Arista Networks Inc (ANET) — Q1 2022 Earnings Call Transcript

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 first quarter ending March 31, 2022. 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 second quarter of the 2022 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, supply chain constraints, component costs, manufacturing capacity, inventory purchases, and inflationary pressures on our business, product innovation and the benefits of acquisitions, which are subject to the risks and uncertainties that we discuss in detail in our documentation 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. Thank you, everyone, for joining us this afternoon for our first quarter 2022 earnings call. In addition to the pandemic, we are now facing global uncertainty with the war in Ukraine and increasing inflation trends. On behalf of Arista, we express our deep concern over the tragedies in Ukraine and thank the Arista employees for their thoughtful donations to the Ukraine Humanitarian Causes matched by the Arista Foundation. Back to Q1 2022. We delivered record revenues of $877.1 million for the quarter, with a non-GAAP earnings per share of $0.84. Services, software support renewals contributed approximately 19.2% of the revenue. Our non-GAAP gross margins of 63.9% were influenced by continuing supply chain constraints and elevated costs. Vendor decommits for certain components increased sharply in March 2022, with no clear relief in sight in the near term. In terms of Q1 2022 verticals, Cloud Titans was our strongest and largest vertical, followed by enterprise, cloud specialty providers, and financials tied at third place, and service providers at fourth. Our geographical mix included strong performance in the Americas at 76%, with the international contribution at 24%. According to market analysts, in calendar 2021, Arista gained market share in overall data center, high-performance switching growing to approximately 19% market share. We are proud to maintain our #1 position in 100-, 200-, and 400-gig switching and achieved this growth despite all the challenges of the supply chain. Arista has also been a pioneer and leader in client-to-cloud networking. Our customer relevance is increasing with new logos in tech enterprises, health care, education, and retail as well as the provider sector. Our million-dollar logos have doubled in the last three years in all categories. These categories include: greater than 1 million, greater than 5 million, greater than 10 million, and greater than 25 million customers. What's clear as I personally engaged across the worldwide base of CIOs and CXOs is that the planning horizon for networking has really changed. Our visibility and demand have never been stronger. The nature of our discussions with our customers is also most strategic in nature. Let me try to illustrate with a few examples. For example, on AI Spine, we're building upon our cloud heritage to expand scale for AI workloads that are both data and compute-intensive, thereby requiring 100-, 400-, and 800-gig performance in the future with rich EOS software features. In the cognitive campus, threat hunting is now becoming integral to the network instead of being an afterthought and is changing the way cybersecurity and virulent threats can be detected and managed holistically. Network security is clearly migrating to secure zero-trust networking based on AI and AVA sensors. In larger enterprises, network as a service based on CloudVision is enabling our customers to manage their data sets across client-to-cloud domains with superior, unmatched automation and analytics. Our decade-long preferred partnership with Cloud Titans, especially Microsoft and Meta, has helped forge joint engineering products for hypercloud scale all the way now to Edge use cases. Clearly, Arista is in the midst of an exciting and growing total available market or TAM expansion to fulfill our customers' quest for proactive, predictive, and prescriptive networking, all this with the right data-driven foundation and architecture. This Network Data Lake EOS stack that we launched last November, is resonating well, and it's validating our customer network design. In conclusion, I'm so proud of Arista's progress and traction and look forward to multiple years of double-digit growth. And with that, I'd like to turn it over to Ita for Q1 2022 financial specifics.

