Welcome to the third quarter 2015 Arista Networks Financial Results Earnings Conference Call. [Operator Instructions] As a reminder, this conference is being recorded, and will be available for replay from the Investor Relations section at the Arista website following this call. I will now turn the call over to Mr. Chuck Elliott, Director of Investor Relations. 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 third quarter ended September 30, 2015. If you would like a copy of the release, you can access it online at the company's website. During the course of this conference call, Arista Networks management will make forward-looking statements, including those relating to our financial outlook for the fourth quarter of the 2015 fiscal year, industry innovation, our market opportunity and the impact of litigation, which are subject to the risks and uncertainties that we discuss in detail in our documents filed with the SEC, specifically on 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 in 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, Chuck. Thank you, everyone, for joining us this afternoon for our Q3 2015 earnings call. I am pleased to report our sixth consecutive beat as a public company. Consistent with prior quarters, customer demand for our 7000 Series, the U.S.-based products drove results that exceeded the consensus. From a geographic perspective, our customers in the Americas generated 78% of our sales, and 22% was derived internationally from the EMEA and Asia-Pacific theaters. We are witnessing balanced traction across our familiar 4 verticals: our Cloud Titans, Financials, High-Tech Enterprises and Web and Service Providers. Revenue grew 40% year-over-year to a record $217.5 million. Service contributed in the double digits at 11% of overall revenue, including software subscriptions. We delivered non-GAAP gross margin of 65.5%, resulting in a non-GAAP earnings per share of $0.59, thus growing EPS in excess of 40% year-over-year in our competitive and dynamic industry. We now have over 3,500 customers with our continued trend of 1 to 2 new customers per day. This quarter, we had a number of key highlights. Our progress with the HP partnership via demonstrations at VMWorld 2015 of converged solutions with HP Openview continues. We introduced next-generation leaf switch platform based on Broadcom's much anticipated Tomahawk silicon for flexible 10, 25, 40, 50 and 100-gigabit Ethernet switches, with hitless Smart System Upgrade capabilities. Our new products were endorsed by a number of our ecosystem partners at an NYC event we held on September 14, 2015. We do expect the next year to fuel the demand for 25, 50 and 100-gigabit Ethernet upgrades. We unveiled a strategic security architecture for cloud networking called Macro-Segmentation Services, or MSS for short. Our CloudVision MSS has been endorsed by many new and existing security leaders, including VMware, Palo Alto Networks, F5, Check Point and Fortinet.…

Thanks, Jayshree, and good afternoon. This analysis of our Q3 results and our guidance for Q4 '15 is based on non-GAAP, and excludes all noncash stock-based compensation expenses, legal costs associated with the ongoing lawsuits, and the release of a GAAP tax reserve as described below. A reconciliation of our selected GAAP to non-GAAP results is provided in our earnings release. Total GAAP revenues in Q3 were $217.5 million, up 40% year-over-year and comfortably above our guidance of $208 million to $212 million. We experienced good momentum in the quarter with revenue contributions balanced across our key verticals. Service revenues continue to tick upward at 11.1% of revenue for the quarter. International revenues came in at $47 million or 22% of total revenue, down from 23% last quarter. We experienced continued growth in EMEA, offset by some lumpiness in APAC related to some large deals recorded last quarter. While we continue to focus on expanding our international footprints, you should expect our geographical revenue mix to fluctuate on a quarter-over-quarter basis depending on the timing of U.S. and international deployments. Overall gross margin in Q3 was 65.5%, down slightly from Q2 at 65.8%, and just above the upper end of our guidance range for the quarter. Operating expenses for the quarter came in at $83.1 million, up from $74.7 million last quarter. This increase in spending was largely related to increased personnel and prototype expenses for R&D. We continued to grow our sales and marketing headcount in the quarter, but these increase personnel costs were offset by some reductions in demo and other marketing expenses. Overall spending was 38% of revenue consistent with last quarter with expenses growing at the same rate as revenue and investments being funded by top line growth. Our operating income for the quarter was $59.5…

Thank you, Ita. We are now going to move to the Q&A portion of the Arista earnings call. [Operator Instructions] Thanks, all.

