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

Good afternoon for those of you on the East Coast here. I'm Samik Chatterjee, the networking and hardware analyst at JPMorgan. For the next session today at our conference, we have the pleasure of hosting Cisco. I do want to mention this is kind of the first year we managed to get Cisco to our conference since the years I've been here just because it always used to conflict with the earnings. So it's good to have you guys here. And finally, the earnings currently doesn't conflict as much, which is great. So definitely, want to thank our conference guys to -- in part to kind of lead that change. Jonathan Davidson, SVP and General Manager of Cisco's Mass Scale Infrastructure is with us for the fireside chat. Jonathan, thanks for taking the time to do this. Definitely appreciate the time here.

I'll get you kicked off your with 1 question and for the audience, obviously, you have the option of sending in your questions through the ask a question queue on the website. So please feel free to do that as well. Jonathan, good timing, you just reported results last week. Significant acceleration in service provider order trends. Fiscal third quarter up 17% versus 5% the quarter before that. Maybe let's just start off there. What drove the acceleration? What are the trends you're seeing? And then we can get into some of the more broader topics.

Absolutely. Happy to go through that. And thank you very much for the invitation. We're happy to be here with you. Before I jump in, of course, maybe to give you a little statement here about the forward-looking statements that I may be making that obviously, our actual results may differ materially from these forward-looking statements and are subject to the risks and uncertainties found in our most recently filed 10-K and 10-Q. I also want to just give you a bit of -- a little bit of a background on the area of organization that I'm responsible for. What the Mass-Scale Infrastructure Group is because it doesn't really lend itself towards understanding. So I'm responsible for the overall service provider mobility business, also optical, cable, our routing business, our automation business for that as well as our cloud-based IoT business. And then, of course, our optics portfolio as well. And we have seen, to your point, there's tremendous traction in response. Going back to our December 2019 launch, where we announced really a new architectural approach to what we were calling Internet for the future. This is where we announced our new market-leading silicon systems, our new approach for the market, including our -- not only our traditional systems business, but also our disaggregated approach. Our customers can purchase our silicon stand-alone, our software stand-alone, to work on top of white boxes, our systems without software or, of course, our traditional systems business, and certainly, our Optics stand-alone business. And just a couple of months ago, in March, we had an all one launch for the Internet for the future, where we really rolled in the Acacia piece, which further enabled this architectural approach, or what we're calling router optical networks. And sure we'll talk more about Acacia, but the acquisition furthers our commitment to this disaggregated approach working with all customer types of business models, selling to not only our traditional customers, but also our traditional competitors. So we're very excited about that as well. Now directly to your questions around what were the drivers? Certainly, we talked about that, but we saw great consistency across the different theaters. We're -- Americas as well as EBR as well as APJC, with not only our traditional service provider customers, but also what I would consider the web or hyperscale customers as well. And just tremendous traction, not only with the kind of existing portfolio that's been around for many years because of the leading edge -- really innovation, not only in the software perspective, we call IOS XR, but also with the silicon for our 9000 platform. We've got market leadership position there. And then, of course, the new platform, Cisco 8000, the one we launched in December of has got traction across traditional service providers, MSOs, the cable companies as well as the web and hyperscale accounts. And they're all in deployment now in this phase and obviously continuing lots of trials and new use cases and new customers as well.

Great. John, thanks for that background in terms of kind of your responsibilities. I know a lot of investors don't really associate the title immediately with what the responsibilities are. Let's go through some of these important kind of portfolio changes one by one. Did want to start with Silicon One. Obviously, it's been in discussion for routers. You recently announced that the Silicon One is now adopted on switches as well. So where are we in terms of the road map of adopting Silicon One across the portfolio that you wanted to kind of go across? How crucial can it be? Because you mentioned web scale growth as well and kind of 1 of the drivers, how crucial can it be, particularly when it comes to like web scale customers and your leverage to growth there?