Thanks, Jayshree, and good afternoon. This analysis of our Q1 results and our guidance for Q2 '22 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 Q1 were $877.1 million, up 31.4% year-over-year and well above the upper end of our guidance of $840 million to $860 million. Our enterprise business continued to contribute healthily to our overall revenue growth in the first quarter, combined with accelerated shipments to our cloud titan customers, some of which were deferred. Supply remained constrained in the quarter with supplier decommits resulting in higher broker purchases and expedite fees in the period. Services and subscription software contributed approximately 19.2% of revenue in the first quarter, down from 21% in Q4. International revenues for the quarter came in at $212.7 million or 24% of total revenue, down from an unusually high 29% in the fourth quarter of 2021. This decline in international mix primarily reflects some volatility with our global customers, including the deferral of some international cloud shipments in the period. Overall gross margin in Q1 was 63.9%, above the midpoint of our guidance range of approximately 63% to 64%. Operating expenses for the quarter were $225.3 million or 25.7% of revenue, up from last quarter at $206.2 million. R&D spending came in at $144.3 million or 16.5% of revenue, up from last quarter at $130.3 million. This primarily reflected increased headcount and higher new product introduction costs in the period. Sales and marketing expense was $66.2 million or 7.5% of revenue compared to $61.2 million last quarter, with increased headcount and strong shipments driving higher variable compensation expenses for the period. As a reminder, we continue to benefit from lower COVID-related travel and marketing expenses. Our G&A costs came in at $14.8 million or 1.7% of revenue, consistent with last quarter. Our operating income for the quarter was $335.6 million or 38.3% of revenue. Other income and expense for the quarter was a favorable $3 million, and our effective tax rate was approximately 20.7%. This resulted in net income for the quarter of $268.7 million or 30.6% of revenue. Our diluted share number was 319.7 million shares, resulting in a diluted earnings per share number for the quarter of $0.84, up approximately 35% 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 $136.2 million of our common stock during the first quarter at an average price of $116 per share. As a reminder, we've now repurchased approximately $209 million or 1.8 million shares against our October 2021 $1 billion Board authorization. The actual timing and amount of future repurchases is dependent on market and business conditions, business requirements, stock price, acquisition opportunities, and other factors. Now turning to operating cash performance for the first quarter. We generated $217.1 million of cash from operations in the quarter, reflecting strong earnings performance, combined with continued working capital investments. DSOs came in at 67 days, up from 58 days in Q4, reflecting the linearity of billings and growth in deferred revenue in the period. Inventory turns were consistent with last quarter at 1.7x. Inventory increased to $694.2 million in the quarter, up from $650.1 million in the prior period, primarily reflecting higher component inventory. Our purchase commitment number for the quarter was $4.3 billion, up from $2.8 billion in Q4. The significant increase in commitments largely represents orders for 2023 and beyond, reflecting overall strength and demand for those periods and our expectation that this long lead time supply environment continues. As a reminder, we continue to prioritize newer, early life-cycled products for inclusion of these strategies in order to help mitigate the risk of excess or obsolescence. Our total deferred revenue balance was $1.1 billion, up from $929 million in Q4. 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 $327 million of this balance, up from $160 million last quarter, represents product deferred revenue, largely related to acceptance clauses for new products, most recently with our large Cloud Titan customers. As a reminder, we remain 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 have resulted in increased customer-specific acceptance clauses and higher product deferred revenue amounts. Accounts payable days were 58 days, down from 63 days in Q4, reflecting the timing of inventory receipts and payments. Capital expenditures for the quarter were $14.9 million. Now turning to our outlook for the second quarter and beyond. Our Analyst Day outlook for 2022 calls for 30% year-over-year revenue growth, somewhat balanced across our market sectors and heavily constrained by the supply environment. As we progress through 2022, demand metrics remain strong across the business with particular strength from our Cloud Titan, other cloud, and enterprise customers. While we've added manufacturing capacity and component supply in response to this demand, supplier decommits make forecasting accelerated shipment momentum difficult. These decommits also have a negative impact on gross margin as we must turn to other sources to try to backfill decommitted components in the quarter. From a business model perspective, this means that in this supply-constrained environment, any accelerated growth, while accretive to the bottom line, may come with a lower gross margin percentage due to increased cloud mix and additional expedite fees. Turning specifically to Q2. We expect revenues of approximately $950 million to $1 billion, including approximately $50 million of cloud-related deferred revenue recognition from the balance sheet. On the gross margin front, an expected healthy cloud mix in the quarter, combined with 200 to 300 basis points of assumed expedite costs would result in gross margins of approximately 60% to 62%. As to spending and investments, we expect to continue to grow our investments in R&D and sales and marketing in line with our baseline investment plan. With all of this as a backdrop, our guidance for the second 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 $950 million to $1 billion, gross margin of approximately 60% to 62%, and operating margins of approximately 37% to 38%. Our effective tax rate is expected to be approximately 21% with diluted shares on a post-split basis of approximately 320 million shares. I will now turn the call back to Liz.

Thank you, Ita. We will now move to the Q&A portion of the Arista earnings call. Thank you for your understanding. Operator, take it away.

I have one clarification and then one question. First, Ita, you mentioned customer acceptance clauses. Can you elaborate a bit more on that since you have some deferred revenue and it seems like there is a lot happening? Specifically, could you define or provide more details on that for midyear 2022? My actual question is regarding your suppliers; what do they tell you is the reason for the decommits? Additionally, how many days before the planned delivery date do they typically decommit?

I'll take the deferred revenue question first, right? I mean, obviously, in time periods where you have lots of new products and you're bringing new products to these larger customers, in particular, to completely new customers, they don't have the opportunity to test everything about the product in their environment. And we can't mimic those large-scale environments. So we have customer acceptances where we give them the time period to accept the products, so we can ship, bill, and collect cash on those, but it's deferred from a revenue perspective because there are some criteria that we need to prove that we satisfy so that they can give us the acceptance.

