Q1 2024 Arista Networks Inc Earnings Call
Okay.
Speaker Change: Ladies and gentlemen, welcome to the first quarter 2020 for Arista networks financial results earnings Conference call.
Speaker Change: During the call all participants will be in a listen only mode.
Speaker Change: After the presentation, we will conduct a question and answer session and instructions will be provided at that time.
Speaker Change: If at any time during the conference you need to reach an operator. Please press the star followed by zero.
Speaker Change: And 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.
Liz Stine: Ms. Liz spine are risked as director of Investor Relations you may begin.
Liz Stine: Thank you operator, good afternoon, everyone and thank you for joining us.
Liz Stine: With me on today's call are J C. La Arista networks, chairperson, and Chief Executive Officer, and Shantou Bright up everything Chief Financial Officer.
Speaker Change: This afternoon Arista networks issued a press release announcing the results for its fiscal first quarter ending March 31st 2024 if.
Speaker Change: If you would like a copy of this release you can access it online at our website.
Speaker Change: During the course of this conference call Arista networks management will make forward looking statements, including those relating to our financial outlook for the second quarter of 2020 for fiscal year longer term financial outlooks for 'twenty 'twenty, four and beyond I total addressable market and strategy for addressing these market opportunities including AI.
Speaker Change: Good customer demand trends and supply chain constraints component cost manufacturing output inventory management and inflationary pressures on our business lead times product innovation working capital optimization and the benefits of acquisitions, which are subject to the risks and uncertainties that we discuss in detail in our documents filed with the SEC.
Speaker Change: 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.
Speaker Change: 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.
Speaker Change: Also please note that certain financial measures. We use on this call are expressed on a non-GAAP basis and have been adjusted to exclude certain charges.
Speaker Change: 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 Jason.
Jason: Thank you Liz Thank you everyone for joining us this afternoon for our first quarter 2024 earnings call.
Jason: Amidst all the network consolidation Arista is looking to establish ourselves as the pure play networking innovator for the next era addressing at least a $60 billion pen in data driven client to cloud AI networking.
Jason: In terms of Q1 specifics, we delivered revenue of 1.5 dollars 7 billion for the quarter with a non-GAAP earnings per share of a dollar at 99 cents.
Jason: Services and software support renewals contributed strongly at approximately 16.9% of revenue.
Jason: Our non-GAAP gross margins of 64, 2% was influenced by improved supply chain and inventory management as well as a favorable mix of the enterprise.
Jason: National contributions for the quarter registered a 20% would be Americas strong at 80%.
Jason: As we kick off 2024, I'm, so proud of the Arista teamwork and our consistent execution, we have been fortunate to build a seasoned management team for the past 10 to 15 years.
Jason: Our co founders are very engaged in the company for the past 20 years.
Jason: China is still actively programming and writing code, while Andy is a full time chief architect for next generation AI Silicon and optics initiatives.
Jason: Holbrooke, our recently promoted Chief development Officer is driving a major platform initiatives in tandem, but John Mccool and Alex on the hardware side.
Jason: This engineering team is one of the best in Tech and networking that I have ever had the pleasure of working with.
Jason: On behalf of a instead, though I would like to express our sincere gratitude for onshore for Donna 16, plus wonderful years of instrumental service to the company in a diverse set of roles.
Speaker Change: I know he will always remain a well wisher and supportive of the company, but also I'd like to invite you to say a few words. Thank you Geoffrey.
Speaker Change: The rest of the journey has been a very special one.
Speaker Change: We've come a long way from our startup date to over $80 billion company today.
Speaker Change: Every milestone and prevent the ups and downs are all interested in my mind.
Geoffrey: I've had a multitude of roles.
Speaker Change: And lawn and growing more than what I could have ever imagined.
Speaker Change: I have decided to take a break and spend more time, but probably.
Speaker Change: Specially when the kids are young.
Speaker Change: I'm also looking and exploring different areas in the future.
Speaker Change: I wanted to thank all of you on the call today.
Speaker Change: Customers investors partners and all the Wedbush shows over these years.
Speaker Change: Artist isn't.
Speaker Change: Isn't just the workplace it's family to me.
Speaker Change: People around you that make life one.
Speaker Change: Special Thanks to Arista leadership, Chris.
Speaker Change: Chris Austrian John Mckew Mach.
Speaker Change: VTOL and Chantelle, Mark talk say few Holbrook can do though and many more.
Speaker Change: Above all there are two ways special people I want to thank.
Speaker Change: And the impact just fine.
Speaker Change: Our vision.
Speaker Change: Vision passion guidance and listening to me.
Speaker Change: And of course Jeffrey.
Speaker Change: So it hasn't been just my manager, but also by mentor and coach for over 15 years.
Speaker Change: You for believing in me.
Speaker Change: As always continue to be in the rest of Wedbush.
Jeffrey: Back to you Jeffrey.
Jeffrey: Sure. Thank you for that very genuine and heartfelt expression of York huge contributions to Arista. It gives me goosebumps hearing you understand <expletive> memories, we will Miss you and hope someday you will return back home.
Jeffrey: At this time Arista will not be replacing the C O role and instead flattening the organization.
Jeffrey: We'll be leveraging our deep bench strength of our executives who stepped up to drive on new Arista to dog O initiatives.
Jeffrey: In particular, John Mccool, our Chief platform Officer, and Ken Kaiser a group Vice President have taken extended expanded responsibility for our cloud AI tightening initiatives operations and sales.
Jeffrey: On the non cloud sorry, two seasoned executives have been promoted Ashwin Kohli, chief customer officer, and Chris Schmidt Chief sales Officer will together address the global enterprise and provide opportunity.
Jeffrey: Our leaders have grown up in Arista for a longtime with long tenures of a decade or more.
Jeffrey: We are quite pleased with the momentum across all three sectors cloud and AI tightened enterprise and providers customer activity as high as Arista continues to impress our customers and prospects without undeniable focus on quality and innovation.
Jeffrey: We built a programmable network onto these based on a universal leaf spine topology. We are also congest constructing network as a service suite I'm overly such as zero touch automation security telemetry and absurdity.
