Arista Networks Inc (ANET) 2021 Q3 法說會逐字稿

Welcome to the Third Quarter 2021 Arista Networks Financial Results Earnings Conference Call. (Operator Instructions) As a reminder, this conference is being recorded and will be available for replay from the Investor Relations section at the Arista website following this call.

Ms. Liz Stine, Arista's Director of Investor Relations. You may begin.

Thank you, operator. Good afternoon, everyone, and thank you for joining us. With me on today's call are Jayshree Ullal, Arista Networks President and Chief Executive Officer; and Ita Brennan, Arista's Chief Financial Officer.

This afternoon, Arista Networks issued a press release announcing the results for its fiscal third quarter ending September 30, 2021. If you would like a copy of the release, you can access it online at our website.

During the course of this conference call, Arista Networks management will make forward-looking statements, including those relating to our financial outlook for the fourth quarter of the 2021 fiscal year, longer-term financial outlook for 2022 and beyond, our total addressable market and strategy for addressing these market opportunities, the potential impact of COVID-19 on our business, product innovation, the imply of supply -- the impact of supply shortages and manufacturing constraints on our business, including lead times and inventory purchases, and the benefits of acquisitions, which are subject to the risks and uncertainties that we discuss in detail in our documents filed with the SEC, specifically in our most recent Form 10-Q and Form 10-K, and which could cause actual results to differ materially from those anticipated by these statements.

These forward-looking statements apply as of today, and you should not rely on them as representing our views in the future. We undertake no obligation to update these statements after this call. Also, please note that certain financial measures we use on this call are expressed on a non-GAAP basis and have been adjusted to exclude certain charges. We have provided reconciliations of these non-GAAP financial measures to GAAP financial measures in our earnings press release.

With that, I will turn the call over to Jayshree.

Thank you, Liz, and welcome to your first earnings experience. Thank you, everyone, for joining us this afternoon for our third quarter 2021 earnings call. Today's call will be followed by our Virtual Analyst Day at 3:00 p.m. Pacific Standard Time. We delivered record revenues of $748.7 million for the quarter, with record non-GAAP earnings per share of $2.96. A-Care services and software renewals contributed approximately 21.5%.

Our non-GAAP gross margins at 64.9% was influenced by enterprise and cloud titan momentum. We remain pleased with our healthy customer growth, including record million-dollar customers and new customer logos in our mainstream enterprise.

In Q3 2021, cloud titans was once again our top vertical, with enterprise being a close second, followed by financials and specialty cloud providers tied at third and service providers at fourth place. All verticals contributed to Arista's diversity and growth. International contribution was strong at 25% with the Americas at 75% for the quarter.

No earnings call these days is complete without supply chain commentary. We are clearly in the midst of an acute supply chain crisis with increased prices and long lead times. We are changing our Arista mindset from our historical build to forecast and orders to build to invest, doubling our purchase commitments in excess of $2 billion and planning for the next 1 to 2 years.

Lead times of many components have extended to 50 to 80 weeks with price hikes ranging from 15% to as high as 200% across our entire supply chain of copper, steel, substrate, printed circuit board, memory, silicon, ICs, connectors, freight and labor. Arista has been deliberate and thoughtful about price increases so far, as we shared with you but we have recently announced increased list prices effective November 4, 2021, averaging about an approximately 10% to offset the very high escalating costs.

Customer demand remains strong for Arista products as we're gaining market share in 100- 200- and 400-gig high-performance switching according to market analysts. We truly appreciate our customers and partners for their patience and understanding as we navigate these turbulent times throughout 2022 as well.

Recently, we have witnessed the progress of our routing products with key customers and the acceptance of our routing edge use cases. Similar to cloud titan, carriers and large enterprise customers are deriving immense benefit from Arista's EOS and rich routing features. We deliver simplification and unified service delivery with the support of segment routing with traffic engineering and EVPN as well as rapid failover techniques. This provides that ideal alternative to today's complex legacy router deployments with much more improved total cost of ownership and CapEx benefits.

Since its founding, Arista has pioneered the transformation from router to routing with our leaf/spine, R-Series platforms. Arista's third-generation R3 series based on EOS 4.26 delivers 3 new edge use cases this year. The first one is a multi-cloud edge that brings provisioning and programmatic traffic steering. The second is a metro edge for single protocol adoption across multiple edge VPN services into the metro ethernet fabric. And the final use case is a 5G RAN edge with the 5G edge disaggregating the radio area network with scale-out routing.

Continuing our theme of big bet wins, I would like to highlight worldwide examples of our strength with specific customer names in routing and campus adjacencies. The first customer was CDLAN, an international service provider in Italy that adopted Arista for their routing transformation. Arista solution led them to take a fresh approach to routing for next-generation edge and backbone, reducing the complexity of protocols. This delivered L2 and L3 services with EVPN on a segment routed backbone, along with modern operations and superior services and experience.

The second customer was Connecticut Education Network, who standardized on Arista's R-Series with Arista EOS being instrumental in their transformation of the MPLS VPN edge providing 100-gig density in with rough scale, stability and manageability. The advantages and the relationship with CEN across service and engineering affirm their decision to choose us at Arista.