[Operator Instructions] Your first question comes from the line of Mark Sue of RBC Capital Markets.

If I look at your results in the pipeline, it seems the runway is still lengthening for you and in terms of the opportunity. So in terms of how you're planning your growth related to a headcount and opportunity and TEM expansion, are we at the point where we could see additional verticals being added to Arista? Or are we at the point where we could add more channels to Arista? Just how we could kind of frame the outlook considering you're right around touching your run rate of $1 billion now?

We don't see any dramatic change in our run rate. We're very pleased with what I classify as 2 categories of customers we're serving. One is the cloud customers, which is our top 4 verticals. And the other is the converged customers through our partnerships, we can handle a lot of verticals that are not mentioned in the top 4, but are still addressed through our partnerships with HP, VMware and a number of others. So our 2 thematic focuses are really cloud platforms, where our top 4 verticals are something we directly and intimately deal with, and then converged platforms, where we work many times through technology partners not just channel partners and they address a number of verticals. As I said on the call, one of the meaningful verticals for us that we look to address and today, we largely group into high-tech enterprise is media and entertainment. We see that as very important vertical that appreciates the performance, the software's capabilities and the cloud characteristics are very similar and they're handling a lot of performance and workstream. So our pipeline is very solid, and we continue to focus beyond the 4 verticals, but the 4 verticals have a lot of penetration left in them.

Your next question comes from the line of Inder Singh of SunTrust.

I wanted to just ask you about your operating margin performance, which really has been outpacing, I think, some of the guidance that you've been giving over the last few quarters may be more. It's been running 200 basis points or so higher than your guidance, yet you feel that there's some cautiousness or something that you are trying to factor into your guidance. Is there something that in particular that causes you to be that cautious? Or is it just conservatism in terms of your guidance around operating margin?

Thank you, Inder. First of all, I would say that we want to make sure as a company, and this is our philosophy, I think, I've shared it with you all many times that we are peddled to the metal on R&D and we leave no stone unturned in terms of innovation. So we continue to allow a generous investment in R&D, much higher than the industry averages, as you know, and in terms of percentage of revenue. Now we haven't always been able to meet it, because we don't want to reduce the caliber of our hiring. Our often prototype expenses that we foresee in a given quarter slide out, because the chips are not available. So we're not going to waste money just because we want to spend money. So our R&D will continue to be high in the 20s, in percentage, as a function of revenue, but obviously, as our revenue is growing fast, we're not always spending as fast. But we plan to spend it. We continue to want to invest there. Sales and marketing, a little bit different. As you know, we've had a very targeted sales and marketing focus. In fact, I will tell you that in terms of quality, we're putting a higher priority on this and flushing out the nonperformers. And over there, our spending has been low and once again, below the targets we intended for 2 reasons. We want to keep the quality high, but also there, we're leveraging a lot of our technology partners. So we want to plan for the spending, but not necessarily spend for the heck of it. And that's what you're seeing in the last few quarters. Ita, maybe you want to add something there.

Yes, the only thing I'd add in to is, some of that out-performance is, obviously, been coming from the gross margin line as well where we've been forseeing and -- hitting the upper end of that range. And I would say for this quarter we provided some color in my script, but I do -- I think we see at least based on what we can see now that there is a mix towards those larger customers this quarter. So I would expect the gross margin line to really come in line with the guidance just because we see that trend already.

Good point.

Your next question comes from the line of Jeff Kvaal of Nomura.

I have a question about the broader outlook for some of your top verticals. I think some of your end customers or some of their peers have started to talk about being a little bit more careful with their CapEx spending over the course of the next few years. Could you help us understand what you are seeing from the cloud Titans, in particular? And, in particular, I guess, it would be helpful if you would allow us to understand which applications are the ones that typically drive the most purchasing for you, whether it's search or video or advertising or public cloud or what have you?

