Raymond James Financial, Inc. (CSCO) Earnings Call Transcript & Summary

Well, folks, thanks for joining us. This is Simon Leopold, Raymond James data infrastructure analyst, hosting a fireside chat, no fires here though, at Mobile Congress with Jonathan Davidson, who is the EVP and GM of Cisco's Mass-Scale Infrastructure. But before we get into our discussion, we have with us Emily Hunt from the IR team at Cisco, who has some very important opening comments.

Thank you, Simon. I'd like to remind the audience that today's call relates to Cisco's Mass-Scale Infrastructure business only. No new financial information regarding Cisco's overall performance is intended or implied. We may make forward-looking statements, which are subject to risks and uncertainties outlined in detail in our documents filed with the SEC, including our most recent filings of Form 10-K and 10-Q. Actual results may differ from statements made today. That's it.

That was very good. Thank you, Emily. So I've got a set of questions prepared that I will go through with Jonathan. [Operator Instructions] With that, Jonathan, welcome. Appreciate it. [indiscernible] jetlag. [indiscernible] just now adjusting. So it's 3 in the afternoon here in Barcelona. It's more crowded than I thought it would be. So I'm a little bit surprised, but not a lot of the analyst community is here. So I appreciate the opportunity. Maybe we could start by just setting some of the context and just let folks know maybe the boundary conditions, your responsibility, your role and the scope of what we're going to talk about today.

Yes, absolutely. So pleasure to be here.

And folks, if you could mute your lines.

[indiscernible] Okay, we're back. So this is Jonathan Davidson. -- so thank you for your question. Let me go ahead and share a little bit about what I do for the company. As you mentioned, I lead the Mass-Scale Infrastructure group, which we call internally, we call it MIG, because it's fun to talk about planes right now though.

Yes. Not a good time at the moment.

Yes. So under that, I'm responsible for the engineering and product management with regards to most of the -- what we would deem as critical infrastructure products that communication service providers, also known as telcos, web sale companies, hyperscale companies as well as very large enterprises build their own infrastructure. We build the routing portfolio, the IP, optical, cable access technologies, all the automation portfolio as well as the mobile IoT and mobile core portfolios for those customer bases.

So maybe one way to sort of help frame this is Cisco often talks about the market verticals, not just the product segmentation. And so there's been some recent disclosures. But for your business unit, I presume you've got a greater bias for service providers. Could you talk about your mix of market verticals for your business unit?

Sure, absolutely. So inside of the Mass Infrastructure Group, which roughly translate externally from a financial world into Internet for the Future is the new subcategory. So you could see how that is doing publicly now since Q1 of this fiscal year. The -- you're right, we are definitely organized technology area and then those technology areas applied to specific verticals. We don't break out the verticals by subgroup, but certainly, will classify anything in the SP/webspace is the predominant portion of our business. So when we break out SP, unless it's denoted, SP -- looks like [ Rachel ] you're still having challenges, if you could mute or...

I think [ Rachel's ] coming in with a morse code system.

Yeah. So we don't break it out by our subgroup [indiscernible] for the future, but we do break things out by as I described earlier. So what I would say [indiscernible].

Could you guys mute if you're not actively speaking? Go ahead, John.

All right. So what I would say is we don't break it out specifically. However, we do have the predominant portion of our business is to service providers and web-scale companies. And from that, the rest of it is focused on the enterprise, and we don't touch typically commercial expenses.

And given that we're coming from Mobile World Congress, you had a couple of announcements this morning. So maybe if we could get some highlights of what Cisco is talking about at this trade show.

Absolutely. What would be really helpful to describe a few things that we actually announced today, which is around the private 5G as a service, which is a technology that we have been working on for the last couple of years. We actually had our first announcement as part of that back in February of this year, talking about how it is in conjunction with WiFi 6, the ability to offer outcomes to our enterprise customers. And today, we had an extension of that announcement talking about the specific verticals that we were talking to and specific customers in those specific verticals as well as one of the service providers that we've been working with in the U.S., which is DISH. However, we certainly are working with more carriers around the world.

And so when we think about 5G, people don't usually think about Cisco as a 5G play because we're, I think, generally in the investment team, we focus on sort of the radio side of the world. Maybe set a little bit of context in terms of how you see 5G within the Cisco portfolio?