Speaker Change: We'd like to invite can do that our founder and CTO and recently elected to the Arista Board to describe our enterprise strategy as we drive to our enterprise campus school of $750 million in 2025, what would you again, thank you Jay and thanks, everyone for being here I'm getting to the CTO of Arista networks.
Speaker Change: Excited to talk to you today about net D L.
Speaker Change: The Arista Networks' data leak and how it supports our network as a service strategy.
Speaker Change: From the inception of networking decades ago networking has evolved rapidly changing data.
Speaker Change: Is that about how the network is operating which paths through the network, our best and how the network is being used historically.
Speaker Change: Well historically.
Speaker Change: Most of this data was to simply discarded.
Speaker Change: Org change the state and that which was collected can be difficult to interpret because it lacks context network addresses important numbers by themselves provide little insight into what users are doing we're experiencing.
Speaker Change: Recent developments in AI have proved the value of data.
Speaker Change: To take advantage of these breakthroughs you need to gather and store large datasets LIBOR suitably for machine learning.
Speaker Change: Arista is solving this problem with NEC deal, we continually monitor every device not simply taking snapshots, but rather streaming every network is that every counter every piece of data in real time.
Speaker Change: Archiving, a full history and net deal.
Speaker Change: Alongside this device data, we also collect flow data and inbound network telemetry data gathered by our switches.
Speaker Change: We enriched this performance data further with user service and application layer data from external sources outside the network.
Speaker Change: Enabling us to understand not just how each part of the network is performing but also which users are using the network for what purposes and how is the network behavior is influencing their experience.
Speaker Change: Net yellow is a foundational part of the stock.
Speaker Change: Stock, enabling advanced functionality across all of our use cases for example in AI fabrics no deal.
Speaker Change: Enables fabric wide visibility integrating network data and mixed data to enable operators to identify Ms configurations weren't misbehaving hosts and pinpoint performance bottlenecks, but for this call I want to focus on how net deal enables network as a service.
Speaker Change: Work as a service or NAS is a risky strategy for up leveling our relationship with our customers taking us beyond simply providing network hardware and software by also providing customers or service provider partners with tools for building and operating services.
Speaker Change: The customer selects the service model configure a surface instances and Arista is C V. NAS handles the rest equipment selection deployment provisioning billing monitoring and troubleshooting. In addition, <unk> provides end user self service, enabling customers to manage their service instance, provision new endpoints provision new virtual.
Speaker Change: Apologies.
Speaker Change: Topic prioritization policies set access rules and get visibility into their use of the service and its performance.
Speaker Change: One can think of Nash as a point cloud computing principles to the physical network reusable design patterns scalable autonomous operations multi tenant from top to bottom with cost effective automated end user self service.
Speaker Change: And we couldn't get to the starting point without that deal as Youll provides a database foundation of Nash service deployment and monitoring.
Speaker Change: No no no. This is not a separate SKU, but really refers to a collection of functions in our vision for example, Arista validated designs or a P. D. As a provisioning system. It's an early version of our Nash Service instance configuration tool.
Speaker Change: Our Agni services provide global location independent identity management needed to identify customers within das.
Speaker Change: Our product or Universal network Absorbability will ultimately become the service monitoring element of mouse and finally, our <unk> solution has security integrated through our Z T N or zero trust networking product that we showcase at RSA. This week.
Speaker Change: Thus, our Nast vision simultaneously represents a strategic business opportunity for us while also serving as a guiding principle for our immediate co vision development efforts.
Speaker Change: While we are really excited about the future here are our core promise to our investors and customers is unchanging and uncompromised.
Speaker Change: We will always put quality first we are incredibly proud of the amount of success customers have had deploying our products because they really work and as we push hard building sophisticated new functions in the <unk> areas, we will never put our customers' networks at risk by cutting corners on quality.
Speaker Change: Thank you.
Speaker Change: Thank you Ken for your tireless execution in the typical Arista way.
Speaker Change: In an era characterized by stringent cyber security observer ability is an essential perimeter and imperative we cannot secure what you cannot see we launched cloud vision, who know in February 2024 based on the U S Network data Lake Foundation that Ken just described for Universal Network Observer ability.
Speaker Change: Cloud vision, who know delivers fault detection direction and recovery. It also brings deep analysis to provide a composite picture of the entire niglet with improved discovery of applications host workloads and it systems of record.
Speaker Change: Okay switching to AI of course, no call is complete without that.
Speaker Change: Generative AI training tasks simple they are made up of many thousands of individual iterations any slow down due to network and critically in fact, the applications performance, creating inefficient waste states and idling away process and performance by 30% or more.
Speaker Change: The time taken to reach coherence known as job completion time is an important benchmark achieved by building proper scale out AI networking to improve the utilization of these pressures and inexpensive gpus.
Speaker Change: Arista continues to have customer success across our innovative AI for networking platforms.
Speaker Change: Some blog from one of our large cloud and AI tightened customers Arista was highlighted for building a 24000, no GPU cluster based at our flagship 7800 AI spine.
Speaker Change: This cluster tackles complex AI training test that involve a mix up model and data parent organization across thousands of processors and Ethernet is proving to offer at least 10% improvement of job completion performance across all package sizes voices instead it back.
Speaker Change: We are witnessing an inflection of AI networking and expect this to continue throughout the year and decade ether.
Speaker Change: Ethernet is emerging as a critical infrastructure across both front end and back end AI data centers.
Speaker Change: Yeah, I application simply cannot work in isolation and demand seamless communication among the compute nodes consisting of back end Gpus and AI accelerators, and then as well as the front end nodes like the Cpus alongside storage and IP and systems as well.
Speaker Change: If you recall in February I shared with you that we are progressing well in four major AI Ethernet clusters, that'd be one versus Infiniband recently in all four cases now migrating from trials to pilots connecting thousands of GPU as this year and we expect production in the range of 10-K 200.
Speaker Change: <unk> Gpus in 2025.
Speaker Change: Internet at scale is becoming the de facto network at Premier choice for scale out AI training workloads.
Speaker Change: A good AI netbook needs a good data strategy delivered by our highly differentiated Eos and that's what data Lake architecture.
Speaker Change: Therefore, becoming increasingly constructive about achieving our AI target of $750 million in 2025.