Peering between ISPs using 100-gig and MPLS PEs to replace their large legacy routers. The third customer was Zenlayer, an international customer in Asia Pacific, who is delighted to partner with Arista and build their next-generation cloud edge and routed backbone for the infrastructure growth. Arista's rich routing stack brought programmatic traffic engineering at the core of the segment routing without sacrificing quality, performance and reliability.

And finally, in the campus, we continue to make progress towards our goal of doubling to $200 million in the cognitive campus in 2021. An example of this is an international customer win in Australia, the Australian Securities Exchange, providing cognitive campus for its corporate sites in Sydney, Melbourne and Perth. The new campus network is based on Arista's wired platform the 720XP Series, and it's built on a multiyear relationship we have built between Arista and ASX utilizing EOS and CloudVision for real-time insights across all devices in trading and nontrading environments.

In summary, Arista's customers strongly endorse our client-to-cloud strategy to unify silo data sets consistently. We believe we are well positioned for the next phase of growth in data-driven cloud networking with proactive platforms, predictive operations and a prescriptive experience. We look forward to sharing more of this and our vision and our goals with you at our Analyst Day later this afternoon.

I will pass it over now to Ita Brennan, our Chief Financial Officer, for financial specifics. Ita?

Thanks, Jayshree, and good afternoon. This analysis of our Q3 results and our guidance for Q4 of 2021 is based on non-GAAP and excludes all noncash stock-based compensation impacts, certain acquisition-related charges and other nonrecurring items. A full reconciliation of our selected GAAP to non-GAAP results is provided in our earnings release.

Total revenues in Q3 were $748.7 million, up 23.7% year-over-year and above the upper end of our guidance of $725 million to $745 million.

Shipments remained constrained in the period as we continued to carefully navigate industry-wide supply chain shortages and COVID-related disruptions. Services and subscription software contributed approximately 21.5% of revenue in the third quarter, down from 22.3% in Q2. International revenues for the quarter came in at $191 million or 25% of total revenue, down from 27% in the second quarter.

The shift in geographical mix on a quarter-over-quarter basis reflect a continued healthy performance from our cloud titan and in-region businesses in EMEA with some volatility in our APAC business. Overall gross margin in Q3 was 64.9% at the upper end of our guidance range of approximately 63% to 65%. We continue to recognize some incremental supply chain costs in the period, and these were offset by a healthy mix of revenue from our enterprise customers in the quarter.

Operating expenses for the quarter were $192.4 million or 25.7% of revenue, up from last quarter at $189.8 million. R&D spending came in at $125 million or 16.7% of revenue, up from last quarter at $119.6 million. This reflected increased headcount and employee-related costs and higher new product introduction spending in the period. Sales and marketing expense was $55.8 million or 7.4% of revenue, down from $57.9 million last quarter with lower demo and other variable expenses in the period. As a reminder, we continue to benefit from lower COVID-related travel and marketing expenses.

Our G&A costs came in at $11.6 million or 1.5% of revenue, down slightly from last quarter, but in line with normal quarterly seasonality. Our operating income for the quarter was $293.7 million or 39.2% of revenue. Other income and expense for the quarter was a favorable $1.3 million, and our effective tax rate was approximately 19.7%.

This resulted in net income for the quarter of $236.9 million or 31.6% of revenue. Our diluted share number was 79.9 million shares, resulting in a diluted earnings per share number for the quarter of $2.96, up approximately 22.5% from the prior year.

Now turning to the balance sheet. Cash, cash equivalents and investments ended the quarter at approximately $3.4 billion. We repurchased $134 million of our common stock during the third quarter at an average price of $357 per share. As a recap, at the end of Q3 2021, we had repurchased $897 million or 3.9 million shares against our Board authorization to repurchase $1 billion worth of shares over 3 years beginning in April 2019.

In October 2021, Arista's Board of Directors increased the authorization by adding an additional $1 billion to the repurchase amount. The actual timing and amount of future repurchases will be dependent on market and business conditions, business requirements, stock price, acquisition opportunities and other factors.

Now turning to the operating cash performance for the third quarter. We generated $273 million of cash from operations in the period, reflecting strong net income performance and continued investments in inventory and supply chain. DSOs came in at 49 days, up slightly from 47 in Q2, reflecting the linearity of billings in the period. Inventory turns were 1.7x, consistent with last quarter. Inventory increased to $575.7 million in the quarter, up from $543.2 million in the prior period as we continued to buffer certain components and products.

Our purchase commitments number for the quarter increased to $2.1 billion, up from $1.1 billion in Q2. This reflects a combination of increased lead times for many components and improved demand visibility. We continue to prioritize newer early life cycle products for inclusion in this strategy to help mitigate the risk of obsolescence.

Our total deferred revenue balance was $800 million, up from $746 million in Q2. The majority of the deferred revenue balance is services related and directly linked to the timing and term of service contracts, which can vary on a quarter-by-quarter basis. Approximately $113 million of the balance, up from $90 million last quarter represents product deferred revenue largely related to acceptance clauses for new products, most recently with our larger cloud titan customers.