Yes. So a couple of things. So back in December of 2019, we launched One chip in the Silicon One portfolio at 10.8 terabits. And then over the last, really, 16 months or so, we announced an additional 9 pieces of Silicon in the Silicon One family. So now it goes from 3.2 terabits all the way up to 25.6 terabits. And the reason why that's really exciting because it covers both switching use cases as well as routing use cases. And when you get to the weber cast, specifically the hyperscalers, having consistency of feature set and consistency of operations is really important. So you can go all the way from the top of rack switch all the way to the top of the data center switching hierarchy as well as into the wide area network and use one piece of Silicon with a common feature set. So that's really exciting for them and as well as for us, obviously. But additionally, they can also choose which OS they want. So if it's a more simplified use case, they may use an open operating system or their own operating system. And then when you move higher and the complexity gets higher, they will typically then transition over to our own OS, iOS XR, because that gives them all the features and capabilities and functionality that they would need for those more complex use cases. But the underlying silicon is the same. And that really resonates with the customer. So the Silicon One piece is critically important to our continued growth in the web scale market.

Got it. So digging into the numbers there of it, when we think about the web scale order trends, you reported 25% order growth in the last quarter, 50%, I think, on the trailing 12-month basis. Firstly, there's been some discussion that some portion of it is being driven by campus equipment rather than data center sales to those web scale customers. Can you break that down it any further? Like how broad-based is that trend versus being driven by campus equipment?

Yes. So we don't break out the campus versus the data center side, but I can tell you we have seen just tremendous momentum on just all use cases inside of the web scale on what we call the production network. So instead of saying campus and data center, we say kind of the IT side versus the production side. So the production side would be inclusive of not only the data center network, but also, they have global wide area networks as well the hyperscalers do. And there's a tremendous mentor infrastructure go there. So we kind of classify IT and production and on the production side, there's quite a few use cases, some of them are upwards of 10 different use cases on the production side. Some of them have as few as 4 or 5 use cases. But our goal is to win every single one of those use cases on the production network. And of course, we've had ongoing strength from the IT side of the hyperscalers, it's just like we do in all the traditional enterprise accounts.

Got it. The other discussion point regarding web scalers is we've moved to the upgrade to 400-gig. And there's an upgrade cycle both in WAN and then kind of inside the data center as well. When we look at your order trends, how much that is a reflection of the 400-gig upgrade cycle versus continued momentum, for example, in something like 100-gig, which is the previous generation in that sense?

So what we've seen is that customers are -- even the customers who are buying 100-gig from us today. That equipment is already purpose-built for 400-gig. So they're not -- although they may be buying a 100-gig port to build out their 100-gig data centers that might be just refreshing an existing equipment. That equipment is also purpose-built for 400 gigs, so it's ready to go at the 400-gig pace. That said, we have seen 400-gig momentum certainly pick up in our fiscal Q3, and we expect that to accelerate over time. Right now the year-over-year percentages are not worth repeating because they're just ridiculous when you have almost nothing last year, and then they're really think this year, the percentage numbers are -- they don't make sense because they're so large. So I think we expect to see continued 400-gig growth. I'd say it's not only in the web scale accounts. So we're seeing a lot of data center growth around 400-gig, specific to 5G build-outs as well. So this goes to wherever we see 500 -- 5-gig really growing, we see the movement to 400-gig happening as well. So obviously, 5-gig, in certain countries in Asia, has been taking off well ahead of the curve of most other countries. And so there, they've had to go and make sure their data centers are 400-gig ready to prepare for the onslaught of additional bandwidth and traffic.

Got it. Got it. Let me remind investors, if you have any questions, please feel free to send them in. It's the ask a question button on the website, so you can use that. Jonathan, probably I'll stick to that overall 400-gig for a bit more here. The recent order trends, you've obviously highlighted the strength in 8-K outdoors, which I would assume is a lot of the WAN use cases, correct me if I'm wrong. Now when you think about the switching upgrade and the 400-gig switching upgrade inside the data center, is that also part of the order book today? Or is that largely still in front of us in terms of the web scale customers and their inside data center switching upgrades?