Okay, thanks, Jeff, I'll try my best. First of all, the interest from the cloud Titans has been very high, and the growth has been unabated. And that's why we are forecasting a very strong cloud Titan quarter for Q4. We are in 7 -- 6 out of 7 cloud Titans in the U.S, and we continue to make some progress internationally as well. And while not all 6 show up every quarter, and especially the last 3 quarters have been very balanced, we think we will see some lumpiness in the direction of more cloud Titan spending if not less the next few quarters. So we're not seeing what the others are seeing. The acceptance of the cloud Titans to Arista has been particularly consistent and strong. It varies by quarter. It's always lumpy. And as Ita pointed out, if they buy too much, then it will affect our gross margin, but it's been healthy. Now in terms of applications, they tend to be very dense computing, very large scale out storage. They are definitely more geared to the public cloud. There are some hybrid cloud applications as well, but I would say, more on the public side than in the private.

Your next question comes from the line of Michael Genovese of MKM Partners.

Could you talk about the importance of 25 and 50G? And is there any of that uplift in average selling price, perhaps in the fourth quarter guide? Or is it really a 2016 event to see 25 and 50?

Michael, I think all these speed upgrades take longer to happen than when the products are available. So the effect of 25 and 50, I'm going to forecast is more real in 2016 than Q4. We'll see some small perturbations, but to put this in perspective, it's taken 10 years for 10-gig to become mainstream. I'm not predicting 25 and 50 years for 25 and 50-gig, but it does take time. In terms of pricing, you do not see significant price upticks. You'll see mild maybe, but most people expect higher performance for, I guess, less ASP degradation is the way to say it.

Yes, I mean, as long as if you haven't moved on yet, I will ask a follow-up. Could you just on update on the lawsuits, the 3 that you were going on in the last quarter?

Sorry, what was the question? We didn't understand it.

I'm sorry, the question is about the lawsuits, if you could give us an update on the 3 cases that were in the courts last quarter or, and what's going on with them?

I am going to turn over to my expert General Counsel, Marc, are you ready to answer that question?

Sure, I'd be happy to answer. So as you point out, we had several cases that went on last quarter. The first was in the ITC, we called the 944 investigation. That trial was completed in September. Since one of the patents accused in that case was dropped, that means there are 5 patents that are remaining in that case that need to be decided upon. We expect Judge Shaw to issue his initial determination on that case on January 27, 2016. What happens next is of that the ITC commission then reviews that initial determination and issues a final determination on May 27, 2016. That then goes to the U.S. Trade Representative who makes the final decision on the case on July 27, 2016. The second case in the ITC is the 945 investigation, that actually goes to trial on November 9, on Monday, lasting through November 20. Following that trial, the administrative law judge McNamara, is scheduled to issue her initial determination on that case on April 26, 2016. That then goes to the full ITC commission, he reviews the initial determination and issues their final determination on August 26. And then finally, the U.S. trade representative makes the final decision on October 26 of next year. That's the status of each of the 2 ITC cases. As a reminder, there is also the copyright case in that, and that's ongoing as well, and that is scheduled -- that continues to be scheduled of August 2016, and we're currently in discovery in that case. Just to aid the investor community, what we're doing is we're going to put out on our website, our investor website, a document that sets forth all of the key dates in the trial. So you can refer to that if you need the specifics. The last case that we worked on was the OptumSoft case. The first phase of that lawsuit, the trial occurred in September. There is no scheduled date for a decision by the judge, but we would expect one probably within the next 6 weeks or so towards the end of November, and we'll obviously be waiting on that. The second phase of that lawsuit is currently scheduled for April of 2016.

Your next question comes from the line of Vijay Bhagavath of Deutsche Bank.

Quick question. On DSOs, I'd like to get any color on that? And then also on gross margins, it was slightly [indiscernible] expectation on your guidance, was that mix-related? Or is that just price aggressiveness into the cloud Titans?

Yes, so I think the DSOs is really just mechanical collections activities. In the quarter, we had some turn over in the team, et cetera. So I expect that to come right back in Q4 and, in fact, we can see that happening already. So there's nothing more there than just sheer execution. Yes, in terms of gross margin, I think, for Q3, we were actually at the upper end, above the upper end of our guide. And then in Q4, we are guiding to the lower end of that typical range, and that is really all based around the mix of Titans that we expect to see in the revenue in Q4.