As a reminder, we are currently in a period of significant new product introductions, combined with a healthy new customer acquisition rate and expanded use cases with existing customers. These trends in conjunction with reduced levels of upfront in-person testing have resulted in increased customer-specific acceptance clauses and higher product deferred revenue amounts. Accounts payable days were 47 days, down from 54 days in Q2, reflecting the timing of inventory receipts and payments.

Capital expenditures for the quarter were $45.9 million, including approximately $40 million of CapEx related to the purchase of land to construct a new data center and hardware engineering building in Santa Clara. We'll provide more details on this project over coming quarters.

Now turning to our guidance for the fourth quarter and beyond. As outlined in our guidance, we now expect to achieve year-over-year revenue growth for the full year 2021 of approximately 25%. This reflects continued healthy demand across all market sectors, tempered by the impact of the difficult supply environment. On the gross margin front, industry supply constraints and elevated logistics costs continue to pressure gross margins, with customer price increases as a potential offset. Based on our current outlook, we continue to reiterate our overall gross margin outlook of 63% to 65%, with customer mix remaining the key driver of volatility on a quarter-by-quarter basis.

Turning to spending and investments. We remain committed to growing our investments in R&D to support innovation across the business and sales and marketing to support our go-to-market expansion. Finally, we also announced today that Arista's Board of Directors has approved a 4 for 1 stock split. Each Arista shareholder of record at the close of business on November 11, 2021, will receive 3 additional shares for every share held and trading will begin on a split-adjusted basis on November 18, 2021.

With all this as a backdrop, our guidance for the fourth quarter, which is based on non-GAAP results and excludes any noncash stock-based compensation impacts and other nonrecurring items is as follows: revenues of approximately $775 million to $795 million, gross margins of 63% to 65%, operating margin of approximately 37%. Our effective tax rate is expected to be approximately 20.5% with diluted shares on a pre-split basis of approximately 80 million shares.

I will now turn the call back to Liz. Liz?

Thank you, Ita. We are now going to move to the Q&A portion of the Arista earnings call. (Operator Instructions) Operator, take it away.

(Operator Instructions) Your first question comes from the line of Samik Chatterjee with JPMorgan.

Congrats on the strong results, really impressive. So let me keep it broad-based, Jayshree. I think you mentioned strong demand that you're seeing, and I think you highlighted cloud customers in the press release. But just generally, if you can talk to how broad-based is the demand that you're seeing across cloud, and then what is the kind of magnitude of demand that you're seeing from enterprise customers? And how do you think about sustainability of that level of demand, which you're seeing this year? How do we think about sustainability of that into next year?

Well, thank you, Samik, for the good wishes. It's a proud moment and I really congratulate my entire leadership team and my employees for getting us here. I think demand is very strong, as I mentioned in my earlier script, across all 5 verticals across all 3 product lines and across all 3 sectors as well. So I would not -- I would tell you, we are growing in that what Ita highlighted as our 25% annual growth, every sector was growing.

And so that's, in some ways, I feel bad that I even have to rank and rate them. But if you ask me to highlight some of the growth vectors, I would say, obviously, cloud titans are back. We had a rough fell. If you remember, 2 years ago, Halloween was not a treat. It was a trick. And it's just come back and it is a volatile sector, and it's positively volatile right now. So we're enjoying the growth of cloud titans.

We're also enjoying many pieces of our enterprise market growing. And there are really some verticals there that are doing very, very well, not just the financials, but different parts of the enterprise. I think it's fair to say Arista has arrived in the enterprise. We've been growing double digits for a couple of years, and we expect to continue to see double-digit growth in the enterprise sector. And this is by far our largest momentum of all the verticals, I would say.

But not -- as I mentioned, a lot of routing use cases, these routing use cases are not only in the cloud titans, but are obviously also in service providers and enterprises as well. So we're just enjoying a very diversified momentum of our business at the moment.

Your next question comes from the line of Fahad Najam with MKM Partners.

I wanted to ask you a question on the visibility. You mentioned that certain components have lead times have extended from 50 weeks to 80 weeks. I presume your customers in turn are giving you forward-looking guidance as well. So can you give us a sense on the visibility you're seeing? And help us quantify that in any way you can.

Sure. I'll say some few words, and Ita, if you could add to that. I think because of these kind of long lead times on our components, first thing Ita and the team are doing, Ita and I'm sure the entire team are planning ahead. And we're no more like we said, building to forecast orders but really building to a future demand. So visibility becomes very important in that case because it's no more 1 or 2 quarters. The cloud tightness visibility has improved a lot this year.

Typically, it used to be 1 to 2 quarters. Right now, it's more like a year or more. So this is the best visibility we've ever had with the cloud titan that's allowing us to build the inventory and to build a plan and to get ahead, if you will. In the enterprise as well, nobody's lead times are very good right now, and we're no different. Although we thought and we believe we have a head start by starting on this problem as early as last year. We have hundreds of suppliers, and we've had to increase our strategic interface with these suppliers to -- and make, again, bets on them long term.

So visibility in the enterprise is also 6 months to a year, visibility in the cloud and specialty cloud providers is now exceeding a year. So in general, we're now able to make a plan to buy components well ahead of the purchase orders and forecast. Ita, do you want to add something more?

