Welcome to the second quarter 2021 Arista Networks financial results earnings conference call. During the call, all participants will be in a listen-only mode. After the presentation, we will conduct a question and answer session. Instructions will be provided at that time. If at any time during the conference you need to reach an operator, please press the star zero. 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. I will now turn the call over to Mr. Charles Yager, Director of Product and Investor Advocacy. Sir, 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 second quarter ending June 30th, 2021. If you'd 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 third quarter of the 2021 fiscal year, longer-term financial outlooks for 2021 and beyond, our total addressable market and strategy for addressing these market opportunities, the potential impact of COVID-19 on our business, product innovation, and the benefits of acquisition, 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.
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, Charles. Thank you, everyone, for joining us this afternoon for our second quarter 2021 earnings call. I hope you're all being safe and vaccinated in these post-pandemic times, especially with the resurgence of the Delta variant. I would also like to take this opportunity to warmly welcome Liz Stine, our new Director of Investor Relations Advocacy, working closely with Charles. Liz is a long-time Aristan with deep networking expertise, and most recently ran our South Central region as the Systems Engineering Manager. Welcome, Liz. Back to Q2 2021 specifics. We delivered revenues of $707.3 million for the quarter, with a record non-GAAP earnings per share of $2.72. A-Care services and software renewals contributed approximately 22% of revenue. Our non-GAAP gross margin of 65.2% was influenced by an enterprise momentum and software and services contributions.
We are pleased with the healthy customer traction, including new customer logos and record million-dollar customers in the mainstream enterprise. In Q2 2021, cloud titans were our largest vertical. The enterprise was a close second, followed by financials and specialty cloud providers tied at third place, and service provider at fourth place. International contribution was strong at 27%, with the Americas at 73% for the quarter. We surpassed a cumulative of 50 million cloud networking port shipments this quarter, and we consider this to be a key golden milestone. In light of the industry-wide supply chain shortages and escalating cost of components, rate, and expedite logistics, I would like to invite John McCool, our Senior Vice President and Chief Platform Officer, to shed some more light on our manufacturing execution. Welcome, John.
Thanks, Jayshree. The continued industry-wide impact of COVID on global supply chain output, combined with increase in demand for electronics across all segments, is expected to remain for the foreseeable future. Component lead times are the highest we've seen and have roughly doubled from pre-pandemic norms. Most notable are semiconductor lead times, which have extended in the range of 40-60 weeks. Factories are operating near full capacity, limiting flexibility for changes in demand. We expect extended lead times and escalating product costs due to expedites and elevated component increases in 2021 and 2022. To mitigate these headwinds, we've taken a number of steps in Arista manufacturing. First, we improved manufacturing procedures to maximize capacity and material utilization. We are increasing our purchase commitments for 2022 forecasts to adjust for increased component lead times. We placed additional emphasis on inventory for our new products to offset supply constraints.
Finally, we are working closely with our strategic suppliers to plan for capacity expansion programs. Clearly, we are redoubling our efforts and execution in this challenging macro environment and look forward to supply chain improvements in the second half of 2022 and beyond. Back to you, Jayshree.
Thanks, John. We really appreciate the diligent and disciplined work that you and the entire manufacturing team have stepped up to. We welcome Susan Hayes, your newest Vice President of Manufacturing, as a strong addition. We also thank our customers for their patience and understanding during our lead time constraints, and will strive to keep doing better as we recover in second half 2022. Our enterprise customer momentum has never been stronger. Continuing on our theme of enterprise wins, I would like to share with you three examples of strength and success. The first win was in the retail sector for both data center and campus. We began with the data center win with Arista's hallmark EOS for DANZ, or data analysis switching and routing. We expanded in 2021 with cognitive campuses for both Power over Ethernet wired switches and our wireless, providing a natural expansion into these use cases.
Retail markets cannot tolerate downtimes with the magnitude of IoT proliferation they have. Arista was able to perform real-time upgrades without downtime across 100+ stores and warehouses. In the absence of retail remote staff and hands, we drove automation across all these stores with open APIs and CloudVision. Our second win was a major U.S. financial for both data center and routing. Arista continues to expand its enterprise routing use cases, not just supporting with data center with Arista EOS features, but also a more software-led, simplified, automated deployment of core routing in the spine using standard-based routing protocols with VXLAN, EVPN, and Multicast. Our customer was able to rapidly migrate from legacy to Arista within a few months, enabling billions of transactions.
An international media and entertainment campus win included a customer who wanted an extension of their data center in the campus with simple operating systems, easy to scale, common spine deployment using CloudVision. This customer took a build-as-you-go approach for flexibility and visibility. This also included device access for back-to-work applications with our Cognitive Wi-Fi. In all of these three examples, enterprise customers are hungry for an alternative, and Arista was chosen as the disruptor with superior product capability and a cohesive client-to-cloud strategy to unify siloed data sets consistently. Arista's innovations, combined with high quality and support, is becoming the gold standard for customers to build cognitive cloud networking. According to industry experts, Arista continues to gain switching market share across large enterprises and providers. We are proud to be the number one market leader in 100 G Ethernet ports for the fifth consecutive year.
