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

Good morning. The conference of this call and the views expressed on it are solely for the use of clients and invitee guests of Cowen & Company. Members of the media are not invited to participate. Any publication, distribution, reproduction, posting, sharing or transmission of this information without the express permission of Cowen is expressly prohibited. As a reminder, we are not interested in receiving and you should not disclose any confidential information. In the event that you inadvertently disclose such information, please notify us as soon as possible. I would now like to turn the call over to your host, Paul Silverstein with Cowen. Thank you. You may proceed.

Thank you, Simi. Good morning, everybody. Thank you for joining us for the 18th edition of our networking leaders, legends, luminaries and visionaries fireside chat series. Before I introduce the speakers, Marty Palka from Cisco wants to do a disclosure. Go ahead, Marty.

Thank you, Paul. This audio cast is educational in nature with no new financial information discussed. We will be making forward-looking statements. Actual results may differ materially from these forward-looking statements, and they're subject to the risks and uncertainties found in our most recently filed 10-Q and 10-K. I'll turn it back to you, Paul.

Thanks, Marty. So it's my pleasure to be joining today by 2 distinguished leaders within Cisco. Gee Rittenhouse, Senior Vice President and General Manager at Cisco Security Business Group; and Rakesh Chopra, Cisco Fellow and Cisco's Common Hardware Group Architecture and Platforming. We're going to structure this as 2 separate 45-minute conversations starting first with Gee, who leads Cisco's global security business, including strategy, portfolio and architecture. Previously, he was the Senior Vice President of Engineering for Cisco's Security Business Group, where he was responsible for development efforts across the business. In 2014, Gee was named one of Cisco's rising stars by CRN after he was recruited by John Chambers to help Cisco shape its cloud and visualization strategy. As Vice President, General Manager of Cisco's Cloud and Virtualization Group, he accelerated the development of the company's software stack, NFV and Software as a Service consumption model. With more than 20 years of experience in the telecom industry, prior to joining Cisco, Gee served as President of Bell Labs, the iconic institution, where he led the extensive transformation of the organization that followed the merger of Bell Labs and Alcatel's Research group. I'll introduce Rakesh after my conversation with Gee, which will get lasted about 45 minutes before we cut over the second session. And Gee, you once again, thank you for joining us this morning. It's great to have you with us.

My pleasure, Paul. Thank you for the invitation.

Why don't we start -- let's start off with a big picture in terms of the market -- the security market in past, present and future? Where is the market? Where does security come from? Where is it today? Where is it going over the next 2 to 3 years and over the next 3 to 10 years?

Thanks, Paul. So the security market arose from the fact that there are active adversaries who exploit vulnerability in systems, businesses, software, hardware, et cetera. And so since its very beginning, there has been this codependency between the market itself and the way attackers exploit some of these weaknesses. And so over time, what you found was there'd be a weakness, let's say, in a device operating system. And so you see the emergence of antivirus software or you see there's a weakness in the complexity of an enterprise and so a firewall emerge. And on and on and on, it went throughout the decade. And so the industry continues to find these point products, if you will, or these businesses around particular exploits. And over the decades, what has happened is this proliferation of these technologies has led to a massive amount of complexity in protecting the enterprise as we find customers have tens, if not 100 different vendors in their environment, trying desperately to protect. And so you see this continued expense given to adopting these technologies. But quite frankly, we're not that much better along. And so what we've seen over the last 2 or 3 years, and Cisco has been helping to lead and shape this discussion is a consolidation of these technologies, trying to produce a security environment that's simple because it's integrated and broad as well as deep, so that you can protect against these particular vulnerabilities. And I think this is going to be the trend going forward as well, broad platforms that try to integrate large amounts of the security portfolio to simplify the task of doing this for end customers.

So Gee, I know you just introduced SecureX, unfortunately, in February. And by unfortunately, I mean, obviously, right in front of the global pandemic hitting hard and we're going to come back to that later on in terms of that key platform for integrating your front end as well as the back-end integration that you've done. But let me stick on the larger picture, in terms of trends and market opportunities. So ongoing public cloud adoption, the ongoing embracement of multi-cloud IT environment, and the organizational trend toward allowing and providing connectivity for any user on any device to any application network from anywhere, obviously, has changed the security landscape. What has been the impact? Can you provide more insights on what's required for true security?

Yes. And so what has -- in the past, we were able to surround the enterprise with a layer of protection. So that inside that layer, you were somewhat protected. And then outside was the Internet and all of the diverse ways of getting attacked. The cloud, of course, has completely removed this perimeter. And quite frankly, for the last 5 years or so, there's been a recognition that, that perimeter was breaking down any way. And so the emergence of new technologies along the lines of what the industry called Zero Trust and these kind of other capabilities, but the pandemic has accelerated that. Customers were on their cloud journey, they were moving to the cloud. It was fairly prescriptive. The road maps were clear and they were taking it step by step. And then immediately, overnight, the entire workforce had to work from home or the entire workforce had to have access to these other applications and whatnot. So we've seen over the last 6 months, an acceleration of a trend that started years ago, but -- well, unlikely, we won't return back to those previous times. And so the trend has accelerated, and we feel that this has long-lasting impact to the industry. Even though the kind of immediate effects of the pandemic to the underlying top and bottom line of the business may have been -- maybe past us, the effect will be long lived.

Gee, to what is then Cisco has shifted its security portfolio to the cloud and away from premise-based solutions?

We've been doing this on this path for about 5 years. It started with -- in earnest with our acquisition of a company called OpenDNS. OpenDNS was unique to Cisco and a center point of our strategy for a couple of reasons. First of all, as you point out, it is a cloud-based platform, cloud native. We currently support 150 million users on it, over 200 billion DNS queries every day. So massive cloud scale platform. But more importantly, for us, the cloud transition forces you to think about security in a fundamentally different way. But again, traditional security is as you look at it from a network perspective, you put yourself in the middle of the transaction, inspecting packets, inspecting files and things like that, to see if there's malware or attacks or whatnot. But of course, when you move to the cloud, like we do today, if you go to salesforce.com or ServiceNow or something, you're not going through your corporate firewall. You're addressing that directly. And so this whole issue of a bump in the wire or man in the middle type of security doesn't fully apply to a cloud-type infrastructure. And so we recognized this early, and I think we could take credit for being one of the first to think about this. And so OpenDNS, not only being a cloud-based platform, was the first in Cisco's history to provide security outside of that bump in the wire, outside of inspecting the packets and inspecting the files. And this led us to, again, continue to push how do you provide security in this new environment, but it also enabled us to do it in a way that is really easy for our customers. Within the matter of minutes, you can set up OpenDNS -- or now it's been incorporated into a product line, we call Umbrella, and a large scale enterprise, the size of Cisco, in minutes. And this kind of speed and level of protection is really important. Another example, more recently, is our Duo acquisition, where we recognize that in a cloud environment, your attack surface is much broader, and we as humans continue to be the weakest link in that attack, and so we tend to click on things. The Internet has trained us to click on things. So we click on things all the time. And so selling credentials is a very large attack vector. And so again, we support this cloud transition with multifactor authentication and perimeter-less security provided by Duo and our Zero Trust. And then finally, we have taken large parts of our portfolio, the firewall, our cloud proxy, our web proxies and things like that and moved it into the cloud as well, to support corporate hygiene, compliance and governance. So we've been on this journey for a long time, and it's now a significant part of our business.

