2023-06-06 14:16:06 ET
Arista Networks, Inc. (ANET)
William Blair 43rd Annual Growth Stock Conference
June 6, 2023 11:40 AM ET
Company Participants
Ita Brennan - Chief Financial Officer
Conference Call Participants
Sebastien Naji - William Blair
Presentation
Sebastien Naji
All right. Good morning, everyone. My name is Sebastien Naji from William Blair. I'm pleased to have Ita Brennan, CFO of Arista Networks with me. Before we begin, I'm required to inform you that a complete list of research disclosures or potential conflicts of interest, are available on our website at williamblair.com.
And with that out of the way, Ita has a little bit of a presentation.
Ita Brennan
Yes. I would also refer you to our kind of risk factors, et cetera, in the SEC filings as well. So we have a couple of slides. Just for anybody who doesn't know who we are, I will go through those pretty quickly so that we can move off to the Q&A. So Arista, for anybody who doesn't know who we are, I mean, we are a software driven networking company. We did roughly 4.3 billion in revenue last year. Market cap, depending on the day, so we're between 45 billion and 50 billion. We grew revenues last year, roughly 50%, we grew bottom line, EPS roughly 60%. So, it's a highly cash generative business model. And obviously, when we have strong growth in the business, a lot of that flows through the bottom-line and you see that bottom-line growth as well.
The company was founded in 2004. First products shipped in 2008, a big focus on cloud, what we all know to be kind of cloud data centers and supporting those customers as they were building out their cloud business models. We IPOed in 2014, and joined the S&P 500 in 2018. So very rapid progression, always with this focus on products for customers, engineers, talking to engineers, making products that kind of solve problems for customers. And on profitable growth, right? The company was profitable right from the beginning. So we were very keen focused on trying to maintain that balance. Obviously, we want to grow the business aggressively, but we also want to make sure that we are driving profitable growth.
We have been expanding our presence across markets, and we will look at those in a little while. Here, you can see references here from Gartner and Forrester. Again, it's all about a focus on product and solving customer problems. Jayshree Ullal is our CEO, has been our CEO since prior to the IPO. Andy Bechtolsheim and Ken Duda, our founders. Andy more on the infrastructure, hardware, architecture side of things, and Ken leads the software team. 90% of our R&D is software and software team. So the software is a big part of the overall solution that's delivered to customers, leveraging merchant silicon, in particular, Avago as a silicon partner on the hardware side.
So large, fast growing market opportunity. We sell essentially into two markets. One into the data center switching market, which is a -- it has been a fast growing market. It grew almost 20% last year. It's expected to grow somewhere low mid-teens this year, and then continues to grow beyond that. A large piece of that being driven by increased spending with the cloud customers, but also investment by large enterprises, et cetera, in that data center market. And then on the campus side of things, which is another $12 billion, $13 billion market, we are very early into that. We did roughly $400 million in campus revenue last year. So a big opportunity there to continue to grow and earn our fair share of that campus market as well. So early in some very large markets and also with the data center market, in particular, a fast growing market growing somewhere from $20 billion, $22 billion now to $30 billion by 2027. So good growth opportunities in that market as well.
So what drives? Again, if you look at it over time, we have obviously been growing roughly 21%. Growing earnings a little bit faster as there is leverage in the business, right? But we have seen this very kind of consistent ability to grow the business over time in this kind of 20% plus five-year CAGR. And that's been kind of a pretty consistent metric when you look at the business from when we went public through today, and we will show you that in a second on the slide.
This is our TAM chart. So we continue to expand our TAM. We are expanding TAM in two ways. We are expanding it in terms of the products that we are delivering to customers, and then we are expanding in terms of the parts of the market that we address and that we play in. We started out very much focused on cloud and on products that satisfied cloud customer needs. But then over time, those cloud products have become much more relevant to other parts of the market, right, to enterprise customers. A large enterprise customer today faces a lot of the same challenges that a cloud customer faces, and wants to achieve the same total cost of ownership and service capabilities as a cloud customer does. So there's been a -- the market, if you like, has come to us in terms of this cloud networking solution, as enterprises become more and more cloud-like in terms of the metrics that they are trying to drive and the automation that they are trying to drive.