Yes. Look, there's a few things to look at with 5G. So First of all, the 5G is -- obviously we're here at Mobile World Congress. There's a lot of people talking about phones, people are talking about radio infrastructure. Cisco does not participate in either one of those. However, one of my customers said to me once many years back, which was mobile networks aren't mobile for very long. They have to get through a wired infrastructure. And that's where we see the -- one of the opportunities for Cisco is in the build-out of new backhaul infrastructure, new metro infrastructure to handle all of the capacity and bandwidth necessary. And we've been seeing build-outs of 5G infrastructure for several years now, but we're certainly at the point where most carriers in -- around the top, I would say, GDP countries around the world are in the process of deploying their 5G infrastructure. That's kind of part 1. Part 2 is we're working with customers on how they can monetize their 5G investments. We're doing that through our mobile IoT control center business, which has nearly 200 million IoT devices on it. We were the first to actually qualify on a 5G SA. We were the first to qualify a 5G connected car. And we will actually use that knowledge to actually extend that into that private 5G domain that I was just describing, which is yet another area where our CSP customers can generate net new revenue. And then last, but certainly not least, is you would think of the core of the mobile infrastructure is the mobile core. In 4G world, it was called the EPC, and this is where all of your billing and your subscription and the really experience for your mobile phone exists. And so we are -- continue to be a market leader in that space. We have over 1 billion consumers on our 4G network. T-Mobile in the U.S. with over 100 million subscribers is working on our converged 4G, 5G core. And we're very excited about the technology we've built in that space.

And I guess one of the things when we talk to industry participants about the -- basically the generations of wireless technology, they've been these 10-year cycles. And so 3G to 4G. We're only a few years into 5G. And what do you see as really the differences? And where do you think we are in this fifth generation?

Yes. We think of 4G is predominantly has been about the ability to provide new connectivity and really 5G is about experiences. And with the massive amount of bandwidth that has been made available or in certain countries are still being made available depending upon your city or geography, there's going to be a whole new set of experiences that can come along with that. For example, private networks for cellular spectrum is not new. That's been around -- we have over 100 private 4G customers in the world today. However, we certainly expect to have significantly more than that in the 5G world because of the time that we've taken to build the infrastructure necessary to offer this as a cloud-based service. We think that, that in conjunction with the additional capacity with the conjunction of also the need for high-value asset tracking is really, really, really important.

And I guess one of the terms that we've been hearing that maybe isn't strictly 5G related, but has come up more in the context of 5G networks is mobile edge compute. And that feels like that's something that's probably more Cisco oriented that you probably have some role. If you could maybe elaborate on what you think that term means and what it means to Cisco?

Mobile edge compute, specifically?

Yes.

Yes. So I think mobile edge compute is often a overused term. I'd like to just refer to it as we want to be able to move storage and compute closer to where either the humans or the things need the service. And there's lots of different use cases for that. So first of all, and the most critical, obviously, is a lot of the Internet traffic today is video. So offering a better video service to customers can be achieved by moving the CDN closer to the customer themselves. So that would be the content delivery network. Once you have that computing storage closer to the consumer, you can actually start to do more interesting things with it. You could deliver gaming services more closely or you can actually think about the traffic going in the other direction. Instead of the traffic going towards the consumer, the traffic can come from things up to the edge and you could actually do processing at the edge to do one of the few things. One, decrease the amount of bandwidth going back to the core of the network or create intelligence at the edge of the network so that those things can take the next appropriate action. And so we're big believers in the edge. We're big believers in the fact that you need to move storage and compute closer to the users. In fact, here in our booth, we're showing in partnership with several different technology partners, a set of use cases that can be utilized more effectively for delivering these types of services and applications.

Now one of the interesting aspects that we hear in terms of 5G use cases is around private wireless networks. And so it does come as if operator spending doesn't grow that fast. And so many of the players in this market space have talked about private networks as the real growth driver and sort of the inflection. How do you view these kinds of opportunities predominantly, I presume, with enterprises for private wireless networks? What does this mean to Cisco?

Yes. So let's just start from the top. As mentioned, we have over 400 different 4G enterprise deployments. Now it sounds like a lot, but in some cases because people didn't really think that there were private 4G networks, but it's mostly mining operations and things where they're far away from public infrastructure. When you move to 5G, we think that there is an ability for 5G private networks as well as WiFi 6 to live in conjunction. So if you take a manufacturing facility, and you think about mass scale customization. So I need to potentially reconfigure my factory every few days in order to run a different set of mass customization of products for a set of consumers. Doing that when everything is fixed, not like fixed in place, but their connectivity is fixed makes things significantly more challenging. So we need to be able to understand how we can go and deliver these services in a way that's not only simple to deploy, but actually enables the carriers to go and actually offer this to all of their customers. They have the spectrum necessary. They have the ability to do the on-premise RF management. And now we're giving the ability in a very simple way, offer a net new service to their customers that they have not been able to offer before. So we think that there's a lot of potential around private 5G, but we're still very early in this market.

