Fortinet, Inc. (FTNT) Earnings Call Transcript & Summary

Good afternoon, everybody. Thank you for joining us. We are very pleased to host the management of Fortinet. My name is Shaul Eyal, and we cover Fortinet with an outperform rating. This afternoon, we have Keith Jensen, CFO of Fortinet. We are also joined by John Maddison, Chief Marketing Officer, EVP of Products; and Peter Salkowski, VP of IR. Peter, do you want to kind of just give the -- give your kind of brief slide before kind of we kick it off?

Yes. Thank you, Shaul. I'm trying to get to the -- hold up. So I'll a call everybody's attention to the safe harbor slide. If we make any forward-looking statements today, that are only as of today's presentation. We take no obligation to just update any forward-looking statements after today. With that, I'll turn it back over to you and let you go with the Q&A.

Thank you so much. So again, thank you for joining us this afternoon, this morning and then the West Coast. The way we're going to probably kick it off with a fireside chat and also feel free to utilize those joining us, the chat box, the Zoom chat box, drop me your questions. We will be more than happy to accommodate it, bring as many as possible, time permitting. So with that, gents, again, thanks for joining us. Maybe just starting off by providing us with a general update of what Fortinet and management team have been seeing in terms of market dynamics, industry trends during third quarter to date. Again, we're not asking for a mid-quarter update or guidance, but kind of maybe from a high-level view building on the very healthy second quarter results, what is it that kind of has been happening within the past few weeks or so? What's driving the markets?

Thanks, Shaul, and thanks for your caveat about not asking for a mid-quarter update. I appreciate that. I think for Fortinet, yes, we -- maybe because of our diversification, we can look at and answer that question with maybe a little bit of a different lens. And by that, I mean, 70% to 75% of our business is international. We're very diversified across the world. And if you look at our customer segments, it's probably 1/3, 1/3, 1/3 roughly between SMB, mid-enterprise and enterprise. And I think what we're seeing, as we look -- as we came into the third quarter and we went through July and we looked at our pipeline, I would say that each of those geographies, if you will, are doing better than the second quarter. And each of those customer segments are doing better than the second quarter. I think, for us in particular, we saw very strong health also in retail in the second quarter with our SD-WAN solution and we expect it to be more of the same. But I think really the headline for everybody, from our viewpoint, is it really comes down to the ability of countries, geographies and such to reopen their economies and stay open. And really, that's the unknown that we're all fighting our way through.

Absolutely. And speaking of economies being reopened, you were actually amongst the first companies going back to the first quarter, reflecting on the first quarter. Actually, if you talk about this incremental spend, emergency type of spend and clearly, like we have seen that in great results at that time, have you seen that maybe moderating a little bit or being rationalized a little bit? And I think maybe kind of trying to put a context on that. We have seen a very healthy number of deals greater than $1 million booked in the second quarter. We're going to touch on the SD-WAN kind of probably separately in just a few minutes. But maybe just also an update on that front as well, Keith, if you may?

Yes. I think we came into the Q1 earnings call with a -- feeling a bit of responsibility to share with the world, if you will, what we were seeing because of such a dynamic situation. And also we have the ability, I think, to really identify 3 key products. We call them Authenticator, Token and Client or respectively 40 this, 40 that and 40 this with our naming convention. And those 3 products, we really saw jump up fairly dramatically compared to their historical norms. We wouldn't have expected to get about $10 million of billings out of those 3 products in the quarter, and we got closer to $20 million. And then we updated that number on our second quarter call by saying, yes, and to your point, Shaul, it's still outperformed and outperformed significantly, but that overperformance probably really moved from about $10 million down to $5 million. I don't know that we think that we really saw a lot of firewall buying, if you will, to increase capacity or VPN. Perhaps some of our competitors did. Were there corner cases of things being stood up, study-from-home application or work-from-home application? Sure. But I think that was really just kind of a shift in use cases. And as we got through April, we really did not see one of the things we checked was if there had been an unnatural pull forward, we would have expected to see a slowdown in April, and we did not get that slowdown.

