Arista Networks (NYSE:ANET) Q2 2026 Earnings Call

0:00 Welcome to the Second Quarter twenty twenty six 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. As

0:21 a reminder, this conference is being recorded and will be available for replay from the Investor Relations section on the Arista website following this call. Mr. Rudolf Barraggio, Arista's Head of Investor Advocacy, you may begin. Thank you, Regina. Good afternoon, everyone, and thank you for joining us. With me on today's call are Jayshree Ullal, Arista Networks' Chairperson and Chief Executive Officer and Chantal Braida, Arista's Chief Financial Officer. This afternoon, Arista Networks issued a press release announcing its fiscal second quarter results for the period ending 06/30/2026. If you want a copy of this release, you can find it on our website.

0:58 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 twenty twenty six fiscal year, longer term business model and financial outlook for 2026 and beyond, our total addressable market and strategy for addressing these market opportunities, including AI, inventory levels and management, lead times, purchase commitments, component supply and product innovation, 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 ks, 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 representing our views in the future. We undertake no obligation to update these statements after this call. This analysis of our Q2 results and our guidance for Q3 twenty twenty six is based on non GAAP and excludes stock based compensation expense, intangible asset amortization, gains losses on strategic investments and income tax effect of these non GAAP exclusions, including the recognition of direct excess tax benefits associated with stock based awards.

2:00 A full reconciliation of our selected GAAP to non GAAP results is provided in our earnings release. With that, I will turn the call over to Jayshree. Thank you, Rudy, and welcome, everyone, to our second quarter twenty twenty six earnings call. Arista is experiencing significant demand to achieve our first $3,000,000,000 quarter in revenue. To put this in perspective, just five years ago, our entire year was $2,900,000,000 in 2021. We have also entered the prestigious Fortune 500 list in '26, and in addition to that, we are now included in the Russell 50. Our AI fabrics momentum with EtherLink switches now exceeds a 100 cumulative customers from the initial four to five customers I spoke of in 2024.

2:44 Arista has developed innovative features to enable smart system upgrade, SSU, deep analytics, load balancing at scale for AI training workloads. Our latest member is the 7,060 x e seven for 100 terabit capacity and 1.60 terabit throughput, as well as the first liquid cooling options, highlighting our continued scale out leadership. Over a decade ago, we pioneered the use of leaf and spine topologies in the data center and cloud networks. We are now building upon that with our lossless and high performance AI fabrics. The maximum possible scale for an AI network generally depends on two things, the number of tiers in the network and the number of ports per device, often known as radix.

3:33 Increasing the tiers and ports is expensive and power hungry. Our customers deploy and often choose the Arista flagship 7,800 AI spine to achieve that high scale without adding additional tiers. Scale across is an important application. The scarcity of compute capacity, physical space, and gigawatts of power mandates that the AI infrastructure must be designed thoughtfully. The Arista 7,800 platform continues to be the flagship spine for distributed scale across applications, providing traffic isolation, contextual routing, and security.

4:11 The scale across switching and routing TAM is forecasted to be roughly 15 to $20,000,000,000 in 2030, and Arista is well poised in this segment. Our scale across AI innovations deliver programmable and deterministic routing, SR, Refix, Multiplane Forwarding, Multi Tenancy and Traffic Engineering, as well as load balancing across the regions. We are capable of providing near instantaneous recovery in the event of transient congestion, packet loss, or a physical failure of AI clusters independent of their geographical location. This scale across use case is expected to be approximately 30% of our overall AI target of at least 3,600,000,000 in 2026.

5:01 While compute and AI accelerators supporting billions of parameters often grab the headline, we did introduce a suite of EOS Extensible Operating System innovations to enable a robust AI network capable of

5:14 diverse models and accelerators. At this point, I'd like to invite Ken Duda, our President and Chief Technology Officer, to highlight some of these AI innovations. Thanks, Jayshree. Look, I have never witnessed the combination of rapid innovation and scale deployment that we are seeing in AI networks. I'd like to call your attention to three innovations: SSU, MRC, and SRv6. First, SSU. SSU is Arista's smart system upgrade, the ability to upgrade switch software without any disruption. Frequent upgrades are a hard reality today, especially as AI both uncovers security vulnerabilities and creates tools to exploit them. While many competing systems require a full reboot to address these issues, leading to expensive and disruptive downtime, Arista's EOS handles these upgrades seamlessly.