You can see on here is our core business, which is really where our data center switching business is. Then we have our cognitive adjacencies, that's where a lot of our campus products are. And we have been expanding kind of around that enterprise opportunity over time, including network and detection response. We have some new WAN products that we just launched that we'll talk some more about, over the next couple of years you'll see us expand, but really building out kind of that solution for an enterprise customer. Used to be, we could only sell an enterprise customer data center opportunity. Now we have the ability to address campus, Wi-Fi, all the way through to routing, and routing use cases for those customers. So that expansion has helped us become a more credible provider for those enterprise customers. Our TAM, overall, if you look at it grows from roughly 31 billion in 2023 to 51 billion in 2027. Some of that is natural growth in that core market. And some of that is also addressing more and more of those adjacencies and services and software offerings.
When you look at the business and the pieces of the business, I always think about it as kind of we have different building blocks. We have a cloud business, enterprise business, and then the providers business. If you look at last year, the cloud was roughly 46% of the business, enterprise was 32% and providers was 22%, right? So we're selling into those different pieces of the market. And we're selling them these different products over time.
We're very focused on kind of growing our software and services business. There's a services piece of that. It runs roughly mid-teens percentages of the business. It's a very profitable services business. It's not -- so don't think about it as a professional services business, but it's more of a support and software support business, highly profitable, generates roughly 80% gross margin on that services business. So that's a key part. We're doing some automation, some new products around, kind of how do we better support enterprise customers using technology as solutions to make them more efficient and to help them run their networks.
You can see, we obviously have the enterprise WAN. It was a new product that we announced at the last Analyst Day. We've yet to set some targets for that. But we'll do that here over the next while. So it's always growing each of these building blocks, and using these building blocks to drive the overall business model.
There's some very good offsets in terms of -- the cloud business is, obviously, high volume, maybe a little bit lower gross margin but less of an investment from a sales and marketing perspective. You have the enterprise business where you've got higher gross margins, but we need to spend more to sell the products. But net-net, from an operating margin perspective, the overall business kind of matches up pretty well from the different trails.
So we've talked about this a little bit, every -- we look at it over five year period, the business has grown greater than 20% in every five year period. We have a target of 20% growth for 2020 to 2025. We do see if the cloud business can be somewhat cyclical inside of this, which is why you need to look through a longer time period. And when you do that, you can see that it's actually a pretty consistent business and that we've done a good job of being able to go back and win cloud business and win new business with these customers in every cycle.
This is just again on the business model. Obviously we need to -- when we think about capital allocation, we invest in the business as our primary goal and then we look to return cash after that. R&D is the lifeblood of Arista, probably always will be. There's a big focus on the product, making sure that we have the right resources to develop the product. So we'll continue to kind of grow that R&D investment as we grow the top-line. Sales and marketing, very much focused on the enterprise piece of the business. There's very little sales and marketing that goes into the cloud part of the business. So if you take all of those sales and marketing spend that we have and addressed it to the enterprise, we're spending roughly high-teens, 18% plus or minus of our revenue on the sales and marketing for that enterprise business. And then G&A and the other pieces are just around being efficient, and being as efficient and the best of classes we can be.
The go-to-market, that's what we just talked about. We really have kind of two business models inside the business. You have a high volume cloud business, low go-to-market investment, and then you have a more traditional enterprise business where we are spending most of our sales and marketing. So it's more of the sales and marketing you could get allocated to that enterprise business. So sometimes people look at our business model and think 7% of sales and marketing is too low, but, really, you need to address that to that enterprise, and then it's in the high teens, which is still efficient but definitely a respectable investment to drive the enterprise part of the business.
This is our business model. From our last Analyst Day, you can see our target model for 2023. We are kind of recovering from a gross margin dip, because of all those supply chain activities that we were engaged in. Hopefully, we hit a low point in Q1 of this year at 60% gross margin, and we will build our way back up as we go through 2023. And hopefully, depending on the customer mix in the business, we can do a little bit better in the longer-term model.