Absolutely, absolutely. And I've mentioned already kind of touching indirectly on the SD-WAN and the fact that, I think it was about like 12, 13 deals greater than $1 million were SD-WAN-driven. So given that we also have John joining us this session, maybe John, how do you see the SD-WAN opportunity from your perspective? You guys are being kind of out there. You guys have done tremendously well. You're like the #3 -- not like, that you are the #3 within the SD-WAN market from a market share perspective. We are seeing consolidation recently that could actually benefit you in the near-term Silver Peak kind of being acquired. We typically know what's happening given the initial stages of integration, a little bit of a disruption. So maybe kind of just broad terms SD-WAN and then kind of bringing it down to the Fortinet opportunity.

Yes, SD-WAN, we feel, is really critical technology. And a lot of people apply it to the branch. So it's the SD-Branch and SD-WAN for the branch. But in our opinion, SD-WAN technology goes all the way from home. In fact, I'm broadcasting here with my SD-WAN connectivity at the home. And it's for the branch, it's for the campus, it's for the data center, and we just recently announced SD-WAN for cloud. So it's not just something for the branch, it's technology that needs to get integrated across the entire product range. Because in the end, it's going to decide how your users and devices get to their applications, which are -- and both of those things are moving around quite a lot. So that's why we took the strategy of integrating it into our product line versus building it separate or acquiring it separate because, in our belief, it's so fundamental that needs to go inside our capabilities and inside our products. And as we go forward, what's also interesting is this convergence theme, people refer to it SASE. Now we've had this since the beginning of the company, and that is the convergence of security and networking. We've been doing it through our appliances through our ASICs and SPUs, where we're able to run networking and security at very high performance. So not only is SD-WAN a great and expanding marketplace because people need more flexibility, but secure SD-WAN, the ability to provide that security is also very critical. And as we go forward, we'll see more flexibility of where that security is, whether it be on the appliance or in the data center, virtual or in the cloud. And so it's not just one market anymore. These markets are starting to get connected because of the convergence and consolidation.

Got it. And as we are thinking about some near-term disruption that we have been seeing as of late, John or Keith, anything so far or like early days to maybe gain a little bit of an advantage from somebody else's pains, near-term pains, maybe long-term pains?

Well, I think the game is convergence in this platform approach. So -- and when I speak to customers, first of all, the COVID has accelerated their digital transformation or innovation especially in certain industries like retail. They want to go faster. But to get there, they can't glue together all these pieces from networking vendors and security vendors. They want one solution that comes and provides that connectivity, high availability and security. And so we're seeing a lot of customers decide on a converged platform approach versus trying to glue individual together from the networking or security side.

Understood. Maybe Keith, talking a little bit about ASPs and typical discounting rates, growing contract terms, given the current environment. Can you maybe provide us with a little bit of an update kind of on the discounting dynamics that you have seen a little bit that you talked about it during kind of the second quarter update and contract duration kind of being just kind of, I think, a little bit up, but maybe customers are beginning to find a higher level of kind of confidence beginning to lock down contracts now that the things are slightly more -- or the picture is becoming a little clearer?

Yes. I think the -- to kind of parse it into 2 pieces. On the discounting side, if you go back, I think, for 2 quarters in a row, Q4, Q1, maybe even a little bit longer than that, we talked about discounting being a tailwind. As we got into the second quarter, I think this is just a reflection of the worldwide economy, discounting became a small headwind. But I tried to put that in the context of product gross margins being up, I think, 320 basis points year-over-year. And that's really some structural advantages that we have in the cost of the products. But it was also -- it would have been higher without the discount. And really that was kind of the point that we're trying to make. In terms of contract terms, I think we were sequentially up a month, flat year-over-year, something like that. But I do agree with your premise that as the economies are starting to reopen, when I look at the July activity and the SD-WAN projects that have come online, I am seeing that there is a comfort zone, if you will, on contract duration. I certainly don't feel that we're feeling any downward pressure on contract duration. And maybe that's just because of the sheer breadth of our portfolio of solutions. I can have some cloud offerings or what have you or SaaS offerings that have shorter durations. And the other end of the spectrum is probably SD-WAN, where customers are trying to lock in that MPLS savings for a much longer period of time.