Well, I think you just answered the next question, which is the large part of it, which is how is Cisco positioned to address these ongoing trends? And maybe the piece you haven't really answered, what do you need to get right from here to drive at least market growth and even better to gain share? What are Cisco's greatest opportunities? And what are Cisco's greatest challenges?

So I apologize for you and the listeners that I keep going before cloud and after cloud, if you will. But before cloud, when you were protecting your assets on your premises inside of your enterprise, you would be able to buy a firewall, let's say, you buy Cisco's firewall, but you buy your web proxy from somebody else, you buy your e-mail proxy from somebody else, you buy your DDoS from somebody else, on and on and on, it goes. And you just stack these different vendors and these different capabilities at the edge that the perimeter of your enterprise, and yes, you have integration issues to deal with that I mentioned, and all of that is true. But at least it's -- you're able to kind of, let's say, by a best-of-breed approach and integrate it together. As you move to the cloud, it is very difficult to buy a cloud firewall from some vendor, a web proxy from another, an e-mail proxy from another and stitch this all together. The amount of encryption and traffic flows and network design to stitch cloud SaaS platforms together is prohibitively complex. And so what we see in the environment going forward is that you have more of a winner-take-all strategy per customer. And so a customer will buy the whole stack from a particular vendor instead of trying to stitch these different clouds together. This will have a couple of effects on the industry. First of all, we expect to see a consolidation in the industry. And I think this is proving out as you look around because independent small vendors can't just insert themselves into a traditional security stack by just putting their -- plugging in their appliance. So you're going to see a consolidation. And you'll see a consolidation around a few large security vendors that the industry recognizes and trust. The second thing that you will see in this is that before you had a security stack and a networking stack. And in fact, the security industry has always had this interesting relationship between networking, who wants to connect everything together, and security which wants to block things and whatnot. So the dicotomy has always manifested itself in the industry and that is breaking down because now you are seeing more and more security built into networking and more networking built into security. And so you can't have a cloud network -- a separate cloud networking stack and a separate cloud security stack and try to stitch those together. That cloud of integration is expanding or that spear of the integration is expanding. So this plays -- so Cisco has recognized this. We have, for many years, not only building out the broadest portfolio in the industry but spending a lot of time deeply integrating this portfolio together. And it is a hard thing to do and it takes time. And so we feel that we're ahead in that integration. And I'll wait for your SecureX question to elaborate on that. I think the second thing that benefits Cisco is the fact that we are recognized as the world leader in the networking space. And so as we see this convergence of these 2 stacks, we feel that we're uniquely positioned to take advantage of this over the next 5 years as we see more and more of this integration happening and we recognize that while there's complexity on the integrating the security stack, there's also complexity on the networking side, and combining those 2 together will really provide a unique value proposition for our customers.

Gee, once again, you're anticipating me. Maybe just a quick word on the COVID-19 impact before drill down of the portfolio return to a number of these issues you've raised. I trust COVID has just accelerated what was already well underway, but let me let you respond.

Yes. So COVID did 2 things. And first, if anybody on the call has family or friends affected by it, I wish you all the best. It's a very, very serious social condition and a pandemic. But from a business perspective, it did 2 things: one was, overnight, it structurally changed the way the enterprise operated at fundamental levels. So I think all of us are familiar with kind of a modern enterprise, 5%, maybe 10% of the employees work from home. The -- as the company was moving to the cloud, there were well-known steps on how to do that, starting with perhaps your IT stack and then your active directory and then things move along. And of course, overnight, you go from 5% or 10% of the employees working from home to 100%. And now when that happens, as a CIO or a CISO, you realize that, "Wait a second, I don't have the capacity to create 100% or allow 100% of my workforce to VPN in from home. I don't have the VPN capacity." Those who were terminated, their VPN stacked into their enterprise. Our firewalls are a market leader in this space, didn't have enough firewalls. Like, they physically didn't have enough equipment to terminate that kind of traffic. So of course, they moved to cloud-based VPNs, but that's a new technology introduction to them that they had to work through. The second way that you do it is you ask yourself the question, do I have to terminate all traffic? Do I have to send all traffic through my VPNs and get to my corporate network, we use Webex or video systems or my Salesforce is going to -- Salesforce.com. Like, can I split the traffic and say, "There's going to be traffic that I trust and I'm going to let them go directly into these external platforms, and there's traffic that I need to bring to my business." And -- so fortunately, again, we're a leader in this space, but -- so we are able to provide the technology, but the business processes of how to decide which things go where and the processes that you have to do to generate all of that are complex, and so the businesses did that. And this is why I say that even though that initial hump of just a lot of work and a lot of material being shipped and whatnot, once you've done that, you won't go back. You won't say, "Okay, I split all my traffic. And now that it's over and people are back at the office, I'm going to allow all my traffic to go back to VPNs or something." Like, those business processes are now ingrained in the enterprise. And because they continue to push forward the narrative of a digital transformation, we expect them to continue.

Gee, I'm going to ask multiple questions since there's a lot I want to start to get through in the next 20 minutes. I'm going to apologize, but ask you most of your questions and ask you to respond as you'd like. But you pointed out a number of current trends in terms of -- we've seen this vulcanization of security over the past decade plus that doesn't appear to have made organizations more secure. And so you -- and you're not alone. We hear the same messaging, I think, from Juniper, Palo Alto for a whole host of others in terms of going through an integrated approach. And you, of course, from the breadth of your security portfolio as well as having the NOI switching, routing, wireless, et cetera, in terms of the NOI network, no one has your footprint. And so you have the ability to -- at least, you're in the position -- you have that privileged position in terms of doing integration. I think you had a scope that no one else does. But if that's the question -- historically, we've seen -- there's always been this tension between best-of-breed and integrated approach and reluctant to rely on any single vendor. That's been a constant criticism or constant, I guess, a roadblock in terms of folks embracing Cisco more than whatever extent they had. And at the same time, we've now seen Palantir and a host of other companies, CrowdStrike, Carbon Black, et cetera, become very prominent vendors measured by revenue growth, market cap, et cetera. And so it gets to this question of, you're talking about moving integration, what -- are you seeing customers actually embrace that integrated approach as they look to become more secure? Are there any data points you can cite in terms of the journey and how you're positioned for that journey? Let me ask you to address SecureX. Again, my understanding is you introduced in February right before the pandemic hit hard, is the integration of your front end and with meaningful benefits in terms of visibility, security, et cetera -- in simplicity, scalability, et cetera. But let me -- address that however you would like.