And I guess two follow-ups to that. One is, how big is this market in the sense of -- keep hearing about this as sort of the holy grail of 5G, but it hasn't come through. Is it just that new audience [ meeting ] us and the investment team, we're not patient enough? Or is there something more to it? What makes this really worth the investment?

Well, I've never heard that the investment community is not patient. That's the first time for me. So I'm surprised by that potential option. Seriously, though, when we're thinking about this, first of all, you have to build the network [indiscernible] a telco or a CSP, you have to build a network to win the consumer. And right now, it's a competitive imperative, 5G is exciting and consumers are going to make decisions based upon the 5G service that they're going to deliver. However, most consumers are not willing to pay more for their 5G service. So the profitability of 5G by many people's beliefs is going to be around net new service creation. And one of those big net new services to be created is going to be around private 5G. How big the market is? There's lots of analysts out there throwing out a lot of numbers. I would just say that it's still very early, but we believe that it's going to be an important market. And we put a lot of energy over the past 2 years from an engineering perspective to build something that can be deployed and scaled worldwide.

Now the other aspect that we're hearing more and more about in 5G is around fixed wireless access. So this is using wireless as an alternative means to provide consumer broadband. So a number of operators have talked about initiatives. Could you put this in context for Cisco?

Absolutely. Well, the technology stack that we've built, whether it's an on-premise technology stack through our traditional mobile core model or whether it's through private 5G as a service, there's not a whole lot difference between a private 5G service and a fixed wireless access from a technology perspective. And so we certainly believe that when, I would say, post the consumer's network being deployed, we're about to see a lot of interest. And we're seeing that in certain geographies like in the U.S., where fixed wireless access in certain specific geographies where a telco may not have fixed -- or may not have a fixed wireline network where they're using fixed wireless to go and capture market share. And we think that is absolutely going to continue and then that market is going to become more competitive over time.

So I want to pivot away from 5G specifically and talk a little bit about how your customer verticals have shifted, particularly with the advent of public cloud adoption. And so that represents, in some ways, partners and competitors to your traditional customers. How do you see the advent of the public cloud entities affecting your business overall?

So when we talk public cloud, we're talking hyperscalers for the most part. And so hyperscalers have had a significant positive impact on Cisco's business over the past several quarters. I think Chuck said in the last earnings call that in the last rolling 4 quarters that we've seen 100% growth year-over-year. I think that speaks to the fact that it's also very material, which is the next question. When you think about the overall percentage, I think we're right around 30% for web as it applies to the overall SP business. Because, as I mentioned earlier, web is part of SP and how we track it. And so it certainly is material to the business as well. Now private networks are still going to be built. There are workloads that only -- maybe not only for security or privacy reasons people want to keep on-premise, but also for cost reasons, people want to keep those workflows on-premise. That doesn't mean that there's not going to be significant growth in the public hyperscale space. We cannot expect to see that continue from a workload perspective, but we also expect to see the private data center space to be very healthy for the foreseeable future.

And it seems as if the terminology has shifted in the last year or so from hybrid cloud to multi-cloud. And I feel like there's a bit of a nuance that might not be well appreciated between hybrid and multi-cloud. Could you maybe unpack that a little bit for folks?

Sure. Well, the applications are distributed. And so while you might have your front end on 1 given public cloud, you might have your storage on another public cloud and then the actual customer data might actually be a private data center. So the multi-cloud aspect is really important. And this is actually one of the things where we've been spending a lot of time specifically around what we call full stack observability. So when you're disaggregating your applications in such a manner, you need to want to make sure you have the appropriate security, you need to make sure that you understand the performance of those applications to your business. And this is where things like AppD come into play. But also you need to understand is that if there's a problem with an application, is it the application, or is the network between my user and the application. And this is where things like ThousandEyes become so important to understand. Think about the global footprint of all of the networks around the planet plus the hyperscalers and what's causing my application problem? Is it one of the transit providers? Is it my last mile provider? Or is it happening inside of the hyperscale network that's happening inside of the public cloud network? Is it a congestion issue or is it a problem with my database, right? So we can help the application owner very rapidly find out where the problem actually is and to resolve that problem very quickly.

