Cisco Systems, Inc. (CSCO) Earnings Call Transcript & Summary

Good morning, everybody. Derrick going to close the door in a few seconds. We had a dinner last night, and I promised them I'll keep all the good questions for today. So I get a lot of great questions. I'm very happy to host Scott Herren, CFO; and Mark Patterson, next CFO for a call.

What could be better than 2 CFOs.

I don't know. We have 2 CPA.

And an engineer, correct.

I'm both.

Thank you. I want to start with a question that is always kind of on people's mind when there is a change of CFO. In your mind, what is the role of a CFO, Mark and -- what are your targets? Like what are you -- if you think about your next position and what you need to accomplish and what you want to accomplish, what are the things that you have in your mind?

Yes, great question, and 1 I'm getting a lot these days. Yes, certainly, I'm focused on -- particularly coming from a strategy role, like I have been in as Chief Strategy Officer, really focused on prioritization. And right now, I would say, I've been at Cisco for 25 years. I feel like we've got perhaps more opportunity ahead of us than I can recall in all the years that I've been at Cisco, to be honest with you. So really making sure that we're funding what really matters and putting the fuel behind what we need to do in terms of innovation to capture the growth that's ahead is a big thing. Another for me is just coming into the role, I think for any leader, you want to do a lot of listening. And I'm going to spend a lot of time in venues like today, listening to investors and analysts also spending a lot of time with customers. In my 25 years at Cisco, I spent 11 years in the sales organization. I think the value of really spending time with both customers, but then also with partners as well is really key for us. And so that's certainly something I'm focused on. The third area I'd just say is what you'd expect in terms of financial discipline, really transparency. And I mean transparency from an investor standpoint, but also talk to the team, the executive team a lot about it. I want to be very transparent and give -- I think if we give the team good data and they're able to make more informed decisions, we'll do much better as a company, too.

I want to start maybe with the State of the Union. Just to understand how is the current environment? April, a lot of companies said that the business environment was weak, maybe not orders, but business environment was weak. How is the market your customers reacting to the uncertainty in the economy?

I mean -- I'll start, Mark, and you add your comments on top. We haven't really seen a change. And it's one of the things that, obviously, we put a lot of focus on. So our fiscal quarter ended at the end of April. And if you remember, the cyclical tariffs were announced on April 2. And I thought there could have been one or 2 things that could have happened, I guess, 3 things. The third is what actually did happen. One could be people try to front run the tariffs, and we see a wave of orders. The second could be all this uncertainty, I'm going to pull back and retrench a little bit. And the third is just business as usual. And really, what we saw was more business as usual. And I have to say I inspected this from pretty much every angle. You could internally, more trying to say because we had a good quarter. We had 20% product bookings growth, even ex Splunk, it was 9% of product bookings growth. So it was a good quarter. And I looked at linearity within the quarter. Did we see a spike in orders post April 2? We did not. I looked at -- do people place orders in that second in that third month in the -- after April 2, with ship dates requested way out. That didn't change. We looked at channel inventory. It actually was down. We looked at webscaler inventory. It actually was down. So we looked at the length of time between when we ship a Meraki box because we know serial number XYZ left on the state, when did it actually get implemented because they have to go to the cloud to do the activation, that didn't change. It's actually, where it was pre-pandemic. So we didn't see any signs of pull ahead in demand that fueled that 9% organic product bookings growth, we also didn't see things slide out. We didn't see projects push. We didn't see pause. So we really didn't see -- I think of the 2 things that I was worried about and neither one of those is what we saw. I think the world has been dealing with so much uncertainty, since the onset of the pandemic and then all the geopolitical issues that have cropped up since then. And people have gone from, geez, it feels uncertain, I need to take a break and pause to -- I got to run my business, right? There's never going to be a time when things are fully settled, and I just have to continue to run my business, and that's really what we're seeing.

Got it. The stock had a tremendous move, great move. You're now at the high end of the historical valuation, but we're seeing other stocks like IBM that traded even to higher level than historical levels. And the question that I'm sure everyone has on its mind is now the question I'm getting is always what would drive the stock to higher valuation. And I want to focus on networking because networking is your core business, it's the majority of your revenue...

2/3 of our revenue, yes.

Right. And it's the majority of your revenues, and there is a cycle.

Yes.

So I'm going to ask you, both of you, I mean, you're coming from strategy, same thing. The question I have is how long is the cycle, the longevity of the cycle? Are we talking about 2026 CapEx up for Cloud Titans and that's it or 2025, I mean? Or can you talk about the 3-year cycle, a 5-year cycle? What are the things that you see in front of you for the networking cycle?