Understood. Going back, revisiting kind of the second quarter kind of metrics, actually. The SMB side of the equation actually showed some notable strength, specifically in the Americas region, which, in a way, I think, much different from some of the narrative that we have been hearing from some of your peers. So maybe can you help us kind of reconcile a little bit as to kind of you guys have done absolutely well, some other companies kind of have been like more mixed, is it a matter of execution? Or is it a matter of kind of market share dynamics that kind of where we're beginning to see out there?

Yes. I think you've kind of touched on some of the key drivers behind it. I suspect that also coming into the quarter, like a lot of people, I thought the SMB was going to be a bit of a challenge on a worldwide basis. We saw the U.S. perform exceedingly well on the SMB section. And part of that is maybe just the definition of the SMB. For us, those could be companies that have up to 500 employees. And sometimes when we hear SMB, we think about the dry cleaners in the corner or something like that. These tend to be the entities that have a little more substance behind them. That said, the sheer -- how fast that market actually is in the United States, it really becomes a quantity play. And by that, I mean, I think our U.S. channel leadership has done a very good job of bringing in perhaps lower-level salespeople, giving them a playbook to run over and over and over again, partnering very closely with the channel partners and providing the incentives to the channel partners. There's thousands of those in the U.S. And as long as you target those channel incentives the right way and work with the partners, I think the U.S. team has really shown the ability to drive growth in that area.

Understood. Maybe one for you, John, Keith also, if you want to also comment on that front. I want to talk a little bit about the recent acquisition that you've just announced, OPAQ Networks. We are familiar with the concept or hot topic of Zero-Trust Network Access. We understand Fortinet, you guys already kind of have -- you guys have like the FortiSASE, but kind of wanted to hear from you about the acquisition, what do you think of achieving longer-term from this kind of tuck-in acquisition? Any color on integration time line? Maybe even some billings or contribution that we might be seeing down the road hitting the financials?

I'll leave the billings to Keith. But I think there's still a lot of confusion over cloud because people say cloud and they go -- they get everything mixed up in a joint bundle. Is it security from the cloud, like SaaS delivery? Is it security for the cloud? Is it SaaS protection itself? And I think what's very important to us and very important to customers is that let's put protection for the cloud to one side. This is -- these are things you build for the big public cloud vendors or the application security and just talk about cloud delivery, security delivered via the cloud. We think this consumption model has been changing over the last 20 years. Some markets, for example, have gone almost 50% cloud delivery like e-mail. It makes sense to be close to the application. We also think that there's a trend going on right now, where some of the network security, customers want that flexibility to be able to deliver it at the cloud edge versus the WAN edge versus the data center edge business. So it just -- the OPAQ acquisition allowed us now to have a delivery form factor for our core business firewalling as an appliance, as a virtual machine and as a cloud delivery right now. And I've speak to customers over the last few years and then a lot of them will say, "I'm not going to cloud tomorrow or next year, but I'd like to know that you've got that in case I decide at some point that part of my network needs to be protected by firewall as a service." Well, some of my employees, which are totally off the network and they'll be off forever, I won't want to deploy security-defined perimeter for them to provide security that way. So I think it gives us now a full spectrum of delivery mechanisms. Again, a lot of our customers are still very much in hybrid mode and will be in hybrid for a long time. But they like to match their vision in the future that if they continue to put certain applications in the cloud, they may want to put that protection in the cloud. And the OPAQ acquisition now, again, gives us those 3 form factors, appliance, virtual and cloud delivery depending on how the customer wants to build their infrastructure around.

Understood. Understood. And John, while we're kind of diving a little bit on the product front, I want to specifically ask about your FortiGate 4400, the hyperscaler product recently launched really focuses 5G networks. I think also it is being powered by the recently launched NP7 processor. How should we be thinking about this really hyperscale products? On the one hand, we like it because probably they add to kind of 7-digit transactions longer term. But maybe kind of help us understand where is the demand coming from? Is it kind of data-driven, volume-driven? The fact that there's kind of much more to process and analyze and go through it, obviously, security being top of mind here. So kind of just try and frame kind of the whole kind of product set for us.