Now in this space, I want to follow up in that. It seems as if Cisco is one of the very, very few companies that touches multiple aspects of the cloud, of the infrastructure. And so you can make this argument of having complete visibility this idea of a full stack, yet it seems as if in many aspects the buyers are buying point solutions. Why is the market behaving the way it is in solving the problem with point solutions instead of going to this full stack? Is it too immature? Is that an issue?

I think the full stack observability is a term hasn't existed for very long. But I think it exists this way because we've started from solving point problems. And I think people are just in the last 12 to 24 months realizing that there is a broader problem that needs to be solved, and that is the disaggregation of the application. And now that we live in this multi-cloud world, how do I make sure that I'm delivering the customer experience. And the delivery of that customer experience isn't just the application on the phone. It's all of the capabilities behind that, and it's been very challenging in the past to be able to decompose and understand where the challenges might have been. And so we're -- we think that as time moves forward, the application developer is not going to be a network expert. They're not -- today, they're not going to be. What we need to be able to do is help them narrow down very rapidly what the challenges are so they can fix the problem. So we see more convergence in this space over time. specifically around helping people understand that full stack.

And how do you see this changing your line of business, if at all?

For the mass-scale infrastructure, we are going to do our best to continue to give out the telemetry needed so that we can provide the right level of data so that you can look at that full stack. We are still going to focus on moving bits very rapidly at a very cost-effective place and also make sure that it's automated. But we're also having conversations with people about how we can make sure the network is more driven by APIs as well. So if I see that there is a challenge on the network, I want to be able to dynamically change pass as the application owner that sits on the edge of the network. And today, I don't have that ability. It's a very coarse-grained control. And if I see that there's a challenge, I have to call someone or open a ticket and have them physically reconfigure a device so that, that traffic can get moved somewhere else. That should be more dynamic in nature. And today, it requires blackout situations in order for network traffic to move. But we think in brownout situations that there should be a better response to the availability and resiliency and SLAs inside of the global infrastructure that could be managed more tightly than what's happening today. So that's my job is to provide that visibility and that transparency, that API capability through our automation suite but then it's up to the application developer to take advantage of it.

So I want to pivot to several questions that were addressed on the earnings call. I want to put the answers or have answers that are in the context of your business unit. So I hate to use this phrase supply chain. But basically, every earnings call we've gone through, we have to talk about supply chain. Pushed it off until now, but inevitably, people want to know. So could you talk specifically to your business unit, what you're experiencing in terms of supply chain constraints and what you're doing to respond to those?

Absolutely. Well, with my -- the supply chain constraints in my business are really no different than the rest of Cisco. And I think Scott and Chuck have both said it very well around what the challenges are, it certainly is, over the last 18 months, been the most challenging supply chain of my career being in the space for over 27 years. And so we certainly are seeing more predictability come into the space, but we certainly would like to see significantly more improvements. And I think that goes for anybody who's in the hardware domain, whether you're in mass-scale infrastructure or any other parts of Cisco's business.

Great. It looks like we've got a question, a hand raised. So if we can unmute Sami Badri, do you want to ask a question?

Yes, please. I -- Jonathan, just first, I want to give a shot to make sure that you've answered all the comments on supply chain before we kind of just jump on something else you talked about earlier. Have you kind of given the remarks just to check?

Yes, I'm all done on the supply chain side.

I want to go back to some of the product order mix and some of these technologies and new opportunities you're talking about with 5G, mobile edge, private networks et cetera, and even just the idea of a killer app. When we go back to your product mix and specifically service provider product orders, what percentage of those orders are actually going towards new applications, new technologies versus what percentage of those orders are going just to continue to support older infrastructure, older apps?

So a great question. We don't necessarily break it out that way. It would be difficult to do that in any kind of meaningful fashion. So what we stay focused on is obviously keeping our existing customers very happy. So that they continue to fill up line cards into their existing [ cases ]. But we also are laser-focused on new franchises and new use cases. And so I think I've said this before, take hyperscale as an example. When we go and win a use case in a hyperscale account, we're typically winning 1 use case like a top-of-rack switch or an aggregation switch or a data center interconnect router or a peering router. There's a number of different roles, and you have to win each of them one at a time. And then when there's technology transitions, there's opportunity to -- for typically where they will look at new vendors. So we live in a good time right now where the people are making their 400-gig decisions. And so we've been laser-focused on making sure that we have the right technology at the right time to do that. But that's happening in the SP domain as well. I mean we're at Mobile World Congress. Obviously, a lot of those conversations are happening here, talking not only about 5G and what we can do in helping customers with ORAN and our backhaul, but also just how they're transitioning to a 400-gig infrastructure in what we're calling routed optical networking as well.