I think we're making good traction inside of the hyperscaler data centers. It is a -- this is the core of their business. So obviously, that gets rolled out over time, but definitely a tremendous amount of interest and innovation around that particular area. And given the Silicon One portfolio enables you to be a router or a switch, there's really not whole a lot of difference other than buffer sizes. So customers are able to choose really what they want, big buffer, small buffer, simple operating system or more feature-rich operating system. And they can have that level of capability and consistency without having to change platform types. And that they view that as a real strategic value-add as well because it makes their lives a lot easier to operate and manage their infrastructure.

Got it. So let's talk about disaggregation a bit. That's obviously been one of the, I would say, like themes as you kind of move towards Silicon One. So when you're looking at the demand drivers here, how much of the momentum is coming from like selling a full solution, hardware, software, the complete stack versus hardware stand-alone or software stand-alone?

The majority of the top line revenue is going to come from selling fully integrated systems because the dollar value of everything together is there. That said, the flexibility that we're offering of selling silicon, they can build their own systems, they can integrate their own software to buy our systems, put your own software on it to fully integrated systems has really resonated. And so the -- what we've seen from a hyperscale perspective is they're oftentimes picking 2 out of those models. And they'll pick a different 2 of those models. But the fact that they're able to have a single interface point with a single company to solve both of those areas is tremendously differentiating for Cisco. And we think that our flexibility of not only the business model, but the fact that we're investing so much in innovation across each of those various areas is a huge benefit to Cisco and to the customers at the end of the day.

Got it. Maybe if I can just follow-up on that point. You said the customers are opting for 2 of those kind of alternatives or business models. Is the monetization of the software here on a stand-alone basis, fairly similar to hardware stand-alone monetization? Or does the subscription really lead to a much lower monetization on an annual basis?

Yes. So the -- today, we're -- so A, we're happy to sell the software to work on white boxes. But that's been more -- we see more interest on that side from a traditional service provider side than from the hyperscalers. The hyperscalers, when they use ROS, they typically want to just use our fully-integrated system. So the -- out of the 3 models that I described of Silicon-only systems with no software and fully-integrated systems, what you'll see is the hyperscalers often will pick fully integrated systems and then 1 of the other 2 models for their deployments depending upon the problems that they're trying to solve for. But we're also very happy -- the one exception I would add is for our software running virtually inside of the hyperscalers if they have the -- they have that SaaS or PaaS type of business model or IaaS business model, then the ability to go and sell that to the end enterprise customers, but that's not part of the production network. That's part of more of IT for enterprises through them. So I'm not counting that as part of this discussion, just to be super, super clear.

Okay. So when you did launch the development or announced the development of Silicon One, a large part of the investment community really perceived it as a move to participate in the growth that white box solutions were seeing in the industry. Is that kind of the right way to think about it? Is that the strategic direction that like Cisco was really intending? Or like how should investors really think about that development of Silicon One, particularly when it comes to the threat of white box?

Yes. So 2 things. One, we -- Cisco has always been building our own silicon for as long as I can remember. Going and really having a clean sheet architecture for building Silicon One enabled us to have significant innovation in silicon. And to your point, we position that so that we can not only use it in our internal systems, but that we could also go and address this white box market. Because as you well know, and I think not only everybody on the call is aware, that white box market has been growing every year for probably the last 7-plus years. And what I would say there is, we view the white box market share as an aggregate of the data center market share. It's really a mechanism for tracking the hyperscale networking needs because they are, by far, the largest users of a white box switching. And we see -- this is an ability for us to really expand our total addressable market by going after that market space.

Okay. Got it. Reminder again, if anyone has any questions, I don't see any yet, but if anyone has any questions, please feel free to send them in. Jonathan, wanted to change gears here a bit. Talk about Optics, which you now have the responsibility as well for Acacia, and I think there are probably 2 parts to this, but let's talk about Acacia and how critical it is to Cisco's optical framework before we move to how critical it is to switching and routing as a product solution as well?