Yes. Well, first of all, for the OPAQ acquisition, one of the things we like about it is, I mean, you can do this convergence through appliances or you can do it by the cloud, if you haven't got appliances powered by ASICs, and that's what a lot of competitors do because they're not going to go and start building ASICs. What we're going to do is take ASICs and cloud together to give a huge advantage. But coming back to your second question on the FortiGate 4400F, think about networking back in 2000 and where it's come. Networking has just got faster and faster. In fact, we're being asked now for 400-gig ports. If you go back to 2000, you were looking to get 100-meg ports. Now networking has not really changed a lot, routing and switching. The functionality -- there's a few new changes here and there, but it's just got faster and faster. Security has got more and more applications built onto it. So it started as a firewall, then you got a VPN, then you've got UTM, Next-Gen firewall, SSL inspection, now SD-WAN. And our whole strategy is that you need to keep offloading the processor. So think about gaming systems from NVIDIA with GPUs, where the GPU offloads the rendering from the CPU, so you can run it very quickly. It's the same concept for us. We offload networking functions on offload security functions, so we can add more functionality on to our Next-Gen firewalls. And this gives the customers now 15 to 16 different use cases, they can use our firewalls for. So it's not just about getting faster, although we are getting very fast. It's also -- I was speaking to a customer this morning, who said, "Oh, I've just switched on your web filtering", "Well, I've just switched on the SSL inspection." They were separate products that just get consolidated and converged in. From a next generation, what we call network processor 7, these are big investments for us. They take a lot of money and a lot of skill set to get done. It's a 200-gig firewall in a chip. It's faster than most of our competitors appliance -- high-end appliances, just on the chip. That's the sort of advantage it gives us. And what we're seeing is, right now, there's still people building their own hyperscale data centers. There's still 5G networks being built out there and most firewalls cannot operate in these environments. They're just not built to do it. They are CPU-based systems. Really, they're basically software-based systems, running on a standard CPU. They're not optimized to run 400-gig, 100-gig interfaces. They're not optimized to do Carrier-Grade NAT. They're not optimized . They just can't do all these things at that speed. And we think the 4400F -- by the way, we did the 4200F earlier, but this 4400F really does bring huge speed, 1 terabit per second firewalling and all that functionality offloaded the CPU, IPv6 and all that capability that runs now at just core network speeds. Will we ever catch up with the fastest routers and switches? No. Because it's doing simpler things. But for a lot of the core, what I call staple firewall functions, we're now running at very high speeds and allows us to be put in situations that you can't put normal firewalls in or it allows us to go into a maybe 2 or 3 use case situation and run at a price point 10x faster for the same amount of dollars from a price point perspective. So we think it gives us a huge advantage for those applications, but it also across the board gives us a huge advantage.

Got it. So as we think about expecting more maybe from a puzzle type of perspective, we have the SD-WAN piece. We have OPAQ. Clearly, we have kind of the fabric also that we haven't kind of talked about yet. But John, anything else you might be thinking that's still missing a little bit because you guys have done a lot. The platform without a doubt has broadened considerably, by the way we're seeing that in results. But to your view, is there anything that you think that might be needing like a little bit of a tweak?