Got it. One other question, if I may as well. You made a reference earlier to, at some point, investments need to go from wireless into wireline. When you think about what all the carriers and other service providers are trying to achieve, where are they holding that "accountable?" Where are they in this transition? Have they shifted now into more infrastructure-based investments to do more with data centers and routing? Or are they still so focused on radio access network, wireless, et cetera.

Yes. So I think my exact point was that mobile networks aren't mobile for very long. They become wireline networks and carriers have to invest in both the wireless network portions or the radio as well as the wired network, almost simultaneously. Because if they have a new radio and the radio is advertising that it has 5G connectivity and you go and you do a speed test and it's slower than your 4G network, what you have is a cranky 5G customer. And so it's important to not only have a new radio but also to have a new backhaul or metro network in order to give the right level of capability that the customer expects when they buy into this new experience. So hopefully, that answers your question. If not, let me know.

So one of the things that Cisco has been reporting quarterly is the order growth. So that's been quite extraordinary, presumably your business unit is seeing similar kinds of order patterns. When we think about the lead times, what kind of lead times did you have historically, so pre-pandemic, and how does that compare to the lead times and visibility you have today?

Yes. So the lead time is really -- it depends upon the individual products. So if I have, let's say, for example, one of my routers is the NCS family. But inside of the NCS family, I'm going to have literally hundreds, if not thousands of different product IDs or SKUs for customers to order. And the lead time can vary based upon each and every single one of those SKUs because it really comes down to the specific supply chain challenge. So if there are 1,000 parts that go into a line card, and I'm missing 1 of them, I don't get to shift that line card. So it really is where are the challenges that we're facing and what can we ship? I'll give you 1 great example where we have a cell site router product, and we were unable to -- we're having a supply chain challenge with 1 particular component that prevented us from being able to ship copper ports. Most of the ports are fiber ports. And we have 1 customer that -- we said, "Hey, we're going to have to push out because we got decommits from one of our suppliers." And they said, we don't need the copper ports. Why don't you -- can you rebuild the product for me without those? And we said we can't rebuild it, but we can take those copper ports off. We have to change the software. Have to change a little bit of hardware, but we can then ship to you those things next month, and they said, fine, ship them. So we are trying to be as dynamic as possible, which is putting a lot of strain on the hardware team and a lot of strain on the software team to be able to go and meet the customer's exact requirements. And we're being very flexible and thank you to all your customers out there who are being very flexible to help us with that particular challenge. And again, that was a transient issue. That was a transient issue for a few months that we were able to resolve, but these things pop up and they go away pretty rapidly.

So is it unreasonable for analysts to ask about sort of the average lead times because I appreciate it's going to vary by product. And I feel like I've heard anecdotes that lead times have gone from 12 weeks to 52 weeks for some things off in 6 months. So I'm looking for some quantification of what's typical.

Yes. So certainly, some of our suppliers are in the 50, 52 week and in some of the case, beyond. So that means that we're having to place orders at that lead time. So we're placing order for that lead time, that means that for some of our components potentially that are lead time from products. If they exceed our already forecasted demand, they're going to be outside of that window as well. So think of it as a rolling demand that you have to have. And if you misjudge on 1 component, then that stops up the whole challenge. So lead times are definitely at an all-time high. And it's always fair to ask those questions. It's never good -- I'm never going to tell you it's not fair to ask a question, it's a very difficult question to answer with any sort of specificity other than in this space, the lead time is the longest that I've had in my career.

I see we've got a question on the line from [ Peter ] [indiscernible]. [ Peter ], did you want to ask a question?

Yes. Thank you, Simon. So some people talk about optical becoming a larger portion, if not the majority of cloud networking spend in the future. When I look at this 100% order growth over the last 12 months for cloud, does that already reflect this higher optical mix? Or are those optical orders this richer mix of optical? Is that still in front of us? And I actually have a quick follow-on also related to optical. Just as we think about 5G, you talked about bandwidth, metro network built to the traditional wireline. But when I look at the 5G RAN, there's actually more fiber in that as well. With the ORAN coming on and some of the more standard components, sort of hope to use more standard components. I'm curious, does Cisco have more opportunity in the RAN because of your position in optical and frankly, opening up your silicon to kind of white box networking devices. That's it.