Yes. I mean I think the only thing I'd add, Fahad is it's hard to be too quantitative when you think about demand and bookings just because obviously, the lead times and the times frames are very different, right? So I think from -- just from a business perspective, we'll continue to focus on the revenue and the revenue metrics and then the bookings numbers will kind of ebb and flow. But obviously, right now, you are getting a lot of visibility to what's happening with customers just because we need that to be able to drive the types of purchase commitments et cetera we're driving.

Your next question comes from the line of Rod Hall with Goldman Sachs.

Yes. And again, I'd like to echo the positive comments. These are phenomenal results in this environment. I guess my question is regarding the COGS and the cost of some of the products you're getting from Broadcom, other companies we've talked to during this earnings season has talked about really high expedite fees. And just wondering if you're seeing those? And how they're factoring into the forward cost in the business. Like are you able because of this visibility to set your prices at a level that compensate for it? Will we see higher COGS unwind, maybe the early part of next year. I'm just curious whether you're seeing those expedite fees and then how they might affect margins at some point?

Rod, I think everybody is seeing -- I don't want to talk about a particular supplier, but we are seeing expedite fees and incremental costs kind of across the supply base, right? And you haven't seen those in the gross margins in the income statement to date just because the mix has been more enterprise heavy. And that's been kind of offsetting that, right? I think we are -- as Jayshree mentioned, we are in the process of instituting some price increases, et cetera, to help offset some of those costs. So that will help.

I think we're comfortable thinking about that 63% to 65% range is still being reasonable. But you will see some more volatility quarter-by-quarter just as the mix of the business. The customer mix is still going to be the biggest driver. The other costs we're managing with some of the price increases, et cetera. But when we have a heavier kind of cloud mix in a particular quarter, et cetera, we will see some lower gross margins than what we've seen over the last couple of quarters.

And thank you for the good wishes, Rod.

Sure, Jayshree. No problem. Just a question. Our customer -- what about the cloud customers, are they willing to accept a little bit of price increase, knowing that things are getting more expensive. Just curious what the conversations are like there.

I would say all our customers are very understanding, but nobody is willingly accepting price increases, including the rich cloud titans.

Your next question comes from the line of Jim Suva with Citigroup.

Truly spectacular. And I got to just ask about the build to forecast versus build to order. Kind of when did you implement that? And what was the reaction of some of your customers? Or is it more internal? And what I'm wondering is how much further you may be ahead of some of your competitors? It sounds like it's actually maybe something that you're looking for, for like the next couple of years if you have lead times going so long.

Right. First of all, I just want to give a big shout out to Anshul, John McCool, Susan Hayes and the entire manufacturing team. If you -- let me just step back, Jim, and thank you for the kind wishes. The traditional model for everybody has been built to forecast, Lead times are based on supplier commits. There's some buffers, but most of it is just in time, right? And very rarely does anybody pay for expedites.

If you look at supply chain in 2022, first of all, expedites are a way of life. It doesn't matter which vendor it is. You have to plan not just weeks ahead, but months ahead. Often, you can get decommits from suppliers. There's shortages across the board. There's lots of orders with no buffers. Everybody is coming at them sometimes and we used to think the high-tech industry is special. But some of the components we're talking about, we compete with the automotive industry and the consumer industry, which makes it tougher.

And it isn't surprising at all to see expedites involve not just CEOs, but heads of countries literally. That's how rough it is. So it's a very oversubscribed process. We thought we got a head start by starting when was it Ita, late last year when Anshul and the team put together a plan. So we've definitely had a head start. And if things have gotten better, we would be well ahead of everyone. But just things keep getting worse.

So now the head start is good, but we have to add to that head start and hence, the doubling of the inventory. And without naming any vendor, I'll just say we've increased the strategic nature of our relationship with not just 1 or 2 vendors, but 25% of our vendors. So this is a much larger relationship pool, and we're committing to them long term. They're committing to us, but both of us have to be patient and understanding of the short-term troubles we have.

And again, a really big congratulations to you and your entire entity.

Thank you. Big kudos to my team.

Your next question comes from the line of Paul Silverstein with Cowen.

Two earlier questions, if I may. Was there any 200-gig from Facebook and any 400-gig revenue from Microsoft in your third quarter? And do you expect in the fourth quarter? And Jayshree, would you care any comment on the outlook for next year? I know supply chain is challenging, but...

Yes. I'm just checking to see, but there was 400-gig revenue overall. As I told you last time, we have increased our customer logos in the 400-gig category from 75 customers last year to the first half was 150, and we're trending to about 300 customers. So they're definitely 400-gig. I need to double check on whether there was any 200-gig. Let me beg off that question. Ita, do you have...

Yes, I was just going to say, Paul, some of the commentary on the deferred is probably relevant here, too, where we talked about the deferred balance becoming more cloud titan heavy this quarter, whereas before, there had been more other verticals, et cetera, right? So I think that's -- so we may not have had revenue, but we probably had activity. We deferred the revenue.

So we have some deferred, too.

So that's what I thought. But just to be specific Jayshree, the question specific to Facebook and Microsoft, I assume a lot of that revenue is being deferred or maybe it's not. But that's the specific question.

Yes. I think we grew our deferred revenue when we mixed definitely towards cloud, and that includes new products, right? So that's kind of...