6:11 We ensure our customers stay secure without sacrificing even a single minute of valuable XPU cycles. Second, to maximize XPU utilization, you need MRC, or multi path reliable connection. See, in first generation AI networks, every packet on an XPU to XPU flow has to take the same path.

6:35 That means if two flows hash to the same link, they both run at half speed. MRC enables senders to spray a single flow across many paths through the fabric, where receivers reassemble any data that arrives out of order, eliminating the performance hit from fabric hash collisions. But how is the sender supposed to control which paths the flow will use? And that's where the third innovation comes in. SRV six or segment routing, It's not new, but using it to load balance an AI fabric, that's the game changer. The sender tags each packet with a stack of SRV six segment IDs, dictating the exact path the packet will take. The system then uses real time congestion signaling to dynamically shift packets away from hotspots.

7:23 Because Arista EOS provides a single unified operating system, we support this SRv6 intelligence all the way from the scale out fabric to the long distance scale across routing. It gives our customers the combination of high quality, top performance and operational simplicity that Arista is known for. Thank you, Ken. Undoubtedly, the Arista EOS architecture shines with its inherent state sharing programmability and single binary. And all of this, as you just pointed out, is foundational for reliable accelerator operation and optimal compute utilization. Arista's multipath and multi plane monitoring with explicit SRv6 probes ensures that reliable performance for accelerator communication.

8:10 We have collaborated with leading customers to build that AI fabric with that operational excellence. Well, as you know, during the last six months, I've been quite vocal and candid in sharing our supply chain challenges and concerns, not only affecting us, but affecting the entire industry. While the industry wide supply tightness and rising component costs persist, Arista has taken individual and aggressive proactive steps. Arista is making solid progress here in addressing our tight supply chain. After all, we have a lot of experience that comes from many years. During the lawsuit crisis, we had to build a manufacturing site in The United States in 2016, during COVID challenges in 2021, and here we are back again in 2026.

8:56 Our manufacturing rigor is based on a three pronged approach. The number one is the people and leadership. Recently, we've hired a terrific global operations executive, Yohenia Karelis, with over thirty five years of engineering, networking, and manufacturing expertise. And she has built an outstanding suite of leaders for new product engineering, contract manufacturing, global supply chain, direct fulfillment, and logistics. We're also partnering closely with our key suppliers to meet our prioritized forecast, and I'm so thankful to them for their continued cooperation. And finally, Arista is leaning in with our increased multiyear purchase commitments, now almost tripling from a year ago at 3,600,000,000 to approximately $9,700,000,000 by the end of Q2 twenty twenty six.

9:44 To describe that relentless execution, I would like to invite Todd Nightingale, our Chief Operating Officer and President. Over to you, Todd. Thanks, Deshri. Arista has spent the last six months improving our supply chain to keep growing product demand, and we're seeing significant improvements. We've secured multiyear agreements with leading vendors of strategic components, qualified new suppliers in key areas to limit risk, and built out supply chains for next gen AI technologies. Our capacity has been increased in both manufacturing and distribution, we've negotiated better component delivery terms to drive up both factory efficiency and capital deployment. Relationships with our strategic silicon vendors continue to be strong with really excellent collaboration in both supply chain and technical engagements.

10:32 Our memory supply has been secured for 2026 and we have extended visibility well into 2027 across DDR4, DDR5 and NAND memory. And importantly, we've increased our resiliency through optionality and expanded vendor qualification. For PCBs and optics, we're now able to build capacity in a twelve month window and have strengthened our engagement and commitments from key suppliers. We've improved our lead times and inventory management of thousands of component SKUs, improving sub component pipelining and multi sourcing and providing increased flexibility with reduced inventory risk. In a new area, we've now established a liquid cooling supply chain capable of driving and delivering the next generation of AI infrastructure. This includes cold plate, quick disconnect and tubing vendors with capacity agreements for cutting edge new AI technology.

11:21 And to match our capacity with customer demand, we've increased both our manufacturing and distribution capacity.