Yes, Paul, I think you raised some of the key topics. So let me try to net out the essentials of this. As an enterprise, as a Chief Information and Security Officer, your job is to minimize the security risk that your enterprise is exposed to. And these are Board-level discussions, and your job is on the line almost every day as a result of this. So the one thing that you will not do is compromise your security posture inside your organization. You will spend more to ensure that you have the best protection possible if you're allowed to. You will spend to the extent that you're allowed to, and you will not compromise, like, simplicity over the best protection that you can get. So there is no way -- regardless of the customer value prop that you provide through an integrated strategy, there's no way that they're going to bring in a number 6, 7, 8-kind of security vendor into that environment and expose themselves. So first and foremost, you have to focus on efficacy and making sure that in all the categories that you participate in, you are recognized for your efficacy. Protection is your first and second and third job. Now fortunately, in all of the markets that we participate in, we're recognized for efficacy. We have Talos, the largest research -- nongovernment research group in the world that takes all of this portfolio and gets all of that data, massive and massive amounts of data, curates it and pushes it back into the portfolio. So we're broadly recognized to do that. And I think any vendor that wants to do this kind of platform play has to recognize that across everything. So what we do see in the industry is our vendors that start off with one particular technology--it could be a firewall, it could be an endpoint, it could be whatever it is--and then try to expand that into a platform play. And they have -- that's a really tough road to do. Because there are large parts of the portfolio that have significant threat vectors that devalue that platform if you don't have it. So like if you have a great firewall, you protect people with the firewalls, but you don't have an e-mail proxy, which is still the #1 threat vector in our industry, it's hard to imagine that your platform will offer any kind of simplicity or whatnot that allows you to collapse the number of vendors in an enterprise. We've also seen that -- and by the way, the industry has recognized this for a long time, and so the way they try to -- the way the industry has tried to deal with it in the past is to get a SIM, security information and event manager. There are lots of third-party SIMs out there. And they took in broad third-party integrations and every firewall and every proxy. And what ended up happening was the number of alerts just explode, and it's not really very useful at the end of the day, except to have a trusted data repository that you can go back and inspect what happened from a compliance perspective. So the way we look at this problem is that the integration that you do, that you provide has to provide to the end customer, a level of value that overcomes this desire to integrate best-of-breed. And if that level of value is not recognized by the customer, or the level of protection that you provide as a result is deemed insufficient, it will not work. So we spent many, many years building out this broad portfolio, as you mentioned. We spent many years integrating it together to have a very, very high efficacy, and we're recognized as leaders in this space. But we also found that, that was not enough. The value proposition of that for our customers was not enough to shift the discussion to an integrated strategy. And so over the last few years, we've integrated the front end of the portfolio together to provide that level of simplicity, integration and customer value that the customers can see, and that platform is called SecureX. It is the -- it has the fastest rate of adoption of any product in our history. So we announced it at RSA, in February, the industry response was very positive, and I think we struck a cord that the industry recognized that this has always been an elusive goal. We formally made a GA in a Cisco Live U.S. And subsequent for that, almost every month, we have another 1,000 customers adopting it and using it every day. And what we've seen from this is a couple of things in the business from that. First of all, by tying this all together, we monetize this through enterprise license agreements and those are the commercials that we run, and those are now the fastest-growing part of the portfolio. And so you see that customers are buying more of our portfolio together as a result of this. The second thing that happens is the usage continues to go up. And so not just the adoption of the platform but the daily usage goes up. So people are using this as their kind of log-in page, seeing what's going on in the industry, in their environment and kind of going from there. And then finally, what we've seen is kind of an independent third-party community forming of customers who have integrated other products into this platform, and it's an open platform, so we encourage that. And so you see people cheering ways of integrating in others for this. And so I think it's really -- we're making a real inflection point in the industry, and we expect this to continue to accelerate. I will say, in kind of conclusion for this, is that I don't believe that there will ever be like a sole security vendor, the one that kind of rules them all for the reasons that you articulate, customers want some diversity. But I do see that there are segments opening up. And so our customers, even when they buy the whole portfolio of Cisco, still have half-a-dozen trusted security vendors in their environment. It can be their cloud provider. It can be Microsoft or their application suite provider and others. So you kind of broaden the aperture out and you see very large players, each playing their role inside the enterprise, and I think that will likely continue as well.

Gee, by definition, is SecureX the most important security introduction for Cisco in the past decade? And what's next? And what more you need to do? I trust as with any major product, there's always scalability, stability, et cetera issues that need to be ironed out over time, and that's true virtually every vendor and every product. What needs to be done? And is that the most important platform? And what's next? What else is coming that you can talk about and would care to talk about here now?

Sure. So SecureX is the biggest thing that we've done in our history, and again, ties the portfolio together. I think it would benefit the callers on this line to understand the business model of SecureX because it also is unique in the industry. Most people, when they have a platform play, charge for that platform. You swap out your SIM or you swap out different pieces of equipment and have to integrate this in together. And so what happens is the way it's monetized or the value that it rolls out to the customer is there's a large upfront cost, you have to adopt this platform, and then the marginal cost subsequent to that is low. As you integrate other services, they're already pre-integrated. So large upside expense, then hopefully you get that back in lower marginal investments. SecureX has flipped that upside down. So SecureX comes with free of charge with every purchase you make of Cisco security products. So if you buy a firewall, you get SecureX. You buy our endpoint, you get SecureX. You buy our Cloud Umbrella, you get SecureX. And the way we want to -- the way we see this and the virtuous cycle that a platform has to provide is that as more customers use SecureX, we see their work -- we see their workflows, and we can start to integrate those workflows with confidence. Most people, if you follow the security industry, know that when you do e-mail, if you have a suspicious file attached, it automatically gets sent to a particular technology called the Sandbox, and we look at that file, we inspect it and make sure it's superclean before we let it get to you. Otherwise, we delete it, automatically. But you don't do automatic -- you can't have automatic responses get more complex. Because if in the presence of false-positives, if all of a sudden, you think there's an issue and you respond to it, but that issue didn't exist, you could do damage to the enterprise, as if it did. So if you thought they was an attack and you put a block on your firewall that took out your network, and there wasn't an attack, you've just behaved as if there was, and that's not good. So what our model is, is to get more and more people to use it, get the adoption up that we can automate more and more functions on behalf of the customer and make their experience easier, make them -- enable them to do more with less. That's the fundamental value proposition that we put in the industry. That is why you would buy SecureX. And as a result, the Cisco portfolio because it's making your day-to-day life easier, and you're not substituting that ease-of-use for protection. You're simultaneously getting the protection and the ease-of-use. And so while other people try to monetize their platform, that introduces friction into the marketplace, so they don't get the adoption, they don't get the usage, sufficient to generate the analytics and the machine learning and everything to simplify that. In fact, what they do is, they say, "On our platform, you can build your own workflows, and here's editors to do that. And we've made it easy-to-build workflows and stuff like that." But I think the rest of the industry has failed to recognize is customers don't want to do that. They want that to be taken care of automatically for them so that they can work on the more complex stuff. And I think that's the fundamental insight that SecureX has provided, and I think it's where we are unique in the industry.

Gee, just to confirm, you mentioned upfront that it is -- you are seeing -- I guess the question would be, are you seeing meaningful pull-through, meaningful impact on the rest of your portfolio or a good part of that portfolio in terms of incremental revenue since the launch of SecureX?