And Sami, to answer your question, the decommits come literally the week we are expecting the components. So they surprise us, right, when we're looking to build them, which is why we struggled, frankly, in the back half of the quarter not getting the components we needed and just having much of our contract manufacturing capacity waiting on key components. And the only way to resolve that was to pay extra expedite fees by orders of magnitude to get them. Sometimes we could get them, and sometimes we couldn't. So we believe this very constrained environment of components combined with decommits is going to continue into Q2 and who knows about Q3. And John McCool, as our Head of Manufacturing and Platforms, you have more to add to that?

Sure. Just the nature of the decommits, I think that we see very part specific reasons for each of those decommits. Some can be tester capacity, yield, logistics issues as the suppliers work through that. So I wouldn't say there's any generic means for those decommits. It really depends on what part and what supplier.

Jayshree, in your last update, you mentioned that you had exceptional visibility like never before, indicating that you could see demand for the entire calendar year 2022. I'm assuming you are already discussing the demand outlook for 2022 with your customers. Can you provide some insights into the dynamics you are observing with them regarding 2022? Also, could you clarify what you mean by having the best visibility ever?

Thank you, Fahad. Last quarter, our visibility was very strong for 2022. This quarter, our visibility continues to be very strong for 2023. But I would also add that, in particular, the Cloud Titans and some enterprises are starting to plan for 2023. They have to start thinking about it now that we're in Q2 here. So all this to suggest that the demand is strong for this year. We expect demand to be strong at least for the first half of next year from a planning purposes. Things could always change, orders could always be moved around. But in all my career across Cisco and perhaps even others I have never seen demand be so clearly purposed and strong. Anshul, do you want to add more to that?

Sure, Jayshree. I think customers are no longer struggling to understand that the supply chain is constrained. They very well understand it. Many of these cloud companies build their own compute and storage as well. So they're planning along with their own teams for 2023 right now. And their businesses are good, which means the underlying demand is strong.

If I could follow up on the deferred revenue from the cloud, was it just one cloud titan customer? Can you help us understand if that's related to your cloud customers' ability to take in your supply? Any insights on the dynamics would be helpful for us to model the Titan revenue going forward.

Yes. No, I think it's across multiple customers. And again, it's very standard for us if you look back historically to have periods at the beginnings of product life cycles where we're shipping the product. They're deploying the product, but we're just waiting for that kind of acceptance so that we can actually take the revenue from a rev rec perspective, right? So it's not unusual for us to see this. You'll have seen that over time. What we said was we built deferred in Q1, and then we will draw down about $50 million in Q2, but we're still up roughly $110 million for the first half from a product deferred revenue perspective.

I have a quick question and a follow-up. Can you provide more details about the vendor commitments and what you're experiencing from your contract manufacturers compared to 90 days ago? Has the recent lockdown in Asia or other areas affected the situation, or is it mainly due to reduced availability and mismatched components affecting complete sets? Also, regarding the expedited costs you mentioned, does the 200 to 300 basis points impact on gross margin include the shift towards a more hyperscaler mix in the Q2 guidance, or is that a separate cost in relation to your gross margins for the remainder of the year?

Yes. I think the 200 to 300 basis points is really looking at kind of an estimate, if you like, of what we think those decommits could cost us. So that's separate from the customer mix, et cetera. It's really driven more by kind of looking at kind of the decommits that we saw end of last quarter, beginning of this quarter, and then what's the impact for that. So it's separate from the customer mix.

Yes. So I would say on the supply chain piece, I think we've kind of seen a more pointed or focused issue really around semiconductors in general. That's still led by the supply-demand imbalance. And I think in terms of particulars on decommits, again, each part, each device has a separate story. We've seen some suppliers that are trying to increase test capacity and don't have test equipment. They're waiting for orders that are also constrained by semiconductors, some perturbation with the lockdowns in China for raw material and equipment. So it's across the board and very specific to each device.

And I think it's safe to also clarify that it changed a lot since our last call. So we have a couple of vendors that are causing a lot of gaps for our decommits. We don't want to name them because I know they're working hard to improve their commitment to us, but 2 or 3 vendors have had significant issues.

Was there any competitive issue in the decommits? Or that wasn't the issue, meaning that maybe there's some allocation issues between yourself and some of your competitors?

No. Not that we know of.

Thank you, Ita, for explaining the deferred revenue release in the guidance. That information is useful. I was examining your deferred revenue over time, and it seems quite high. I am interested in your thoughts on how long it might take to bring it back to a more normal level. While I understand it's difficult to predict, do you think this could occur this year or might it take up to two years? Additionally, we are unclear about the order backlog. If possible, could you provide some details on that?