We see 2021 as the first year of inflection for higher speeds, ranging from 100 to 200 to 400 G after 18 months of trials. We have now shipped more than 2.5 million ports of high-performance ports in the first half of 2021, according to analysts, placing us also at number one leadership in the combined 100G , 200 G, and 400 G Ethernet high-performance cloud switching. In summary, Arista is well positioned for the next phase of our growth in cloud and data-driven networking. We do this with proactive platforms, predictive operations, and a prescriptive experience. We believe we are poised to achieve increasing market share with greater business diversification.
While the path forward is paved with supply chain obstacles, volatile customer demand, and your normal typical competitive tactics, I believe Arista, which means to be great in Greek, will live up to its name and with our A game, our A customers, and our A team. With this, I'll pass it over to Ita, our Chief Financial Officer, for financial specifics.
Thanks, Jayshree, and good afternoon. This analysis of our Q2 results and our guidance for Q3 is based on non-GAAP and excludes all non-cash stock-based compensation impacts, certain acquisition-related charges, and other non-recurring items. A full reconciliation of our selected GAAP to non-GAAP results is provided in our earnings release. Total revenues in Q2 were $707.3 million, up 30.8% year-over-year, and well above the upper end of our guidance of $675 million-$695 million. Shipments remained somewhat 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 22.3% of revenues in the second quarter, up from 21.4% in Q1. International revenues for the quarter came in at $193.2 million, or 27% of revenue, up from 25% in the first quarter.
This shift in geographical mix on a quarter-over-quarter basis reflected strong international deployments by our cloud titan and specialty cloud customers, combined with healthy performance from our in-region businesses. Overall gross margin in Q2 was 65.2%, above the upper end of our guidance range of approximately 63%-65%. While we recognized some incremental supply chain costs in the period, these were more than offset by a healthy mix of enterprise and software revenues for the quarter. Operating expenses for the quarter were $189.8 million, or 26.8% of revenue, up from last quarter at $180.9 million. R&D spending came in at $119.6 million, or 16.9% of revenue, up from last quarter at $110 million. This reflected increased employee-related costs and higher new product introduction spending in the period.
Sales and marketing expense was $57.9 million, or 8.2% of revenue, down from $59.5 million last quarter, with lower demo-related 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 $12.3 million, or 1.7% of revenue, consistent with last quarter. Our operating income for the quarter was $271.7 million, or 38.4% of revenue. Other income and expense for the quarter was a favorable $1.7 million, and our effective tax rate was approximately 20.7%, reflecting an improved geographical mix. Other income and expenses for the quarter included approximately $2 million of interest income, offset by some unfavorable FX amounts. This resulted in net income for the quarter of $216.8 million or 30.6% of revenue.
Our diluted share number was 79.71 million shares, resulting in a diluted earnings per share number for the quarter of $2.72, up approximately 29% from the prior year. Now turning to the balance sheet. Cash, cash equivalents, and investments ended the quarter at approximately $3.3 billion. We did not repurchase shares of our common stock during the second quarter. As a recap, we have now repurchased $763 million or 3.6 million shares against our board authorization to repurchase $1 billion worth of shares over three years commencing in Q2 2019. We will continue to execute opportunistically against the remaining mandate. Turning to operational cash performance for the second quarter. We generated $263 million of cash from operations in the period, reflecting solid net income performance and continued investments in inventory and supply chain.
DSOs came in at 47 days, down from 51 days in Q1, reflecting the linearity of billings in the period. Inventory turns were 1.7x , down slightly from last quarter at 1.8. Inventory increased to $543.2 million in the quarter, up from $483.2 million in the prior period, as we continue to buffer certain components and products. Our total deferred revenue balance was $746 million, up from $720 million in Q1. The majority of the deferred revenue balance is services related and is directly linked to the timing and term of service renewals, which can vary on a quarter-by-quarter basis. Approximately $90 million of the balance, up from $70 million last quarter, represents product deferred revenue, largely related to acceptance clauses for new products across various customers and sectors.
As a reminder, we expect 2021 to be a year 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, may result in increased customer-specific acceptance clauses and increased volatility in our product deferred revenue amounts. Accounts payable days were 53.7 days, up from 52.3 days in Q1, reflecting the timing of inventory receipts and payments. Capital expenditures for the quarter were $4.5 million. Now turning to the outlook for the third quarter and beyond. We reported strong year-over-year revenue growth of approximately 29% for the first half of 2021, reflecting healthy demand across all our market sectors, combined with favorable comparisons from the first half of 2020.
While we expect continued strength in demand as we move through the second half, we will likely see some deceleration in year-over-year revenue growth given the top-line recovery experienced in the back half of 2020. Turning to gross margin. Industry supply constraints continue to pressure component costs. Some of these incremental costs will initially be recorded as inventory and only be recognized on the income statement when the products are sold in future periods. With this as context, we will continue to reiterate our gross margin outlook of 63%-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.