Speaker Change: In summary, as we continue to set the direction of risk into it auto networking, our visibility to new AI and cloud projects is improving and our enterprise and provide a activity continues to progress well.
Speaker Change: We are now projecting above our analysts day range of 10% to 12% annual growth in 2024, and with that I'd like to turn it over to Chantelle for the very first time as Arista CFO to review financial specifics until there's more warm welcome to you Chantelle. Thank you Jason and good afternoon.
Chantelle: Dallas as of our Q1 results and our guidance for Q2 'twenty 'twenty four is based on non-GAAP and excludes all noncash stock based compensation impacts certain acquisition related charges and other nonrecurring items a full reconciliation of our selected GAAP to non-GAAP results is provided in our earnings release.
Speaker Change: Total revenues in Q1 were 1.571 billion up 16, 3% year over year and above the upper end of our guidance of 1.52 to 1.56 billion.
Speaker Change: This year over year growth was led by strength in the enterprise vertical with cloud doing well as expected.
Speaker Change: Sort of as a subscription software contributed approximately 69% of revenue in the first quarter down slightly from 17% in Q4.
Speaker Change: International revenues for the quarter came in at $316 million or 21% of total revenue down from 22, 3% in the last quarter. This.
Speaker Change: This quarter over quarter reduction reflects the quarterly volatility and includes the impact of unusually high contribution from our EMEA region customers in the prior quarter. In addition, we continue to see strong revenue growth in the U S with solid contributions from our cloud Titan and enterprise customers.
Speaker Change: Gross margin in Q1 was $64, 2% above our guidance of approximately 62%. This is down from 65, 4% last quarter and up from 63% in Q1 FY2023.
Speaker Change: Our year over year margin accretion was driven by three key factors.
Speaker Change: By chain productivity gains led by the efforts of John Mccool My campus and his operational team a stronger mix of enterprise business and a favorable revenue mix between products services and software.
Speaker Change: Operating expenses for the quarter were $65 million or 16, 9% of revenue up from last quarter at $262 $7 million.
Speaker Change: R&D spending came in at $164 6 million or 10, 5% of revenue down slightly from $165 million last quarter.
Speaker Change: This reflected increased head count offset by lower new product introduction costs in the period.
Speaker Change: <unk> of prototypes and other costs associated with our next generation products.
Speaker Change: Sales and marketing expense was $83 $7 million or five 3% of revenue compared to $83 $4 million last quarter with increased head count costs offset by discretionary spending that is delayed until later this year.
Speaker Change: Our G&A costs came in at $16 $7 million or one 1% of revenue up from 0.9% of revenue in the prior quarter.
Speaker Change: Income from operations for the quarter was $744 million or 47, 4% of revenue.
Speaker Change: Other income for the quarter was $62 $6 million and our effective tax rate was 29%.
Speaker Change: This resulted in net income for the quarter of $637 7 million or 46% of revenue.
Speaker Change: Our diluted share number was $319 9 million shares resulting in a diluted earnings per share number for the quarter of $1 99 up 39% from the prior year.
Speaker Change: Now turning to the balance sheet cash.
Speaker Change: Cash cash equivalents and investments ended the quarter at approximately $545 billion during the quarter, we repurchased $62 $7 million of our common stock and in April we repurchased an additional 82 million for a total of $144 $7 million at an average price of $269 eight.
Speaker Change: <unk> per share.
Speaker Change: We have now completed share repurchases under our existing $1 billion board authorization, whereby we repurchased eight 5 million shares at an average price of $117 20 per share.
Speaker Change: In May 2024, our board of directors authorized a new $1 $2 billion stock repurchase program, which commences in May 2024, and expires in May 2027.
Speaker Change: The actual timing and amount of future repurchases will be dependent upon market or business conditions stock price and other factors.
Speaker Change: Now turning to operating cash performance for the first quarter, we generated approximately $513 $8 million of cash from operations in the period, reflecting strong earnings performance, partially offset by ongoing investments in working capital.
Speaker Change: <unk> came in at 62 days up from 61 days in Q4, driven by significant end of quarter service renewals.
Speaker Change: Inventory turns were one <unk> flat to last quarter inventory increased slightly to $2 billion in the quarter up from $1.9 billion in the prior period.
Speaker Change: Reflecting the receipt of components from our purchase commitments and an increase in switch related finished goods.
Speaker Change: Our purchase commitments at the end of the quarter of $1.5 billion down from $1 $6 billion at the end of Q4.
Speaker Change: We expect this number to level off as lead times continue to improve but will remain somewhat volatile as we ramp up new product introductions.
Speaker Change: Our total deferred revenue balance was 1.663 billion up from 150 $6 billion in Q4 fiscal year 2023.
Speaker Change: The majority of the deferred revenue balance is services related and directly linked to the timing in term of service contracts, which can vary on a quarter by quarter basis.
Speaker Change: Our product deferred revenue balance decreased by approximately $25 million versus last quarter. We.
Speaker Change: We expect 2024 to be a year of significant new product introductions, new customers and expanded use cases. These trends may result in increased customer specific acceptance clauses and increase the volatility of our product deferred revenue balances.
Speaker Change: As mentioned in prior quarters, a deferred balance can move significantly on a quarterly basis independent of underlying business drivers.
Speaker Change: Accounts payable days were 36 days down from an unusually high 75 days in Q4, reflecting the timing of inventory receipts and payments.
Speaker Change: Capital expenditures for the quarter were $9 $4 million.
Speaker Change: Now turning to our outlook for the second quarter and beyond.
Speaker Change: I've now had a quarter working with Jay Sri the leadership team and the broader Arista ecosystem and I am excited about both our current and long term opportunities in the markets that we serve.
Speaker Change: Passion for innovation, our agile business operating model and employee commitment to our customer success are foundational.
Speaker Change: We are pleased with the momentum being demonstrated across the segments enterprise cloud and providers with this we are raising our revenue guidance to an outlook of 12% to 14% growth for fiscal year 2024.
Speaker Change: On the gross margin front, given the expected end customer mix combined with continued operational improvements we remain with a fiscal year 2024 outlook of 62% to 64%.