Your next question comes from the line of Sanjiv Wadhwani of Stifel.

Just one clarification and a question. The clarification, any 10% customers in the quarter? And the question I had was, Jayshree, when you look at 25, 50 gig, understanding that it's a 2016 event in terms of demand, is it mainly going to come from the cloud titan? So do you see a lot of your financials and other verticals also picking up 25, 50 gig?

Thanks, Sanjiv. So first of all, there was no 10% concentration this quarter. As we said, we don't really reflect any concentration on a per quarter. We like to report that on the year. But there was none, to answer your question directly. And in terms of 25, 50, the biggest uptick we do see and interest we do see is for storage and compute applications in the cloud. So that is the primary application. But we don't preclude interest in many cloud-like enterprise and financial and web and service provider customers. So we do see a sprinkling of interest across all 4 verticals. But I would say, the concentration of interest is coming from the cloud, to answer your question. Now one common interest we see across all 4 verticals is the 100 gig. Everybody's looking for a 100 gig spine from Arista, and Arista is very uniquely qualified in building the best spine platform.

Got it. Jayshree, one quick question on Microsoft, given that they haven't been a 10% customer for 2 quarters now. Do you expect it to be up for this year compared to last year?

I've always projected that they will be flattish and as the year ends, we will definitely give any guidance and information on customer concentration.

Your next question comes from the line of Ryan Hutchinson of Guggenheim.

Jayshree, my question is on guidance. So just, obviously, no slowdown in sight with strong growth from the cloud titan. But as we think about '16, specifically Q1, should we take into consideration typical seasonality witnessed by other networking, security and storage vendors, especially given the above consensus guidance that you gave this afternoon? And the reason I asked is, prior to this release, consensus estimates imply flattish to slightly up revenue entering the new year, and I just want to make sure that we're appropriately modeling Q1 at the right level.

Thanks, Ryan. I appreciate it. Yes. So I think we have signaled a strong Q4 guidance. Obviously, Ita and I will talk about Q1 guidance in Q4. However, if you look at our last few years and history, we've always had a seasonally weak Q1, and we've generally tended to be flat to down. So I wouldn't read anything into it except when we have a strong Q4, we generally have a weaker Q1, and then we pick up the pace in Q2, Q3 and Q4. So I think that's just good modeling for you.

Your next question comes from the line of Alex Kurtz of the Sterne CRT.

So Jayshree, just back on the litigation issues. How are you working with the sales organization to mitigate the impact on this, around pipeline with existing customers, new customers and just sort of how people are thinking about big purchases going into the first half of '16 around this specific issue?

Yes, no, thank you Alex and thanks for your wishes. I've said this before, and I'll say it emphatically again, the lawsuit has not had a dramatic negative impact on our sales momentum and customer revenue. That's because our customers are smart, and they understand that while there are risks, that there is deep appreciation for our technology advantages and also our commitment to assure continuous supply through workarounds in a variety of ways. So the way we would address it is as factually as we address it with you all. We explained the risks. We explained our advantages. We explained the possibilities of workarounds, and get them comfortable. Marc, I don't know if you want to add something. You've joined me on many of these customer calls.

No, I think that's right, Jayshree. I think -- yes, I think the focus right now is to continue to defend ourselves in the case and to create a mitigation strategy in the event that there's a negative outcome.

Yes, thank you.

And your next question comes from the line of Hendi Susanto of Gabelli & Company.

I would like to clarify your expectation of strong cloud titan performance in Q4. Will that be concentrated in Q4 or will be a multiple quarter rollout? Additionally, like, what are some of the major applications that the cloud titans will be engaging in Q4?

Thanks, Hendi. I think I answered the cloud application question as best I could. A lot of focus on scale-out storage, compute and public cloud deployments, largely leaf and spine, with really dense computing, thousands of servers, petabytes of storage, et cetera. In terms of any further guidance beyond Q4, the best I could say is they're confident of $1 billion run rate in 2016.