With regard to cash flow, we expect to fund approximately $40 million of CapEx in the third quarter for the purchase of land to build a data center and engineering location in Santa Clara. We will provide more details on this project over the coming quarters. Our outlook discussed above and our guidance for Q3 reflects our current understanding of COVID-19 and its impact to our business and supply chain. This remains an inherently uncertain situation, and we will need to continue to monitor and attempt to mitigate new challenges as the situation unfolds. With all of this as a backdrop, our guidance for the third quarter, which is based on non-GAAP results and excludes any non-cash stock-based compensation impacts and other non-recurring items, is as follows. Revenues of approximately $725 million-$745 million, gross margin of 63%-65%, operating margin of approximately 37%.
Our effective tax rate is expected to be approximately 21.5%, with diluted shares of approximately 80 million shares. I will now turn the call back to Charles. Charles?
Thank you, Ita. We are now going to move to the Q&A portion of the Arista earnings call. Due to time constraints, I'd like to request that everyone please limit themselves to a single question. Thank you for your understanding. Operator, take it away.
We will now begin the Q&A portion of the Arista earnings call. In order to ask a question during this time, simply press star, then the number one on your telephone keypad. If you'd like to withdraw your question, press the pound key. We ask that you pick up your handset before asking questions in order to ensure optimal sound quality. Your first question comes from the line of Meta Marshall with Morgan Stanley.
Great, thanks. I've gotten a couple of questions on this today, but just if you could give a sense of if the supply chain constraints are worse in any portion of the portfolio, and then maybe as it relates to that, where the inventory is. I guess just trying to get a sense of, is it on the high-speed products or more of the campus service provider portfolio? Thanks.
Thanks, Meta. I think just about every component is affected in our supply chain. I'll let John or Ita comment. We're affected on chips, memory, copper, passive components, freight, logistics, expedite fees. I don't know if I can pinpoint. It affects all our products. The lead times vary, as you heard, they've all doubled. We've been experiencing anywhere from 20 weeks to 60 weeks or 40 weeks to 60 weeks, like you said, John, right?
Yes.
Depending on the part, we are experiencing component levels of increase across the board. Campus, routing, switching, data center, you name it, and they're at the component level. We're going to try and absorb as much of it and offset as much of it as we can, and not pass it on to our customers if we can help it, except in modest levels. I don't think it's anything more than across the board.
Okay. Great. Thank you.
Thanks, Meta.
Your next question comes from the line of David Vogt with UBS.
Great. Thank you for taking my question. A competitor just recently noted, as you guys are talking about, that they're seeing some orders placed a little bit earlier suggesting that there's been a little bit of a pull forward given sort of the industry constraints. Can you guys just explain a little bit how you're thinking about visibility relative to your order book and backlog given the strength this quarter and sort of supply constraints? Does this lead to better visibility, obviously, over the next, let's call it two to three months? What does that mean for the fourth quarter without getting into specific guidance? Thank you.
Sure, David. Well, as you know, we've always had limited visibility, the last few quarters have noted that our visibility has gone up due to our lead times. I think there's a direct proportion to long lead times, slightly longer visibility. Particularly with the cloud titans that Anshu works closely with, we have been able to get visibility beyond the one to two quarters that we normally get, we do have visibility into 2022. I think it's directly tied to them planning better and realizing that with longer lead times, they need to know what they're going to do in 2022 for us to supply product. Anshu, you want to add some more to that?
Sure. Jayshree, I would say we're not seeing order pull-ins or if customers are doing it, they won't tell us. It's much more the non-binding demand signals we get is where the discussions happen with customers. That's how we get our visibility from them.
It's just prudent planning, David, from a customer.
Great. No, that's helpful. I appreciate it. Thank you very much.
Thank you.
Your next question comes from Amit Daryanani with Evercore ISI.
Good afternoon. Thanks for taking my question. I guess my question's really around the 400G opportunity. I think in the last couple of years, you've spent a lot of time explaining to folks how your software stack is really differentiated against white box risk. I would love to get your perspective as the underlying technology gets more complicated, with 400G and 800G, what's the potential for cloud titans that have historically relied more on white box to start talking to Arista? I would love to know if you're seeing a shift in customers that have skewed more white box and always build their own product now looking to perhaps buy.
Sure. Amit, the companies that build their own white boxes are fairly sophisticated, so they're not short on talent. If they are committed to a mission, they'll go ahead with it. I don't believe if any changes happen in the industry, it'll be simply because the next gen is harder to build. It's much more about the collaboration we have with them and the co-development partnerships we have there. We've talked about this in the past. Nothing has changed on that front. We are moving along on our mission in the few situations where customers consider supplying from the outside, and that will take a few years to actually materialize. No big change here expected.