Speaker Change: Now turning to spending and investments we continue to monitor both the overall macro environment and overall market opportunities.
Speaker Change: Inform our investment prioritization as we move through the year.
Speaker Change: This will include a focus on targeted hires in leadership roles R&D and go to market team as we see opportunities to acquire strong talent.
Speaker Change: On the cash front, while we will continue to focus on supply chain and working capital optimization. We expect some continued growth in inventory on a quarter by quarter basis, as we received components from our purchase commitments.
Speaker Change: With these sets of conditions and expectations our guidance for the second corner, which is based on non-GAAP results and excludes any noncash stock based compensation impacts and other nonrecurring items is as follows.
Speaker Change: Revenues of approximately $1 16 to $1 $65 billion gross margin of approximately 64% and operating margin at approximately 44%.
Speaker Change: Our effective tax rate is expected to be approximately 21, 5% with diluted shares of approximately 325 million shares.
Speaker Change: I will now turn the call back to <unk> for Q&A. Thank.
Speaker Change: Thank you Chantelle, we will now move to the Q&A portion of the Arista earnings call to allow for greater participation I'd like to request that everyone. Please limit themselves to a single question. Thank you for your understanding operator take it away.
Speaker Change: Thank you.
Speaker Change: 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 and then the number one on your telephone keypad.
Speaker Change: If you would like to withdraw your question prestige star and the number one again.
Speaker Change: We ask that you. Please pickup your handset before asking questions in order to ensure optimal sound quality.
Speaker Change: And your first question comes from the line of Akshay <unk>.
Akshay: <unk> with Citi. Your line is open.
Speaker Change: Hi, it's Adrian for <unk>, Thanks for taking the question.
Adrian: I was hoping you could comment on your raised expectations for the full year with regards to customer mix. It sounds like from your gross margin guidance, you're seeing a higher contribution from enterprise, but I was hoping you could comment on the dynamics Youre seeing with your cloud Titans. Thank you.
Speaker Change: Yeah. So I Shantallow Nice described you know when we when we gave out guidance in November we didn't have much visibility beyond three to six months and so we had to go with that the ax.
Speaker Change: Tim any in Q1 alone and I believe it will continue in the first half has been much beyond what we expected and this is true across all three sectors cloud NII Titans providers and enterprise. So we're feeling good about all III and therefore have raised our guidance earlier than we probably would have done in may I think we would've I.
Speaker Change: Do you like to look at two quarters Shunto, what do you think but I think we felt good enough yeah. No I think we saw because of the diversified momentum and the mix of the momentum that gives us confidence.
Speaker Change: Great. Thanks.
Speaker Change: Thanks Adrian.
Speaker Change: Okay.
Speaker Change: And your next question comes from the line of stomach Chatterji with J P. Morgan Your line is open.
Samik Chatterjee: Hi, Thanks for taking my question I guess, a gesture of trying to.
Samik Chatterjee: I appreciate the sort of raising the guidance for the full year sure.
Samik Chatterjee: But when I look at it on a half over half basis in terms of what you're implying.
Samik Chatterjee: I am doing the math correct, you're implying about a sort of five 6% half over half growth, which when I go back and look at previous yours, Youre, probably theres only one year out of the last five or six that you've been in that sort of range or below that nobody, although george being better than that I'm. Just wondering you mentioned the Q1 activity that you've seen across the board why are we not seeing it.
Speaker Change: A bit more of a half over half uptick than in sort of the momentum in the back half. Thank you.
Speaker Change: Okay.
Speaker Change: Thanks, you know, it's like anything else the numbers are getting larger and larger so activity has to translate to larger numbers. So of course as we see it improve even more we will guide appropriately for the quarter, but at the moment, we're feeling very good just increasing our guys from 10 to 12 to 12 to 14.
Speaker Change: As you know Arista doesn't traditionally do that so early in the year.
Speaker Change: So please read that as confidence, but you know cautiously.
Speaker Change: Confident or optimistic confident but nevertheless confident.
Speaker Change: Thank you.
Speaker Change: Okay.
Speaker Change: And your next question comes from the line of Ben Reitzes with Melius Research. Your line is open.
Benjamin Alexander Reitzes: Then if you're talking we can't hear you.
Benjamin Alexander Reitzes: Operator.
Benjamin Alexander Reitzes: Ben.
Speaker Change: So we will move onto the next question Mr. Right. Just if you can hear US Please re hit star one.
Speaker Change: And we will move to our next question from George Notter with Jefferies. Your line is open.
George Charles Notter: Hi, guys. Thanks, a lot.
George Charles Notter: I want to key in on something I think you guys said earlier in the monologue, you mentioned that Ethernet was 10% better than Infiniband and <unk>.
George Charles Notter: My notes are incomplete here could you just remind me exactly what you were talking about there what is the comparison you are making to infiniband and.
Speaker Change: Just anything I'd love to learn more about that.
Speaker Change: Okay George.
Speaker Change: Historically as you know when you look at Infiniband and Ethernet in isolation.
George: There are a lot of advantages of each technology traditionally infiniband has been considered lossless and Ethernet is considered to have some loss properties. However, when you actually put a full GPU cluster together, along with the optics and everything what and you look at the coherence of the job completion time across all package sizes.
George: Data has shown that this is data that we have gotten from third parties, including Broadcom that just about in every package size in a real world environment independent of the you know comparing those technologies. The job completion time of Ethernet was approximately 10% faster. So you can look at these things in silos that you can look at it in a practical cluster and in them.
George: Practical Costa we are already seeing improvements on Ethernet now don't forget. This is just Ethernet as we know it today once we have the ultra Ethernet consortium and some of the improvements youre going to see on packet spring and dynamic load balancing and congestion control I believe those numbers will get even better.
Got: Got it I assume you're talking about rocky here as opposed to just straight straight up Ethernet is that correct.
George: In all cases, right now are pretty at UEC, we're talking about a R. D. M. L. Ethernet exactly Iraqi version, two which is the most weight employed Nick you have in most scenarios.
Speaker Change: But yeah.
Speaker Change: With kind of a rocky missing 10% improvement imagine when you go to U V C.
Speaker Change: I know you guys are also working on your own version of.
