, you know, to your business at the moment, that stuff that's in production?
Yeah, good, great question. I'll ask or answer it in a few different ways. One is there's how much dollars is associated with actually winning sort of silicon-only business. That's a reasonable TAM and a reasonable revenue source. I think the more important piece of it is actually the additional business that you pick up as a consequence, i.e., once web scalers trust you and trust your technology, they're interested in buying more and more technology from Cisco. What we see happening is customers starting to deploy Cisco Silicon One in a silicon-only model that employs in various roles in their network, and then they say, "Hey, you also have a Cisco 8000.
I'll go ahead and deploy that in other segments of our network." You see the effect sort of ripple through into other what business lines at Cisco as well.
Okay. Just also, wondering, with regard to Silicon One, and your products, if you wouldn't mind just describing a little bit, you know, in the enterprise and with these web scalers, what is the competitive advantage that, you know, is really coming through and shining with regard to your merchant silicon competitors?
Yeah, yeah, great question. Several pieces of this. I'll start with sort of the web scale one, which I think is probably the most competitive in terms of silicon-only based style deployments. For a long time, web scalers have been relying on sole source technology. They're very hungry, especially after sort of the constrained environments of the last few years, when silicon was becoming harder and harder to get a hold of, to find an alternate off-ramp from sole source technology. That's a great tailwind for us. They're looking to find somebody else besides their existing provider. The second is, the existing providers all have very fixed functionality built into their topology based on the narrow scope that they were solving before.
We have a set of flexibility and programmability in our devices that allows our customers to add value on themselves on top of the sort of raw infrastructure, which they find very interesting. The third is around power efficiency. At the scale of these customers operate their network, efficiency is a huge, huge deal. So with Silicon One, we have an incredibly efficient infrastructure, and so we can drive power out of their network by deploying us versus deploying other people's infrastructure. The third is around that notion of convergence, which is many of these customers design their own hardware and design their own software. Let's start at the software piece for a second. There's a software development kit that runs our SDK, that runs on top of individual silicon that exists in their infrastructure.
Competitors offer one SDK for switching silicon, another SDK for routing silicon, another SDK for advanced switching silicon. It's all disjoint and all very convoluted. If you're porting your NOS that you've written as a web scaler on top of someone's infrastructure, the return on that investment is limited to the functions that you can meet with that silicon. With Silicon One, because we've converged across all of these different business models and different networking roles, once you port our SDK, you can then deploy it across your entire network. There's a very, very big lever for our customers to use that convergence. That same advantage that we see as Cisco being an equipment provider, having a converged networking infrastructure, our end customers can do that as well when they're writing their own software or developing their own hardware.
Hi, Samik from J.P. Morgan. Just a couple of questions. When we look at the D.C., AI D.C. switching market today, obviously InfiniBand has that early advantage, right, early mover advantage. Now, what does it take for Ethernet to sort of eclipse that? In your mind, is it just sort of now going through the design wins and sort of ramping on that? Because you also mentioned there's the sort of trying to avoid sort of being sole source to the existing supplier. Is it more about that it's not just your innovation that makes it happen, that you need more suppliers to come through with similar innovation for really that to change?
The second part there, like when you are going and sort of talking to your customers, is the competition that you're competing against NVIDIA on InfiniBand, or is the primary customer in that competitor in that discussion, a Broadcom at the silicon level or rest at the device level? Like, who is the customer really then sort of comparing you to?
Yeah. It's a great question. Several pieces to it. I think forgetting about Cisco for a second, the raw scale of the AI ML networks are 100% unsustainable with a sole source technology. Even though I work on this stuff day in and day out, I sometimes have a hard time wrapping my head around the size and the scale of these clustered and these networks. It is absolutely huge. That fact alone means that they have to move away from InfiniBand. Every customer out there at this point is saying, "I've got to get off of this sole source technology. I'll use it for now because I have lineage to HPC, but it's not sustainable." Obviously, it has to change.
That's compounded by the fact that as more GPU vendors, either built by the customers themselves, the web scalers themselves, or by other people in the industry, when they're putting interfaces on, they're not putting proprietary InfiniBand, they're putting Ethernet. That, again, is sort of the clear sign that the transitioning is happening with Ethernet. You know, I think another way to think about it is, if you want to look at a proof point that the transition to Ethernet is well underway, look at what companies who build InfiniBand switches have recently announced in terms of Ethernet-based switches for AI. That trend is very, very clear. It has already been deployed in production at large-scale AI ML clusters that already exist with Ethernet. That transition is happening or has happened in some cases in some customers.
That will flush through. The key question is: Who makes the best switches? Who makes the most efficient switches, and what value add can you add on top of that sort of open infrastructure? That's sort of, I think, where Cisco differentiates itself versus all of the other folks coming to play in sort of the Ethernet switch market.
All right. Time for one more question. Anybody else has any burning questions?
Just one last one. Scheduled fabric, I guess kind of the question is, you know, doesn't that get you down the rabbit hole of another proprietary stack? Just kind of what work is being done to kind of work with the standards body or, you know, how do you see converging towards one standard? You know, that's kind of what made Ethernet take off in the first place, was that it was a standard. How do we kind of converge back to a standard?
It's a great question. Again, I think there's a continuum here, right? I think there's the perhaps the fallacy that we have a tendency to fall into as an industry is thinking that all customers make the same trade-off decisions, that they view every decision as equal. What we see in our customer base today with Cisco Silicon One in the AI/ML space is we have a group of customers who are firmly in the fully generic Ethernet camp. We have customers who are fully in the enhanced Ethernet camp, and we have customers who are firmly in the fully scheduled camp. These are all major large customers, all who have really spent their time thinking through for themselves, what do they value the most?
I think there's not, might not be sort of one answer to rule them all across the industry. There is work going on between multiple customers and multiple vendors trying to sort of come up with an infrastructure that works, across multi-vendor to allow this to sort of scale to the points that we need. I think there always will be a set of customers who want to get that maximal performance benefit, especially if it's reconfigurable over time, to something else more generic for them as well.
Okay. All right. Thank you, Rakesh. Really appreciate.
You're welcome.
Yeah, you're welcome.
Thank you, everybody, for doing that.
All right, I'll just go ahead and close this out. Thank you, everyone, for attending. I think you can hear the four main key takeaways here. That we're cutting through the complexity that you hear that is out in the market with simplicity, with integrated platforms. Second, you hear about those massive growth opportunities that we're going to go after that will drive our long-term growth, cloud, security, AI, plus all the other ones you heard, whether it's FSO, hybrid work. You've heard about the game-changing innovation we're driving across our portfolio, how that's going to lend itself to growth. Lastly, the differentiation, I think is pretty clear, as you heard from the speakers, as you heard from Rakesh with Silicon One, and I think we're very well positioned.
If you have any further questions post our event, the IR team will be here to take any questions. While many of our speakers, including Scott, wanted to be here to address your questions, they all had speaking engagements on stage with our customers, so they were not able to be here. We're here to take your questions and/or please let us know, and we can set up some separate discussions, wether with Rakesh, Liz, or any of the speakers you heard today. Thank you all for coming. We really appreciate it.