Amit, I think the way to think of this is if you want commodity 100G without our software stack and you just want to buy basic vanilla stuff, you'll continue. If you really appreciate our collaboration, our engineering, our innovation, and our software stack, that's when Arista gets chosen.
Perfect. Thank you.
Thanks, Amit.
Your next question comes from the line of Aaron Rakers with Wells Fargo.
Yeah. Thanks for taking the question and congratulations on the quarter. I wanted to actually ask about the progression of the subscription business within the model. As we look at the services line, is there any way or how we investors should start to think about that as being an increasingly visible growth driver for the company? Thank you.
Thanks, Aaron. Thank you for the wishes. Wish we could ship more. I think the real issue to think of subscription as a long-term indicator of not necessarily just revenue, but the stickiness of our business, right? When you look at CloudVision, CloudEOS, what we're doing with network detection and response with AI-driven security or Big Switch acquisition with DANZ Monitoring Fabric. These are all, if you will, layered icing on the cake, and the icing contributes good margin and obviously has a long-term one-year, two-year, or three-year subscription. The revenue shows up a little bit later, but the bookings are very strong, as you can imagine, from that. We'll continue to see strong software service renewals as well from our A-Care. The two together, we believe, we've always said will be a 20%-25% contributor to the business.
Your next question comes from line of Samik Chatterjee with JP Morgan.
Hi. Good afternoon. Thanks for taking my question. Jayshree, you did comment, starting off that you're seeing strong demand from enterprise customers. Wondering if you can just talk about what you're seeing from the other customer verticals, like is demand accelerating? Similarly, for example, the cloud titans, as well as some of the telcos. Keeping that backdrop in mind, does look like the third quarter guidance is for like a 4% increase when traditionally we've seen like a high single here. Should I just infer that's all supply Any moderation relative to historical trends?
Well, first of all, thank you. Historically, I'd say, I think all our five verticals are doing well. In that, I would highlight Cloud Titans and Enterprise as stronger, but that doesn't take away from the contribution and success we're having with specialty cloud, financials, or the service providers. I guess it's a case of rising tides raise everything. All of them are contributing well. Relative to our Q3, demand is strong. I wish we could ship more. I wouldn't compare it necessarily to our last Q3 in the post-pandemic era. I would just say on the basis of large numbers, we're doing well.
Yeah, I think if you look at last year, given the year that it was, we saw a very significant uptick in Q3. It was up 12% or something quarter-over-quarter. That was just more of a semblance of what was happening with COVID at the time.
Right. Previous years weren't like that.
Yeah. I think the Q3 guides 21% at the midpoint year-over-year. That's a good place to start, right? Yep.
Thank you.
Thanks, Amit.
Next question. Your next question comes from line of Fahad Najam with MKM Partners.
Thank you for taking my question. In terms of the pricing environment, can you comment on, are you planning on passing on any of the incremental costs that you're experiencing? Can you also talk about the dynamic you're seeing in the market in terms of pricing? Are your competitors increasing their pricing as well? Are they using this as an opportunity to absorb the cost and maybe aggressively price? Maybe a comment there. We appreciate it.
Well, Fahad, I'm not in a position to talk about competitors. Maybe they'll share that with you. They certainly don't with us. I will say that we're going to try our best to absorb the cost. On selective models, we will have to, where the increases are significant, increase prices slightly. We don't expect the impact on that on our backlog or existing inventory. The real impact of any changes we make will impact our gross margin or our price changes next year.
Appreciate the response.
Thank you.
Your next question comes from the line of Jeff Kvaal with Wolfe Research.
Yes. Thanks very much. I am wondering, you started to talk a little bit about deceleration in the back half of the year, which makes complete sense. I'm wondering if there are any one-time tailwinds that you are expecting this year that we should be considering when we start to think about the shape of 2022. Obviously, you all don't want to talk too much about it, but anything that we should be considering as we have to think about it?
Yeah, I think it's a little early to start getting too specific about 2022. This year did have an unusual slope to it just because of the pandemic and how 2020 played out, right? We're putting up some very good numbers now this year, that's obviously setting a good base for us to grow up next year. I'd bear that in mind. The business is solid. There's strong demand. I think we're executing well. It's broader, right? It's across the verticals. We'll probably take a shot at giving you some bookends in next quarter for 2022. It's just a little too early yet.
Okay. Thank you, Ita.
Your next question comes.
Thanks, Jeff. Nice job, Anshul, Chris, and the whole team for creating demand.
Sure. Thanks, Jayshree.
Your next question comes from line of Paul Silverstein with Cowen.
Thanks. A clarification question. Just did I hear you or Ita say that you're expecting 20% growth now for top line for the year?
No. You did not hear us say that.
Ita would never say those things, Paul. Ita Brennan, what did you say?
We talked about the midpoint of the guidance for Q3 year-over-year growth. Right?
Well, let me ask the question, because last quarter, I think you responded in response to a question, you said for this year, you're definitely expecting 15% growth for the year. That was in the Q&A. Any thoughts for what you're expecting now for the year, and can you also comment on how much the supply chain is costing you in revenue and in margin structure?