11:29 We have now three contract manufacturers and three distribution facilities providing geographic diversity in The U. S, in Asia, and in Mexico. By focusing on vendor stability and diversity, risk mitigation and predictable delivery terms, innovation for new AI products and capacity across our factories, we are making significant improvements

11:50 and significant capacity increases across our supply chain. Thank you, Tricia. Thank you, Todd. My God, it's so gratifying to hear the great strides and progress that you have made in such a short time. Great job by you and the team. Given our improving stance in supply chain, we are excited to increase our guidance for the third time this year to $12,600,000,000 revenue in 2026. We are now projecting 40% annual growth, which is an incremental $2,100,000,000 over our Analyst Day goal of $10,500,000,000 and an incremental $1,100,000,000 over our recent projections of $11,500,000,000 in May 2026.

12:30 Our renewed enthusiasm in fulfilling demand in the second half of twenty six is expected to apply across all our product sectors in a widespread manner, including the back end AI fabrics, the core data center front end, and campus and routing adjacencies. And with that exciting guidance, I'd like to turn it over to none other than our chief financial officer, Shantel, for more financial specifics.

12:55 Thank you, Jayshree. It's great to see the supply chain ecosystem gaining traction to meet our customers' demand. Let's review how that is translating into our financial performance and outlook. To start off, total revenues in q two were just over $3,000,000,000, up 37.7 year over year and above our guidance of 2,800,000,000. This significant growth was driven by our AI and enterprise customers. Congratulations to the employees on our first $3,000,000,000 quarter.

13:21 International revenues for the quarter came in at $697,800,000 or 23% of total revenue, up from 15.5% last quarter. This quarter over quarter increase was primarily influenced by strong organic growth across our international regions, combined with a shift in the geographic mix of sales to our large global customers.

13:40 The overall gross margin in Q2 was 63.4%, down from 65.6% in the prior year, driven by end customer mix and up from 62.4% in the prior quarter, benefiting from both tariff refunds and customer mix. Operating expenses for the quarter were $411,000,000 or 13.5% of revenue, up slightly from the last quarter at 396,800,000 due to an additional investment in liquid cooling, high radix switching, and AI optimizing software.

14:09 Our r and d spending came in strong at $278,100,000 or 9.2% of revenue, up slightly from the last quarter at $271,500,000. Arista continues to demonstrate its commitment and focus on networking innovation. Sales and marketing expense was $109,800,000 or 3.6% of revenue, down slightly from 3.8% of revenue last quarter, representative of the highly efficient Arista go to market methodology.

14:36 Our G and A cost came in at $23,100,000 or 0.8% of revenue, up slightly from $21,800,000 last quarter, reflecting our strong base cost productivity within the pure play networking business model. Our operating income for the quarter was $1,500,000,000 or 49.9% of revenue, an incredible financial outcome for the company.

14:57 Other income and expense for the quarter was a favorable $120,300,000 and our effective tax rate was 20.3%. Overall, this resulted in net income for the quarter of $1,300,000,000 or 42.9% of revenue. Diluted earnings per share for the quarter was $1.02 based on 1,276,000,000 diluted shares, representing a significant 39.7% increase from $0.73 in the prior year.

15:23 Now turning to the balance sheet. Cash, cash equivalents and marketable securities ended the quarter at approximately $13,300,000,000 up from $12,400,000,000 at the end of Q1. In the quarter, we did not repurchase our common stock. Of the $1,500,000,000 repurchase program approved in May 2025, dollars 817,900,000 remain available for repurchase in future quarters. The actual timing and amount of future repurchases will be dependent on market and business conditions, stock price and other factors.

15:52 Now turning to operating cash performance for the second quarter, we generated approximately $1,100,000,000 of cash from operations. This was driven by a robust earnings performance coupled with an increase in deferred revenue. DSOs came in at sixty eight days, up from sixty four days in Q1 due to the timing of customer shipments and invoicing.

16:11 Our inventory turns remain at 1.7 for the quarter. We ended the quarter with $2,500,000,000 in inventory, up from $2,400,000,000 last quarter. Inventory level fluctuations are expected to continue as we work through the balancing of component timing and availability. This could result in quarters of elevated inventory balances affecting the timing of cash flow from operations ahead of ahead of the deployments.

16:35 Our purchase commitments at the end of the quarter were $9,700,000,000, up from $8,900,000,000 at the end of q one. As mentioned in prior quarters, this expected activity mostly represents purchases for chips related to new products and AI deployments. Our total deferred revenue balance was approximately $6,900,000,000, up from $6,200,000,000 in the prior quarter. The majority of the deferred revenue balance is product related. Our product deferred revenue increased approximately $600,000,000 sequentially versus last quarter. We remain in a period of ramping our new products, winning new customers, and expanding new use cases, including AI.