Your next question comes from Simon Leopold of Raymond James.

This is Victor Chiu in for Simon. I wanted to ask a little bit about the VMware, HP and the Dell partnerships that you mentioned before. Can you just give us more color around how that's progressing and may help us assess the prospects and timing for each. And maybe I think it would be helpful if you could give us maybe a sense of the ranking of them by significance?

Well, I never like to pick favorites among my children so don't make me do that. But definitely, I would say our progress with HP and VMware is stronger than the others. Palo Alto, I would say is the other -- these 3 have been ones we've been working very closely with. And each of them have their unique nuances and technological advantages. With VMware, we have 6 or 7 levels of integration. Most people just focus on NSX and VXLAN, but we're very partnered with them on their cloud initiative, on their monitoring initiative, on their ESX and mainstream vSphere initiative. So our partnership with VMware continues to be very robust, especially as it relates to virtual to physical to cloud networking. And more recently, we also introduced security initiatives with both VMware and Palo Alto. In terms of HP, a real focus on converged infrastructure. We see a class of customers that just want their compute, their storage and their network together and really bringing these best-of-breed components, and this definitely includes Arista as their networking component and HP as their compute component and sometimes storage. So that would be the 2 examples I would give you. In terms of Dell, we've done some work with them on CloudVision integration and ASM integration. And things have just gotten started over there, so it's very early and time will tell.

Okay, great. And the timing for the other ones?

I'm sorry, timing for...

The, I guess, timing for when VMware, HP, et cetera might become more material?

Oh, more material. Well, the -- we started to see wins with both VMware and HP in 2015, and I believe we'll continue to see that in 2016 and '17. So the timing has already begun.

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

This is Fahad in for Paul. Can you provide some quantification as to the level of diversification, amongst the top 4 verticals, what was the distribution across these verticals?

We don't intend to break it out, Fahad. But as I've said, they were very balanced so you can assume all of them were double-digit percentages.

Okay. And one question on the competitive front. As you introduced CloudVision and this Micro-Segmentation service, it appears that you're increasingly getting the NSX space or the VMware. How do you manage any potential channel conflicts? And do you envision any material -- as you move forward with your vision, those coming to conflict with your biggest channel partners?

No, not at all. There is absolutely no overlap and competition with VMware's Micro-Segmentation and Arista's MSS. So just to be clear, VMware's Micro-Segmentation is fine-grained security and firewall policy at a virtual machine level. Arista doesn't do any of that. That gets mapped into their v switch, and Arista picks up where they leave off and really provide east-west security mitigation and risk compliance between the VMworld and the firewall policies that may be coming from Palo Alto or F5 or Fortinet or Check Point. So there's a lot of talk about security threat. But even after you catch them, you have to prevent them from spreading. And that's really where Arista comes in. So we're really bridging the gap between NSX, Micro-Segmentation and existing firewall rules and policies, which tend to be more north-south, and have done some very deep integration with both NSX and Panorama from Palo Alto.

Your next question comes from the line of John Lucia from JMP Securities.

Jayshree, you indicated Arista benefited from software subscriptions in the quarter. What was the makeup of those subscriptions? What services were those? And then can you talk about, just in general, how you expect software subscriptions to trend as a percentage of revenue in the long term for Arista? And then I have a quick follow-up, how many salespeople did you add in the quarter?

I'll take your second question, first. We don't tend to give out specific numbers, so that's easy, I won't tell you, but thanks, John. But on the software subscription, I think this is a long-term vision and strategy that we're developing. This is very well understood in the security world, but it's only just now forming in the cloud networking world. And CloudVision is clearly our first example of that. And what -- the uptake you're seeing, I mean, obviously, we're getting strength and strong acceptance of CloudVision, both in terms of customers and bookings. But by the time we realize this into revenue and billings, we'll have a delayed factor because these are long-term subscriptions. So I expect in the long term, long term envisioned by 1 to 3 years, for CloudVision and software subscriptions to be a greater component. Until then, I think we will be in the 10% to 12%. But this could give us a reason to get into the teens in the future.