Speaker Change: Of Ethernet.
Speaker Change: Presumably it blends into the UEC standard over time, but what do you think the differential might be there relative to incentive and do you have a sense on what that might look like.
George: I think they have metrics, yet, but it's not like we're working on our own version of our Internet. We are working on the UEC compatible and compliant version of Ethernet and there's two aspects of it and what do we do on the switch and what others do on the Nick right. So what we can do on the switch I think will be you know we've already built and architecture, we call. It the easy link architecture.
George: It takes into consideration the buffering the congestion control the load balancing and largely we will have to make some software improvements. The next especially at 408 hundred is where we are we are looking to see more improvements because that'll give us additional performance from the server onto the stage. So we need both hubs to work together.
Speaker Change: Thanks George.
George Charles Notter: Thank you.
Speaker Change: Thanks George.
Speaker Change: And your next question comes from the line of Ben Reitzes with Melius Research. Your line is open.
Benjamin Alexander Reitzes: Gosh I hope it works at this time.
Benjamin Alexander Reitzes: We can hear you now.
Benjamin Alexander Reitzes: Oh, great. Thanks, a lot I was wondering.
Benjamin Alexander Reitzes: If you can categorize kind of characterize how youre seeing nvidia in the market right now and.
Benjamin Alexander Reitzes: Are you seeing yourselves go more head to head how do you see that evolving and if you don't mind also.
Benjamin Alexander Reitzes: I think video needs to more systems based approach potentially with Blackwell.
Benjamin Alexander Reitzes: How do you see that impacting your competitiveness with Nvidia. Thanks, so much.
Speaker Change: Thanks, Ben for a loaded question first of all I want to thank and video and Jensen I think it's important to understand that we wouldn't have a massive AI networking opportunity if nvidia didn't build some fantastic Gpus. So yes, we see them in the market all the time, mostly using our networks to their gpus and and in video as the market leader.
Speaker Change: Are there and I think they've created an incremental market opportunity for us that'd be a very very rejoiced by now did you see them in the market of course, we do I see them on Skus. We also see them on the you know Rocky R. R. D M a Ethernet Nic side.
Speaker Change: Sometimes when you see them, obviously, when when Theyre pushing infiniband.
Speaker Change: Which has been for most part the de facto network of choice you might have heard me say last year or the year before I was outside looking in to this AI networking, but today, we feel very pleased that we are able to be the scale of network or in videos Gpus and next based on Ethernet.
Speaker Change: We don't see a nvidia as a direct competitor yet on the Ethernet side I think it's 1% of that business, it's 100% of our business. So we don't worry about that overlap at all and we think we've got you know 20 years of founding to now experience to make our Ethernet switch and better and better both on the front end and back end.
Benjamin Alexander Reitzes: So we're very confident that arithmetic and build the scale of that book and work with Nvidia scale up Gpus.
Speaker Change: Thank you Ben Thanks, a lot.
Speaker Change: And your next question comes from the line of Amit <unk> with Evercore ISI. Your line is open.
Amit: Good afternoon. Thanks for taking my question I guess.
Amit: Jesse given some of the executed transition you're seeing at Arista.
Amit: Perhaps talk about conviction you can the discussion you've had with the board around your desire or your commitment to remain to see if there's any you can touch on that that would be really helpful.
Amit: If I just go back to this drop completion data that you talked about given what you just said unexpected improvement what are the reasons a customer would still use infiniband versus switch more aggressively towards Ethernet. Thank you.
Amit: Well first of all you know you heard our onshore Oh, I'm, sorry to see onshore decided to do other things I hope. He comes back we've had a lot of executives you know make a U turn overtime and we called them Boomerangs. So I certainly hope that's true with onshore.
Amit: But we have a very strong bench.
Amit: And we've had we've been blessed to have a very constant bench for the last 15 years, which is very rare in our industry and in the Silicon Valley. So well, we're starting to see onshore and make a personal decision to take a break.
Amit: We know he'll remain a relevant share and we know that the bench strength.
Amit: <unk> Hello, onshore will now step up to do greater things asphalt Mike Amendment.
Amit: To the board I have committed for multiple years I think it's the wrong order I wish I answered that state and I'd retired, but I'm committed to staying here for a long time.
Speaker Change: Thank you.
Amit: Yeah.
Speaker Change: And your next question comes from the line of Ann.
Ann: <unk> she gave them with New Street research your line is open.
Ann: Okay. Thank you so much for taking my question.
Ann: Just as soon as they get introduced in network computing capabilities with any switch performing some calculation inside the switch itself, perhaps no. He's not the best time to announce new products, but I am curious about whether this is something the brother merchant silicon and it turned out because instead of them couldn't introduce at some point.
Speaker Change: And finally are you asking what is that new products 40, I got the question.
Speaker Change: No I'm asking specifically about in network capabilities.
Amit: The switch can do some matrix multiply them inside the switching.
Amit: And I was wondering if this is something that the brother merchant silicon Ethernet ecosystem could introduce as well.
Amit: Yeah.
Speaker Change: So just for everyone else's benefit a lot of the in network compute is generally done is closest to the compute layer as possible and you know where you are processing. The GPU. So that's a very natural place I don't see any reason why we could not do those functions in the network and offload the network for some of those compute functions. It would require a little more states and.
Speaker Change: And processing power et cetera, but it's certainly very doable.
Amit: I you know I think it's going to be six of one half a dozen of the other somewhat prefer it closest to the compute layer and somewhat like a network wide for Netflix scale at the network layer. So the feasibility is very much there in both cases.
Amit: Antoine.
Antoine: Thanks, everyone. Thank you.
Speaker Change: Thank you.
Speaker Change: And your next question comes from the line of James Fish with Piper Sandler Your line is open.
James Edward Fish: Hey, Thanks for the question.
James Edward Fish: Onshore, we will Miss having you around echo my sentiments, there, but hopefully see you soon.
James Edward Fish: Phase III, how are you guys thinking about timing of the 800 gig optics availability versus kind of use some systems and you keep alluding to kind of nexsan product announcements for multiple quarters now not just this one but should we expect this to be more around adjacent use cases, the core including AI or software kind of pick up in the product roadmap correction of accounts.