It's our 200-gig and 400-gig.

All right. And would you care to give any comment about the outlook for next year?

We are going to at the Analyst Day. How about then?

I can wait 30 minutes.

All right. And suppose if you're in the East Coast, I apologize for keeping you up late, but we'll make it short and sweet. I think our Analyst Day will be 2 hours.

Yes.

So you don't have to stay up too long.

All right. I appreciate.

Your next question comes from the line of Sami Badri with Crédit Suisse.

Jayshree, you've mentioned a couple of times talking about reference to enterprise wins and that really kind of dialing up as far as momentum. But if you were to bullet point the key reasons why you're winning, you continue to win with what it would sound like increasing momentum.

Can you just highlight them for us because most of the people on this call are used to hearing about very defendable sales channels and many other vendors with very comprehensive solutions. Can you walk us through the key sales pitch and just what is resonating with the enterprise customers?

Right. And again, I'll give some of this at the Analyst Day, but I'll give you an abbreviated version Sami. First of all, I think our relevance in the enterprise customer has increased from data center is really a much broader portfolio that's client to cloud, going all the way from campus, WiFi, (inaudible) design for the data center to routing and a very large dose now of software and services as well, everything from A-Care to CloudVision to cloud in the U.S. to total cloud software as well as our recent acquisition of Big Switch and Awake now contributing as well to segmentation, observability and security as well.

So the completeness and the innovative nature of our portfolio itself. The second thing that's helped is our Power of One, if you will, one OS, one image, one cloud vision. Customers just love the -- not just the innovation, but the quality and support. I'm not having to buy silo boxes, but having an innovative and much better operator experience with a much lower TCO. And finally, at the enterprise customers is that we have now much as we talk about products, we have invested in customers. Our investment in sales led by Chris Schmidt and Ashwin Kohli and the entire team worldwide really began in 2017.

So this is our third or fourth year of enterprise investment. And I think we're now seeing the results of that. The first and second year, we were kind of getting in, and we're just coming into the campus. And now I think we're coming on to our own in a complete holistic fashion.

Your next question comes from the line of Jason Ader with William Blair.

Yes. First, I want to say horrible numbers you guys need to do better. But my question is, can you quantify the backlog or book-to-bill or anything that might help us understand kind of how much of a gap there is between demand and supply. And is there any risk that customers are over ordering right now where you could see an air gap in demand maybe in some time in 2022.

Yes, Jason, I know lots of folks have been talking about bookings and trying to put some boundary around that. I just think it's really hard from a timing perspective. When you have these lead times, of course, you're going to have accelerated bookings and larger bookings. And certainly, we have our fair share of that, right? There's no -- it's always difficult to talk about the business, I think, in that context. So we're more focused on what can we deploy and that's how we're running it internally as well, right?

What are the periods where these bookings will get deployed and building out deployment plans. And that's really what's going to matter. And I think when you think about the business that way, the pull-ins and push-outs of the actual bookings numbers and how much visibility you're getting, et cetera, becomes less important, right? So not kind of duck -- not ducking your question, we have obviously lots of demand. We've talked about the demand that we have, but these are extended lead times. So we're just focused on making sure we understand how it's going to get deployed.

Yes. And I want to echo what Ita just said. We're not going to get excited about backlog. We're a little bit excited about deploying our customers with real revenue. And some of the backlog may materialize and remember, they're cancelable orders, some of them may not. So it's best to be responsible as a company as we always have been and share with you that demand is certainly outstripping supply. No question about that. And we're going to work hard as hell on fulfilling the supply and improving the supply.

Your next question comes from the line of Meta Marshall with Morgan Stanley.

Great. I realize it's kind of difficult to quantify the supply chain impact currently. But if any way you help us with the gross margin impact? And should we see the gross margin step down in the guide is more supply chain related or more related to the mix of revenue types?

Yes. I think the best way to think about it is that we've been operating at the upper end of that range for the last couple of quarters. That's definitely a customer mix effect, right? It's offsetting some of the cost impacts as well. And we've been deferring some of the -- as Paul was talking about, some of the larger customer revenue as well.

So I think as you look forward kind of outside of these quarters, when the mix of the business comes back to something more balanced, I think you will see back towards the bottom end of that range from time to time. I think we believe we'll stay in the range over a long period of time, but there will be quarters where we could be pressuring the bottom end of that range and maybe even break the bottom end of that range.

While for, say, for 4 quarters, I think we can still be okay. So there is definitely a customer element to this. There's a cost increase element to this, and we will benefit from some customer price increases here that will help offset some of that. But I think the days of living at the upper end of that range, I wouldn't assume that we can do that on an ongoing basis as you look forward.

Got it. And not to trouble you in the -- like any forward guidance, but just for the price increases, obviously, you probably wouldn't see most of that impact in Q4, you would expect the price increases to impact more Q1 in 2022. Would you...

Yes, we do expect.

Your next question comes from the line of David Vogt with UBS.

Great. This is a question for both, I guess, Jayshree and Ita. I just want to follow up on the cloud titan CapEx and the hyperscaler CapEx and the visibility. I think it's fairly well documented that the balance of this year into 2022, There's going to be significant data center expansion and availability expansion by the hyperscalers so that's clearly reflected in your confidence.