I would say, Paul, that it's very early. There's been a few months, but we do see those indicators that this is having an effect in the purchasing of our enterprise licenses as well as the number of products included in those enterprise licenses, yes.

Okay. And I guess this gets to -- sorry about this earlier. I don't think it's stating anything that should be a surprise to you to point out that the investment community looks at dedicated security providers, whether CrowdStrike, Carbon Black, et cetera, and there are dozens of them. And they have the benefit of disclosure on what they do, and you have a very broad, the broadest portfolio in the industry. And you also -- you already said, there's been limited insight into that portfolio. Where does your security revenue come from? How much is firewalls? How much is intrusion protection? Maybe that's not the right way to look at it. But it gets back to this question of having a broad portfolio and being able to do the integration along with the integration into your networking platforms. So that's seamless and getting credit for that. We look at your growth rate, which is not going to be ashamed of, but it's certainly north of 40%, 50%, 60% you see with dedicated point rightful subscribers. How do you maintain -- I guess, the question would be, how do you have best-of-breed on a silo-by-silo basis that you can integrate into SecureX, monetize that as opposed to a customer getting the benefit of SecureX, but relying on other point product solutions, and many of those together with SecureX? And how do you get credit, how do you drive -- are you expecting to drive incremental growth from SecureX? And that we should see a step up, to your point, just launched, but a year, 2 years, 3 years from now, I assume you'd be disappointed if there wasn't a meaningful step-up in your growth rate?

Yes. No, I -- look, Paul, I think you and the rest of the investor community have made that point, and it's very clear. I will break it into 2 pieces. One is on the security business itself, we are absolutely expecting this to drive top line growth. It will drive top line growth in 2 ways. One, I alluded to before, which is that people will buy more of the Cisco portfolio because it is integrated together and they see real value in that integration. We fully expect that. As I said, we're seeing the indicators of this early on, but certainly much less than 2 or 3 years. We're expecting this in a year and 2 and to see the lift of SecureX. Now how that translates into the numbers that we report and things like that? I would just say that Chuck has -- it continues to really fundamentally change Cisco and leave his mark on the company. And recently this year, probably about the time of RSA, Chuck created an entirely new business entity called Security and Applications. And we now have a new leader of this group, Jeetu Patel, who joined us from Box. And the one thing that this group shares together is the fact that we are very much cloud and SaaS focused. So Chuck, through his -- through the clear establishment of this new business that's focused on SaaS and growth, with Jeetu leading it, I think you may be able to recognize more of these things. And I can't speak to the future of what we'll release or whatnot, but the company is definitely shifting or expanding to include SaaS and high-growth businesses and allowing the company to focus on those in a way that perhaps was more difficult when they were buried inside of other larger businesses. So I would just say stay tuned. You've got Marty at the helm, which will take great care of you. But I think I'm encouraged to see the direction that Chuck has taken the company.

Gee, can I squeeze in one last question? Do you have 2 more minutes?

Of course. Yes.

And again, it's going to be a compound question, I apologize. But what does integration mean in practice? When Cisco talks about integrating security with this underlying networking hardware, what does that involve and really mean in practice? In connection with that, has Cisco been able to avoid solution clutter in a security portfolio? There's been a long-standing issue we've heard about from the dawn of Cisco, one of the burdens of having so many different solutions, that sometimes you're competing against yourself and causing customer confusion. I certainly have heard that over the many years I followed the company.

Yes. I think that's a great question. So look, we take a fairly pragmatic approach as well as a measured approach. So we are integrating in our SD-WAN capability into our cloud platform, Umbrella. But we've been in the market with an integrated security portfolio of Meraki and Umbrella, and Viptela, our other SD-WAN solution and Umbrella. Meraki has been over a year. Viptela has been since last April. And so -- and then when we take those 2 capabilities and move them to the cloud, we have a full cloud SD-WAN security platform. But we've been in the market for, in some cases, a year of integrating that together and for -- because we integrated with Meraki first, the low end of the market, the fire -- the low-end firewall, actually, Meraki is larger -- that integration between Meraki and Umbrella, is actually larger than our low-end firewall now. So we see this kind of shift into the cloud, we see it in the dynamics of our business and we see the customer response. And so we have taken our brand strategy, which has 83 separate product brands. They're very complex for our partners, for our customers. We consolidated them down to 7 product families and product brands, and one house brand, Cisco Secure. And we've done the same with the Ofers and whatnot. We're spending a lot of time on the simplicity of our portfolio, but also the way we -- customers consume the portfolio and partners can enable our portfolio and just stay tuned on that. I think that's a very rich area for Cisco because as you reduce the friction, then the scale of Cisco can really be applied and it's -- and be recognized in that way.

All right. I know you've got customers to speak too. I want to thank you for joining us. I greatly enjoyed the session. Look forward to talking to you in the future. Thanks again, Gee.

Thank you again, Paul.

Great. Marty, do you want to read the disclosure before I start off with the audio cast.

Yes, I do. Thank you. Hey, this is a educational audio cast. No new financial information will be disclosed. Actually, we will be making forward-looking statements. So actual results may differ materially from those forward-looking statements. And they are subject to the risk and uncertainties found in our 10-Q and 10-K. Thanks, Paul.

Thank you, Marty. So now we're going to start with Part 2, and I want to welcome Rakesh Chopra. Rakesh, thank you for joining us. Rakesh is a Cisco Fellow in Cisco's Common Hardware Group and has been at Cisco since 1997. During his 22 years -- actually, more than 22, Rakesh has driven a number of industry leading products at Cisco across the enterprise and service provider routing space. Today, Rakesh focuses on system architecture across Cisco's routing and switching portfolio to identify market transitions and ensure that Cisco's products meet customer demands. He works with external suppliers, the ASICs organizations, hardware organizations and the software organizations to realize next-generation products to optimize across company, organizational technology boundaries. We're not focusing internally to drive product development, Rakesh promotes Cisco's technology to its suppliers and customers. Rakesh, my first question for you is, is that an accurate representation of you about?

It's a long one, but yes, it's a pretty accurate one, Paul. Thanks.

All right. Thank you, Rakesh, for joining us today.

Of course, I appreciate it.

First and foremost, let's focus on Silicon One, which you announced back in December 2019, which sounded, at least on paper, very exciting in terms of breakthrough innovations to drive the Internet of the Future in a new approach to developing silicon. Can you describe how you got involved in that effort? And what is the proposition of Silicon One? How does it differ from Cisco's historical internal approach to developing silicon? How does it differ from the approach of Broadcom and external merchant suppliers?