All right. Then that -- they're 2-parter . All right, I'll do my best to answer both parts. So I think, first of all, let's define what optical is. So we define optical as being different from optics. So optics are going to be the transceivers and optical will be the technology that actually moves the bits from point A to point B. Now our belief is that there is going to be a transition in the market towards what we call routed optical networking. And this means that takes traditional transponders and moves them from being a shelf or a separate box or a device and turns them into a pluggable optic which you then plug into a router. That's Phase 1. Phase 2 is you actually replace ROADMs that exist in the optical infrastructure with routers that have those same pluggable transponders. So you could think of it as -- and then Phase 3 is actually putting private line emulation onto that same infrastructure. But it depends on how you slice up the pie. Radio access networks themselves are between $30 billion and $40 billion a year depending on who you talk to. Optical can be between $10 billion to $15 billion a year. And then routing is below $10 billion a year, depending on who you talk to again it's from $8 billion to $10 billion a year from a TAM perspective. Now our belief is that the optical TAM will start to shift over time as routed optical networks becomes more prevalent because it will move from the optical domain into the optic transceiver market. Now it really -- where do people decide to put that -- those dollars? So my competition who can be in the legacy optical space to also build those optic transceivers. They may decide to keep that in the optical TAM, or they may decide to move that into their own routed channel. I can't control how they do it. There's been no clear demarcation in the market. And so it's difficult to say where the TAM is going to go. But what I can tell you is my belief is that the market will shift towards this new architecture because it's significantly more cost-effective to do that. So that was a long-winded answer to your first part of your question. And then the second part of your question was really around is ORAN an opportunity. I think any RAN is an opportunity. The reason why ORAN is an opportunity in addition is because if there are customers who have a desire to deploy ORAN technology, then that means that they're going to be talking to suppliers that are not legacy suppliers who are siloed and are trying to offer a single end-to-end package. Most of the ORAN suppliers, like an AltioStar, they're more willing to partner with a Cisco or other IP-only company. And so therefore, we have -- our addressable market is higher for those customers who are interested in ORAN-based technologies from what we've seen in the marketplace. So we're very we're very bullish on ORAN for the simple reason that it opens up a closed infrastructure of RAN. So a $30 billion to $40 billion RAN market is still closed, it's locked. And even though there are standards, people do not do any interoperability testing between vendors, which is fundamentally changing with ORAN, where the CSPs or telcos today are forcing vendors to interoperate their ORAN implementations, which is great for the industry because that will drive more rapid innovation, which will decrease costs, which means more connectivity from consumers and everybody benefits.

So Jonathan, I wanted to follow up on the first question about optics within the platform. Just to paraphrase to make sure I understand the answer, is that, I think in the past, we used to talk about on Ethernet platforms at 100-gig, maybe those ports those transceivers, where a minority of the total bill of materials of a fully configured switch. At 400-gig, it gets to be almost half of the bill of materials when fully configured. And the point I think you're making, and I want to make sure I understand you, is that you're not -- you don't really care how it gets counted, that it's all part of the solution to enable the switch and whether a bigger portion of the bill of materials is optical transceivers versus the other stuff isn't really a relevant question.

Well, I think it's important to understand because it's -- you'll have to ask your other people like me and other vendors, how they slice and dice things. But I do think that the net effect is what you call. So I would go back to 10 gig going all the way to 400 gig. And back in the 10-gig world, over 80% of the cost was in the router or the switch and the rest of it was in the optic. You go to 400 gig and it's pushing much more towards more than 50%. Now not all optics are created equally. So you have short-reach optics, you have kind of median reach. You've got long reach, you have extra-long reach like you've got -- then you've got something that's at a digital coherent optic which is more expensive. So depending upon all 400 gig ports are not created equally, but it's a short reach port, significantly less cost than a digital coherent 400-gig port. So it depends what you're plugging in to get that ratio right.

Great. No, I appreciate that. I think that's helpful clarification. Now I wanted to also ask about within your business unit, the specifics of Cisco's software transition. And so many of us are familiar with the product cycle around the Catalyst 9000 in the enterprise switching that has a software subscription business model. And we understand, I think, the broader scope of Cisco with more software, more recurring revenue, could you take us down to mass-scale infrastructure and help us appreciate the software narrative within your business?