17:10 These trends have resulted in increased customer specific acceptance clauses and an increase in the volatility of our product deferred revenue balances. As mentioned in prior quarters, the deferred balance can move significantly on a quarterly basis independent of underlying business drivers.

17:25 Accounts payable days are fifty seven days, up from fifty four days in Q1, reflecting the timing of inventory receipts and payments. Capital expenditures for the quarter were $29,700,000 Our construction work to build expanded facilities in Santa Clara remains on track, and we expect construction to be completed by the end of fiscal twenty twenty six.

17:45 These exceptional Q2 results, combined with the ability of the ecosystem to deliver what is required, are foundational to underpin our financial outlook for the company. Reflecting our strong momentum, we are raising our 2026 fiscal year outlook to 40% revenue growth, equating to approximately $12,600,000,000. Within this guide, our 2026 campus revenue goal is at least 1,250,000,000, and our AI fabrics goal is at least 3,500,000,000.

18:12 For gross margin, we are maintaining the range for the fiscal year of 62 to 64%, inclusive of mix and anticipated supply chain cost increases for memory and silicon. We have increased our fiscal twenty twenty six operating margin target now at a range of 48 to 49% while maintaining an expected tax rate of 21.5%.

18:32 Now more specifically, our guidance for the third quarter is as follows. Revenues of approximately $3,300,000,000 gross margin of approximately 63% operating margin between 4849% diluted earnings per share between $1.06 and $1.08 with approximately 1,279,000,000 diluted shares. Our effective tax rate is expected to be approximately 21.5%.

18:58 In closing, the Arista team is energized. We are well positioned for this AI super cycle and for Ethernet networking overall. This is earned through a combination of our innovation, our culture, and our focus. The opportunity ahead of us is tremendous, and we are ready to capture it. Now back to you, Rudy, for q and a. Thank you, Chantal. We will now move to the q and a portion of the Arista earnings call. To allow for greater participation, I'd like to request that everyone please limit themselves to one question only. Your line will be placed on mute after your question. Thank you for your understanding. Regina, please take it away.

19:32 We will now begin the Q and A portion of the Arista earnings call. Our first question will come from the line of Amit Daryanani with Evercore. Please go ahead.

19:57 Thanks a lot. Good afternoon. Congrats on a really good set of numbers here. Jesse, maybe I'll just ask you the opposite of the white box question you typically get. One of the trends we hear a lot from hyperscalers and even frontier labs is that the AI networking is becoming a lot more complex in multiple dimensions simultaneously, not only of denser clusters, but also you need to connect campuses and multiple data centers. As the networking problem seems to getting more complicated, more system level, does that increase the value of the integrated networking platform versus white box approaches?

20:29 And then are you seeing customers rethink the build versus buy decision at this point when it comes to AI infrastructure both across hyperscaler and frontier labs?

20:37 Well, thank you, Amit, for the good wishes as well. I think you're absolutely right that Arista's networking strategy has accelerated. And as you know, there's three types of AI fabrics that are critical for us to participate in, scale up, scale out, and scale across. Especially in scale out and scale across, I think the importance of the supply chain and rapidly being able to connect their processes so they don't remain idle, and the ability to get consistent performance because they often cannot get the power in one location. So they have to distribute and get the same level of,

21:11 capability, security, traffic engineering is becoming critical. So I would say white box is certainly a tactical solution that we tend to see more in use cases that are simple, scale up, or scale out where, you know, the actual amount of software and system requirements are low. But, you know, when you look at traditional network topologies, you can slow down the job completion significantly if you go with, you know, a box by box approach. And so this as you

21:40 rightly point out, the EtherLink system wide portfolio with all of the features Ken alluded to for reliability, MRC, SRV six, traffic engineering, synchronizing elephant flows, low latency. This massive training and inference does put more pressure on the combination of our hardware and software, and we feel we feel very well recognized and ready to achieve that.

22:08 Our next question will come from the line of David Voigt with UBS. Please go ahead. Great. Thanks guys for taking my question. Maybe just one question. So Jayshree, you took up the calendar year '26 revenue guidance substantially from last quarter. But if I remember correctly, you've got the AI target unchanged and only touched the campus. Can you kind of explain kind of the thought process there? What are seeing from other customers and other verticals and other use cases that sort of ultimately result in sort of that algorithm that you just laid out going into the balance of calendar twenty six? Thanks.