Yes. Regarding the deferred revenue, I prefer not to make predictions, but we will indicate when we expect to draw it down in our guidance. That refers to the $50 million. At this point, I don't believe we would need to draw down to meet a 30% growth rate. Clearly, this is constantly changing. However, we anticipate that this balance will not decrease year-over-year in our assumptions for achieving 30% growth. As for the order backlog, I'll let Jayshree provide additional insights. For me, given the effect of time on the backlog, it currently isn't a significant figure for us to disclose. It's not a metric that we consider meaningful to share. I’m not sure if you have any further questions about this.

Yes. No Rod, we've stayed away from order strength and backlogs. These don't mean anything unless we can ship it. So we'll continue to keep telling you about our visibility and demand in a qualitative fashion, but in a quantitative fashion, the only number that matters is shipments.

I guess my question is really around, if I heard the purchase commitment number correctly, I think it was $4.3 billion. It was up a fair bit sequentially, and I'm sure the math is not linear on purchase commitments, but could you maybe just help me connect the dot between the $4.3 billion purchase commitment versus what, I think, TTM cost of goods sold is on $1.2 billion? Are you locking in supply on a multiyear basis? Or do you really see a sustainability of this current 30% growth to be a lot more durable versus perhaps that you talked about at the Analyst Day? Just put in context, it seems like a sizable number over here versus the growth rate.

No. Yes. Yes, Amit, good observation. I think we jumped it from $2.8 billion to $4-plus billion. And you're absolutely right, this is a multiyear commitment now. This is not just for '22. It's also for '23 and perhaps it leads into '24 as well. Given the extended lead times that are only getting worse, and we wanted to make sure we secured our commitments. So we're placing a bet on long-term demand, a multiyear double-digit growth and accordingly planning for it. And so you don't need to read any more or any less on to it, except we're bullish about demand and we're planning for multiple years.

I wanted to ask if you could describe the timeline for your sales to a hyperscale data center. Specifically, I'm trying to understand how long it takes from the start of construction for them to begin purchasing from you for initial deployment and upgrades. Based on your experience, how would you distribute the spending for a hyperscale data center over several years? I hope that makes sense.

Yes, it actually does make sense because I think there's a period of planning and build-out and what they want to do, then there's actually putting a design on what they're going to do. And then there's the deployment. Anshul, you are so smack in the middle of this. This is yours to answer.

Sure, absolutely, Jayshree. When we work with our Cloud Titans, they think of regions. And the build-out in different regions, different geos is sometimes different. So you cannot actually just take a regional buildout and say this will always be the same. In many of the major regions, which are hundreds of megawatts or sometimes gigawatts, the DCI network needs to be built first before racks can be added. In some of the smaller or midsized regions, they can actually start with racks and DCI can grow over time as well. But net-net, these are generally about 2- to 3-year planning cycles for the customer. And as equipment or supply is showing up last minute, they decide where they would deploy it in the network. So we don't really control the last part, but the planning is really 2 to 3 years.

And I'm not going to ask a question about deferred revenue or supply. So you'll be happy to hear that. My question is on the enterprise side. You guys continue to do well there. Wondering if your ability to deliver supply faster than some of your competitors has made a difference. Maybe I don't know even know that's true, but if you can comment on that. And also whether subscription software mandates from some of your competitors is helping you win business? And any examples of that would be great.

Yes. No. Thank you, Jason, for stirring us the repeat question. If you step back and look at our enterprise momentum, I'd say it's really picked up steam in the last 3 years. In some cases, I think we're now larger than stand-alone companies with our enterprise business who have been around 25 years. I think the reason our enterprise business is growing so well is really 3 reasons, and probably very little to do with the fact that we can supply products sooner. They all wish we could do it earlier. It has to do with an architectural approach that is different now, especially post-COVID, for our campus and data centers, where they want to build off 1 EOS, 1 cloud vision, and 1 leaf spine architecture. So they're really coming to us for a different design approach, whereas historically, it was rinse and repeat. So that makes a huge difference. The second is their experience with us in the case of existing customers, where they've had such good quality, lower critical vulnerabilities, and experience with us that they really want to take that across the network, client to cloud in many more use cases. And the third, I think, is what you alluded to. There is a lot of fatigue and frustration from our industry peers who've been going one way and only one way and seeing a better alternative, both from a technology and consumption point of view, where we're not forcing them down a subscription and we're giving them options. They can have it as a service or they could buy perpetually, it's the customers' choice. I think all of this has helped cement our enterprise momentum, at least with the high-end enterprise adopters. We've got a long way to go in the mid-market. But certainly, I would say that's the case for the enterprise high-end customers.

I have a couple of related questions. What is the value of purchase commitments in a time when decommits are happening? I am curious about pricing and its effect on the gross margin, particularly in relation to your guidance on gross margin. Would it make sense to adopt a model for the foreseeable future where certain aspects of your pricing are variable and you adjust prices based on your costs? Alternatively, could you implement a system where there’s an empty box on the invoice that gets filled in based on the components you purchase? I believe clients wouldn't be overly surprised if such changes were made.