Absolutely. So we started right around 4 years ago to transition all of our new products over to this new software-based model, which is pretty similar to what the enterprise team was doing at the time. And I'll go -- if you go into the -- the way things were a decade ago, you would sell a router and generally significantly less than 10% of the router cost that we sold it for was in software, significantly less than 10%. What we've transitioned to is with our new routers that we've been bringing to market over the last 4 years, as we shipped new routing devices or new families, we shifted anywhere between 40% to 60% of that cost of the ASP to software. And then we made a portion of that 40% to 60% renewable a term-based license. And so people could buy a 3-year license or they could go beyond that. On average, they've been doing more than just over 3 years. And so we're coming up -- starting in this -- our fiscal year. So we're halfway through our fiscal year, that those early renewals are coming on now. We didn't do it technology area. We did it new product by new product introduction. So what that means it's a very slow ramp because if we launched a new cell site router, only that cell site router was impacted. And you have the same cycles that you have to go through with any new product. You have to talk to the customer about it. you have to get a design win, they have to go and test it for 12 months, and then they put it in the network and then they roll it out. So that takes 18 months. So you changed the pricing model 4 years ago, it took 1.5 years to really start for that product, and we've been slowly doing it. And ever since that all new products that we've come out with have moved to this model. So it's about -- it will take years for this to fully ship, but we're happy that we're on the path.

And what's the customer reaction been? How do they see it? And I guess the other thing I want to understand is I know the pitch on the enterprise side was it separated the life cycle of the hardware from the software. Is it any different for your line of business?

Well, you have to have a conversation about value. And we never have that conversation with regards to software that sat inside of an embedded router before. And so it also is part of our disaggregation story. So in December of 2019, we came out publicly discussing the fact that you could just buy our software. And if we didn't make this transition, people would say, "Oh, well, I'm just going to buy a white box and your software is basically free, so I'm just going to buy the software. So we -- no, no, no. There is significant intellectual property in our software, we believe it's the best software for network operating systems on the planet. Here is the true value for that software. And the customers fundamentally agree with that. And even the customers who went down the path of white box and they're still willing to pay for our software, they're paying a very healthy premium for Cisco iOS software. And so the conversation isn't a challenging conversation, but the conversations had to happen. But once you got past this, it's like, hey, do you want to pay for the additional innovation. They've got an extremely large software team every day that are building new features and new capabilities into this network OS. And if they don't want the new value, they don't have to pay us for it. But if they want the new value, they have to pay us for it.

I want to follow up on that. Because I feel like in many of the conversations Cisco has talked about Silicon One, the chipsets as a breakthrough business model. We'll sell you our ASICs. I've heard last about this software aspect. And I guess I want to get an understanding of, is it material? Is it a breakthrough? And is it something where the customer can buy your software and run it on somebody else's platforms or their own platform, because that's maybe not so clear.

Right. So 100%, you can go and buy our IOS XR software and run it on a third-party device. Now it's not as simple as buying a PC and getting windows to run on it. It's a complex engineering key. So there's a lot of integration work that has to happen in order to make that work. And so you have to be able to -- and they have to pay us for that integration. It doesn't come for free. So the really the question for the whoever does this model is can they amortize the cost of the software and the integration over just buying an integrated system? And some of them can because they have such scale that it makes sense to do that. And so we absolutely, I would say that it has been a good business for us to be in. And it also just shows our willingness to participate in all different business models. And that flexibility makes customers feel good.

So is it taking too big a leap to tribute the growth you've had in hyperscale with this shift in the business model? Are they synonymous events? Or...

I would say the willingness to transact in any business model definitely has had a extremely positive impact with hyperscale. And I think the facts we reiterated on the earnings call about how many of each one we're working with different business models. But we all of the business models have seen success in the hyperscalers, and we're happy to continue down that model. Because really it comes down to choice and no one wanting to feel locked in to us [indiscernible]. Some people don't want to be locked in silicon. Some people don't want to be locked in on systems. Some people don't want to be locked in on software. So you have -- as long as you can offer flexibility across it, then they feel more comfortable with you.

So I have to imagine there was a lot of internal debate about going down this path, evident that you decided to do the disaggregated model. When you sort of reflect back, what do you see as really the trade-offs because in a sense, you may be giving up opportunities to sell larger solutions, how do you sort of solve this debate internally?