This is yours to ride, so I'll let you.

So let me just preface it that we certainly have a lot of execution ahead. But as we talked about a little bit last night, networking is cool again. And we're seeing a lot of tailwinds, frankly, across the networking business. The opportunity that we have in web scale, I think we all understand is a massive opportunity, and it's new TAM for us. The opportunity that we also have in the AI enterprise adoption and the inferencing infrastructure that we're going to be delivering along with NVIDIA and other partners. I think as another new TAM area that will also have -- we're at the very, very early stages of that, and that will run for quite a while. And we think it's an order of magnitude higher in terms of the actual opportunity size than the hyperscale and the training side is. Now you're starting to see sovereignty -- data sovereignty become a really big deal. We made a number of announcements in the Middle East, both in Saudi Arabia and UAE, around some massive investment that they're going to be putting into AI build-outs and infrastructure, and they want us to help them really design and secure the AI infrastructure for them as we've been selected as a partner in that space. And then the third area is really around more the traditional connectivity, if you will. And the campus space, a lot of people don't understand that it's 3 to 4x the size of our data center business. So significant business has not been a significant grower like the other 2 spaces I've just mentioned, but I think there's a lot of opportunity there. There's -- there's so much going on in terms of building security into the network of the fabric. There's so much going on relative to AI and AI ops and how AI really comes to the campus and what does that look like? So I think you're going to see us innovate there a lot. Again, we've got a lot to do to go execute, but massive opportunity, I think, in front of us.

Multiyear is the right word.

Multiyear, right. It's not in FY '26.

It's not a step function and then flat, it's a multiyear ramp.

This is a multiyear.

And what drives your -- the Southern growth? I've been covering you for over 25 years and networking over the last 20 years, let's say. It didn't grow as much. What suddenly makes this a multiyear cycle?

A great CFO. Well, it's multiyear.

I would tell him even a better replacement.

And even better backfills.

And I think these are the tailwinds. I think back to [ Gaya ], the -- who leads Meta's AI build-out. And one of the most largest, most complicated infrastructure that's being built. And when he was asked on stage...

With Chuck.

With Chuck, he was asked, what is the biggest determinant of the success or failure of AI. He said, no question the network. And so the importance of the network is more important now than ever, I believe that. And I think that the security threats that we're facing the advent of AI in the security space, in particular, and the sophistication of those threats are going to require the network and security to really come together in a meaningful way, where you're going to take security and you're going to build it in to -- melt it into the fabric of the network because there's no other way that you're going to actually be able to protect your entity. And so I think there's just some of that going on that these are tailwinds in that networking space that we just haven't seen in quite a while actually.

Yes. So I'm going to divide up my questions into cloud and enterprises because -- and in the enterprise data center is in campus because I think there are different trends. But in the cloud, have you penetrated all the areas -- have you already penetrated all the areas you want to penetrate? And from here, it's just about deployment? Or are there still customers and projects and things that you can penetrate? How do I -- because this year is a phenomenal year for cloud. I'm trying to understand what would drive growth from here going forward.

Yes. So first off, for those of you who may not know, we set a target of $1 billion in AI web-scale orders and exceeded that -- well exceeded that a quarter early in our fiscal '25 over $600 million of orders alone in Q3. We've said all along, it's not going to be linear in terms of the order growth rates that you're going to see there. But this is one of those areas and you're going to just hear me say there's a lot in terms of we've got to execute. There are no more demanding customers than this customer set and they will very quickly change the direction, increase the requirement, they need better power savings here, et cetera. And so there is all of that. Now having said that, we've got double-digit design wins in the way these things work as you win a design and then ultimately, they put it into production, and that's when you really start to take the orders and ultimately the revenue and only about half of our design wins are actually beginning to be in production, and those are multiyear opportunities in themselves. And so I think there's a lot of potential runway here. But it will take some time for this to actually turn into orders and then ultimately revenue.

But there's still -- I will call it nascent, but early days, opportunity in sovereign web than the NVO. Right? And the announcements that Mark talked about, the partnerships that we struck in the Middle East is a great example of that. There's a huge amount of opportunity that still is pretty well untapped as that gets built out for data sovereignty issues as those get built out.

You're leading the routing market and you're leading the optical pluggable market. You're leading optical in general. In the last quarter, we have seen a change where networking was 66% of revenue -- of your cloud revenues and optical that used to be 50-50, now it's 33%. So 2/3, 1/3. How should we think about the growth going forward in networking? Are you going to lead with the routing and Silicon One? Or how do we think about the balance between optical and networking?