There's always something the customer wants you haven't got, so. But I think the key is for us building that platform has taken a while. Other people build it by just acquiring a bunch of stuff and then they kind of bolt it together. It's really hard to do that. It's really hard actually to build it organically because your products needs to be best-of-breed in their own right and then be part of the platform. So you have to build that functionality. And we've spent a long time building out products beyond our networking capability, endpoint, WAF, e-mail, cloud, secure access and all these things that now come together and allow us to present a solution. One of the great solutions we're doing now to retail is SD-Branch. That consists of SD-WAN. It consists of firewall and it consists of secure access, AP, switching, endpoint, NAC, all connected through segmentation PCI into the data center or on a single platform with a single console. The reason they like that a lot is they've got one person, one vendor to be able to implement to train. And then key going forward is to be able to build automation across that platform, which is almost impossible with different vendors. And so I would say there's always some functionality that's going to be needed across the areas. But if you think about -- and if I come back to the Gartner SASE definition that has 13 different products, SD-WAN being a core piece of it. I say SD-WAN is actually the foundation of SASE. If you don't have SD-WAN in your portfolio, I don't think you've really got a SASE offering. But then it's got Zero Trust. It's got firewall as a service. It's got WAF, CASB. It's got networking -- sandbox networking. So all those things we've got now and tick off. In fact, we've got the full portfolio of SASE as a product spectrum. And it's all totally integrated. So yes, there'll always be some little pieces here they want -- the customer wants, but we think we've got a great full offering across the full spectrum.

Fair enough, fair enough. The topic of the whole kind of partner's landscape, I don't know whether Keith or John, who wants to kind of take it. But clearly, we see -- we understand that Fortinet has an extensive integration ecosystem, completely vendor agnostic. Can you maybe share with us how's that assist in accelerating Fortinet's fabrics adoption rate? Is it strictly partner? Is it also like internal? What's the way to think about it?

Well, I look at it 2 ways. One is the -- our partners, and we're a very channel-driven company, as Keith was saying. We don't want to build systems that don't allow our partners to add value to the offering, because they're critical in that they've got the customer relationship and the customers trust them. And so even our SASE applications allow our partners to provide value inside that. So that's the baseline in our opinion. And then the second piece of that, I think, is the -- is what I call the fabric-ready partner ecosystem, which could consist of somebody who's built an API to our system, all the way to us building deep integration to the orchestration systems of the big platform vendors. So we can see where infrastructure is changing. But both are important. I think what customers don't like -- customers don't like ripping everything now and starting with this huge platform. They like to incrementally add things. And so it's very important to us that we work inside the ecosystem. We have 300-plus fabric-ready ecosystem partners. So it's rarely we run across them. They're not there. But when we do, we try and make sure they're part of our ecosystem, because the customers want -- eventually, they want a platform where they can consolidate and converge, but they also want an open platform where they can bring some of their own things, existing vendors or new vendors into that ecosystem.

Keith, we have a question from the audience. Can Keith discuss the competitive landscape? Has he been observing any changing dynamics?

Yes. I don't think I've really seen any changing dynamics. I mean the one thing that I called out a few times is that we did see one of our competitors, particularly in the SMB space, where we compete toe-to-toe most often, offer some long-term payment plans in Latin America that extended out until the first calendar quarter of next year. And I think you saw us also follow with some payment plan opportunities, again, in Latin America, really because that geography was hit and is still been hit very, very hard by the pandemic and to help them -- help our partners and help our end users manage their capital, if you will. Other than that, I don't think that I've really seen much in terms of the difference at either end of the market, if you will, whether it's the SMB or the enterprise players.

Got it. Got it.

I would say -- I would just comment on that. I think I'm definitely seeing the platform message resonate a lot more with customers. And so the larger platform vendors seem to be gaining traction in my opinion versus a point solution that's got a low differentiation in the marketplace. It's going to be very hard for them to get into some of these accounts.

And one of the questions, John, Keith, without a doubt that we have been getting for the past, probably 3, 4 quarters now is when we look at Fortinet's product revenue growth, it's without a doubt the fastest that double-digit that we had seen when you look at some of the other firewall-related vendors out there. So the question always is it pricing-driven? Is it throughput-driven? Is it some features-driven or all of the above combined? In other words, what is it that you guys are doing that maybe some of the other providers are not doing, which kind of assist in getting to those mid-high teens product revenue growth from a year-over-year perspective as we have seen in recent quarters? And I know also, you touched on in it, John, from an ASIC perspective, that could be the secret sauce, but I want to leave it kind of to you maybe to share your views with us.