Speaker Change: Yeah, Yeah, you know James Youre My German like.
Speaker Change: Like Deja Blue we've had similar discussions on 400 gig too.
Speaker Change: And as you well know in order to build a good switching system.
Speaker Change: You need an ecosystem around it but that's the next the optics the cables and accessories. So I do believe you'll start seeing some early introduction of optical and switching products for 800 gig, but to actually build the entire ecosystem and take advantage, especially of the next I think we'll take more.
Speaker Change: Than you are.
Speaker Change: So I think I'm, probably more into 'twenty, five or even 'twenty six.
Speaker Change: That being said I think youre going to see a lot of systems and I had this discussion earlier, you're going to see a six to one and a half a dozen of the other you're going to see a lot of systems, where you can demonstrate high ratings scale with 400 gig and go east to west much wider and those large clusters that are in the tens of thousands.
Speaker Change: And then once you need once you have gpus that source for 800 gig, which even some of the recent Gpus don't then you'll need not just higher ratings, but higher performance. So I don't see the the ecosystem off 800 gig you know limiting the deployment of <unk> networks, that's an important thing to remember.
Speaker Change: Thanks, James Thank you thank.
Speaker Change: Thank you James.
Speaker Change: And your next question comes from the line of Simon Leopold with Raymond James Your line is open.
Speaker Change: Yeah.
Speaker Change: Hi, guys. This is Victor Chu in for Simon Leopold.
Victor Chu: Do you expect a risk.
Victor Chu: To see a knock on effect from networking in the front end or the.
Victor Chu: As customers eventually deploy more AI workloads based.
Victor Chu: I'm, sorry bias towards emerging and then maybe help us.
Speaker Change: I understand how we might be able to size. This if that's the case Simon.
Victor Chu: Simon that's a question we haven't.
Victor Chu: And to consideration.
Simon Matthew Leopold: This phase III production, but you're absolutely right to say as you have more back end than the backend has to connect to something which typically rather than reinventing IP and adaptive routing you would connect to the front end of your compute and storage and Wan networks. So while we do not take that into consideration in our <unk>.
Victor Chu: $150 million projection in 2025, we naturally see the deployment of more backend cost us, resulting in a more uniform compute storage memory. You know overall front end back end holistic network for AI humming in the next phase.
Victor Chu: So I think it makes a lot of sense, but I just wanted to get the cluster is deployed and then we'll do that a lot of our customers are fully expecting that holistic connection and that's one pfizer being one of the reasons. They look so favorably at us. They don't want to build is just disparate silos and islands of AI clusters, they really want to bring it in terms of a full year.
Speaker Change: Reform AI data center.
Speaker Change: Thanks, so much.
Speaker Change: And your next question comes from the line of meta Marshall with Morgan Stanley. Your line is open.
Meta A. Marshall: Great. Thanks, maybe I'll flip James's question, and just kind of as you know what do you see as kind of some of the bottlenecks from going to from pilots to ultimate deployments you know it sounds like it's not necessarily 800 gig and so you know is it just a matter of time are there other pieces of the ecosystem.
Speaker Change: That are that need to fall into place before some of those deployments can take place. Thanks.
Speaker Change: I wouldn't I wouldn't call them bottlenecks I would definitely say, it's a time based and familiarity based situation you know the cloud everybody knows how to deploy that it's sort of you know plug and play in some ways.
Speaker Change: But even in the cloud if you may recall there were many use cases that emerge. The first use case, that's emerging point, yeah networking is let's just build the fastest training workloads in clusters.
Speaker Change: And you know they're looking at performance power is a huge consideration the cooling of the Gpus is a huge part of it yeah, you wouldn't be surprised to hear a lot of times, it's just waiting on the facilities and waiting for the infrastructure to be set up right, then there'll be Oh, EFS and operating side and Ken has been quietly I'd love for them to try to chime in but does it.
Speaker Change: Tremendous amount of foundational discovery that goes into what do they need to do in the cluster you know do they need to do some fashion and they don't need to do load balancing do they need to do this at layer two layer three do they need visibility features do.
Speaker Change: Do they need to connected across the Wan or interconnect so and of course as you rightly pointed out there's the whole 408 hundred were but we're seeing less of that because a lot of it is familiarity and understanding how to operate the cluster with the best job completion time and visibility manage manageability and availability of the Gpus nobody.
Speaker Change: Can tolerate downtime can I'd love to hear your point of view on this yes. Thanks T III.
Speaker Change: I think that.
Speaker Change: So I'm walking People's deployment is the availability of all the pieces and so there's a huge pent up demand for this stuff and we see these clusters getting built as fast as people can build the facilities get the Gpus.
Speaker Change: And get the network since they need.
Speaker Change: I think that you know we're.
Speaker Change: Extraordinarily well positioned here because we've got years of experience building scale storage clusters of some of the world's largest cloud players and storage clusters are not technical day only clusters, but they have some of the same issues with managing a massive scale backend network that needs to be covered.
Speaker Change: Load balanced neither a lot of buffer to maintain two managed first and so and then some of the congestion management stuff. We've done there is also useful in AI networks and in particular this independent down topic keeps coming up and I just point to point out that Ethernet is about 50 years old.
Speaker Change: Over those 50 years Ethernet has come head to head with the public.
Speaker Change: A bunch of technologies like token ring.
Speaker Change: Solid ATM pretty hippy scalable because youre in interconnect.
Speaker Change: Marinette [laughter].
Speaker Change: And all of these barrels have one thing in common.
Speaker Change: Ethernet one.
Speaker Change: And the reason why isn't it one is because of mid caps law, but the volume of network is quite broad and can the number of nodes.
Speaker Change: Interconnect and so anybody who tries to do something which was not Ethernet historically off with a very large quadratic disadvantage and any temporary advantage. They have because of the some detail. The tech cycle was going to be quickly overwhelmed by the connectivity get done this year with Ethernet and so I think you know exactly how many years it takes for Infiniband to gold waiver fibre channel.
Speaker Change: I'm not sure, but that's where it's all headed.
Speaker Change: Thank you guys. Thank you.
Speaker Change: And your next question comes from the line of Ben Bollin with Cleveland Research Company. Your line is open.