But you noted that you have a little bit more than a year visibility. I mean how should we think about 2023? I know we don't even have '22 guidance yet, but given the strength and the expansion and the availability and the data center trajectory, how do we think about that? And then just as a follow-up on pricing, when you think about the 10% price hike that you're going to implement in a couple of weeks, in your mind, is that sort of more than offsets the supply chain? Or is there a way to quantify how we're thinking about price versus the margin impact from the higher components? Does it reduce it by 50%? Is there some way to think about it that we can model out going forward?

Yes. No. I mean, we've tried to be very transparent with customers in terms of what we're seeing on the cost side and looking for them to help us kind of offset that. So we're definitely not looking to increase margin or make margin on that. We've been very open and transparent with what we are seeing on the cost side and looking for help to offset that. Jayshree, anything you want to add there?

Yes. And to answer your question on 2023, I guess I would say stay tuned for the Analyst Day, we'll try and give you better visibility on 2022 and give you some visionary statements on 2023 and beyond.

Your next question comes from the line of Amit Daryanani with Evercore ISI.

And I'll extend my congratulations as well to you. I guess when I look at your performance in '21 based on the midpoint of your guide for December, I think you would have clearly gained some sizable market share in the year. I'd love to understand, do you think the share gains are coming from white box vendors or coming more from sort of the traditional competition that you have?

And then maybe a second part to this, as you think about the next couple of years, could you see customers that use white box solutions today come to Arista? And if so, what do you think would motivate them to do so?

Both very good questions and related to each other. I would say this year with all the supply chain issues, much more of our share gains is coming from enterprise and cloud titans. Just getting our fair share from our peers in the industry, not necessarily white box. If you fast forward to later years, I do think Arista will have an advantage, not just in product capability, but also in the ability to rapidly supply product probably better than some of the white boxes. And Anshul has often alluded to this. So the make versus buy decision for many of our cloud titans may shift in the direction of Arista rather than strictly white boxes.

We look forward to that. I'm not going to make any guesses on that, but I don't preclude that and Ita has Anshul when he's spoken in the past. So it certainly wasn't part of the market share gains and the growth this year, but it could be next year.

Your next question comes from the line of Pierre Ferragu with New Street.

I am very intrigued by the 1-year visibility you have with your cloud clients. And on that front, I was wondering, first, Facebook, one of your important clients hiked up their CapEx for next year by like 60%, 65% or so last week.

And I was wondering, is that something that is I would say, aligned with the visibility you have or if it came as a surprise. And then along the same line, within that visibility, how do you see the spending of cloud titans changing in terms of how it is split between what's happening inside the data center, which is more on the switching side and what is more happening outside the data center in the DCI and more on the routing side.

Yes. Thanks, Pierre. Both again, very good question. So I'll take the second one first. I think Arista's presence for most part until recently has been intra data center. But what has been phenomenal to watch my cloud titan team do, led by Anshul, Mark and others is the use cases have proliferated, not only outside the data center of a DCI, but routing AI use cases, top-of-rack use cases, special customized use cases. So both within the data center and outside Arista is getting its fair share of opportunity to respond. And we're doing a lot of proof of concepts and testing work with them.

Regarding the cloud titan CapEx spend, we're always surprised when the numbers actually come out because they're in billions. And of course, they're nowhere close to the percentage they spend with us necessarily. But our relationship with Facebook dates back now at least 4, 5 years.

We have done joint development with them in the FBOSS and we've jointly developed products with them. We have shared with you in the past that we are developing our next generation of products with them with 200-gig. So we were pleasantly surprised, but we were not completely surprised.

Your next question comes from the line of Aaron Rakers with Wells Fargo.

Congratulations as well from me. I think the one number that stands out the most to me is your $2.1 billion plus purchase commitments, and I think that's up over 4x relative to what it was exiting last year. I think going into kind of the June quarter, the expectations were that maybe some of these component constraints would start to ease as we move into the mid-part of 2022 and certainly into the second half?

And just curious, your best assessment right now where you stand on some of those lead times starting to normalize or shorten back down? And do I think that you're going to carry kind of this higher degree of visibility well into 2023 at this point?

And Presenters are -- is your line muted?

Sorry about that. I don't know when that (inaudible). Where did I stop. Yes. So I think -- sorry, Aaron, can you help me with how much of that you got.

Actually, I didn't hear any of it. I apologize.

Okay. All right. Okay. Let me start. Yes. So I think, look, we probably have 2 dynamics happening. We've seen a push out of lead times again with the products and the vendors that we were managing directly. And that's probably -- I think now our view is that's probably the end of 2022 before we start to see things get better there.

In addition to that, we've also seen it kind of broaden out to other components, right? And we're now managing vendors directly that would have normally gone through the supply chain gone through the contract manufacturers, et cetera, and we're having to engage directly with those suppliers.

And then that's also driving some increase in those purchase commitments, and we're looking out longer with those suppliers as well. So it's a combination, I think, of both of those is taking that number increase. We are trying to focus on new products and products that have long life cycles. So that gives us a little bit more leeway there in terms of taking a longer view, and we'll continue to do that. But yes, I don't think we've kind of asked the point yet where things are improving.