Sure. I'd be happy to. So I got involved about 6 years ago or maybe 7 years ago at this point. So I was asked to take a step back and think about the transitions that are occurring in both the service provider and the web scale markets, and think about what that means to Cisco's product portfolio. So a group of engineers and myself, we sat down and did an analysis of every product that Cisco currently ships to the market. And interestingly enough, we did an analysis of all of the products that our competitors also ship to the market. And what we found during that analysis was that by about 2019, everything that was available in the market would begin to run out of steam. And so we needed to rethink things in a very fundamental way and do things very differently than we've done in the past. So back in 2013, we kicked off a bunch of investments in hardware to sort of rethink again everything that we do in terms of hardware. So we had investments into printed circuit board materials or heat things, connectors, memories, and most notably, probably, both silicon and optics needed to be done very, very differently. So the result of all of that effort and all of that investment is what we announced back in December of 2019 with the Internet of the Future, where we launched, among other things, the Cisco 8000 and the Cisco Silicon One portfolio. And so when we think about what Cisco Silicon One is, it's really trying to take a step back and rethink what needs to happen from a networking perspective. At the end of the day, we've all been taught for so long to think about having complete unique architecture and disjoint functions deployed throughout the network. And what we tried to do is a very holistic approach of thinking about the problems and how to solve it in sort of one unified architecture and being able to deploy that across the service provider and web scale market. Now that's obviously not an easy thing to do and it took us a very long time to do it, and we have to have a bunch of talented engineers working on the problem. But that's really what we set out to do.

So it begs the question from an architecture perspective, I think you all pointed out in the past, Broadcom has Trident, Tomahawk et cetera. And Silicon One is different, and that is an architecture that allows for fairly rapid introduction of different iterations of silicon for different use cases, whether routing, which is the first use case, you all appear to have chosen with your new 8000 series routers being relying upon the first iteration of Silicon One. What is different in the approach?

Yes. It's a great, great question. So I think at the end of the day, it's certainly not one single thing that we did to enable this. It's really a combination of many, many different factors. So I think when I break it down to my head, really, the first thing to realize is that silicon technology and all technology evolves over time. And that's especially true when you think about silicon process nodes. And when we build an architecture, we build a piece of silicon, it's really trying to figure out what is the optimal thing to do based on the limitations that exist for that silicon node. So most of the architectures that exist in the industry are now sort of 10 to 15 years old. And so all of the trade-offs that they made back then were perfectly the correct thing to do then. But now things have moved, and it's no longer necessarily the correct choice to trade-off. But once you started to invest behind an architecture, it creates what I like to think of as sort of technical depth. And the thing about technical depth is once you have that depth, it's really hard to start unwinding it. And so vendors keep having to iterate on these old architectures. So I think part of the magic to what we've done here is that, frankly, we're just starting fresher. So we're starting in a 16-nanometer process technology node and so we can rethink things in a different way. And when we talk about Silicon One, as you've mentioned, we have all of these different things laid on top of each other. So a really good example of that is sort of what we've done with sort of the fabric element. So when these architectures all existed in the past, there's a dedicated fabric element for the function, and that dedicated fabric element was based on cells, and they take a packet, they slice it into cells and they spray across a bunch of links. And that was the right thing to do because at the time, sending a large packet over SerDes length took a very long time. It was very slow. And memory inside of silicon was very, very expensive. And so you have to do everything that you could to minimize the amount of memory in the silicon. But fast forward to today, all of a sudden, SerDes are now running at 56 gig PAM4 SerDes. There's -- memory is, to some degree, free compared to what it used to be. And so you can now build a fabric element and overlay it on top of what you have for a stand-alone processor. And once you've done that, you can now start to strip away all of the dedicated function that was put in just for that -- those individual pieces of silicon. All of a sudden, you now have stand-alone processors and fabric elements overlaid on top of each other. The next thing that we sort of did was to sort of think about how memory works from a silicon perspective. And every architecture that exists on the market today is based on one fundamental limitation in silicon, which is how many times can you access the memory every single clock cycle. And what we've innovated and invented here in Silicon One completely changes that paradigm, allowing us to get billions and billions of accesses to a memory every single second without any sort of contention. Once we solve that fundamental technology building block, all of a sudden, we can overlay high scale on top of high bandwidth. And the final thing that we've sort of done is that we've really come up with a brand-new way to process packets, where all been used to this notion of sort of a fixed pipeline architecture or run to completion architecture or something in between. What we've achieved here is a new run to completion engine that is purpose-built for networking, and that allows us to actually get the full efficiency of a fixed pipeline architecture, but all the benefits of our onto completion network processor, and that allows us to overlay sort of all of the routing features with the switching efficiency. And then we pull all of these things together, and it allows us to create one architecture where we can have a stand-alone node, a line card processor or fabric element, all at the same time. But this wouldn't be particularly interesting is if you looked at each one of those, and you did a point for point comparison to the next best piece of silicon in the market. If we weren't better than all of those individually, it wouldn't be particularly interesting. And again, I think that's part of the magic of what we've done with Silicon One here.

Rakesh, before specifically ask you about those metrics, my understanding is that you also maintain that Silicon One allows customers to significantly reduce OpEx, which if that is the case, I assume above and beyond the technical merits in terms of packet processing, et cetera, that would be of great interest. But can you -- is that in fact the case? And can you talk about that?

Sure. Yes, it's again another good question. So I think it really breaks down into 2 different independent things, which allow us to think about those savings. So the first is, let's talk about what it means for a customer to deploy a bunch of unique architectures in their network. So if they're deploying one architecture in their DCI and one architecture in their WAN or one architecture in their spine architectures, for example, that all takes a lot of effort for them. Their network architects all of a sudden have to think through how to be able to deploy all of these different entities and work around the limitations that each one of them has. Maybe one is power-efficient, but doesn't have scale or one is programmable and one is low latency. And they build their entire network architecture based on those disjoint realities. And so it drives a huge amount of complexity into that underlying network architecture. And once they figure that out and they deploy it into the field, all of the field technicians that have to surface it, they need to understand all of those different technologies and how to troubleshoot each one of those. Now imagine what happens if you can converge all of that under one single architecture. All of a sudden, you can get technology deployed into the network much, much faster. And any time an issue shows up, it can be debugged once across the entire fleet of the network. Technicians can now all of a sudden not be sort of specialized to troubleshoot in one role in the network, they can troubleshoot any role or any form factor in the network. So that to me is a huge savings for our customers. The second one is a slightly different piece maybe. And it's really about power efficiency. So when we think about Cisco Silicon One and we think about the Cisco 8000, we're really hyper-focused on power efficiency. And we can talk more about why I say that, but I believe it's just so fundamental for us to be focused on power efficiency. And so we've driven extremely low power systems into the market much, much lower than anyone else in the industry can come to. And it matters just because it's just so critical. So every watt that the system draws means that facilities need to be built to deliver that power and cool that power, driving sort of CapEx expenses at the front end. But more interestingly is the amount of -- that our customer spends to deliver and cool the power of these systems on an ongoing expense can be very, very large. And so if you can change the amount of power that the systems consumes, it can really change the economics for our customers.

Yes. I think it's certainly fair to say that most of the investment community or a good portion was aware that power is probably the number 1 factor, at least within the data center, in terms of the gating factor for systems deployment and a key critical focus to your point. Let me ask you, the first iteration of Silicon One again in the 8000 series, which are routing platforms. And just for those in the audience, you may not be aware, correct me if I'm wrong, Rakesh. What we're talking about routers, which, by definition, tend to be external facing. So we're not talking about intra-data center switching. And so my question to you is, I know it's relatively early. I think you've now been GA for a quarter or 2 with the 8000, but it was in customer stance for a year. If I remember the performance metrics measured by throughput and power consumption, on paper, the 8000 is not just better, but dramatically better than your competing solutions or better solutions. Can you address what the early -- what the actual performance, the early data indicates, whether customers are seeing a difference? And I don't know if this is beyond your purview. Was that translating into meaningful traction for 8000 as we look to other platforms that are introduced at some point in the future based on Silicon One in terms of speaking to whether it's real, whether the reality matches the coins?