Benjamin James Bollin: Good afternoon, everyone. Thanks for taking the question.
Benjamin James Bollin: G. III you made a comment that back.
Benjamin James Bollin: Back when you had guided in November you had about.
Benjamin James Bollin: About three to six months of visibility.
Benjamin James Bollin: Could you take us through what type of visibility you have today, and maybe compare and contrast, the different subsets of customers and how they differ thank you.
G: Thank you Ben that's a good question. So let me take it by category like you said in the cloud and AI Tightens in November you know, we're really searching for them in three months visibility six would have been amazing today, you know I think after a year of tough tough situations for us where the cloud Titans were pivoting rather.
G: Rapidly to AI and not thinking about the cloud as much. We're now seeing a more balanced approach, where they're still doing AI, which is exciting but they're also expanding their regions on the cloud. So I would say our visibility has now improved to at least six months and maybe it gets longer.
G: As time goes by on the enterprise.
Speaker Change: No I'm not I'm, not a bellwether for macro but everybody else's siding macro, but I'm not seeing macro what we're seeing with Chris Schmidt and Ashwin and the entire team is you know a profound amount of activity in Q1.
Speaker Change: Better than we normally see in a Q1 Q1 is usually you know come back from the holidays January slow, there's some east coast storms to deal with winter is still strong, but we have had one of the strongest activities in Q1, which leads us to believe that it can only get better for the rest of the year, Hence the guide increase from an otherwise.
Speaker Change: Seven a team of Santana and myself right and then the tier two cloud providers I wanted to speak to them for a moment because not only are they are strong for us right now, but they are starting to pick up some AI initiatives as well so they're not as largest causes the cloud Titans, but the combination of the service providers in the tier two specialty providers is.
Speaker Change: Also seeing some momentum so overall I would see how visibility has now improved to from three months to over six months and in the case of the enterprise obviously, our sales cycles can be even longer. So it takes time to convert into wins, but the activity has never been higher.
Speaker Change: Thanks Ben.
Speaker Change: You.
Speaker Change: And your next question comes from the line of Michael <unk> with Goldman Sachs. Your line is open.
Michael: Hey, good afternoon, and thank you very much for the question. It was very encouraging to hear about the migration of trials pilots.
Michael: With its production rollout to support Gpus in the range of I think you said 10000 to 100000 Gpus for 2025.
Michael: First I was just wondering if you could talk about some of the key determinant about where we how we end up in that range high end versus low end.
Michael: Second point.
Michael: Assuming $250000 per GPU that would imply about $25 billion of compute spending.
Michael: In its target of $750 million would only be about 3% of the high end and I think you've talked about 10% to 15% networking as a percentage of compute historically. So I was just wondering if you could talk about what I may be missing there if there is.
Speaker Change: Anything to call out.
Speaker Change: Assumptions. Thank you.
Speaker Change: Yeah. Thank you Michael I think they could do better next year, but your point is well taken that in order to go from 10 thousands of Gpus to you know 30 1500 thousand a lot of things have to come together first of all let's talk about the datacenter our AI center facility itself, there's a tremendous amount of work.
Speaker Change: Org and lead time that goes into the power the cooling facilities and so now when youre talking to this kind of production as opposed to proving something in the lab and that's that's a key factor. The second one is the GPU the number of Gpus the location of the Gpus the scale of the Gpus the locality of these gpus.
Speaker Change: They go with Blackwell Osha, they've built build with you know our scale up inside the server or scale out to the network. So the whole center of gravity Whats nice to watch which is why we're more constructive on the 2025 numbers is that the GPU lead times have significantly improved which means more and more of our customers will get more gpus, which.
Speaker Change: Which in turn means they can build out the scale of our network, but again a lot of work is going into that and the third thing I would say is this.
Speaker Change: Scale the performance how much radix they want to put in and then I'll give a quick analogy here, we ran into something similar on the claws. When we've been talking about for weight CMP or eight waste in U C. N P or Israel based designs as is often called and the number of Nicks you connected to go eight way, a four way or 12 way our surcharge.
Speaker Change: Off and go to 800 gig the performance and scale will be the third metric. So I think power GPU locality and performance of the network are the three major considerations that allow us to get more positive on the rate of production in 2025.
Speaker Change: Thanks, Michael.
Speaker Change: Yeah.
Speaker Change: And your next question comes from the line of Matthew <unk> with Deutsche Bank. Your line is open.
Matthew: Hey, Thanks, so much for taking the question I got to ask one more on AI, sorry to beat a dead horse, but as we think about the stronger start to the year and the migration from trials to pilot specific in relation to AI is there a ramp towards getting to that 750 mill next year and I guess more importantly is there any material.
Matthew: Contribution baked into this year's outlook or is there any contribution that may be driving the two percentage point increase relative to the prior guide for 'twenty four thanks Chantelle.
Chantelle: I can tell you want to take that I've been talking up AI a lot I think you said yeah I can take this AI question. Yeah. So I think that when you think about the 750 million target that has become more constructive. The G. III is prepared remarks, that's a glide path. So it's not zero and 24, its a glide path to 'twenty five.
Speaker Change: So I would say there is some assumed in the sense of it's a glide path, but it will it will end in 2025 of the 750 and of course, that's not a hockey stick.
Speaker Change: It's not zero this year, yes, Matt for sure yes.
Speaker Change: Okay. Thank you.
Speaker Change: And your next question comes from the line of Sebastian <unk> with William Blair. Your line is open.
Speaker Change: Yeah.
Sebastian: Thanks, Good afternoon, I've got a non AI question here so.
Sebastian: Maybe you can talk a little bit about some of the incremental investments that you're making within your go to market. This year, particularly as you look to grab some share from competitors.
Sebastian: All of them are going through some type of disruption in one or the other acquisitions et cetera, and then what you might be doing with the channel partners to land more of those mid market customers as well.
Sebastian: Yes, Sebastian we're probably doing a little more on investment than we have done enough progress on channel partners to be to be honest.
Speaker Change: But last couple of years, we were getting very apologetic about our lead times our lead times have improved so we have stepped up our investment on go to market, where I'm expecting Chris Schmidt on Ashwin team to grow significantly and judging from the activity they've had and the investments they've been making in 'twenty four 'twenty three and 'twenty four.