Yes, we need to recruit you into our procurement department.

I think my billings team might be a little confused.

Good question, Ita. Where to start? Yes. I think on the pricing side, we'll start to benefit from the last price increase really in Q4, mainly in December, maybe a little before that. That price increase is rolling through. Yes, price increases are challenging. We will consider when we see sustained costs that require us to pass them on, and we will do that. However, price increases can impact the customer. We'll evaluate if we need to implement another one based on how lasting these cost increases are and then make a decision. We should start to see some benefits from the previous increase having worked through the backlog in Q4.

Sorry to bring it back to deferred revenue. But just a point of clarification, Ita. You had noted last quarter that the 30% kind of outlook for the year didn't include kind of recognition of any of the deferred revenue balance. But obviously, the deferred revenue balance kind of grew by quite a bit. And so I just want to clarify that the 30% is without kind of where we were at the beginning of the year on deferred revenue or even where we are now or will be at the end of kind of 2022.

Yes. I think, Meta, the commentary around not drawing it down is really an annual statement year-over-year, right? But what you saw us do for Q1 was obviously grow at $167 million. We'll draw down $50 million of that. But I think the best way to think about it as the first half, right? We'll be up $110 million from our product deferred revenue at the end of Q2 based on our guidance, right? So we're building it, and I don't think we draw it down versus the beginning of the year, at least, and we'll see what happens in the second half. It's very hard to forecast it precisely. But again, I think the 30% did not contemplate us picking anything out of the opening year deferred revenue balance.

Okay. Got it. And then maybe just a small follow-up. I mean, obviously, the inventory balance has grown by a fair amount. Just wanted to get a sense of, are there any concerns around obsolescence of any of that inventory? Or just as it takes longer to get all of the parts? Or is that not a concern we should be mindful of?

Yes. The inventory balance itself isn't significantly higher. It has increased slightly with raw materials, but not by much since we are shipping everything we have. The purchase commitment balance has increased substantially, but that is mainly time-bound. We are focused on 2023 and making commitments accordingly. We are making efforts to select the right products, which may not be perfect, but we are approaching this strategically and collaborating with our customers. That’s what we need to focus on right now.

At the risk of asking questions that you either cannot or will not answer, first off, I'm hoping, Jayshree, that you or Anshul or somebody could provide any incremental insight regarding the Microsoft-Meta announcement. And secondly, looking at the numbers, your first quarter results and second quarter guidance, that's almost 35% year-over-year growth. To do a 30% outstanding guidance for '22, that translates to about 26% growth for the second half of the year, and you're talking about the best demand environment you've ever seen. I appreciate we're only 1 quarter into the year, I get it. But it sure sounds like you're, supply permitting, that you could do well north of 30%. I'm hoping to provide some insight on that. And is the growth just a result of backlog or is it all just organic demand?

So to help you answer the Microsoft partnership, you may have seen the quote in our earnings release. We consider Microsoft a very strategic and preferred partner, and so do they of us. The use cases are expanding. And of course, that, like I've always said, doesn't mean we get 100% of the use cases for the business. They've always had multi-vendor and open. And from time to time, they've chosen other partners for other use cases. So it doesn't change at all our status with them, but obviously means they're going to always be multi-vendor and open. Anshul, do you want to add to that?

Sure. Paul, our relationship with Microsoft is so long and so deep. We worked with them not just for our current product, but for several generations to come. And the discussions on paper span all the way up to 2025-2027 architecture, what can be possible, what we build for them. So that leadership position we've had will continue. As Jayshree mentioned, they can be multi-stores, but we believe we'll stay in a healthy position, and we're not going to get distracted by this announcement. Ita?

Yes. And then, Paul, I guess, coming back to the growth, I mean, yes, you're right, it was close to 35% for the first half. I think we would have an outlook of 30% for the second half. So I'm not necessarily saying we're going to decelerate off of 30%, but it's all about supply, right? If we could solve for these handful of components that are causing these decommits, yes, we could do some more, right? But for now, I think we just have to respect that and the uncertainty of that and be cautious.

Can I just ask a quick clarification?

Yes.

Had a lot of folks on the call, it sounds like people think you're just pulling out a deferred to make these numbers. How much of this is just satisfying backlog? How much of it is also shared in? What's going on in terms of driving the demand?

I don't think we are actually up $110 million in our guidance for the first half, right? So I don't think anyone believes we're using deferred revenue to inflate the numbers, right?

It's organic demand, absolutely. We have backlog. We don't talk about that. And we have an increase in deferred, all 3 or 2. We just have to ship.