Yes. Yes, good questions. So there's a few different things I'll sort of tease apart there. So I think you're correct that when we announced the Cisco 8000, right, we just talked about sort of specifics here, right, we introduced a 1 RU 10.8 terabit router that is based on a single Cisco Silicon One Q100 device. At the time when we launched that in order for any of our competitors to build an equivalent system, you are talking about a minimum of 5 pieces of silicon. And in many cases, many more pieces of silicon to achieve that bandwidth. So it's a very, very big needle mover in terms of the amount of silicon and therefore the underlying costs and the power of the system. So we're talking about 10.8 terabits of routing performance at 415 watts of power. So that is an incredibly low number. If you compare that, for example, to another Cisco product that we shipped, not too long ago, our NCS 6008 with a 1 terabit line card, that's a full rack system at 8 terabits of space and 11 kilowatts of power. So that's a 65x reduction in power. So it is a very, very large movement that we've been able to achieve with Silicon One and the Cisco 8000. And so yes, I think it's -- if we look at that product, it's very, very difficult for others in the industry to compete to some degree just because nobody else has that underlying raw silicon in their portfolio. Now again, we have advertised the notion of having a disaggregated business model. So we do allow customers to procure Cisco Silicon One directly and build their own hardware and write their own software on top of it. So we did announce in December 2019, these products. And to be honest, I think it's still early days, but the amount of traction that we're seeing, especially in the customers who are what I'll call the highest bandwidth customers, the web scale and the large service providers. It's so far a very great start for us. More engaged with most of the web scale customers and have design wins in most of the U.S. web scale customers at this point. But it takes time. So it will be time, I think, before there's significant revenue, but the traction is really good, and the message is really resonating very well. The other thing I can sort of comment on is that we do have customers from a silicon-only perspective who have built with our product and built with other people's products. And the feedback that we get from those customers is really incredibly positive. Because to some degree, I think it almost sounds too good to be true in many ways. I've been in this industry for a very long time. And to some degree it just sounds like a PowerPoint pitch. But when you have 2 products right next to each other and testing them in exactly the same condition, and if anything, I think customers are starting to find that maybe we have been underrepresenting the magnitude of improvement that Cisco Silicon One and the Cisco 8000 gives us.

Rakesh, again, I apologize if this is not under your purview, but is there a rhyme of reason to why Cisco chose to do the first iteration of Silicon One for the routing application of the 8000 as opposed to I think a lot of investors would have thought that the first iteration would have been focused at the intra-data center switching opportunity where top of rack leaf spine, all of the above. Was there something that underlie that choice? More importantly, any insight you could share with us in terms of when we should expect the next rev of Silicon One in the next use case?

Sure. So a couple of points there. So again, I think as an industry, right, we have been taught for so long to think about what I would call hard segmentation, i.e. you need a completely specific architecture for a service provider router platform or you need a specific architecture for a web scale switch. That those are just 2 completely disjoints and incompatible things to try and solve. And again, I think because of the technology that we have managed to create with Silicon One, you can start to erase those hard boundaries. Now when we came out with a Cisco Silicon One Q100 device, it's a 10.8 terabit, 16-nanometer device. You can take it with large external buffers to make it a high-end service provider router. You could also take that same architecture, the same device, remove the packet buffering and have an incredibly efficient switch. Now what's a little bit peculiar about the Q100 device when you think about the web scale switching market is the rate x of the device. So it's a 10.8 terabit device. And as you're well aware, I think if you want to be in the web scale switching market, you really sort of need power of 2, whether it's 3.2 terabit, 6.4 terabit, 12.8, 25.6, they're all built on that sort of rate x power of 2 number. So I think the reason you sort of see the focus on the Cisco 8000 and Cisco Silicon One being the sort of service provider roles is that, that number of 10.8 terabits has allowed us to create great large buffer modular and fixed systems. It's slightly peculiar from a rate x perspective into web scale switching. What I would say to your next question is we have a very, very large investment behind Cisco Silicon One. It is something that we are incredibly excited about at Cisco. And I wake up, frankly, every morning, still super excited, to be involved in this amazing technology. And you will see -- in the not-too-distant future, you'll see things from Cisco that show the power of a converged architecture and what that means. So stay tuned.

Rakesh, have customers had input into Silicon One as one would hope and expect.

Absolutely. I think it's very easy as a technologist to sort of only be sort of inward focused and just think about what can we do from a technology perspective in isolation. But at the end of the day, of course, what only matters here is solving the customer problems. So we have good, strong relationships with customers across the industry, which help define both what we do with Silicon One, what we do with our Cisco 8000 products. So yes, we have very tight engagements. And what I like about the engagements is, at the end of the day, everything in engineering is a trade-off, right? So making sure that you understand what truly matters to the customer is really, really important. Otherwise, we'll end up creating technology that's not solving the problem that they're really looking to solve. And so ever since the early days of both the Cisco 8000 and Cisco Silicon One, we engaged with a small set of customers to understand what are their true pain points and what is it that we can do at Cisco to help them solve those problems.

Okay. I think I recall you mentioning earlier in the conversation that back in 2019, you looked, you and your colleagues and saw that the different approaches were going to run us in 2019 time frame, which based on observation, clearly, networking is undergoing significant change, most prominently driven by the shift of workloads and applications to the cloud and the trend towards work from home, which appears to be giving rise to an era of cloud networking. And I think Chuck alluded to this if not out rightly addressed it on the last earnings call. How does Silicon One address that shift to cloud networking from traditional Cisco and other vendor's traditional model?

Yes. So I think the first thing I would say is, as you think about sort of an ever-increasing bandwidth that we're sort of -- the trend has always been there, right? Bandwidth has always been increasing. I think with the introduction of things like 4K video, augmented reality, 5G, bandwidth is really exploding within the network. But at the same time, many of the customers have a reasonably flat revenue stream that they're trying to manage to. And so if you can't change in a very fundamental way the cost structures for customers to be able to deploy systems, it becomes very challenging for them. So what we're doing with Silicon One is, again, trying to do a very large change from what the industry is used to meet, right? If all of a sudden, we can take a routing system from 11 kilowatts and bring it down to 415 watts of power, all of a sudden, their ongoing expenses for deploying a high bandwidth router have come way down. It also means that you can deploy routing scale, routing programmability, routing buffering in parts of the network that were never possible to have high bandwidth numbers before. Now why does that matter? It matters because a lot of the complexity in the network is driven by the fact of having constrained bandwidth. If you can open up the bandwidth across the network, all of a sudden, much of that complexity starts to a back rate. And you can really just focus on delivering the customer experiences that our customers are looking to deliver to their customers. That allows that all to sort of play out. So that's to me for the service provider. In the web scale, again, everything about web scale deployment is about ease of operations. When you're building networks that's such a gigantic scale, you've really got to get to the point where everything can be easily deployed and easily managed. So if you can give those customers a unifying architecture to be able to deploy across their network and allow them to do it in a way that still is lower power, while picking up programmability and scale. That's a really, really large value proposition for that type of customer. And so I think that's really what we're trying to do is just change the conversation in a very fundamental way. And Cisco Silicon One really gives us that toolbox to have that conversation with the customers.