Speaker Change: We're definitely going to continue pedal to the metal on that I think our investments in AI and cloud Titans remain about the same because you know while there is a significant technical focus on the systems engineering and product side.
Speaker Change: We don't see a significant change on the go to market side and on the channel partners I would say you know where this really comes to play and this will play out over multiple years, it's not going to happen. This year is on the campus.
Speaker Change: Today, our approach on the campus is really going after a larger enterprise customers. We got 9000 customers, probably 2500, that'd be really gonna targets and so our mid market is more targeted at specific verticals like health care education public sector, and then we appropriately work with the channel partners in India.
Speaker Change: Our region in the country to deal with that.
Speaker Change: And to get to the first billion I think this will be a fine strategy as we aim beyond $750 million to $1 billion and we need to go to the second billion absolutely we need to do more work on channels. This is still work in progress.
Speaker Change: Thanks, so much thank you.
Speaker Change: And your next question comes from the line of Aaron Rakers with Wells Fargo. Your line is open.
Aaron Christopher Rakers: Yes, thanks for taking the question, so I'm going to shift gears away from me actually.
Aaron Christopher Rakers: Victory, if we look at the if we look at the server market over the past handful of quarters, we've seen unit numbers down quite considerably I'm curious as you look at some of your larger cloud customers. How you would characterize the traditional server side and whether or not youre seeing signs of them moving past this kind of optimization phase and weather.
Speaker Change: Or not you think a server refresh cycle in front of you it could be incremental catalysts to the to the company.
Speaker Change: Yeah, No I think.
Speaker Change: If you remember there was this one dreadful yeah, where we were at one of our customers skipped a server cycle.
Speaker Change: But generally speaking on the front end network now we're going back to the cloud and we do see server refresh and service cycles continue to be in the three to five years.
Speaker Change: For performance upgrades, they liked III, but occasionally some of them may go little higher so absolutely. We believe there will be another cloud cycle because of the silver refresh and the associated use cases, because once you do that on the server there's appropriately the regional spine and then the data center interconnect in the storage and so many.
Speaker Change: So much ripple effect from that survey use case upgrade that kind of compute and CPU is not changing its continuing to happen. In addition to which we're also seeing more and more regional expansion.
Speaker Change: New regions are being created and designed and outfitted for the cloud Bye Bye Amit you tightened.
Amit: Yes. Thank you.
Speaker Change: Darren.
Speaker Change: And your next question comes from the line of Karl Ackerman with BNP Paribas. Your line is open.
Speaker Change: Yeah.
Karl Ackerman: Yeah. Thank you.
Karl Ackerman: Sure you spoke about how you are not seeing any slowness in enterprise.
Karl Ackerman: Curious whether that is being driven by the growing mix of your software revenue and do you think the deployment of <unk> networks on Prem.
Karl Ackerman: Can be a more meaningful driver for your enterprise and financial customers in the second half of fiscal 'twenty four oral that'd be more of a fiscal 'twenty five of them. Thank you.
Speaker Change: Oh, that's a really good question and I have to analyze some more I would say our enterprise activity is really driven by the fact that Ken has produced some amazing software quality and innovation.
Karl Ackerman: And we have a very high quality universal topology, but you don't have to buy five different noises and 50 different I'm images and operate this network with thousands of people. It's a very elegant architecture that applies to the data center use case that you just outlined for the spine.
Karl Ackerman: And universal spine can apply to the campus it applies to the wide area. It applies to the brand should apply for security it applies to visibility and you bring up a good point that while the enterprise use cases for AI are small we are seeing some activity there as well relative to the larger I Titans, they're still very small but think of them as bad.
Karl Ackerman: Back in the trial phase that was describing hernia trials pilots production. So a lot of our enterprise customers are starting to go into the trial phase of GPU clusters. So that's a nice use case as well, but the biggest ones are still in the data center campus and the general purpose enterprise.
Speaker Change: Yes. Thank you.
Speaker Change: Operator, we have time for one last question.
Karl Ackerman: Thank you and your final question comes from the line of David vote with UBS. Your line is open.
David Vogt: Great. Thanks, guys and congratulations J C. We have a question about I want to go back to AI, the roadmap and the deployment schedule for Blackwell. So it sounds like it's a bit slower than maybe initially expected with initial customer deliveries late this year.
David: How are you thinking about that in terms of your roadmap, specifically and how that plays into what Youre thinking about 25, and a little bit more detail and does that lead delivery, maybe put a little bit of a pause on maybe some of the cloud spend and the <unk>.
David: Fall of this year as there seems to be somewhat of a technology transition going on forge Blackwell away from the legacy product. Thanks.
Speaker Change: Yeah, we're not seeing a pause yet I don't think anybody's going to wait for Blackwell necessarily in 2024, because there's still bringing up their GPU clusters, and you know how our customers divided across multiple tenants the choice of host memory storage architectures optimizations on the GPU.
Speaker Change: Our collective communication libraries specific workloads Brazilians visibility all of that has to be taken into consideration all of this to say a good scale out network has to be built no matter, whether you're connecting to today's gpus are future blackouts.
David Vogt: So they're not going to parse the network because I because they're waiting for Blackwell, they're gonna get ready for the next book, whether it connects to our Blackwell or our current.
David Vogt: Each $100, so as we see it.
David Vogt: The training workloads and the urgency of getting the best job completion time is so important that they're not going to spare any investments on the network side and the Netflix side can be ready no matter what the GPU is.
Speaker Change: Thanks, David This concludes the Arista Networks' first quarter 2024 earnings call. We have posted a presentation, which provides additional information on our results, which you can access on the investors section of our website. Thank you for joining us today and thank you for your interest in Arista.
Speaker Change: Okay.
Speaker Change: Ladies and gentlemen, thank you for joining this concludes today's call and you may now disconnect.
Speaker Change: Please wait the conference will begin shortly.
Speaker Change: [music].
David Vogt: Yeah.
David Vogt: Yeah.
David Vogt: Yes.
David Vogt: [music].
David Vogt: Yeah.
David Vogt: [music].