22:41 Yeah. No. I if if you had to ask me whether our AI number or campus number will go up, I think Shantel, Todd, Ken, and I absolutely believe it will. The question is not whether it will go up. The question is what is that number? And that, I would like to reserve that 1,100,000,000 question to, well, it depends on how we ship. You know, if we ship more front end AI or if we ship more WiFi or wired or EtherLink switches or routing.

23:08 So I'd like to give our customers the priority and Todd's team the flexibility to ship what we can. And that's why we're not holding ourselves to a number. But I think at this point, we can most certainly say all numbers are going up. Our next question comes from the line of Michael Ng with Goldman Sachs. Please go ahead. Hey, good afternoon. Thank you for the question. Maybe I'll ask about the supply chain improvements. Last quarter, you talked about a supplier that was de committed upstream. Maybe you can just expand a little bit on some of the efforts this quarter. I think you talked a little bit about a vendor qualification expansion memory, strengthened commitments from key suppliers.

23:51 Are they getting more capacity? Are you offering more pricing on your side? Are you getting more customer advocacy?

23:57 Any thoughts, on their Absolutely, Michael. So listen, I don't want you to believe that suddenly we waved a magic wand and all our problems have gone away. The industry is going to have a two year problem. I don't think we get out of it as an industry till 2028. But Arista is taking individually and specifically steps in the first half of this year that we believe will have results in the back half of this year. Over to you, Todd. Yeah. On the silicon side, you know, we have very, very tight, relationships with our silicon vendors, and that goes to both, you know, how we work through all the supply chain and how we've improved the delivery on those components, but also the technical engagements for new

24:34 platforms that are super exciting as they get into the field. Both have been going very well. As far as memory, or I would just say memory, PCB, optics, etcetera, you know, we've been able to work pretty hard at adding, you know, optionality,

24:50 qualifying new vendors very quickly through our engineering teams, and that helps us bring multiple paths to delivery possible through the supply chain. We've also worked very hard, especially in the last six months to improve our terms, and that means having shorter lead times and being more nimble. So if our demand shifts from one SKU to another, we're carrying less risk. We're able to be more nimble, and that's been incredibly successful, especially in the last six months. Our next question will come from the line of George Notter with Wolfe Research.

25:25 My question was just on the bigger picture view on all these technology changes going on in the marketplace, CPO, NPO, XPO. Obviously, customers are building next generation racks and switch platforms. And I guess I'm just trying to understand how you see that mix of technology is changing going forward, and how does that benefit Arista or hurt Arista? Like, what's the bigger picture perspective on what you guys are seeing longer term? Thanks. Yeah. Hey, George. First of all, I think it depends on the use case. Let me take the scale up use case, which we are less prevalent in. I think there's very much a philosophy there on copper if you can, optics if you must. So I think you're gonna see a lot of copper in that two meter, three meter distance, well within a rack, that type of thing, and the importance of pluggable optics. But in some cases, there is a number of instances of proprietary implementations of traditional co packaged optics that's been floating around. Arista is not a fan of five different proprietary implementations. They're gonna have, and they're gonna commit, and Andy and the team have been working hard at this, to really solve one, which is an OpenCPO.

26:39 And we don't think open CTO is going to happen overnight, but the idea here is to use socketed optical engines, pigtail fibers, and allow these modules to be fully pretested. And whether they're soldered on the board or nearby, the idea is to have a truly

26:54 open interface that can operate with multiple vendors and multiple switch configurations. And sometimes, as open CPOs are often called NPO too, nearby optics. Right? So we're big fans of that. It's very early stages. It probably comes into examples and trials next year. And probably from our perspective and the industry perspective with all the supply chain shortages, majority of the world will still remain pluggable optics and copper, but there will be some amount of co packaged optics in 2028 and 2029. So that's how we kind of see the world. Our next question will come from the line of Ben Reitsis with Melius Research. Please go ahead. Hey, great quarter, Jayshree. I actually did something weird. I read the press release and there's a line at the end that you're going to host a webinar with Anthropic and Palo Alto. So it got me thinking,

27:46 can you just talk about your progress in landing the AI Labs as customers and how things are developing with that potential 10% plus customer in the fourth quarter? You're sneaking in more than one question there, but thank you for reading the press release. And I think the mythos and the vulnerabilities is very front and center. And first, I just want to acknowledge Ken's architectural design of EOS and how we have the lowest vulnerabilities, and we're absolutely committed to building that rock solid foundation. But to to make AI secure isn't just our responsibility. It really comes across the board. We need to work with the leading model providers. We need to build a robust infrastructure, and we need to work with the best security vendors. So we thought nothing better than getting the best of breed network model and security providers to do this together because vulnerabilities are a fact of life, but how we proactively deal with them with the right technology, with the right systems, with the right process is exactly what that webinar I think it's September 8. Tune in for that because I sure won't do justice to that in this call, but it's very important.