Rakesh, I recognized we're talking about silicon here, but silicon, obviously, there's a little bit of vacuum. I trust you all have done a lot of work in terms of the software piece of the equation. To the extent you have a new silicon architecture, anything you can share with us in terms of what you've done and what remains to be done in terms of enabling customers to take advantage of the significant performance virtues of Silicon One from a software perspective?

Yes. Sure. So ultimately, in my mind, software has several different layers to it. And because we've announced sort of a disaggregated business model now at Cisco, actually the definition of a customer is also varied. So the first would be the type of customer who is going to be consuming our technology as a silicon-only perspective. So for that type of customer, we have created a very clean and easy to use software development kit or SDK, that abstracts the complexity that as all of the functionality of Silicon One into an easy to consume set of APIs. We then layered on top of that SDK, the notion of an SAI abstraction layer to allow customers, for example, who want to run SONiC on top of Silicon One to do so in a very sort of simple and streamlined and standards compliant fashion. Then when we talk about our Cisco 8000 system, whether that Cisco 8000 is running SONiC, that's one way we now offer our products. We also offer it with IOS XR7. So when we built IOS XR7, we sort of tried to take the best of what was IOS XR6 and then sort of improve on top of it. So we took all of the carrier-grade capabilities that exist in IOS XR and continue that into IOS XR7. But we focused in many different aspects. So one is about, again, thinking about the sort of web scale customers. Creating a set of API abstraction layers at multiple levels in that software stack to allow them to write automation software on top of the network and allow them to control every different level of the box and the silicon from outside of the box itself. So creating that notion of a network-based controller over the entire network, allowing them to turn up and manage this equipment very easily. The next is sort of focusing on sort of 0 touch provisioning, i.e., how do you get it to when you take a Cisco 8000 box and deploy it in the network, how do you minimize how painful that is to get that box up and running. So there's a lot of infrastructure that we have put on top of that. And then the third, in my mind, is actually our security stacks. We have this built-in in the hardware of the Cisco 8000 as well as the software on IOS XR7. We have really stepped up the notion of ensuring that what is happening in the Cisco 8000 system is fundamentally secure. So in hardware, we put in the notion of secure root of trust. We can make sure that the code that boots is truly signed and authentic Cisco software. We can even make sure that while the system is running, nobody is actually modifying the memory content of the system. And we can even make sure that when the system leads Cisco manufacturing through the life cycle of the system that nobody is actually modifying any of the major components in hardware. And then IOS XR7 runs on top of that hardware infrastructure and provides all the sort of run time monitoring to make sure that the software is secure at boot up and secure through the life cycle of processes start and stop through that entire life cycle.

Great. Rakesh, we've got some questions that have been sent in to me. But before I ask you those, if you don't mind, I'm going to ask, Simi, to pole the audience to see if anybody got some live questions. And again, I've got some of that have been sent in to me. People -- go ahead, Simi.

[Operator Instructions]

All right. I'm going to go ahead, Simi, unless there's somebody queued up. Let me post you the question that were sent in to me. And again, I recognize that you're on the ASIC side, but clearly the industry is going through integrated silicon and optics. When is that where Cisco is going, when do you expect to get there? How much -- how strategic is that?

Yes. So I think to answer that question or sort of maybe sort of take a step back and look over -- look at history for a second. So if you look at the sort of early days of networking, the way network room was initially deployed is everything connected in the network was done with copper. And over time, what you've seen is as optics has started to sort of mature in the industry, it starts with sort of the longest possible links, the long-haul links first switched to optic first and then everything else was copper. As you fast forward in time and as optics speeds increase, the percentage of the market which is being covered by optical versus copper is increasing over time. And so depending on if you're in the sort of service provider, web scale market, you've got the situation now, where basically the vast majority of the interconnect outside of the rack is all done with optics. And within the rack, there's still some copper cables. So for example, a copper rack switch down to server is still generally passive copper cable. In the enterprise space, really sort of within a floor, it's usually still copper, but within a building, they start switching towards optics. And so it's really about that sort of distance versus bandwidth equation in terms of what is optics and what is cable. Again, as you look into the future, that trend, we believe, will continue, such that you'll get to the point where all of a sudden, the right trade-off now is everything leaving the system is optics, regardless of what industry you're in. And then eventually, everything leaving the silicon package will be all optics as opposed to copper. Now why is that true? It's true because, at some point, you reach the situation where it's actually not possible to escape enough SerDes from an ASIC package. And you have a forcing function to get to continue to drive up that switch bandwidth to use optics. The second is just about efficiency. Every time you send signals between 2 different points, it consumes a lot of power. And as we talked about earlier in the call, to me, power is sort of that fundamental limitation that we need to solve as an industry. And so that's absolutely coming. And I think if you look at what we announced in December of 2019, right, we announced sort of this key technology pillars being software, silicon and optics. And if we look at the sort of silicon and the optics, part of the reason you see Cisco investing so much in those 2 is that those -- that convergence is inevitable. So the real question is your second question, which is when will it happen? Again, if you look back in time, this has been a slightly moving statement. So I think plenty of people in the industry said that at the sort of 12.8 terabit generation, you'll see that crossover between in package optics being sort of the right way to do things. Obviously, we'll pass that word to the generation of 25.6. 51.2 is coming very shortly. And at Cisco, we are very focused on making sure that we have the right technology to enable all of these appropriate market transitions. So you can imagine that just because we have all of that investment in silicon and optics, we are working together to create interesting products for the future. So if you ask me, I think somewhere between the 51.2 and 102 co-generation is probably when you see that transition starting to happen.

Okay. I recognize you haven't closed the Acacia acquisition, and so it will probably be hard for you to address this question from an investor, but I'll ask the question, which is, if you are successful in closing Acacia acquisition, any thoughts as to -- they have their own -- they design, they make their own -- they design their own chipset. Do you change that design to fit in with this consolidated Silicon One approach? Or since they seems already have significant success as a merchant coherent to your key supplier, so you leave that alone? Or again, I recognize you haven't closed the acquisition, that's the question.

Yes. So obviously, we've announced the attempt to acquire Acacia. And so any forward-looking statements here are obviously my personal opinion and not necessarily anything beside that. I think, ultimately, Cisco has a broad portfolio of technology. Acacia is an incredibly successful organization. And Cisco has announced even before the intent to acquire Acacia, the notion of having sort of this disaggregated business model. So you can envision the notion of us continuing to sell our internal technology to external customers in any which way that they want. From an optics perspective, I think we have a lot of technology between earlier acquisitions in-house development, Luxtera, with the intent to acquire Acacia, all of that, to me, creates a toolkit for us to figure out the correct technologies to bring together.