Yes. I do think it's the ASIC secret sauce to be honest with you. I think we have a structural cost advantage when we go-to-market with our proprietary ASIC as opposed to competitors that may be buying off-the-shelf silicon. And if you look back and I was trying to call it out in the earnings call a little bit, if you look at our E-Series products, that have been out now for, let's call it, a year, I mean, it kind of bounces around depending upon the product, it came into existence with a somewhat better cost structure than its predecessor, the chip that it was running on. And then as you go through subsequent iterations and negotiating with contract manufacturers or what have you, you've seen now us string together 3 or 4 quarters of our product gross margin has been back over 60%, and up 200 basis points, 300 basis points year-over-year. And I do believe when we look at the new F-Series of products that are coming out on the newest chip, that whole concept of a structural cost advantage continues. It creates the opportunity to take previously pieces and functions that were just separate from the actual appliance with chip and embed them in the chip and generate cost savings. So it does give you -- as I say, it gives us a competitive advantage that others simply don't have. And maybe John has more to offer.

Yes. I think it's the scale. We just talked about the 4400 at the high end. We just introduced the FortiGate 80F, which is built. So on the entry level, what we do is we take everything and put it on a system on a chip. So we take this content processing, network processing CPU and put it on our own chip that we build. That gives us the ability to build something at a price point and performance that scales across thousands of sites, for example. And what used to be either a very simple router is now an SD-WAN appliance there. So I think the scale going from the very smallest, and I -- we also believe, again, that some of the homes, we are seeing some of our customers ask for that type of installation in their executives or their power users in their homes. I'm broadcasting here on an SD-WAN, as I said. And I think that scale of going from all the way from almost SOHO small business all the way up to hyperscale data centers allows us to have a product for every use case, for every application. And that means we can compete in a lot of different areas, not just one area and allows us -- our products are built for that, and that gives us that capability to win business -- for new product business for a lot of the different use cases, whether they're networking or whether they're secured or whether they're both.

Got it. Keith, as we think about hiring targets for the latter part of 2020, and I think also in that context for like kudos for, I think you mentioned no layoffs, despite kind of the harsh environment that we have gone through over the course of the past few months. What do you have in mind in terms of R&D, sales and marketing or just overall hiring? And maybe how do you look at it from a regional perspective? Or any specific items, products, regions that could be in a need of some assistance or some ramp-up?

Yes. I think we're obviously very pleased with our operating margin performance, our free cash flow performance in the quarter. So it would seem a little bit of odd perhaps to come back and say we were going to do something on the headcount line that was moving in the opposite direction. I think we clearly have an opportunity right now because the travel costs have really been removed from the system and some of the marketing events as well. And taking that savings, if you will, and reinvesting a large portion of that back into headcount, whether that's sales headcount, whether that's marketing or whether that's engineers and hopefully, someday, finance people, too. We'd love to ask some of them. But the real ability here is that you're bringing in salespeople right now and you're generating pipeline right now that hopefully, when travel returns, you now have a group of salespeople with a very large pipeline that are very productive. And we'll see how that works out for us in 2021. But clearly, we are using that arbitrage, if you will, of repurposing travel dollars into future growth.

Got it. Got it. One of the questions from the audience that we had that actually we also planned on asking, Keith, and I know you're being asked frequently, as what would make you reinstate the long-term targets that you guys kind of talked about on your kind of recent Analyst Day? Is it kind of just kind of the improving fundamentals, kind of getting slightly better reach from the channels? What are the measures?

Yes. I think it's really certainty, right? And certainty can come to us in the form of a vaccine. They can come to us in the form of people really embracing social distancing and mask and getting the virus under control. It kind of comes back to an earlier conversation where it's just this uncertainty. It's the question of whether or not countries, geographies, states, if you will, can open up their economies and remain open. And I think we'd really love to see that before we then come back and have a follow-on conversation about long-term target. That said, I'm certainly very pleased where the numbers we've put together in the first half of 2020 in this environment.

Understood. Understood. We have another question from the audience. Can you compare contrast trends within APAC versus EMEA versus the Americas? Compare contrast trends, regional trends, APAC, Americas, EMEA?