Again, we talked about R&D spending. We will always spend probably mid teens in R&D, 7% -- 6%, 7% in the sales and marketing again going to that enterprise business. And our G&A has been pretty consistently below 2%, giving us this operating margin target, plus or minus 40% is our current view, reserving the right that we could invest more, but would obviously be to drive something that was accretive and incremental to the business.
I think that's all we had for slides. So I'll turn it back over to…
Question-and-Answer Session
Q - Sebastien Naji
So maybe just to get started. I want to talk a little bit about core networking demand. The networking space has gone through the wringer over the last few years. Your lead times were extended. Companies built a lot of backlog. Now they are starting to work through some of that. How should we think about what the real demand trends or the real growth drivers in networking are today and over the next 12 months?
Ita Brennan
Yes. I think, look, we have just come through a period where we got a lot of visibility because we had very extended lead times. We were -- it's lead times in excess of 12 months, which really isn't healthy for the business in the longer-term, but that was kind of the world that we were living in. We very much tried to manage through this by putting these -- that demand into periods of deployments and figuring out with customers, when we needed to deploy equipment, what needed to be deployed first, et cetera. And we are still working through that, right? We are still -- we still got some good visibility to deployments in the current periods, as we work through those deployments. The question becomes, when you look forward into 2024, when will you start to get visibility to that? And we have seen an improvement in lead times. We are seeing things get better. So towards the end of the year, I think we could be back to, hopefully, like, a six month type lead time window.
What that means from a visibility perspective is, we will have to get closer to the end of the year to really understand kind of what's going to happen from a demand perspective in 2024. And that's not unusual. That's just going back to how it used to be before all those supply chain activities impacted those lead times.
So I think overall engagement is strong. You heard Jayshree talk about on the enterprise side of the business. We see kind of continuing engagement on new business and new opportunities. I think for the cloud, they are still kind of figuring out what those deployments will look like into 2024. We'll just have to get a little bit closer to that time period before we'll have visibility to that. And again, that's not good or bad. That's just the model returning to something similar to what it was before we had all of the disruption from the supply chain perspective.
Sebastien Naji
And then talking a little bit about the enterprise side of the business. When we do our checks, we hear about continued market share gains from incumbents. Can you maybe talk a little bit about what has allowed you to take share from those incumbents? And what is it about the Arista value proposition that's resonating so much with those enterprise customers?
Ita Brennan
I mean, it all goes back to that single operating system and having one operating stack across all the different pieces of the network, right? So we had the opportunity to kind of go back and rearchitect that software stack, and understand that really there was no reason why you would deploy a different operating system to do routing versus an operating system to do switching versus the operating system to run campus, right? So we had the luxury of kind of starting, again, if you like, being able to kind of develop this kind of state-based software approach, which says, you capture the state of the network in real time. And then you know everything that's happening kind of in the network, you use one software stack across everything. So that's -- I mean, that's hugely impactful from just running a network infrastructure, because you don't have to patch, you monitor, everything is done once and not multiple different times with different software.
So that's a big piece of, what does the customer experience when the adopter is to say versus others. It's that simplicity. It's that scale. It's being able to see and monitor what's happening in the network. And that drives quality, right? So that single operating system, the focus on quality. If you've ever had heard Ken Duda, our Head of Software talk about, he's uncompromising about quality from a software perspective, he's very focused on driving kind of refreshes of the software, making sure that everything remains current, remains modern. It has some really good YouTube videos of him talking about kind of why customer comes first and this and why the product has to be fully baked, we've automated all of our QA. So it really comes back to that simplicity and quality. So when you deploy the network, when you deploy an Arista network, you will experience a visible difference in kind of your quality, your issues, and the performance of the product.
Ken has this thing where he really just wants networking people to be able to go home on Friday, have their weekend, do updates, do upgrades and not have it be something where you have to take down the network to do that, right? It's this idea of that you should be constantly monitoring and refreshing. And not have that be an outage, just not have that be an event, because you can't take down the cloud. And more and more, you can't take down an enterprise network either, right? So I think a lot of it comes back to some of those fundamental building blocks that the team put in place early on.