Sure. Well, look, we see overall, a huge opportunity for not only growing our Optics business, but plugging into non-Cisco hardware system. So our traditional optics business has been very much a -- attach the optic to the switch or router. I kind of -- you might say it's the -- would you like fries with that model. And now we've not only going to continue that model, but also expand that into attaching our Optics into non-Cisco systems. And so just to -- so everyone is aware, and while it reports to me all of the Optics, I don't see the pricing or volumes for specific network equipment manufacturers. So that's firewall. So what traditionally be a traditional Cisco competitor, the sales team reports up through my team, we have that firewall from the traditional Cisco sales team. And I just want to make sure anybody who might be listening from a competitor, I don't see your pricing, I don't see your volumes, I see just aggregate volumes and aggregate dollars amount because we want to make sure that, that is a business that continues, and we service those customers the same way that we service our internal customers. And that's really important for us to be able to continue that model, which is similar to what we're doing from a -- we love to do from a silicon perspective. So if someone else wants to -- network equipment manufacturer wants to buy our silicon. We're happy to sell them our silicon and we'll keep that firewalled off as well.

So when we talk about the Optical segment, traditional Cisco business has been well adopted in the metro, particularly in North America. [Audio Gap]

Understood. Because I think companies are looking at single lambda spectral efficiency as the measure of total efficiency of how to build networks. And while spectral efficiency is really important, in my mind, it's really legacy thinking to view that as the only measurement of efficiency because it doesn't take into account how real-world traffic flows through infrastructure. And we have shown mathematically, and in reality, that the entire network approach that we've been talking about called routed optical networks can drive significantly higher efficiencies in performance. In one scenario, we're able to shrink the number of landers utilized from over 30 to 2 because our approach has significant efficiencies with stat mux gains using IP with this whole model of routed optical network. We really view it as the future here. And I'm going to emphasize this point because our competitors are really creating a lot of, what we call, fud, fear uncertainty and doubt around this architecture because many of them have been making most of their money around, what I would call the razor, razor blade model. Where the razor was this transponder shelf and the razor blades were the transponders. And now we basically are taking those transponders that used to fit into shelves, and now they're just pluggable optics. And so this approach is a razor, razor blade model of transponder. Transponder something is significantly impacted because of ZR and ZR+. And we've seen customers, we've had press releases that have come out by our customers to talk about the efficiency they've seen, the distance, and we expect this to be really the future of how networks are built.

So following up on that, Jonathan, there have been significant challenges or hurdles in getting the ZR and the ZR+ to the commercial market as well. Can you just talk about what those challenges have been? And particularly now that you have ownership of Acacia, how much of a clear road map do you have to getting into commercial production and then kind of adoption, particularly with the ZR+, where most industry forecast seem to assume that, that's where the bigger opportunity lies?

Yes. Well, I think there'll be opportunity with ZR and ZR+ and the fact that it's an open ecosystem, which I think is really important to really break open the closed ecosystem of optical infrastructure. And the way that we think about it is there's a kind of the 3-phased approach to how we see ZR and ZR+ taking place, first of which is kind of the initial application ZR, shorter reach, more in the metro type of arrangements. But then when ZR+ comes online, we see that expanding to more and more use cases over time, including long-haul types of applications. And just like any new technology, there's always time to market concerns with new technology, but we're out, we're ready to go, and we expect, between now and the end of the calendar year for networks to begin to be deployed in earnest with this -- with this technology. And we've seen pretty exciting customer adoption so far. And the amount of customers who are interested in this continue to go up on a month-by-month basis. So we're very positive about where this technology stack could go.

Okay. And final one on Acacia. How should I think about synergies when I take like a pluggable product and that Cisco also has the switches and the routers? How does it enable kind of selling a full solution to a customer, which might be a more high-performance system than some of your competitors can develop? How synergistic is this combination?