I guess if I can just ask you 2 quick ones on the enterprise vertical here. Jayshree, you mentioned the strong momentum you have with enterprise customers. I mean, one of the concerns we've been hearing from investors is about the current macro and how enterprises respond to that in terms of any weakness in the order trends. Maybe if you can clarify if you are seeing any of that in your discussions with these customers? And also if you can give us an update on the campus revenue, which you were looking to double. Sounds from the momentum that you should be sort of on track to better, relative to your expectations. But if you can just give us an update there?

Yes, Samik. We are seeing no change in the enterprise momentum. It's strong. We haven't seen any slowdown, maybe it will later on. You have the impact of inflation trends, who knows? But at the moment, things are looking very strong. Chris Smith, Ashwin and the whole team are just knocking at the door always for products. They're certainly creating the demand with their customers. Our goal is to double this year, and we've only had 1 quarter, but 1 quarter doesn't make a trend, but this 1 quarter alone is showing a trend in that direction. And as I think you asked me on Analyst Day, I shared with you that we will grow at least to 750 million by 2025. I think we can achieve that perhaps if we are in a less constrained environment, we could beat it, too.

Pierre is in a noisy environment. So this is Antoine asking one on behalf of Pierre’s questions. So could you please share your thoughts on NVIDIA's Spectrum 4? How you see this shift sitting in the competitive landscape? And how you're seeing VCR possibly becoming a competitor? And maybe when do you expect to have 51.2 terabits per second chips in your products?

Well, we've always viewed NVIDIA as a partner, especially for the DPUs and NICs when they were Mellanox, and we continue to view that. There are times when companies choose a vertical stack. And I think NVIDIA's focus on Spectrum 4 is more as a vertical stack for their captive customer opportunities. As a horizontal best-of-breed, we don't see them as a competitor at all. And Arista is poised to be best of breed and continues to be the preferred choice with customers. Anshul, you want to add to the 51 terabit road map?

So Pierre, let me just remind you that the 25.60 is just now ramping. So we're very happy with that. And core development projects we take on with our customers are well along as well. But 51 is really next gen. It's at least 2 years out, maybe more for many of the high-volume customers. And the announcement gain doesn't really matter, right? Just because someone announced doesn't mean others are not working on the chip either. We will be ready in time.

A lot of my questions have been asked and answered, but I wanted to go back to kind of some of the architectural stuff that you've talked about, Jayshree, in the past. AI Fabric, you continue to bring this up on conference calls these last couple of quarters. We're seeing obviously some big large deployments at one of your large Cloud Titan customers. I'm just curious if you can offer up any other thoughts around the size of this incremental opportunity, the trajectory of what you're seeing? If you're seeing it become more broader based? Just any context around that opportunity for Arista.

Yes, Aaron, I mention this because I believe it's a strategic innovation similar to how Arista led in cloud networking with the leaf-spine architecture, which became prominent with a front-end network built on that framework. We observe that the back-end network is evolving, primarily focusing on interconnect bus-based InfiniBand and high-performance clusters, often referred to as HBC. The new AI workloads are indeed data and compute-intensive, and cannot rely solely on bus or IO-based methods or just emphasize latency. They are challenging us to push the boundaries of Ethernet to manage predictable latency and the ability to scale the entire network, among other aspects. We're in the early stages of this development. It’s accurate to say this trend is beginning with the initial deployments among cloud customers, much like how cloud technology emerged five years ago. However, I anticipate it will also reach various specialty clouds and large enterprises. This progression will unfold over the next three to five years; it won’t happen instantly.

I'm going to ask one clarification and one question. For the clarification, on one hand, you say the supply chain is worsening. On the other hand, you're providing very strong guidance for the next quarter. How do you reconcile the worsening conditions with such strong guidance? For the second question, one of the top concerns I'm hearing from investors is about Microsoft and Facebook. We know they are strong and are heavily investing in their data centers right now. How much exposure or dependency do you have on this customer concentration? Any information you can provide on that would be helpful.

Right. So supply chain is getting worse because of the extended lead times, not just because of that because we were planning for that, but because of the sudden surprising decommit. So when they're trying to build a product and ship it out and suddenly we don't have these last 2 components, it just freezes the whole supply chain and our ability to commit to revenue. So as Ita said, what do you do at that point? John and the team had to go scaring the face of the earth to get parts that would normally cost X that are now 100X in many cases. And that's a very stressful thing because sometimes, you get them and sometimes, we don't. So our ability in Q1 to shift more was very much there, but a constraint in Q1 to ship more was also problematic because we couldn't get the parts. And that is the story for Q2 and perhaps will go into Q3 as well. Now how does that affect us? We may be executing better than others, but it's affecting us in that because of the elevated cost of these components and expedite, it's showing up as gross margin. So we have a gross margin pressure for the next couple of quarters, both due to the cloud mix, which was your second question, and the commitments from Microsoft and Meta and other cloud titans as well as these expedites that are adding double pressure on our gross margin. So that's what we wanted to take away. We're going to execute as best as we can. Customers come first. We're going to do our best there, even if it means buying these components at very, very elevated costs.