And Aaron, we see this as an important investment to the business. It is a decision that Ita, myself and Anshul have made very consciously to say we got to invest in the business, and we've got to invest in getting product to our customers. So we think this is an important part of our decision-making process because of the prolonged situation here with supply chain.

Your next question comes from the line of Ittai Kidron with Oppenheimer.

Congrats. Great quarter. I guess a couple of questions from me. First of all, with regards to the purchase commitments. Can you give us a little bit more color whether -- this is a response to competitors of yours doing the same with your suppliers? And does this lock in volume? Or does it also lock in price for the components that you're buying?

Yes. No, I think it's totally us working on our own strategy. And as Jayshree mentioned earlier, we started to do that right back at the beginning of last year even. So just a continuation of that. I think the biggest driver is obviously what's happening in the supply chain and understanding what's happening in the supply chain and just the breadth of suppliers that we need to kind of manage directly and start to deal with directly right now. I mean I think that's the biggest driver of the change as opposed to anything that anybody else is doing, et cetera. I'm sorry, what was the second part of your question, Ittai?

Does it lock in just the volume? Or does it also lock in the prices for you going forward?

Yes. I mean, when you make long-term commitments, there is kind of a pricing element to that, that you have a set price base. As things start to get better, we'll see how some of that plays out, right? But there obviously is a pricing in the market today, and that pricing is kind of what you're making these commitments at. And -- but as time -- as we've seen over time in the past, as things start to loosen up and some of that can change as well. But right now, it is a commitment to volume and price.

Your next question comes from the line of Simon Leopold with Raymond James.

As you probably remember, earlier in my career, I was told never to high-five management on a public call. I'll leave it at that. I've done the last. I wanted to see if maybe you could expand a little bit on the campus opportunity, which sounds like it's overshadowed by what's going on in data center. But just want to get a better sense of where you stand in that part of your business and the trajectory?

Thank you, Simon. We will take your virtual high-five. I think the campus business has been very relevant to a seat at the table in -- with enterprise customers. We're now starting to see enterprise wins and logos where we win the campus before we win a data center because many of these campuses don't have large data -- many of these enterprises don't have large data centers. So I think the conversation, the strategy, the ability to bring all of the silo data sets together, whether it's in your campus or data center or on the core or WAN or branch is very important, and customers are looking to us to build their modern and modernize their enterprise network.

So from that standpoint, although the numbers are still small, and we're talking about doubling from 100 to 200. We think it will be extremely relevant strategically with our enterprise customers and to grow, obviously, in the next few years. There's also a component of channels where we're still pretty nascent. And most of our success and engagement to date is direct, albeit fulfilled by channels, but we hope that will change over time, and that will add further strength to our campus.

Your next question comes from the line of Tal Liani with Bank of America.

I have 2 questions. One is just if you can give us an update on campus switching, where you are versus your targets? If you said it, I apologize, I just didn't hear it. And second, I just want to understand kind of about the accounting. Can you go over again the price increases? When are they kicking in? If they kicked in already?

And then what happens with your cost of goods sold since it's on FIFO. I'm assuming it's on FIFO like everyone else. Does it mean that right now, you're still recording cheaper components, so the margins are higher. I'm just trying -- or maybe I'm totally wrong, I just want to understand kind of the margin evolution as component pricing goes up and pricing increases kick in.

Yes. I mean...

Go ahead.

It's some combination of all of those, Tal. There are some expedite costs, et cetera, that end up being period expenses that we've been recognizing. We recognized a chunk last quarter. We had some again in Q3. There are other costs like higher price -- higher pricing increases, et cetera, that will end up being inventoried and will flow with the inventory. And some of that obviously will burn through the inventory that we have in the supply chain, and then we'll start to see those costs.

Those will line up better, hopefully, with some of the price increases that we are passing on to customers as well. I mean this is -- it's going to be complex. It won't necessarily be perfect, right? But as we look at the various different scenarios, there's a better chance of those lining up with the price increases. So that will help kind of offset some of that.

And finally, Tal, on the campus. We committed to double the $100 million achievement we had last year, this year, and we -- here we are sitting 2 months away from the end of the year. So we believe we will lose it and we will have to set a new goal for next year.

Got it. And just going back to the margin. So I know you're probably going to discuss it tonight, but just in general, how do we think about gross margin going forward?

Yes. I mean, it's definitely part of the discussion later.

Right. And I think you already mentioned. What was mentioned already is we continue to believe with the price increase effective November 4, which will really be effective next year by the time customers realize it and see it that we will be able to offset the escalating costs and our gross margin will depend on mix. And as Ita has often said, if we are heavily mixed on the cloud titans, we could be on the low end of the 63% to 65% and pressured on gross margin there. If we're heavily mixed on the enterprise, we could be on the mid- to high end like we have been with the last couple of quarters.

Your next question comes from the line of Ben Bollin with Cleveland Research.

Jayshree, I was hoping you could talk a little bit about how you view the current technology build-out, the new technology build-out for both enterprise and cloud as they start to transition into 200 and 400. How you see similar or different versus what you saw from 2016 to 2018 with more cloud titans building out 100. Any thoughts on duration, behavior? Any puts and takes would be interesting.