You said in titan, you'd be 10% to 12%. Is that like in a year you'd be 10% to 12%? Or is that in the next couple of quarters?

Yes, 10% to 12% of total revenue, as we said...

Including services, right?

Right, including services. So services and software subscription will continue to hover in the 10% to 12% range of total revenue.

Your next question comes from Kulbinder Garcha of Credit Suisse.

Just 2. One of them is basically on the issue of workarounds for the ongoing litigation, where are we in the process? And how are you guys managing that basically? I guess you have to first, well, be informed in terms of what kind of workaround you need? And then how are you going to communicate it to your customers? And what's the time line? Anything on that would be helpful. And the second is probably a little bit easy. On gross margin, maybe this is for Jayshree a little bit, Arista executed very well versus you've always flagged that gross margins might trickle down. There would be this mix shift towards cloud, maybe a little bit in some quarters. But you executed well against that. We've seen a slight decline now over the last year and a bit. I guess what are the positive drivers in gross margin that you're implementing to offset that over the longer term. If you're cloud business is very significant, there should be some positive initiatives, I guess, in place. The software and services seems like a very -- software seemed like a very long-term issue, but is there anything in next year, whether it's a product refresh or other things in the positive side we should think about?

Okay, so there's 2 questions. The first one, on workaround, as you can imagine, we are putting in place contingencies for the workarounds, and there's a very strong engineering focus. Marc, do you want add to the workaround piece?

Sure. Yes, so to Jayshree's point, we actually developed to a greater or lesser degree design around for each of the patents in the event that there's an adverse outcome. Some of them have actually been implemented already. Others are in the process of being implemented, and frankly, a big part of the litigation process of course is to understand the nature of the claims that are being asserted. And so it's a moving thing, as we see how these are litigated. And I would expect significant effort in the first half of next year.

Right. So on -- thank you, Marc. On gross margin, as Ita has often said, if we have a lot of cloud titan spend, then our gross margins of $60 million to $65 million will tend to favor less $65 million and more in the early 60s. If we get the mix right, and that's why you've seen us do so well, because the last 3 quarters, they have been very balanced quarters. Clearly, while the software and services will contribute in the longer term, we also have a third component, which is product mix. And we will continue to do aggressive cost reduction, and the increase of 100 gig into our mix will improve our margins as well. So for all those reasons, I think we'll always be in that $60 million to $65 million range, trying hard to be more in the $62 million to $65 million in the near term rather than less than that. And we feel comfortable that we've got enough levers to do that.

Your next question comes from the line of James Faucette of Morgan Stanley.

I just wanted to ask a follow-up question on the benefits of the 25 and 100G products in 2016. Should we expect that they will grow kind of in line with the natural growth rate that you guys have been seeing that has to do with growth in the addressable market plus your share gains? Or is there an opportunity or how much of an opportunity is there for incremental acceleration on the back of replacing existing parts of the deployed base?

Thanks, James. I clearly -- it's difficult to predict the port mix of what will grow faster and what won't because it's very customer-dependent. But maybe I step back and share with you some of the Dell'Oro and Crehan data. I think what's clear is 10 gig will be more steady. 40 gig may actually come down over time. I don't know if it's exactly 2016, but if I look in 3 years, 40 gig is projected to come down. 25, 50 and definitely 100 gig are the high-growth areas. And so the mix of all of that, we absolutely predict will be double-digit growth for us in the foreseeable future. So we're less hung up on port mix. We look at this as high-performance ports. In fact, sometimes, we don't even know how our customer might break out a port. They may take a 40 gig and break it out into four 10s. And the same thing with 100. They may break it out into two 50s or ten 10s or whatever. But I think takeaway from this is the common denominator is increments of 10 gig Ethernet, and one thing we're definitely seeing in especially our 4 verticals is the need for 100 gig spine as a common denominator. And then whether host connect to 1 or 10 or 25 or 50 seems to depend on the use case and applications.