Actually we are not -- it was not a long debate. It was a pretty quick debate, where we came together and we said, "Hey, we think the market needs this." Everyone said, okay, and we moved forward. Changing your back-end IT systems to work with all these new models took significantly longer than actually making the decision to do it. And when you meet the customer where they want to be, they are going to be happy with you. And if you don't, they're not going to do business with you anyway. And it actually, I think, we've created significantly more opportunity for us by being flexible than if we were to have remained as a systems-only company. So I think it was absolutely the right decision, and I'm glad we went down that path.

Great. So I had a couple more questions just specific to the routing business. So sort of the bread and butter of mass-scale infrastructure. And this maybe sort of touches on 1 of the earlier questions, but it's this idea that traffic has been growing dramatically, and the router market has been flat to down, flat to down, flat to down. And I think the investment community has all been waiting for this inflection point, which we haven't seen. And so it feels as if it's always on the horizon. And I have to imagine that now that Cisco has introduced price increases that foundation of 5G has been placed in the radios. Can you give us some sense of how to think about the growth of whether you want to talk about your business or the market? How should we think about that?

Well, I think we could just go back to the Q1 and Q2 earnings report where the Internet for the Future space is 50% growth in Q1 and 42% growth. And that's revenue growth, even in a sort of a challenged supply chain environment. And I think that speaks volumes about where we are in the industry. And I think Chuck also said on the earnings call about how we saw growth across all geographies as well, right? So we're definitely seeing the fact that we've got a great portfolio regardless of whether you're a CSP or a hyperscaler or a Fortune 100 company that's building out their own infrastructure, and we have the right solution at the right time. And we -- I think -- and I've heard this from customers. I've heard this from -- certainly, those who've been in the industry for a long time, that this is the best portfolio they've seen across the market, across the board from cell site router to core router in the history of the company.

Which leads nicely to my next question, which was you've got the breadth of a portfolio. Are you seeing variations in market segments, specifically core seems like it's been slow. Metro sounds like there's more activity. And maybe this is artificial and it's the market researchers who want to segment the market. But what sort of variation do you see in the different subsegments of the routing market?

Yes. That's difficult to quantify in that as you'd have to go down by region, by -- because everyone's on a different build path just country by country, and isn't even CSP inside of a country. And there's usually 3 or 4. They can all be on different CapEx cycles. They can all be on different renewal cycles. Generally they're going to depreciate their gear for 7 to 10 years. If they're at the middle of that, it's very difficult to go and build out a new infrastructure until you completely depreciated or unless you want to go and push that gear elsewhere inside of your infrastructure. So there's no great answer to it other than I would say we're certainly seeing a lot of customer interest across the scope, whether it's cell site routers, full metro deployments, core infrastructure, just globally, on average, we're seeing interest in each of those areas.

So the other opportunity that maybe makes the current environment unusual is the geopolitical aspect of Huawei being pushed out of certain markets. And so we try to quantify, and it does feel that much of the focus in the investment community is really on the RAN because it's a big market. But they play everywhere. So from your perspective, how do you see your opportunity set evolving based on geopolitical concerns operators may have, customers may have in buying from Huawei?

Yes. We definitely have seen opportunities that have come up because certain countries have made it difficult, if not impossible, to acquire Huawei equipment. And so if 1 of those 3 or 4 CSPs inside of that country had deployed Huawei, then they are going to have to transition them out or they're not going to participate when it's time to refresh the infrastructure depending upon that other country's laws. So there definitely have been new opportunities because they have to change vendors. And so we're going to compete just like we would if Huawei was there, like we're going to compete with our technology and our ability to help our customers satisfy their greatest challenges. And so we think that we're positioned well, whether the political situation is friendly or not friendly. We think we've got the best products in the world. And given a fair place to compete, we think we'll win our fair share of the deals.

But I guess one of the question that has come up is it sounds like a lot of the attention has gone to the RAN and that's been where the swaps have occurred. Have you seen or have you won deals at the expense of Huawei because an operator chose not to buy from them again? Has that happened yet? Or is that still off the card?

Yes. That's been happening for a while now, simply because this geopolitical situation is not new it's been happening for years. Started with the former administration. And as we see it, it looks like it's continuing.

Great. So let me just check back with the audience to see if anybody else has questions before we run out of time. Seeing none. So Jonathan, I want to thank you very much for joining us today. Good discussion. Appreciate you take your questions. Sorry about the various audio problem at the beginning.

We have a learning. We've learned something here. That's good. So thank you very much for the time as well. It's great to chat.

Great. Thank you very much. Thanks, folks.