28:59 In terms of our own approach to models, you know, every day you it's in the news. There's a China model, whether it's a deep sea or whatever. But from our standpoint, the network must be robust no matter what the model. There's gonna be a lot of enterprise AI agents. There's gonna be a lot of diversity of models. It's gonna be training at some point, but it's gonna move to inference. Arista is taking a lot of care to make sure that we build that multi model network that can handle all of the traffic and all of the different priorities and SKUs. So, undoubtedly, as you can understand, some of the work that Ken and the team did on MRC sometimes requires us to do work on the network. But sometimes requires us to deal with the packet spring and how the end host behaves. Ken, you wanna say a few words on that with MRC?

29:51 Yeah. I mean, look, we're we're extremely excited about the growth and diversity in the industry. The more diversity there is among models and model makers and between AI infrastructure providers, the better for us. And we've been a best of breed player from the very beginning, and we continue that tradition as we expand into AI networks. And MRC is one example of that, of the UEC work we've done also is another example of Arista of strongly supporting the industry's growth and enabling the best of breed solutions through open and interoperable standards. Our next question will come from the line of Aaron Rakers with Wells Fargo. Please go ahead. Yes, thanks for taking the question and congrats on the quarter.

30:37 Kind of building a little bit on Ben's question there. Given all the supply chain work that you've done and the improvements on lead times and such, Jayshree, maybe can you revisit your thoughts on adding a 10%? I think in the past, you've talked about one or maybe even two additional 10% plus customers. And in that, does the two Ms remain 10% plus customers as we think about the updated guide for 2026? Thank you. I can't escape that question. Okay, Aaron. Well, thank you for the wishes. I'll bask in that glory for two seconds. But look, we're gonna increase the number by a billion or more. And undoubtedly, we remain very committed to our two longest partners, Microsoft and Meta.

31:22 And I fully expect there to be one, maybe 10% customers. I'll leave it at that just because we want all the flexibility on what's in the product sectors and who are the 10% customers as we ship. So I wanna give Todd the ability to ship lots and lots of product. Our next question will come from the line of Karl Ackerman with BNP Paribas. Please go ahead. Thank you. Jayshree, could you address the growing partnerships you have with NeoCloud adopting custom XPUs? In particular, do you believe there are growing alternatives to NVLink and optical circuit switches for scale ups for scale up that you can address over time? How are you feeling about your relative opportunity in scale up switches versus earlier this year? Thank you.

32:11 Hey, Carl. Thank you. That's a loaded question. So let me kind of parse it as much as I can. So first, we live in a NVIDIA world, and I think we all can safely say it's it's a high percentage of the GPUs we connect to. So there's really two use cases there. One is where NVIDIA provides the full vertical stack, and usually that's an NVLink, so there's very little participation from Arista or anybody else in scale up. And, generally, Arista does better there in the scale out or scale across domain. The second is the non, in NVIDIA accelerators, and you've heard me talk about our enthusiasm with the MI series from AMD. We're excited about the the the the Google TPUs. You know, increasingly, we see that as a formidable training, processor.

33:00 And we're very excited about the range of inference accelerators as well. And we have a number of partners who are working with you can imagine many of our customers are building their own in house. So now parsing your question a little bit. In that sector where it's non NVIDIA, Arista will be excited and will be working more closely both in the scale up and scale out, in some cases, to build custom racks with their custom processors so that we can better tune our network with the behavior for their inference or training engines. In the NVIDIA cases, it's gonna take a bit longer. I feel a little bit like, you know, two, three years ago when we were talking about InfiniBand, and now we don't mention InfiniBand, but it took two, three years to move to Ethernet. So it will take time to go from a proprietary scale up that's been around a long time with NVLink to these other alternatives, even if it even if Ethernet's really good.