I think for the year now, when you layer in Q3 and you think about Q4, you can probably get to a reasonable view of the year, right? I don't know that we want to put a specific number on it, but you can get there, right? Once we give you the Q3 number, it becomes a lot easier. In terms of how much, clearly lead times are very extended, right? It's always hard to compare that to some kind of normal world. We could obviously do more revenue if lead times weren't that extended. How much, it's really hard to put a number on. I don't think we're going to try to put a number on that.
All right. Given that response, can I ask you within cloud titans, since historically you've been highly concentrated in Microsoft and Facebook, but of more recent vintage, you appear to be having some more success with some of the other folks, I think Google in particular? Any insight you could share with us?
I think, Paul, you summarized it very well. Anshul and the team are doing a great job in both diversification across all our sectors, and while Microsoft and Facebook are very strategic, very important customers, we have other cloud titan customers also.
Next question, please.
Your next question comes the line of Sami Badri with Credit Suisse.
Hi. Thank you. A little just a clarification on just Ita's commentary regarding the second half deceleration. Now, Ita, I was hoping you'd give us some sequential guidance. Is there any potential that sequential growth could be negative in 3Q and 4Q? If I just extrapolate out and I forecast a little bit out from where we are today, and even out above, we're getting into the low 20% range, and I just want to.
That clarify if there would be any sequential decline in 2021.
Yeah. I mean, we've given you Q3, right, at the midpoint. There's clearly growth there quarter-over-quarter. Again, we're not trying to guide Q4. Q4 is usually a good strong quarter. I'll leave you to draw your own conclusions from that. I think, we've pretty much laid it out for the year.
Sami, I think the message we'd really like to convey is demand is strong. We're doing well. We need to execute on our shipments. We can only approach our shipments 1 quarter at a time because our supply chain is so constrained.
Got it. Thank you.
Thank you.
Your next question comes from the line of Rod Hall with Goldman Sachs.
Hi. Thanks for the question. I guess I wanted to ask you about the gross margin guidance. When we were talking to Cisco on a similar note, they had said that the supplies impact on their gross margins was bottoming, or on their margins was bottoming in their guide. I'm curious whether this guide for September, from your point of view, is kind of the bottom. In other words, gross margins at least go sideways from there? Do you think it's sort of an unknown at this point, and things could continue to worsen? That's my first question, I have a follow-up.
Yeah, I think, Rod, on the supply chain stuff, we did have some significant supply chain impacts even in Q2, right? It just was more than offset by the customer mix. I think as we look forward, we are paying more for certain components, et cetera, than normal, and that cost will get recognized whenever we ship those components, and the products that are in those components are incorporated into. We will see some drag on gross margin probably for some time, just given the inventory and purchase commitment levels that we have. The mix is healthier. The enterprise mix is helping to offset that, but customer mix is still by far the biggest driver. If we have a quarter where there's a heavy cloud mix, you are going to see pressure on that gross margin number for the quarter.
I think over time, we feel like we can stay in that 63% to 65% range. I think that's a good, safe place to be.
I guess what I'm thinking is you're already kind of signaling some pressure in the next quarter, because that's down over 1% from this quarter. I'm wondering, when you say that, do you think it's probable that you have a materially lower margin than 64%? Do you think when you say drag, do you mean 64% just is a possibility for longer?
Yeah. No, again, I think, the biggest driver will be the customer mix. In a quarter where there's heavy customer mix, yeah, you could be back at 64%, you could even be below that. Again, it'll be because of a particular mix of a particular quarter. I think the range is still very valid. The 65% plus that we saw this quarter is a really good, solid enterprise mix that's helping to offset some stuff. You will see gross margin is going to move around a bit over time, but I think we're still very comfortable in that range.
As Ita said, Rod, a lot of our inventory costs will be realized later this year and next year. The gross margin will actually be more pressured when those higher costs are realized. If we have a really good enterprise mix, we could be on the better side of 63%-65%. If we have a high cloud mix, I don't rule out the possibility of being on the lower side of 63%-65%.
Yeah.
Nobody's predicting the semiconductor supply costs or shortages are going away in 2022.
No, definitely not. That leads me to the enterprise trajectory. Enterprise spending indicators are very strong here as we look into the second half. Just curious what your thoughts on the enterprise pipeline are. Does that look equally strong in your pipeline, or how does that look to you?
Yeah. Rod, I think you've run out of questions, but the short answer is yes.
Okay, thanks. Appreciate it.
Thanks, Rod.
Thank you.
Your next question comes from the line of Jason Ader with William Blair.
Yeah. Hi, guys. My question is for Jayshree. How does this supply chain situation compare to prior periods in your career where you've seen supply constraints?