And your next question comes from the line of Simona Jankowski of Goldman Sachs.

I had a couple of questions on 25 and 50 gig as well. I guess the first one is, it looks like there's going to be maybe 5 or so vendors with products in that category. So just wanted to see how you view the competitive landscape there versus what you've been dealing with so far? And then as far as pricing is concerned, on a per gig basis, it looks like one of the big advantages is that you're going to be able to get much, much lower price per gig. So how do you think about that from an overall revenue and margin perspective? And then just lastly, would you expect any customers to pause and wait for these products? Or do you think 40 and 100 gig is going to go -- is going to continue to see high demand right up until these products become available?

Thanks, Simona. There's actually 3 questions there so I'll do my best to address 1a, 1b, 1c, right? So 1a, 25 and 50 gig, there's 5 vendors -- there's always been 5 vendors for us in every performance metric. I don't think it's unique to 25 gig. Of all the competitors we have, the largest monopoly is, obviously, Cisco. And then as is often referred to in our industry, there is Cisco and many dwarfs [ph]. And obviously, Arista is clearly now identified as the clear alternative in the data center. So we see that playing out for all the other performance metrics as well. We don't see anything unique. And some of the standouts for us is our high availability architecture, our highly programmable software, our ability to do SSU and hitless upgrade, ability to really deliver features at fierce velocity, not compromising the rapid use of merchant silicon. Everybody has access to the same silicon, but we develop a much, much better and much more technical superior product and, many times, command a price premium for that. To come to your second question, on price per port. We have seen ASPs stabilize in the 10 gig, and we do see that 25, 50 and 100 gig will -- the ASPs will depend greatly on volume just like they did on 10. And also, a real opportunity and issue is the optics, the interconnect optics for these 100 gigs, right? So we see that those often can be more expensive than the port itself. So stay tuned for Arista to really offer some important alternatives, both embedded, integrated and different types of transit options. So I think price per port, especially taking into consideration the interconnects between these ports, you will see some ASP fluctuation but even improvement. And then -- I've forgotten your third question.

Oh, just if you might see any kind of pause in demand as customers waiting for these or...

Oh, yes. At the moment, it's hard to tell that we see any pause. The publicity on, for example, Tomahawk silicon in 25, 50 gig has been going on for a year, and that has caused us no pause. I think our customers expect these speed transitions. If anything, it validates Ethernet as the foundation for all networking connectivity. And so they'll use what we have now and wait for what we have in the next quarter. And especially, in our top 4 verticals, I've rarely seen them just pause if they need to absorb technology and keep their networks running.

Your next question comes from the line of Jess Lubert of Wells Fargo Securities.

I also wanted to ask 2 questions. First, I was hoping you could talk a little bit about the outlook internationally, and to what degree you believe the safe harbor rulings may cause some of your U.S. cloud titan customers to accelerate international data center deployments in 2016 and beyond. To what degree you have the infrastructure in place to capture that opportunity? And then second, our research suggests that Arista may be getting close to launching a DWDM line card to directly address the data center interconnect market. So I was hoping to better understand the opportunity you see there and what the margin implications might be for such a solution?

Okay, Jess. Thank you. Thank you for your wishes. Regarding the cloud infrastructure, I think you bring up a very good point. All of the -- in particular, some of the major cloud titans, not all but certainly the major ones we're working with, all have international presence. And Arista has put in an infrastructure, Ita and Marc, I would say much of last year and this year, so we're pretty ready for this. We've been handling this for some time now. We're coming on 2 years, where we've been pretty prepared for this. Our top cloud titans tend to order at the point of location and that's something we're very comfortable with, and we have station depots and support capabilities everywhere. In fact, I would say to you, with the exception of maybe China, that we haven't seen that same type of cloud titans in other international locations, and we look to grow them more perhaps with service providers internationally who may become cloud providers as well. So we believe strongly that we have the infrastructure for it. Ita, do you want to add to the -- anything more to that?

No, I think, clearly, we've been focused very much particularly, starting with Europe and then in APAC, in adding infrastructure in advance of growing the revenue. So I think we're well setup to be able to take any acceleration of business that comes from that.