All right. And before asking this next series of questions from some investors, let me just call it attention. If there's anybody out there that for whatever reason you don't want to have your voice heard, but you have a question, please feel free to send me the question. I'll try to ask it. Rakesh, how many chipsets does Cisco developed historically? Do you have it present, how many different chips? And I trust the Q100 is a single chipset today. So currently, where you're at in terms of products that have emerged from Silicon One, again, you just announced it back in December, so it's very early. But today, it's a single chip.

Yes. So what we announced in December was a silicon architecture, what we call Cisco Silicon One. I think some people probably took that to mean one chip, but it really means one architecture and actually many other -- the connotation one means many other things, but it means one architecture. The Q100 is a single device. So it's a single chip, and it's the first of a family of devices that will be coming as part of Cisco Silicon One.

All right. Any thoughts you share with us in terms of how -- you made the point a number of times, I think, since that launch that the architecture allows for rapid silicon innovation, design, introduction, launch. Any thoughts in terms of what we could expect over the next 12 months, the next 2 to 3 years, the next 5 to 10 years in terms of how many different iterations of silicon we can expect to see as Silicon One in the cadence of launches?

Yes. So I can't give you anything definitive in terms of what to expect into the future. I'll say a few different things. So I think if you look historically in the industry, what you'll see is out of an architecture, basically 1 chip basically announced every sort of 12 to 24 months is sort of the cadence that you see coming out of most architectures. I think Cisco Silicon One gives us a completely unique proposition from a Cisco perspective to change that discussion. So I would just advise you to stay tuned. I think there will be some pretty exciting stuff coming from Cisco. But we're not, at this point, announcing anything publicly. The other thing I would just sort of comment here is that I think there's general misunderstanding perhaps or differences from how companies generally announce. And if you think about what we announced in December of 2019, as we are announcing the Q100 device, right, the first device from Cisco Silicon One, as part of the full Cisco 8000 portfolio as being deployed in full systems in the field. What you see from a more traditional sort of silicon provider is generally they announce the devices when they power on in the lab, and sometimes based on market reality, they announce actually before they tape-out devices. And so that actually creates a 12- to 24-month difference between when we are talking about a product from Cisco Silicon One and when others might be talking about silicon as a stand-alone device. So you guys should just take into account that there's a time shift from when people are talking about products in the field.

Understood. And then going back to an earlier question, how many different individual silicon does Cisco have at present? That could be Cisco Silicon One at present?

Cisco has a very broad level of investment into silicon. We always have, we always will. So Silicon One is focused on the service provider and web scale market at this time. Besides that, we have many, many other silicon architectures. So for example, we have the UADP that we sell into the enterprise campus. That has a large investment from Cisco historically and has a large investment from Cisco currently. We'll continue that into the future. Similarly, we have the cloud scale asset that we sell into the enterprise and service provider data centers. Those, we are continuing to invest in. So Silicon One really is very focused on the service provider and the web scale market.

Got it. I only see there's a couple more, and then we will wrap up. What is your current process node implementation? I'll ask the question. I think I know the answer, but which foundries?

So I can't discuss specific foundries that we work with. What I can say is that our Silicon One Q100 device is 16 nanometer. You can imagine maybe that we have other devices in the works, which are under different process nodes.

Okay. My last question, going back to your response of cloud scale ASICs, I guess, I'm a little bit surprised that, as you said, you pointed out, given that Silicon One addresses service provider and cloud. As cloud scale ASICs, as I understand them, they're your high-end ASICs that provide significant scalability and other benefits for cloud, specifically for that intra-data center switch application. Does that -- how does one reconcile? I thought Silicon One would address eventually that intra-data center switching, or in the case of the QoS approach routing, whatever layer the stack you want to put on it. I thought that Silicon One would ultimately address that. But if you're going to maintain investment in cloud scale ASICs that would suggest perhaps a contrary. Any insight you can share on that?

Sure. So obviously, there are data centers of different types, different customers, et cetera. The Silicon One portfolio is very focused at this time on the service provider and web scale markets. That's the market that we are trying to address the requirements to. The cloud scale ASICs are focused on the enterprise data center and the service provider data centers, trying to drive the feature sets and the scale and the bandwidth requirements for those data centers. So again, it's easy to -- want to believe that one thing that does that entire portfolio. What we're talking about really is Silicon One is focused on that service provider and web scale market. Cloud scale is service provider data center and enterprise data centers.

Okay. One last follow-up here with respect to your comment about process node technology, how challenging to get it down to 7 or 5 nanometers going from the current 16? I mean, clearly, that's the direction you have to go and to be competitive. How big a step down, how big a challenge given this new architecture?

So there is no real challenge besides what I would call standard of challenges of any new process node. The Silicon One architecture is really, in many ways, it's a far more scalable architecture than anything else in the industry. And so as process node continues to shrink, we have a completely unique ability with Silicon One to be able to drive with bandwidth up, while holding the power on a very flat curve. And so I think what you'll see is -- and again, nobody can predict the future. But as an engineer, I'm excited about the notion of what we can do with Silicon One. Of any advantages that we have today in Silicon One actually growing over time rather than the inverse. We're at the beginning of our architectural build out here from a Silicon One perspective. And so I think the future is really, really exciting for us from Silicon One.

Great. I want to be respectful of your time. I want to give you the opportunity if you want to provide any closing summary remarks.

No. I mean, I think maybe the only a couple of sort of things I might just sort of say, is one of the questions that I've gotten from others, frankly, is really just trying to understand what the name Silicon One means because it has an interesting connotation name. And so I just sort of wanted to give my couple of thoughts on that. So ultimately, at the end of the day, I think the Cisco Silicon One name, it's a name with many connotations. So the first is, I think it is one architecture that can be deployed across the service provider and web scale network that we talked about. It's one architecture that can be deployed across all the different form factors of boxes, whether it's a 1 RU piece of box, a big modular line card or modular fabric cards or even a completely disaggregated system. It means that we can take 1 SDK and put it into any operating system. It means that we can have 1 P4 forwarding code deployed across the network. And all of that really just sort of translates to sort of one experience for our customers. So it means that those network architects or technicians, they can really understand one thing and deploy it. And then finally, I think is that you can take this one architecture and have it in any sort of business model that you want, whether you want to buy a full system or a Cisco build box where you bring your own OS or a stand-alone piece of silicon where you build your hardware and your software. All of that is really sort of meant in that connotation of Cisco Silicon One. And I think it's a pretty exciting time for us at Cisco. I think it's -- we've got this amazing technology that allows us to have very different conversations with our customers and what it means for them to deploy. So I really look forward to seeing what the future holds based on this.

Right. I think that's a perfect place to stop perfect summary. I want to thank you, and I want to thank Gee for having taken time out and running the business and sharing your insight and signing with us today. And I want to wish everybody out there, please stay safe. I hope you and your loved ones are staying safe and sound. Once again, Rakesh, thank you. Thanks, Gee. Marty, thank you so much. Take care.

Thank you so much. I appreciate it.

That concludes the Cowen conference call. Thank you. Enjoy the rest of your day.