Well, I'm glad you're asking the oldest person here, or close to the oldest. In my career of several decades, I've never seen it be this bad, never. This is the worst I've seen it. There's been some pretty big ups and downs. More than the worst that I've ever seen it, I think it's also going to be prolonged. I guess we were all hopeful we would all recover from the COVID-19 pandemic, but everything from copper shortages to wafer starts, to assembly, to manpower, people, logistics, freight, just about every aspect of it is challenged, too. Anshul, do you want to add anything more to that as the youngest person here?
Well, as you said, things are very, very constrained. I think what's happened is the world supply chain never planned for this big a mismatch in supply and demand. As a result, when you run into a crunch, people try to book ahead and plan to rebuild buffers and so on. This is not an industry where you can react in one quarter. This will last a long time. The semiconductor industry is predicting maybe recovery in 2023, who knows what our global demand will be at that time. We are in it and prepared for a longer run here.
Is it inevitable that at some point we're going to see a demand air pocket, Jayshree, just because everyone will have ordered what they needed ahead of time, and we'll hit some type of air pocket?
I don't know the answer to that, Jason. I think, given the business diversification we have, the planning of one type of customer will not affect the planning of others. I'm hoping that some of them are planning for 2022, and some are planning for ahead in 2023, and so the air pockets will balance off, if you will.
Yeah. The visibility, Jason, I think to what the demand is for and when it'll get deployed, at least for the larger customers, is pretty good, so that helps, right? You're deploying over time still, right?
That's right. It's very much, as Ita says, a land-and-expand situation. You don't just land and then buy nothing for a long time.
Very good. Thank you.
Thank you.
Your next question comes from the line of Jim Suva with Citigroup Investment.
Thank you. I only have one question, and that's related to your strength in revenues that you've just seen and kind of the outlook, that strength. Is it being driven across all your end markets, or are there kind of one or two that you really want to call out to us as, if you were to look back, say, six or nine months ago, you would not have guessed it would've been so strong because you really have posted really impressive results and a great outlook. Thank you.
Thank you, Jim. Again, a lot of credit goes to Anshul and his team. Anshul, I would love for you to answer this question, but let me just preface it by saying all five verticals are very strong. They're all growing double digits, and in fact, I feel bad rating any of them as second or third or fourth or fifth place. If you ask me what I'm most positively surprised by, you might remember a year ago, Jim, we were less bullish on the cloud titans. We thought they'd even be flatter down. Anshul, you want to add to that?
Sure. Well, what a difference a year makes, despite all the COVID-19 impacts. The cloud titans are on a strong upgrade cycle with 400G and planning their transitions. Not just 400G, but a mix of 100G, 200G, 400G. The customers themselves are surprised that the end demand is so strong for their businesses. That's been a good upside for us. Same as super enterprise. We've talked about enterprise for quite some time, but we're quite happy with the adoption by our customers, both in our primary data center market, but also expanding into campus, into monitoring, into security. We have a long way to go. It's a very large TAM, and we're just getting started there, so that keeps us very positive on growth in that area, too.
Thank you, and congratulations.
Thank you, Jim.
Your next question comes from the line of Simon Leopold with Raymond James.
Thanks for taking the question. At the beginning of July, the U.S. government announced the cancellation of the JEDI project with a plan to put it out for rebid. I'm just wondering how you would factor that into your own forecasting and modeling, given that you sell into one of the, essentially, participants in that project. Does this change your thoughts on this year and longer term? What do you do in terms of a project when it comes to forecasting your overall business? Thanks.
Thanks, Simon. I don't think we really factored it in very much except making sure we had all the certifications. You want to add to that more, Anshul?
Simon, this was a contract award 10 years. It's not as if we saw any significant change in the last year related to that. In addition, the contract, the JEDI program, might be gone, there's so many other programs the government is doing with various cloud companies, it just gets mixed into the noise. There will probably be big upsides. I don't think there's any big downside either.
How big is government typically for you?
No, in this case, this wasn't going to us, right? This was going to one of our customers, which is why we don't see it directly.
It's not really government, it's cloud titan.
Yeah. No, I guess in general, because government sounds like a good vertical. I don't know that you've talked about it in the past.
No, we haven't. We could do a lot better. Government is not big for us, but it's mixed into our enterprise momentum. It has improved year-over-year, but in terms of any kind of contribution or large concentration, we have a long ways to go.
Thanks.
Next question.
Thanks, Simon.
Your next question comes from the line of Tal Liani with Bank of America.
Hey, guys. I'm trying to ask a question without asking it directly. The Street is looking for a big win with one of the cloud titans. This question I have is, what's your outlook for cloud titans? Are there big kind of wins or big announcements in the pipelines that you're looking for? Is there anything you can share with us about the puts and takes? What could happen with cloud titans over the next few quarters? Thanks.
Sure. Tal, I thought you'd say you're very happy with the results so far, and there's no more upside expected. Look, we're doing well with our customers. Execution has been good. You know it is a competitive market, and despite that, we continue to do very well with both our hardware design, our supplier products, and especially our EOS software. There's no big change. I mentioned this last year as well. We are maintaining status quo in terms of customers. If the customers grow, we grow with them. We obviously don't control that aspect and can't forecast on their behalf. Otherwise, it's more of moving on the next-gen architecture with these entities. While you think of them as cloud titan, internally, we think of each one as a market, right? They're that big, and there's so much complexity.