These are two excellent customers to have, and we are pleased with our relationship with them. We don’t control the market; they are among the largest cloud companies globally, and we lead in cloud networking. While we are exposed to them, these investments are highly leveraged in various areas such as product development and creating a roadmap, as well as achieving economies of scale in our product line and manufacturing. This results in numerous benefits for both us and our customers. We are glad that they are performing well, growing, and increasing their capital expenditures, which also benefits us. It was mentioned previously that both companies are expected to be 10% customers this year, and we are satisfied with this outcome.

Congratulations, Jayshree, either to your teams and Anshul. I have one question, and that is about the lag time between your pricing actions and the orders. So Ita mentioned a few times about the December price increases. But I just wanted to see, wouldn't it be logical that all customers are putting in a lot more orders now to give you more visibility? Because selfishly and rightfully and smartfully and economically, it would be better pricing knowing that prices are going to go up in the future? Or Ita and Jayshree, you're saying that you've adjusted prices since that December price increase? I just wanted to get some color on the lagged pricing orders.

Thank you, Jim, for your congratulations and well wishes. We implemented a pricing increase that we discussed in November. The immediate impact of the pricing is challenging to gauge. We typically provide advance notice, so existing orders are not influenced by the new prices. Orders already in progress and our backlog remain unaffected by the pricing changes. Therefore, while we are securing new orders at the updated prices, most of what will ship in the first three quarters and a significant portion of the fourth quarter will still reflect the old pricing. Our current orders will display the new prices, which will take effect in late Q4 or 2023. We are considering a second price increase due to the considerable pressure we face on costs. However, implementing it now would not show immediate effects. Does that clarify your question?

It does. And my point is, compared to November, things have really changed. It's been 6, 7 months since now. So I guess it sounds like you're kind of contemplating, but I just was wondering if it was more dynamic since last November with your pricing.

Jayshree and Ita, I was hoping you could touch on your thoughts around the durability of the orders that you're seeing right now? Jayshree, you commented a little bit on hyperscale. But any thoughts around how far in advance you're seeing orders from cloud titan and enterprise customers? And then also, I'd be interested in your thoughts on the type of financial commitments you're seeing from the different verticals and how you're monitoring and managing with risk or perceived risk of excessive bookings or pull forward? That's it.

You're inquiring about the durability and authenticity of our orders, correct? Firstly, while we fulfill through channels, many of our orders stem from close relationships with our customers. Regarding your second question, we truly believe that our orders are not double booked; there may be a few instances, but the vast majority of our cloud orders—be it from enterprise or service providers—result from ongoing dialogues and relationships. Therefore, we have no reason to think there is widespread double booking at this time. There might be a small percentage, but not significant. As for durability, our customers are typically planning on a 1- to 2-year horizon, and we consider the durability of our orders during this period to be robust. Naturally, there is always a risk of cancellations and adjustments, but generally, our customers have remained committed and consistent in wanting our products and are willing to wait for them. So, the durability and authenticity of our orders appear to be strong.

What are you informing your customers about the lead times for your longer lead time products? How does this compare with what your customers are saying about the lead times of your competitors? Additionally, Ita, could you explain the 200 to 300 basis point impact, presumably for Q2 and Q3? Do you believe this impact will begin to ease after Q3?

Yes. I mean, look, it's tough, right? I think we're probably with that at least Q2, Q3, and then we should start getting some relief from the pricing and other things in Q4. But I think I'd hold that through Q2 and Q3.

Yes. And on lead times, it really varies by product and it varies by decommits right now. So we thought we were doing super well, and we were – I was king of the jungle and on top of the line if you asked us this last November. But I think things have degraded for all our peers and for us. So lead times are definitely measured in many weeks and many months.

Do you strive to have shorter lead times than your competitors?

We strive to execute better, and I think we have done better than our competitors, and we hope we continue to do so. What we don't do is promise one lead time and then come up with another, at least not intentionally.

Do you have customers leaving your competitors for you because your lead times are shorter?

They do try. I have had a number of enterprise customers come to us and say, 'I got all of this I can give to you if you can ship now.' But again, we're not able to ship now either. So much as they try, the best we can do is face some kind of use cases for them in different products. But it isn't the case of direct substitution. It's a case of switching from one vendor to another and still having a plan across a period of time.