Sebastien Naji
And as part of this initiative to take share, gain share within the enterprise customer base and expanding to campus and routing, you've had to sort of change a little bit of your go to market motion. Can you talk about what you've done in terms of getting closer to your distribution channels, partners, and what maybe there's still left to do?
Ita Brennan
Yes, I mean, we're still heavily direct sales model focused today, because we're still focusing on that large enterprise piece of the market, right? So if you look at the -- obviously on the data center side, most companies that are deploying data centers are large enterprises. If you look at the campus market, about half of that $12 billion is in kind of large enterprise band. So that's our target market today and it's still a direct sales, largely a direct sales model where we have expanded kind of geographically. We have expanded coverage, and that's been the focus. We have been adding sales, resources and kind of expanding that enterprise coverage model. And consistent with this big bet kind of focus, look for large enterprises where the network is critical and then deliver kind of performance to those customers.
As we move kind of out of that large enterprise, that's where the channel really becomes more important. We need to start that work today. We have started that work today. It's more relevant in the campus part of the business and will be more relevant in the campus part of the business. So the data center is still a very technical sale where you are going to have to have a certain level of engagement. But on the campus side, there is more opportunity, I think, to build a channel-led business. So we are signing up new partners. We are developing partner programs, et cetera, but it's early, right? And, again, that's not for today's kind of performance. It's for the future. So we do have some time to do that, but we need to be incrementally kind of building out that channel over time.
Sebastien Naji
Okay. Great. That's helpful. And then I want to turn to the inevitable question around Generative AI. Can you maybe talk a little bit about what is Arista's opportunity here and what differentiates your product set that will allow you to win your fair share of this Gen AI market?
Ita Brennan
Yes, I mean, look, for us -- and again, we've talked first, I think, about AI and the AI spine that is that 7800 product set back in November of last year. So this has been something, again, that's been part of technology discussions and roadmap discussions for quite a while now, right? And building the product set and the capabilities to support different AI use cases was part of that roadmap. And Hugh Holbrook at the time had presented just kind of what Ethernet needed to do and be in order to support kind of more of the AI high performance type use cases. So it's something that's been -- that we have been working on for some time. We see it as something that really underpins kind of spending in from these large customers and potentially in the enterprise as well, but really from these large customers, as they start to kind of build out AI -- pervasive kind of at-scale AI footprints, right?
There is been lots of discussion about InfiniBand versus Ethernet. Our view is that as you scale into these large, vast kind of footprints that that's Ethernet's job and that the Ethernet can perform very well there. There is obviously going to be some technology software capabilities for traffic monitoring, other things that we will do to support that. But that's -- I think in terms of a longer-term opportunity, you really see -- we really see AI as a driver of spend and a driver of technology, a driver of why you need 800 gig, a driver of why you need 1.6 terabits. It contributes to all of that, right?
So I think, yes, net-net we see it as a positive. Again, in the context of short-term business, we will need to work through and see exactly kind of how it plays out in the near-term. But certainly as an overall business driver, we see it as being positive.
Sebastien Naji
Okay. Great. Maybe last question before I open it up to questions from the audience. The cloud titans have been a huge driver of growth for you. And we learned in 2019 that that's sort of a cyclical business. So what confidence that you -- do you have that they are going to continue to spend? And maybe could you frame it for us versus the CPU based spending versus the GPU based spending in terms of where you think most of that growth is going to come from in the next few years?
Ita Brennan
Yes, I mean, look, we have been I think at pains to make sure everybody thinks about the cloud piece of the business as being potentially cyclical, right? We were surprised back in 2019, that was the first time that you'd seen kind of some ebb and flow in their spend, right? So I think it behooves us to think about the business in that context, right? We've been working hard to grow the enterprise piece of the business. That's much bigger now than it was then. And that's a much more consistent kind of steady growth driver. And we'll continue to look for ways to diversify the business for that.