No, there's no big announcements we're making or about to make or have. There's so many key milestones that need to be achieved before that can be done on anything that's dramatic or a big shift.
If I can squeeze in another one. If not, it's okay. Can you talk about the contribution of cloud titans to this quarter and also last quarter results? I'm trying to understand in relative terms, how important is it to the overall growth?
Oh, we decided to go annually rather than quarterly, Tal. It's staying within the annual targets we gave. I believe we said it was 34%-39%. We bracketed it so that's the contribution from cloud titans, 34%-39%.
What about the growth?
The growth percent.
Meaning how much of the growth is it? Is it proportionally much higher than the 35%-39% of the growth?
No. We don't comment on.
It's very hard to look at it on a quarter-by-quarter basis. The cloud and cloud contribution to growth is always going to be important. Enterprise is becoming more important, right? I don't think we want to try and do it on a quarter-by-quarter.
Yeah, I think it'll make more sense annually, Tal, because we're so constrained by shipments.
Right.
It would be a false signal if we said it was high, low, or medium. I think on an annual basis, it's a much better number, and the growth will be good. It'll definitely be double digits.
Still the number one vertical and, no, it's important for sure. Don't let anybody think that it's not an important part of the mix.
Thank you, Tal. I think we're ready for the next question.
Thank you.
Your next question comes from the line of John Marchetti with Stifel.
Jay, I was wondering if you could just spend a minute or two just talking about what you're seeing specifically in that 400G market. Obviously, we go back to the early part of the year, and it seemed like we were pumping the brakes a little bit on expectations. That seemed to change a little bit last quarter. You seemed a little bit more bullish there. Just curious if you could spend a little bit of time, and I'm assuming that is still primarily a cloud titan market, but just anything you're seeing there would be helpful.
Well, it's actually more than cloud titans. I think Arista has been trialing 200 and 400G since Q4 2019, if I remember, when Andy was here and we talked about the whole portfolio of products. This combination of 10, 100, 200, 400, we expect to see a significant increase. Just to give you some context here, we had about 75 customers last year, and we expect that to triple to quadruple this year in customers. It's way more than the cloud titans.
Your next question comes from the line of Ben Bollin with Cleveland Research Company.
Good afternoon. Thank you for taking the question. Jayshree, you talked a little bit about getting some additional visibility from hyperscale partners versus history even into next year. With that perspective, could you talk about how you see the evolution of 200 and 400 as they roll that out, how that develops, and maybe how you think it compares with what you saw with 100 in terms of, I don't know, pace or speed of adoption?
Yeah, no, that's a really good question. I feel like 100G pace and adoption was remarkably fast. It was probably over an 18-month period, Anshul. Correct me if I'm wrong, 2016 to 2018, something like that, right? The 200, 400G has been more gradual, and in very many use cases, it's coupled with 100G. That's why it's so hard for us to separate out. We, I think, are finally at this inflection point or pivot point where we expect to see more 200G and 400G shipments in our cloud titan customer base in the second half of this year. I think this will continue and will be vibrant for at least three years to come, maybe longer. Is there a long tail to that, Anshul, you want to answer?
Ben, I think it's important to note that when the world went to 100G in 2016, 2017, 100G was very much backwards compatible with 40G. It cost the same as 40G, and it was the same power consumption as 40G. It was a no-brainer.
A no-brainer, yeah.
Upgrade. 200G and 400G are not the same. The companies that are going there are the ones who absolutely need it. There were other technologies involved, like the ZR optics needed for data center interconnections themselves, which were a constraint last year and earlier this year as well. Which is why this adoption curve is very different, and it's not at every layer of the network. I would not model that on the 2016, 2017 cycle. These are different upgrade cycles. Which is why you have to come back to what Jayshree was saying, to combine 100, 200, 400. We are winning a lot in 100G using new next-gen products as well, and that is very much material and matters.
One other addition I'd like to make on the non-cloud titan side, Ben, I know you're more focused on that, is many times they make sure they buy a product that's 200G, 400G or even 800G capable, but they'll deploy only 100G, right? What's clearly happening now is customers are getting ready for the higher speeds independent of how they deploy them in timing.
That's great. Thank you.
Thanks, Ben.
Your next question comes from the line of Jon Lopez with Vertical Group.
Hey, thanks very much. Ita, I believe you said in calendar Q1 you grew product deferred by about $40 million or so sequentially, and I thought I heard you say it's another $20 million or so this quarter. I guess the two questions I had there, one, do you ascribe any of that to component constraints, or is that all product feature sets? Secondly, do you expect that to flow into the income statement in the second half of this year, or is that stuff you'll carry with you into 2022?