33:52 But I expect us to do much better there. Our next question will come from the line of James Fish with Piper Sandler. Please go ahead. Hey, thanks for the questions. Just on the 1.6, I guess how many customers are currently qualifying or deploying that now that it's become out there in the market? And are you guys still thinking about that as production scale next year and a growth driver for that point? And then just secondly, Chantal, on the annual guide, you did raise the core to an acceleration now. Is there a way to think about how you guys are seeing the upgrade and refresh cycle impact your business versus kind of drag along between front and back ends? Thanks. Great. On the 1.60, if I I'll go back. History is always a great lesson for what will happen. If I look at the 800T, we were in trials in '23, '24, and the real ramp came to us in '25 and '26. So it takes about six months, maybe even a year sometimes. And in this case, I think it will definitely take a half a year to get to production with the advent of liquid cooling and all the infrastructure the customer has to put in.

35:00 So we will be in trials in the second half of this year. In terms of customers, it will be single digit but very large customers. And I think the real production will still be in 2027. And you were asking Shantel a question on refresh. Over to you, Shantel. Yeah. Thank you. Thank you. Yeah. I think, James, nice to hear from you. I think from the, you know, the piece that is in campus and that isn't the AI, the in between goal that you're referring to, we are pretty excited, and it will come down to out of that 1,100,000,000 raise, like, how that gets divided out. But where are we excited? We're super excited by the new logo acquisition. We're very excited about what we're seeing on international growth. We're excited on the land and expand that we're seeing going both from campus to data center and data center to campus. So there's a lot there to be, you know, working through, and we'll see how much we get to through the end of the year. Yeah. I was hoping Tal would be here, and we could tell him we're excited about the front end data center, and the enterprise is doing really well. But, Tal, this one's for you. This answer from Chantel.

35:54 Our next question will come from the line of Matt Nickman with Truist. Please go ahead. Hey, thanks so much for taking the question. Congrats from my end on the quarter as well. Question on gross margin. So they improved about 100 bps sequentially despite some higher costs. So A, maybe Chantal, you help us think about the tariff refund benefit in the quarter? And then just sequentially, was the lift more mix related? Or was there any sort of pricing relief that's flowing through to offset some of the higher costs? Thank you. Yeah, it's a great question. I would say, you know, of the kind of the, I think versus guide is probably the best way to do it and versus guide you'll see there's probably 20 to 30 bps of tariff refund and the remainder more of a mix equation. I would put it at 20 to 30 for the quarter versus the guide, probably in that range of maybe 30 to 40 bps if you look versus, you know, prior year, but it's in that kind of range and the rest would be customer mix oriented. I wouldn't put anything in there yet due to a price increase in that equation.

36:51 No. As we said before, the price increases will really help or affect us only towards the end of the year or more like next year because we're still going through a lot of backlogs. That's right. Our next question will come from the line of Simon Leopold with Raymond James. I think Tal is a software analyst these days. Anyway, I wanted to follow-up on the incremental $1,000,000,000 for the year. In that, I think you've reiterated the campus and AI. So I'd like to try to get a better understanding of the composition or the source of the incremental revenue if it's not the AI or or campus portion?

37:38 Thank you. Oh, okay. Actually, that's a great question, Simon, because in a nutshell, answer is this is our core Arista product line we're talking about. Campus and AI are the newcomers to it. So if you take out the campus and AI, which we believe will grow, then one of the things that our new software analyst, Tal, has been questioning me on is why is your core data center front end so flat? And so I'd like to first make the observation that it isn't going to be flat, but we couldn't answer that question in q one. We needed to go through the quarters to understand how the demand would translate into shipments. So we fully expect some growth in the front end and core data center and the enterprise team. And I just wanna give a shout out to Krishmit and Ashwin, the campus team. I wanna give a shout out to Kumar.

38:26 They've been, on record ground, and they've been growing tremendously well. So I expect great contributions from the enterprise into that incremental 1,100,000,000 number in addition to AI. And likewise with routing, routing is an adjacency both for edge routing use cases, but also some of the DCI and scale across. This is an incredibly large TAM and getting larger. So we expect that 1.1 to consist of all our product sectors, and we'll know exactly what it is once we ship it. Our next question will come from the line of Ryan Kuntz with Needham. Please go ahead. Hi. It's Jeff Hopson on for Ryan. Thank you for the question. I was just curious of kind of where we are with the scale across build outs.