Yes. So the view reservoir we're talking about with routed optical networks. So there's the simple use cases where it might be more of a data center interconnect. And then you look at the whole network approach in a metro or in a global network. And this is where we really see the efficiencies moving forward for really, at some point in the future, when you get to Phase 3 of this approach is you actually replace all of your ROADMs as well with this router plus ZR, ZR+ optics. And so each 1 of your hubs is able to have just a phenomenal advantage for stat mux gains with this new architecture. So Phase 1, there's a lot of savings. You get to Phase 2, and there's more savings and more use cases and you get to Phase 3 and the market is big. And the way we've calculated it is that we see over $1 billion of potential addressable market could be addressed with ZR and ZR+ optics over the next several years. So we're we're excited about where things are at and where they can go. And obviously, the market will dictate the outcomes, but the feedback we've got so far has been positive.

Great. That's helpful. Let me just again, just take a pause here and see if anyone has any questions, please send them in. We'll definitely go ahead and ask it on your behalf here. Jonathan, let's again switch gears here to talking about 5G. And there's been increase in, I would say, excitement around 5G just given all the North American service providers recently talked about their CapEx plans. So on a broader level, how is Cisco's portfolio going to benefit from that? What are the different levers in the portfolio to really then capitalize on the spend coming through from the service providers, particularly if I focus on North America first?

Yes. So we -- and I've shared this before, but we view investments traditionally following the 5G deployments. And it could be ahead of the boat like with the bow wave or it could be slightly behind it. And there's really 3 major decisions that we see as being made and 3 major investment areas. One is obviously the radio access network, and we can talk a bit about that. The second is really in the IP and optical infrastructure and automation for that infrastructure. We have a phenomenal portfolio, not only in the IP side, but the optical side and then, of course, how we see those coming together with these routed optical networks. And then of course, there's the mobile core side of things as well, where we helped a majority of the world's folks move from 4G to 5G SA, and now a lot of them are making 5G SA decisions. And we see really a lot of decisions being made over the next 12, 18, 24 months in each of these various areas. But the decision time lines really come down to areas where countries have the highest ARPU for mobility plus the spectrum is available in that particular country. And there's lots of different mechanisms for spectrum to become available. So, to your point, in North America, we now have the point where spectrum is very available, especially after the last C-band auction. And we have very high ARPU in North America. So that is driving a lot of the investments. So where we see the -- really the investments taking place, first is around people need to build out their IP infrastructure. So it starts at the tower, starts at the cell site, and moves back. And typically, it's not just about the cell site router and what we've then would call the pre aggregation and the aggregation, but it's really -- it's rebuilding the entire metro network. So I'm in the Bay Area here near San Francisco. So the entire metro region needs to be rebuilt to handle the expected amount of bandwidth that's going to be needed with 5G because I don't know about you, when you get your first 5G phone, first thing you want to do is a speed test, and if that speed test is disappointing, you're going to say, why did I just pay $1,000 for a new phone if it's the same speed is my own phone. So it's really important that, that user experience be great. And in order for it to be great, you actually have to go and invest in those 3 major areas. So we expect -- and what we have seen is customers have definitely started to invest around the world in those 3 areas. Oftentimes, they sign the first contracts around radio access, and those are multibillion-dollar deals. So those -- you see press releases for those. Often, you don't see press releases for the IP and optical infrastructure upgrades because that's more of the course and speed of things. And then you may or may not see press releases for the mobile core side because it's a significantly lower investment area than the IP network as well as the radio access network. So that's, that, I'm happy to answer more questions on 5G that you might have.

Yes. No, that's helpful. I mean, do understand some of the opportunity here is on the routing portfolio. So I wanted to also get your thoughts around 2 competitive threats that have increasingly being discussed. Firstly, white box. I think some of the domestic telecom operators have, at least sounded out that they could potentially use white box routers, not that we've picked up anything in terms of deployment. The other one that I did also want to check on that front was increasingly some of the optical system companies talking about having packet optical pollutions that start to imitate or look essentially substitute somewhat what Cisco does on the routing front in an IP optical network. So I wanted to get -- go through both of those and kind of see how you're thinking about those competitive threats?