It's not really related to the supply side stuff. It's more about new customers, new use cases, new products, right? That's what drives that number. Even in Q2, if you look at it, some of that stuff was recognized, and then we added other new stuff. There is some churn underneath of that balance, right? I don't expect us to take any net revenue, if you like, out of that bucket in Q3. We'll see what happens in Q4. That's a bit further out. I think given the comments in the script around how that balance builds, I think we've got lots of new stuff happening. I think it's probably unlikely that we would be depleting that balance this year.
Thank you.
Your next question comes from the line of Alex Henderson with Needham.
Great, thanks. I was hoping you could talk a little bit about the environment in terms of pricing. We've heard from some players in the market that Cisco is talking about 5%-10% price increases on a variety of products, including switching and routing-related products, and passing through more price increases than what you alluded to. You obviously suggested that you are going to be relatively careful with that and not push through a lot of pricing on your customers. Is there a delta developing there? Within that context, could you talk a little bit about the strategic approach of Cisco trying to bundle together the Acacia products and whether that architectural change to a switch routed optical platform is something that you concur with or which you think is a fool's errand? Thanks.
Wow, that was a loaded question. I'm going to break it into two, Anshul. Think about the Acacia one while I discuss the delta on pricing. It's difficult for us to parse our competitors' pricing. I can tell you philosophically, we got backlog, we got products in flight, and we got customers who are making their budget plans. We're going to try as much as we can in the short term to absorb the cost. That said, we will have some slight price increases, probably in the range of 5% on selective models that will affect our customer base in the tail end of this year and next year. It will still affect our gross margins if the mix changes dramatically to cloud titans, as Ita alluded to.
All this to say, Alex, we're going to be very thoughtful because it's pain for us, and we're going to try and mitigate the pain for customers and only apply it where there's a real shortage and a real significant cost increase. Anshul, you want to take the Acacia question?
Sure. Alex, the way our customers buy ZR optics, as you know we've been involving these architectures since the 2016, 2017 timeframe for 100G, 100 km optics. The cloud companies especially learned this in the '90s and early 2000s. They do not like bundled solutions. They like to disaggregate. They disaggregate every layer, not just white boxes. The optics and the switches are not allowed to be built in a bundled manner. They will run a separate RFP for each one. We have to be independently competitive at each layer to win. Which is why this is our concern. As we mentioned, optics is a large TAM, and other companies can participate. We don't have to. We have our work cut out.
We try to qualify as many optics vendors. At any given time, you have a team, Anshul, that's qualifying a number of vendors, right?
Right. If there's a ZR optic out there, we've qualified it.
Yeah, I've noticed. There's a lead time on your qualification, too, I understand.
Yes. Very long.
Thank you, Alex. Next question.
Your next question comes from the line of James Fish with Piper Sandler.
Hey, thanks for squeezing me in here. Jayshree, I was going to comment before that age is just a sign of wisdom, so I wouldn't say old, but just wise. With that being said
Thank you, James.
I'll just let you and Anshul have that debate, who's wisest though offline.
I think I win that one.
Go on.
You guys are starting to see some 5G core spending, or at least the industry is, pick up, and we even saw AT&T and Microsoft kind of announce something at the core. Carrier grew pretty nicely it looks like if you use the midpoint of those ranges you gave. Does your relationship with Microsoft help extend your reach into the carrier environment as they look to take some of their core to the cloud? Will it be more of a fulfillment via support for Azure, is the way to think about it? Thanks, guys.
Yeah, no. I think in the short term, it's very much a relationship tied to Azure. There's nothing that precludes, particularly with their investments in 5G and some of the recent acquisitions we made. As you know, service providers are a long-term test of our patience, but when you win them, you win them well. We are seeing our own personal wins independent of Microsoft, very much tied to upgrades to 5G, where we are the core routing and platform and spine. Specific to Microsoft, I think it'll take time. It'll probably emerge in the next several years. Today it's primarily Azure.
Okay, we have time for.
Thank you, James.
One more question.
Last but not least.
This question comes from the line of Kyle McNealy with Jefferies.
Hi, thanks a lot for the question. This is Kyle in for George Notter. Curious if there's any change to the mix of customers or ports using CloudEOS versus the traditional standard EOS going with a hardware sale. Maybe still quite small, but any quantifiable color you can provide around what the mix might currently be? Could that help your gross margin at all or make you a little less sensitive to the supply chain issues that you're seeing in some areas?
Thanks, Kyle. CloudEOS is an example of a very strategic software platform. It's very, very tiny in ports, and it has more to do with our multi-cloud hybrid network strategy, where almost every one of these enterprise customers also has some premise strategies. I would say the strong influence of millions of ports is still on the mainstream platforms, and CloudEOS is on top of that to connect into multi-cloud.
Okay, great. That's helpful. Thanks.
Thanks, Kyle.
This concludes the Arista Q2 2021 earnings call. We have posted a presentation which provides additional information on our fiscal results, which you can access on the investor section of our website. Thank you for joining us today.
Thank you for joining, ladies and gentlemen. This concludes today's call. You may now disconnect.