Yes, just to put this in perspective for you, Marc just shared some data with me. We're in 68 countries today with over 40 service depots, Jess, so...

So...Did you see some of those rulings accelerating the activity over there?

Do I see some of them accelerating and growing? To be honest, they don't give us the vision and longevity of that to tell you beyond a few quarters. So I can see it, but it's difficult to predict precisely till they tell me.

And then on the DWDM line card for data center interconnect?

Well, yes. So you did ask a question on a future product. We'll comment on it when we introduce it, how about that, Jess?

Okay. Can we at least assume that the margins will be consistent with the corporate average if you do introduce something along those lines?

Yes, I don't think we're ready to comment on that yet. Let's get some products out, and then we'll, obviously, update you on how it fits into the model.

Thanks, Jess.

Your next question comes from the line of Rajesh Ghai of Macquarie.

I wanted to rephrase a question that's been asked before on this call, and it's really to address a concern that I hear about cloud titans, some of them launching their own switches, their own white-box switches. So what percentage of the data center networking spend are the cloud titans that Arista typically target? In other words, how much of the spend requires a performance latency in Arista switches and how much of the spend could be targeted using white boxes?

Thanks, Rajesh, that's a good question. I'll share with you some of the public data, and then, obviously, whatever is confidential, I can't. But if you look at some of the major cloud titans, like Google or Amazon as 2 examples, they have a precedence of building their own switches long before Arista was even founded or came into the picture. So even when cloud titans invest in their own switches, they complement these with Arista product for a very good reason. And generally, the good reason is they build their own to get economies of scale for single-use cases. But they use Arista's to get the programmability and scale out, particularly in the spine or spline categories. So we do complement each other, and we also have a habit of working with them on switches they have. We work very closely with FBOSS, and we intend to work fully closely with Microsoft ACS initiative or other cloud titans as they present themselves. So we don't look at them as threats. We look at them as opportunities to work with them to make them better.

Your next question comes from the line of Alyssa Johnson from Pacific Crest Securities.

Can you remind me, was there a certain vertical that's kind of driving your expectations for strong Q4?

Yes, I think Ita mentioned it, Alyssa, and that would be the cloud titans.

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

I wanted to go back to the international side of the business. I think if I look at the numbers in absolute terms, your international business has been pretty consistent for the last 3 or so quarters. I guess, I'm looking for more evidence of traction there. I'm wondering if you can give us the number of the VARs or distributors that you guys now are selling through internationally, the number of salespeople. Maybe if you exclude the cloud titans and you sort of normalize those international numbers, is the balance of the business growing? I guess, I'm just kind of looking for evidence of more and more progress?

George, good question. So we have legal entities in 18 countries. We have VARs. Our VARs have increased from -- almost doubled to about 52 channel partners. And these are selected partners. We don't just go sign every Tom, Dick and Harry, right? And we do have presence in, as I've often shared with you, the developed countries. We haven't really put as much focus on the emerging countries. In the absence of cloud titans internationally, I think we're doing very well. We don't talk about breakdown in salespeople, but we are in 62 countries with the 60 plus channels partners we have. So this is also an important factor. But I can tell you, we're hiring disproportionately internationally from a sales point of view than we are in the U.S. relative to the contribution. So we are investing ahead of revenue internationally. And hopefully, those are good statistics for you to start to extrapolate. But I've said this before and I'll say it again, George, I think the success of U.S. is really hiding the success internationally.

Got it. Was the international business up sequentially ex the cloud titans?

Yes.

All right. This concludes the Arista Q3 2015 Earnings Call. I also want to mention that we have posted the presentation, which provides additional perspective on our Q3 2015 fiscal results, which you can access on the Investors section of our website, as well as our General Counsel, Marc Taxay, mentioned, we will be posting shortly after the call a summary of the key dates in our legal proceedings. And finally, thank you to everyone for joining us today.

Thank you for joining, ladies and gentlemen. This concludes today's call. You may now disconnect.