Sure, Ben. I think the cloud titan behavior will be different than the enterprise behavior. On the cloud titan, you're going to see a much more rapid inflection to 200-gig and 400-gig, especially in the spine layers and the uplinks at the top of rack. And they've always been an early and fast adopter of speeds and technology, especially within a data center or even data center to data center. So we are expecting an inflection of 200- and 400-gig.

That basically has started this year was very challenged with the ecosystem and availability of optics and even switches the last year. The year of inflection, in my view, is really late this year and goes well into 2022. On the enterprise too, we expect to have by the end of this year, 300 customers of 200 and 400-gig, primarily 400-gig, I would say. And as you know, our customer base is more than 7,000. So obviously, 100-gig and 400-gig will continue to coexist and live happily ever after together. But we will start to see the uptick and adoption of 400-gig in the high end and early adopters of enterprise as well, like we're starting to see this year.

So I think the next few years can be best characterized as inflection of new higher speeds like 200 and 400 and the continuation and adoption of 100-gig in the mainstream enterprise.

Your next question comes from the line of Erik Suppiger with JMP Securities.

Yes. Congrats. And curious, how has the constraints affected the 400-gig market? It sounds like you've been doing well there, but has that been a factor? And does that make a difference from a market share perspective? Do you have any advantage or disadvantage in terms of access for your 400-gig components?

Erik, I'd say we are as constrained on 400-gig components as we are 100-gig components. So it's been a factor for all speeds, which is constrained. What can I tell you? So -- but our market share continues to be strong in both. We're doing well in both. Our flagship platform, the 7800, which is especially 400-gig dependent is one of the most popular products. At the same time, our [shift to 80] and 100-gig versions of 75 and 7800 are very, very popular, too. So supply chain is bad for everything. It's not necessarily picking one speed over the other.

Are the optics anything different?

Optics is actually better than last year in terms of the ecosystem for 400-gig coming up, but not different other than that. It's actually improved for 400-gig.

Your next question comes from the line of George Notter with Jefferies.

I guess I wanted to go back to your statement earlier, I'm paraphrasing, but I think you said you were running the business to demand rather than orders or something to that effect. Could you go back and kind of expand upon that? I'm just curious about what you meant on that. I assume you're trying to look through customer order books and try to see what they really need as opposed to excess ordering. Maybe you could just expand on that.

It's actually the other way around, George. What we've said is all the just-in-time and build to forecast and being extremely disciplined about buying inventory only when the customer puts in an order has gone up the window a little bit. And because of these long lead times, we're having to plan to order well ahead of the customer orders or forecast. That's what we meant.

So it's build to purchase orders to our supply chain rather than build to customer purchase orders. If that makes sense. Since they're still constrained on long lead times. So we're making a bet that the supply chain constraints, which I hope will eventually improve, will favor those of us who make this kind of purchase commitments. So we're having to get in there early and fast even before the customer orders come in.

Out of curiosity, do you have any flexibility on those purchase commits? I mean are they cancelable on your side?

Most of the semiconductor components are noncancelable, but that's just the way the business is run. We've been careful to choose components that we don't need to cancel like picking new products, which -- and taking common componentry across them. So we believe there's limited risk in the investment we've made.

Your final question comes from the line of Jon Lopez with The Vertical Group.

And I apologize because I'm going to stay on the same topic. But hopefully, it will be the last one we can cover the rest of the stuff on the Analyst Day. This has been alluded to a few times. If we look at your largest competitor, they've also like roughly doubled their purchase commitments fairly recently. You're now doing the same.

The dollars collectively are like many multiples larger than either of you have ever carried. I guess my question here is, to what extent do you think inventory is actually evolving into a competitive weapon as we think about '22 and '23? And maybe like to what extent does it introduce risk, like if you can't get supply as fast or in the same quantities that you're envisioning now, can that influence your revenue outlook in '22 or in '23?

Yes. No, I think there's no doubt that supply is shaping our revenue line right now, almost more so than demand, right? I think that is a factor, right? We are constrained. So supply is definitely a factor. I think -- in terms of looking at the purchase commitments, we're working very carefully with these suppliers. And again, we're expanding kind of the breadth of what we're doing and we are expanding kind of the lead times and the length of time that we're covering with those purchase commitments. And I think that's important. And again, we're doing it on new products, newer products that have significant lives ahead of them. So really, the risk we're taking somewhat is tying up some cash, et cetera.

It's not because of the life cycle of the products and stuff, it's not really an obsolescence risk, right? But it could take some time to burn through that inventory if things change. But I think it's well -- it's a bet that's worth making. We're making it obviously in consultation and discussion with customers, et cetera. But it is kind of a longer lead time and then a broader set of suppliers than we'd normally carry. That's why you're seeing that big uptick is we're not only doing it for the owners that we used to buffer in the past directly, but we're also now doing it for components that would have come to us through the CMs in a normal supply environment.

This concludes the Arista Q3 2021 Earnings Call. We have posted a presentation, which provides additional information on our fiscal results, which you can access on the Investors section of our website. Thank you for joining us today.

This concludes today's conference call. Thank you for participating. You may now disconnect.