But the reality is, we don't know that it will be cyclical, we don't know what 2024 will look like just yet. I think the drivers are there for them to continue to spend, but not at the level we just grew triple digits with that piece of the business last year. So we do expect some moderation off of those levels.
Sebastien Naji
Any questions? I think we have time for a few here.
Unidentified Analyst
[Question Inaudible]
Ita Brennan
I think like on the one hand, it's positive to hear them acknowledge that Ethernet has a major role to play in kind of the scaling of these AI footprints. Clearly, they do have and have had some Ethernet assets for some time, and you'd be surprised if they don't try to leverage those in some way. Again, I think our job is to continue to build kind of the best products for the customers that we have kind of these engagements with, and continue to execute in the way that we have over in the past. And I think if we can do that, we are -- the team is probably one of the better Ethernet switching teams in the industry, right? So we just need to continue to execute and stay focused on solving these problems for these customers and then we'll go from there.
Sebastien Naji
Any other questions? All right, then, maybe a few more for me. So you've talked about Microsoft and Meta as two of those big cloud titan customers. Can you maybe talk about what Arista is doing to gain more wallet share with some of the other hyperscale companies?
Ita Brennan
Yes, I mean, look, we have, obviously two other hyperscale customers that are our customers and would be sizable customers in the normal scheme of things if they weren't completely dwarfed by Microsoft and Meta. So those engagements are there. We continue to kind of further those engagements from a technology perspective. There's that collaboration between the teams is very important. We continue to work at that. We continue to build that. Is there some step function change in how they decide to kind of run their businesses and how they decide to deploy? I mean, that's their decision to make at the end of the day.
Again, our job is to stay close, to be part and demonstrate the technology, demonstrate the products. I think Anshul had referenced in some of the earnings call that we feel like the technology is showing well, but that's not the only criteria, right? At the end of the day, it'll be their decision as to whether they want to move more towards a branded solution. If they do that, I think, we have capabilities that are relevant there, but it's not an easy decision, then it's really their decision.
Sebastien Naji
And then maybe turning to just your strategy, your thoughts around capital allocation. Arista has more than $2.5 billion in cash right now. How are you thinking about returning that to shareholders versus potential M&A? And just what's your overall strategy there?
Ita Brennan
Yes. I mean, we continue to look at M&A and look for opportunities. They've tended to be smaller more tuck-in team based, technology based. Those are more -- they're just more natural given platform nature of EOS. So we will definitely continue to do that. We have had a return program. We have been returning 50% of the cash flow that we generate. Since the inception of the program, I think we are on there. We're just close to ending our second $1 billion authorization, and that's been offsetting dilution. That was what the program was designed to do. If you look back, that's exactly kind of what it's done to date. So I think that's kind of the parameters around the program for now. We're working through some of the working capital. We are building some inventory because of some of the supply chain actions, so we will work through that. And I think once we get to the other side of that, we can, again, look at the program and see kind of do we want to change it or modify it at that point. But I think we want to get through some of the working capital build and start to see that kind of improve, and then we will go back and look at that program again.
Sebastien Naji
Okay, great. Maybe just last question here in the last few minutes. Just overall, what do you think is the most underappreciated aspect of the Arista story today?
Ita Brennan
I think sometimes the enterprise piece of the business kind of gets overlooked a little bit, just because it's not -- we spend so much time talking about cloud, which is exciting and important, and is at our core. But at the same time, this enterprise opportunity is very meaningful. I mean, the size of those of those two markets combined, the early stage that we are at, I mean we can continue to execute systematically there and continue to have it be a big contributor to growth for a long time. It's not as exciting or as the cloud, but it's a real backstop to kind of growth in the business over time, and it's much steadier. And we are making a real difference to customers who are deploying the solution, and the opportunity to keep doing that is pretty significant.
Sebastien Naji
Okay. Great. Well, thank you, Ita. Thanks for being with us. We're going to have a breakout session upstairs in Adler.
For further details see:
Arista Networks, Inc. (ANET) Presents at William Blair 43rd Annual Growth Stock Conference (Transcript)