Yes. I think the -- we came into the second quarter with a new world, right, with the pandemic world. And our expectations at a high level was that -- were that APAC was further along in the pandemic and the arc of it or the path of the pandemic, and we expected APAC to do well in the quarter. In Europe, we thought it was kind of next in line. And then the Europe, and then the U.S. and Latin America would follow. I think we've very much got that in terms of what we -- our expectations coming into the quarter and what we got out of it. And I think as we kind of look at it now, I think -- I do think that APAC has continued to be a step ahead of the rest of the world. I think Europe is doing well with their ability to contain things. And I think the U.S. has actually reached some sort of normal now, and we feel good about what we're seeing out of the business in terms of the U.S., specifically in this early part of the third quarter.

Understood. John, you mentioned before kind of when touching on the whole kind of product portfolio and I think also the shift to cloud. You touched on the concept of automation. And in automation over the course of the past without a doubt 18, 24 months is also a topic that has been getting much more attention. I can tell you from conversations with system administrators I would imagine you hear the same thing from people in the field, you hear it from customers, the need to automate pretty much everything, which is happening within the SoC, security operations centers. So where do we stand in that respect? How is Fortinet looking at it? Are we just at some sort of an initial inning? Is there much more to come in automation? How do we think also in applying and using some maybe AI and machine learning capabilities to facilitate that process?

Yes. I think we're just starting. I think most of our investment up until now has been in AI for better security protection, detection. So all our systems -- for example, we have a huge installed base, 450,000 plus customers. A lot of those customers will send telemetry back to us. It gives us a view of the world of every country, every industry, every vertical, every threat vector, and we bring each of those threat vectors into our data centers. And we apply machine learning against a big dataset. Files, for example, 2 billion good files, 2 billion bad files, we get a pretty accurate result, but that's not what I call protection-type stuff. We then apply AI to look across the different threat vectors to identify what the campaign looks like because they're using multiple threat vectors and the multiple infection points. So the AI now is providing more of a detection, zero-day targeted attack-type information that we apply to our AI-based systems out there. So there's still more work to do there because the AI needs to get more data always, but it's providing some great results from a protection and detection perspective. Adding on top of that then is how do you -- if you do find something, how do you respond quickly? And so there's been some -- in the enterprise, I would say, those people with a mature SoC, who've got the expertise, have started implementing source systems out there. We acquired a source system recently, which has to be multi-vendor because you got to look across different vendors there. So I think that's starting. It's starting to make its way down into the enterprise mid-marketplace as we get more accustomed to building those scripts out. So there's a security angle to automation and AI. I think the area that's going to be most interesting going forward is AI netOps, AIOps, where you're then applying this now to an operational model to either give you more high availability, better save money on routes, save money on hiring people to kind of watch a screen and all those things are going to -- as we start to apply to the networking capability. And so if I have an SD-WAN link and I can see that I've got 3 or 4 of my ISP links down in a particular region, I can redirect that. I can add capacity, add capabilities automatically without having to get somebody as human involved in that. So I think security kind of started, I think, a while ago. I think AI operations, which means you need to have that capability built into your products and into your orchestration systems is coming. But the key is if we've implemented it across 5 or 6 different vendors, you always have to work to the lowest common denominator, which is always a problem. So if you have a platform approach where you have a single system, you can apply that automation then because each of the systems understands the automation language and scripting language, and we're definitely seeing customers now roll out rudimentary automation. But as they go forward, they'll look more advanced from a network ops perspective.

Understood. Keith or John or Peter, any closing remarks? Something you want to kind of share with us before -- because we are just kind of getting to the point that we need to conclude our session. Any closing remarks?

Now I would say -- I speak to about 10 customers a week. It's -- I've actually -- our EBCs or we call them virtually BCs have doubled and tripled since the lockdown, maybe people have a bit more time to have the conversation. And there's just a huge amount of interest in getting to their digital models, which means they need to implement and upgrade their infrastructure and are now see security teams and networking teams working together even more so than a year ago.

Awesome. Gents, thank you so much for joining us. We appreciate it. Take care. We'll talk to you soon. Bye, everybody.

Thank you.

Thank you, Shaul.

Bye now. Bye.