39:11 Like is it concentrated to maybe a couple of the cloud titans, or can we see a a world where this extends out to some of the the Neo Cloud customers as well? Thank you. Yeah. That that's a good question, Jeff. We definitely see the the the why why is scale across such a big thing? I was talking to one of the industry luminaries, and he goes, well, this will be temporary as because once people get capacity of compute and power, they'll go back. And the answer is no. They won't. Because they're gonna be in a constant state of scarcity for power, space, and compute. Getting all those three to lock in, whether you're a cloud titan, an AI titan, or a new cloud, is gonna be very, very difficult. So currently, our scale sale of cost is dominated by the cloud and AI titans, but I see no reason why it wouldn't apply to and we're already seeing it apply to several neo clouds. So I would say less in the enterprise and definitely more in the neo clouds and titans.

40:12 Our next question will come from the line of Antoine Gabin with New Street Research. Please go ahead. Hi, thank you very much for taking my question. Maybe a quick follow-up actually on the scale across. So like how large is the overall opportunity today relative to the €1,200,000,000 and the 30% of the €3,600,000,000 that referred to? And how much do you think can capture of the 15,000,000,000 to €20,000,000,000 in 2030, given like all the other color that you gave on this RBC and the value it delivers, you know, when customers select Arista for both scale out and scale across? Thank you. Yeah. Antoine, the market research is suggesting this mark I'm gonna work backwards to answer your question. The market research is suggesting that it's gonna be about a $15,000,000,000, if you add some of the optics, maybe $20,000,000,000 in 2030.

41:03 So I would say in 2026, as we capture 1.2, this market is still maybe 3 to 4,000,000,000 billion dollars but quickly going to 15,000,000,000 in the next four years. Our next question will come from the line of Meta Marshall with Morgan Stanley. Please go ahead. Great, thanks. Jayshree, you've talked in the past about kind of new customers who have come in who have maybe wanted to start down the road of Blue Box and then discovered with the complexity that they actually need to go down the road of EOS. Can you just talk about kind of the latest trends that you're seeing there, particularly as you continue to add customers? Thanks. Yeah. Good question, Meta. We continue to see the importance of Blue Box because of, I would say, three things, and then I'll tell you the counter view on why a white box may still be useful. The three things they value greatly with The US is operational excellence. They don't have to put a lot of staff. Most of these neo clouds don't have staff. And even if they're paying a little more for the CapEx, it more than makes up, for the operating costs that they would incur.

42:13 So that's a huge piece. The second is the AI features themselves. You heard Ken talk about a few of them. There's a tremendous depth and breadth to EOS that they can't recreate in a white box or they can figure out how it is. And the third is reliability and vulnerability, which is becoming top and center. You can't put these things in the middle and have them blow up. So and, you know, a combination of both EOS and NetDI, our diagnostics layer, which plays in both the hardware and software has been very compelling. So we continue to see that while it's interesting to talk about these open gnosis, that more and more customers want either EOS itself in its entirety or a hybrid combination often of a open NOS and EOS. Now that doesn't mean it's everywhere. If somebody wants to just go put in, you know, commodity hardware and get the ultra low price, such a market exists in every large market, and it's not it's it's in it's there in campus, it's there in data center, and it'll be there in AI too. So there are customers who will just prefer to go with the white box because it's ultra cheap or that have large amounts of staff like the titans often do, and they can manage it and deal with it, and they've built with it built with it.

43:26 But having said that, I think in this in this current scarcity of supply chain and people and needing to deploy AI fast, Arista's really winning out with the system wide approach on EOS and good hardware. Hey, Regina. We have time for one last question. Our final question will come from the line of Atik Malik with Citi. Please go ahead. Hi, it's Adrienne for tightness in the overall supply environment and your demand that you're seeing, I'm wondering how far out your visibility with customers is extending. Yeah. Hi, Adrian. Hope you're doing well. I think with the you know, we still have the same kind of two quarters of visibility that we refer to. But what we would say is given where we are at this time of year, you know, Jayshree and our guiding based on what we're confident we can get the supply for.

44:20 And if the supply was to release a little more, there is an opportunity to do better in the year. So we'll have to wait and see. This concludes Arista Networks second quarter twenty twenty six earnings call. We have posted a presentation that provides additional information on our results, which you can access on the Investors section of our website. Thank you for joining us today and for your interest in Arista. Thank you for joining, ladies and gentlemen. This concludes today's call, and you may now disconnect.