So we were sort of more lopsided towards the optics, early on in optical. And we had always said that based on our pipeline, based on the design wins that we were looking at that we would see that shift to more 2/3. And I think you're going to continue to see going forward. Again, it won't be linear, but I think that you'll continue to see more of a shift towards systems, if you will. And silicon and the 8-K that we sell into that space.

Right. Okay. Enterprise cloud. You have partnership with NVIDIA. Describe what you found in NVIDIA and why you think it's a great partner or a good partner?

Yes. So 2 things, I would say. So when we look at the enterprise space, these are companies that don't nearly have the sophistication that the web scalers have. And there's an awful lot that goes into the build-outs and enterprise AI and the inferencing capabilities that they need to not to mention the data requirements that they have, the legal requirements they've got to work within and everything else that they sort of need to deal with. So our objective here is really to make it as simple as possible and allow companies to be able to take advantage of what AI brings and allow them to be able to deploy as quickly as possible and as seamlessly and simple as possible. So this partnership with NVIDIA, we think does just that. It's all new TAM for us. It's a partnership that brings the NVIDIA GPUs together with what we're going to try to do is build Scott and calls it inferencing in a box. And so these AI pods and the hyper fabric that we're going to build, will essentially take the NVIDIA GPUs, our silicon, our network game, third-party storage, security, collaboration and really put that together in a simple pod that they can basically plug in and start to do inferencing with in the enterprise space.

The power of that is not just that it's easy to order and easy to deploy. It is both of those things. What enterprise customers in particular, want is to know that it's going to be fully supported, right? It's kind of a leading-edge technology that they don't necessarily have the depth of skills that they need in and with the enterprise reference architecture we have with NVIDIA now around these AI pods, both Cisco and NVIDIA stand behind supporting that. I think that's equally important.

The alternative is to go to HP, Dell, Arista, buy switches, make the network your own. Go deep a little bit about the value of reference design. Meaning give us maybe an example just for people that are not partitioners to understand, what are the challenges that a customer might have in deploying inferencing network and how you help them with the partnership with NVIDIA versus the alternative?

Yes. I think -- I mean, if you think about it, it gives a customer comfort that they can use NVIDIA in the back end and all the Cisco equipment in the front end. And the software that they're used to that they can put security with it. It's the networking that they're used to and trust from Cisco, but also the GPUs that they want, and it's going to -- they just know it's going to work. It's going to be easy, and that reference design, I think, is what provides the customers that comfort.

Fully tested and fully certified. I mean it's not an easy process to go through. But then at that point, the peace of mind as a CIO that's going to your -- the lifeblood of your company is going to revolve around your AI apps as you look into the future, to know that if something does go wrong there, you've got the strength of both Cisco and NVIDIA standing behind you to get that righted. That brings a lot of peace of mind. I think that's as important as just to make it easy to purchase and implement.

Yes. Cloud is happening now. When will enterprise happen? When do you expect to see AI deployments by enterprise?

Yes, I think this is the question on everybody's mind. I think there's consensus that it's a massive opportunity, but just kind of when we'll see it pick up. The biggest thing that we can do is really just look at our pipeline and what we're seeing in terms of opportunities. And we recently just stated, it's not in the billions yet, but it's in the hundreds of billions. And it's the growth of that pipeline, I think we're pretty pleased with and the kinds of discussions that we're having with customers would seem to indicate that over the next 1 to 2 years, you're going to start to really see that business pick up.

It's a multiyear build-out. You train the model. You may augment it, you may tweak it, but you train the model, and then you use it at inferencing for years to come, right, as AI kind of transcends the organization. Even in the finance team, I've talked about this -- I talked about this with you last night, I've got a dozen different use cases, AI-based, Gen AI-based use cases that we're working on right now that are in kind of from proof of concept, they're now in what I'd call beta. We don't use that term internally, but think of that as beta mode that will then go into production. Think about -- take that now from just finance to marketing, to sales, to customer service and customer support. There's going to be a huge demand on that, but it's not going to happen overnight. It's going to be a multiyear build-out.

I hosted a dinner with CECLs at RSA and I started with the most generic question on the planet. What is the current spending environment? And the first answer was, you're not asking the right question." The guy told me, he said, "This is not the right question. He said let me give you the first question." He said, every CEO -- he said, every CEO sits down today with his -- all his executives, whether it's marketing, finance, R&D and says, how do you use AI to save 30% of your costs? And...

We just did that same thing...

So what are the -- he said what are the security implications? What are the networking implications? These are the right questions. So anyway. I'm at the age that I can take insults. Campus. What is the opportunity in campus? I think that the last refresh was 6, 7 years ago, what drives campus -- let's talk generic, like let's talk general, what drives campus growth in general? And what -- why are you expecting now some kind of a growth in campus?