Sure. So this gets to the software-only model. And our fundamental belief is that we have a significant market lead as it comes to a stand-alone network operating system, which we call IOS XR. It's running the majority of the world's infrastructure around Internet connectivity and private networks and even in the government space of multiple governments around the world. So it's a highly pressure-tested OS, and we have -- as part of that December 2019 announcement, we have gone and made that available to run on top of white boxes. We have signed multiple contracts with Tier 1 providers around the world. And I have to go look at the numbers. A lot of the other white box software vendors don't really publish the numbers, but I would imagine that we're bigger than all of them combined in this space. That said, it's still relatively a small number of customers because when you think about it, you really need to have the scale to support white box because you need to have the integration between software and hardware, which means you have to pay someone into the integration. You can pay Cisco, you can pay someone else. Then when it comes to a simple operational things like a technical support, who you're going to call for technical support. Is it a hardware problem or a software problem? Do you call Cisco be the software vendor or do you call the hardware vendor. Another simple thing would be, if there's a hardware problem, who handles the RMAs for that hardware? And hardware does periodically go bad. So you need to make sure that you've got the ability to self spare is what we would call it. If the customer wants to handle their own sparing and oftentimes, there'll be a 2-hour or 4-hour sparing response, or do they want to pay a third-party to handle those RMAs because most white box hardware vendors, they never want the boxes back. You buy them, they're yours, they don't handle RMAs, and that's how it works. So it's important to understand not only the technology pieces, which, obviously, works, and we make sure that does worked well, but also the business model pieces and the operational pieces to make sure that when you add up all those components together, there has to be some outcome for the customer who's going to do it. So that requires scale. So if you've got massive scale, then that could be an approach that makes sense for you. But if you don't have a massive scale, I almost guarantee you're going to pay more going down that path.

Okay. That's helpful. And I don't see any questions here, but I know I have like almost 5 more here. Let me get to the last one that I did want to get your thoughts on. A lot of discussion around virtualizing the RAN or even doing open RAN implementation. How does it -- how should we think about Cisco -- the impact to Cisco here like Cisco is one of the incumbents in the service provider networks, not necessarily in the RAN segment itself, but a lot of the kind of software, packet core as well. So how should I think about impact to Cisco? Does it increase somewhat the value just if we virtualize the RAN? Just help me think through that.

Yes. Great question. Look, from ORAN, vRAN perspective, we view this as only a positive driver for Cisco. We certainly have incumbency in a lot of service provider networks on the IP side, the optical side, the automation side. And this movement to really create an open ecosystem around radio access networks is really important because that's really the final closed ecosystem of infrastructure. I think we've done a good job with the whole market over the last 5-plus years. And really opening up optical, there's more work can be done in optical, but that's making progress, at least in the right direction. But radio has been a closed ecosystem forever, and we believe that an open ecosystem drives additional innovation. And typically, innovation means that you have a lower cost structure for those who deploy that infrastructure, which -- that means you can deploy more infrastructure, you can close a digital divide, it's all beneficial here. So the opportunity has basically been for us to go and drive the orchestration software around the deployment of these ORAN, vRAN ecosystems. Because you think about it, you need to have those data centers, those telco clouds. They need to be extended out much further into the network, and you need to make sure you've got the right level of orchestration software as well as those smaller data centers to enable this to work. And this is where we have been going and investing in companies to drive this ecosystem. We've been funding not only internally end-to-end test beds, but also third-party test beds to ensure this ecosystem actually works holistically. So that makes it easier for service providers to go in and deploy this new technology. So we're excited about this, and we think that it's a lot momentum now than just 12 months ago, and we see that momentum is going to be continuing to grow over the next 12 to 24 months.

Got it. It looks like we're almost up on time, Jonathan. So I do want to take the opportunity to thank you for participating at the conference. And thank you to everyone who dialed in as well.

All right. Well, thank you very much for the opportunity.

Thank you.