I guess I had another question on purchase commitments. The $4.3 billion is a big number, an impressive number. And certainly, in this environment, it's very understandable. I guess if I play the other side of this, how do you see the risk of getting caught with lots of high-cost componentry in the case where the supply chain ultimately corrects and prices normalize? Is that something you guys think about? Is it a risk in your mind? Or just simply worth it in terms of having more opportunity to gain share right now?

I think it's simply worth it, George. I've seen a few of these in my career. And the first thing you do here is you really procure your newest products, your components that are least likely to get obsolete. And I think we've been very smart and sensible about it. Second thing is don't confuse purchase commitments with purchase arrivals. They've not arrived yet. They're going to take multiple quarters or years to arrive. So think of this as a multiyear purchase commitment that could arrive in '22, '23 or, in some cases, '24. And the third thing is we look at this as a wise investment for a lot of common components that will be in our new products as well. So all in all, no regrets. There may be some perturbation on some components that arrive and don't arrive, but we feel good about this being one of our best investments for the short and long term.

Got it. And is there any safety valve or net for you guys in terms of the ability to push out those deliveries or cancel orders? How do you think about that?

Regarding pushout, we'd rather they don't. We don't want that safety valve at the moment. So most of the orders in the semiconductor industry are generally noncancelable, but they can be managed from a time basis point of view.

Thanks for squeezing me in here given most of questions have been asked and answered, but I wanted to actually circle back to Suva's question on pricing, not so much on the magnitude of the price side, but what are you seeing with customers across the verticals of what you're implementing that can give confidence that you're not getting a pull-in of orders of potentially more pass-through being needed, as these are smart buyers that can see the supply chain is likely getting worse, if you're seeing it too. And then additionally, why not implement noncancelable terms like others in the space have?

Yes, those are both good questions, James. I'll address the first one. Noncancelable orders are more complex than they sound because they rely on contractual agreements. We will consider implementing them where possible, but generally, we tend to operate under long-term contracts. What was your other question?

Pricing.

Right. Pricing and one, the impact of pricing? Or what was the question again?

Yes. It was more about the factors that give you confidence regarding the absence of order pull-ins as we take the next step up.

I think the key factor is not about pulling in orders, but rather better planning. They are clearly evaluating their purchases. With lead times being six weeks, they can focus on this year without concern for next year. However, they now need to factor in their budgets and plans for both this year and the next. Therefore, I believe you are witnessing demand that spans not just this year but also, as Anshul often notes, potentially additional quarters. From this perspective, they are planning with a longer-term outlook.

My question is about software. The business line experienced a slowdown in growth to 15% from over 30% last quarter, which was a strong quarter. This could be due to timing. I would like to explore further what percentage of this line, which I consider important, consists of subscription and ratable software compared to services. Additionally, I would like to know what the largest software solution is and what the highest percentage of software solutions are.

It's certainly a work in progress, and we recognize the need for improvement. I categorize our software into three main areas. First, we have the perpetual licenses, which are crucial for our products, and we're performing well in that area. These include routing licenses, automation licenses, analytics, and so on. Next, there's the subscription revenue, which, as I've mentioned before, doesn't simply transform our existing business into subscriptions; these are typically new offerings such as DANZ Monitoring Fabric, observability solutions, AVA sensors for threat detection, and CloudVision for network-as-a-service. While these new additions are performing well, their revenue still lags behind bookings because they involve multiyear subscriptions and remain relatively small for us. Together, these two categories represent about 10% of our business, and we aim to double this in the coming years. Finally, there’s the services category. When our product performance is strong, the service revenue percentages can decrease, as seen in this quarter. Typically, Q4 is our peak period for service renewals, usually seeing figures in the mid to high teens. Collectively, these three segments form our recurring software components, which are vital, but they can seem less significant when our product performance is exceptionally strong, as it is this year.

That's very helpful. Just one last quick question. Regarding the deferred revenue delays that Ita mentioned, you talked about new products causing these delays. This has occurred in the past, and there has been a significant increase. Did I hear you correctly mention that it was a new cloud customer? There aren't many in the cloud titans category. Is that accurate?

No, it's not related to the cloud. We're mainly focusing on our existing customers, exploring new use cases, and launching new products. Additionally, on the enterprise side, we are onboarding new customers where we are implementing our solutions for the first time.

We have time for 1 last question.

Just a clarification on the second half outlook. Is that dependent on the supply chain getting a little bit better from where it is right now? Or does that assume that there's no changes to the decommit environment?

Yes. I mean I think, look, we're going to take this quarter by quarter here. I think our reference to the full year, et cetera, is assuming a continued constrained environment. I don't think we think the world changes that much, right? But we'll take it quarter-by-quarter.

Thank you, again. This concludes the Arista Networks First Quarter 2022 Earnings Call. We have posted a presentation that provides additional information on our results, which you can access on the Investors section of our website. Thank you for joining us today, and thank you for your interest in Arista.