Yes. I think -- so again, just to refresh people's mind, the size of this business 3 to 4x the size of our data center business. I think that the growth that you're going to see is certainly the return to offices has helped fuel some of that. But that's not really the big thing. I think the big thing is it's around AI and the advantages that AI can give you, thinking about agentic AI, what does it look like when AI comes to the campus? And how does that impact your network, how does that impact security, et cetera. I think building security deep into the fabric of the network in the campus as well is going to be another thing that's going to drive ultimately the refresh, if you will. And then just simplicity. I mean, the kinds of advantages that AI can drive, I mean, think about having a CCIE assistant, a Certified Cisco Internet Engineer, right, that as an assistant for you to help just configure, operate, manage your network. And so I think that will be another thing that will really help drive the upgrade cycle here as well. But as you mentioned, this will be a multiyear opportunity and one that we look forward to.

And these older out-of-support devices that are still embedded in some networks and probably without the knowledge of the network engineer create security vulnerabilities, right? That's where the bad guys search for once they get into the network, search for one of these older devices because they're out of support at that point. That's another thing that is in today's world, with the geopolitics being what they are in nation-state actors has become really important to discover where you've got those older devices so that you can replace them quickly.

And campus started growing already. So are these the drivers for campus growth right now? Or are there different drivers?

One of the big things that we've talked about is that when we spend time with customers, and you've alluded to some of this already, their top 3 concerns around AI are is my network ready, am I really ready for low latency and the traffic flows that are coming or do I have assets that have been sitting here, and then I've been sweating them for 10, 12 years, and there's a lot of that. Is my security posture appropriate for the advent of AI and the applications I'm going be running? Am I ready for agentic AI and machine-to-machine collaboration that's going to happen? And so security and trust is obviously a big thing. And then the third thing is really around talent. Do I have the talent that has the AI skills that can actually take advantage of this? And so I think that you're going to see -- you're already starting to see in the campus space, some of the growth you're talking about being just that, upgrading their networking infrastructure, upgrading their security posture and really being ready for the applications that they know that they're going to deploy in a big way.

Yes. So we spoke about security, but security wasn't successful so far. You refreshed the portfolio roughly 2.5, 3 years ago over the years, refreshed firewalls, and we are still not seeing much growth. What needs to happen when you analyze the market? What needs to happen for security to take off? And maybe you can actually -- maybe you can start by talking about the portfolio because some parts are growing. Some parts are not growing.

Yes, exactly. That's exactly where I was going to go Tal and thanks for that question. You asked the right question this time.

Now he's kissing up.

Today is my Birthday.

Smart man, happy birthday.

We're going to round up happy birthday when you...

But if you look at our security portfolio, we began refreshing it, as you said, about really almost coming up on 3 years ago. Pivoted a lot of our internal resources without growing our spend envelope, pivoted a lot of internal resources to growing that business, which had underperformed pretty dramatically for a few years. That's allowed us to bring in new talent. That talent has attracted other talent. So the talent level of the organization has come up quite a bit over that time frame. We've been going about systematically starting with the firewall, but refreshing the products across the board. Our firewall line, if you think of ultra high, high, medium, low, ultra low, we've refreshed the middle 3. High, medium and low. By the end of the year, you'll see the ultra-high refresh and ultra low, not long after that. That's a foundation, I think, of any cybersecurity company. So we've done that. And where we have refreshed, we're seeing really nice growth. So to your point, there's elements of our -- we talk about it in aggregate. And you're right, it's not -- it's still not growing at the rate it needs to grow. But when you peel it apart and say, the older products that we have not yet refreshed or not dedicated the resources to -- actually are a big drag on the growth rate. The new products we've launched and the refreshed products that we've launched are growing quite nicely in that space. And the other thing to bear in mind, we talk about it, of course, on a revenue basis. And a lot of that is ratable. A lot of what we sell in security is ratable. So it's built up. You've seen our RPO grow to $41 billion at this point. We've got it. We've built up $42 billion, a huge amount of RPO at this point. So some of that security growth is not yet evident in the revenue line. It's sitting in RPO. There's more work to do. I'm very encouraged by the success we've had and the traction with the new products around secure access with just competing head-to-head with others in that market and winning nicely. Our XDR product, which has grown over 1 million endpoints already, just launched last year. Hypershield, which is what Mark was talking about earlier, melting and in this case, in the case of Hypershield, basically firewall capabilities, melting it into the network with a processing unit added to the switch. So it's just software sitting on the switch and it runs obviously with the Silicon Connect. Those are huge kind of significant step-ups in capability. And then what's still nascent, but has generated a huge amount of interest is our AI defense product. So there's definitely more work for us to do and start there in security for sure. None of us are satisfied, Mark's not, I'm not with where we are in security today. But if you look at the areas, where we focused and invested it's actually growing quite nicely inside there. So there's better days ahead.

I have more questions, but I want to open it up for questions from the audience. If anyone has any question, raise your hand, we have a microphone.

Yes. We need to get a mic down here.

Maybe I'll just ask you. To double click on security, you guys mentioned that in the medium term, we should still have confidence in driving the securities growing like in the mid-teens level. There's some skepticism around that. So can you talk about what drives our confidence in the visibility for that growth?

Yes. I'd say the first thing is the base of that will be Splunk and Splunk continues to grow double digits, right? So start there, layer on what I just said about the new and refreshed products and the rate and pace they're growing. Unfortunately, they're not -- they have to accumulate mass to move the needle on the total -- on our total security number. So Splunk growing double digits, the new and refreshed growing very nicely, just short of double digits from an order standpoint right now. That's what drives that longer term. And then obviously, it needs to -- with it being ratable revenue, it has to go first into deferred revenue and then bleed back out into the revenue stream. So there's a little bit of a delayed effect on that, but that's what drives that. And then some of the innovation. Hypershield is being able to melt the security capability into the network itself is an architectural change that will be a multiyear tailwind that's generated a huge amount of interest. There's a -- Chuck talks about meeting with the CIO of a large customer who's not particularly -- was not particularly happy with us on the security front. And she began the meeting by saying, if you're here to talk to me about selling me more networking and more firewalls, I'm not interested. If you're here to talk to me about building popular security software into the fabric of the network, that's what I want to hear. And of course, that's what Hypershield does. So we ended up having a great meeting with someone who wasn't a huge fan of ours coming in. I think that's the future of our security business.

Anyone else? Margins, healthy margins. What happens to margins going forward? What are the puts and takes for -- and I'm talking about operating -- you can answer it the way you want it?

I'll start at gross margins. And Mark, you can probably talk more about the op margin line. Obviously, tariffs will be a headwind -- where tariffs show up for us is in the gross margin. It's in the cost of goods sold. So it impacts gross margins, and we need some level of stability in what's actually going to be implemented. We need to get through this period of what's actually going to go into place. And that includes not just the deal-by-deal structure that's kind of in play right now, but also the investigation, the Section 232 investigation around semiconductors and other components. That would impact everyone. It would impact everyone that's selling hardware, it would impact automotive, it will impact appliances like semiconductors and everything. So that's a bit of an open switch. So as those things work their way through, though, we've built -- and you remember this, Tal, because we talked about a lot during the pandemic, a really flexible supply chain, right? I have the final assembly and test here, which is what typically will dictate country of origin, country of origin is what dictates what the tariff is. At that tier, we've got 8 to 10 different points in different countries of final assembly around the world. This is a space where our scale is a significant advantage to us because we needed that many points of final assembly to be able to support the volume of product that we sell. So there's a lot of steps that we can take if the regime will just settle down and the details will get published on which HTS codes, harmonized tariff schedule codes, are going to be affected. Once that happens, there's a lot that we can do. So tariffs right now, we've built in everything that is currently being discussed, including the reversion of the reciprocal tariffs on July 9. That's all included in our guide for them. So that's the headwind on gross margins. The tailwind, obviously, Splunk is a high gross margin product. As we continue to build out our overall software portfolio, that will continue to be a tailwind to gross margins. We sell a blend of different gross margin products, and we always have across the portfolio. So there are some lower-margin areas and some higher-margin areas. I don't see that mix, particularly being a particular benefit or negative to us over time, cutting.

Got it. You want to talk about operating?

We're out of time.

I'll give you still like the 30 to 60 seconds.

The quick answer is you can continue to expect operational discipline from us. We've got a lot to invest in right now. So on the OpEx side, my big thing, as I said earlier, is just going to be around prioritization and starting with what really matters most starting with that first and not saying, "Well, boy, I can't afford this, and it happens to be one of the biggest things that we need to go do." So we're really focused on how do we move talent around? How do we move resources around and reallocate to just be more nimble in terms of the needs of the business and not always be asking for more and more money. We do have a lot to invest in. So having said that, but you're going to continue to see the discipline from us and the focus on where we can, obviously, building leverage into the model as well, which is what our long-term model is for. Certainly, we've got uncertainty around tariffs, but we'll deal with that as well.

Great. Thank you.

Thank you.