ANET: Conference call review

Hi all, part of my (new) process is reviewing conference calls, rewriting questions in my own words, and trying to get a longer view. So hope this is useful.

ANET is one of my favourite companies. My personal suspicion is that the legal issues were a bigger overhang than they let on, taking up valuable engineering resources and delaying a bunch of orders. So getting rid of Cisco, even at a cost of $400m is worthwhile.

Some things:

  1. Aristas deferred revenue is a. Services, short and long-term, recognised when service is provided, and b. Product, “acceptance clauses with customers and the delivery of new features… timing of that depends on when you ship and it depends on when the customers accept”.

I was always a bit confused about Arista’s deferred revenue.

  1. A quote from Q32017 “next two years that [software subscriptions] can become approximately a 5% of our revenue”. This is definitely worth watching.

  2. Wide spread across the five verticals, solid continuing demand.

  3. 100g will continue to grow over next several years, even once 400g becomes available.

  4. Good Q22018 results in 100g → “competition did not respond to 100g as well as we did”.

Ultimately, ANET has stayed very on-message since Q3-2017. They only report CloudVision licenses at end-of-year which will be a critical bit of information.

My personal viewpoint (FWIW) is that Arista’s special sauce is underrated. I like this answer from Q32017 about what Arista actually does:

…Everyone gets same merchant silicon… “but somehow, we’ve been able to get more routes, more data paths, more convergence, more scale, more services out of it.”…“we’ve been doing across three generations, Petra, Arad and Jericho to really maximize the feed, speed, scale, and performance. And I cannot underscore that enough.”

“we complement that with our own co-processors, silicon drivers, et cetera, and a set of capability that the silicon may not have to enhance it. And again, this is very important, because in routing, there’s routing features, and then, there’s use cases that we go into. And so, we’ve been going into some very specific use cases in the content, media, cloud, routing aggregation, high-performance core, even some 5G wireless.”

Feel free to skip my notes below (copy into or to see pretty version)


Conference calls

Q3 2017

Our customers in the Americas made up 71% of our total revenue, while the international theaters contributed higher than normal at 29%

On deferred revenue:
“Our deferred revenue balance was $565.1 million, up from $554.5 million in Q2. This balance continues to be made up of short and long-term service contracts and product-related deferrals associated with acceptance terms and future deliverables. Our product deferred revenue declined slightly in the quarter, contributing to our revenue outperformance.”

GD: So they’ve been paid for stuff that they haven’t delivered, as well as pre-paid service contracts.

Guidance (non-GAAP):
“Revenues of approximately $450 million to $464 million; gross margin of approximately 63% to 65%, and operating margin of approximately 30% to 32%. Effective tax rate is expected to be approximately 27%, with diluted shares of approximately 80.1 million”.

Q: Cloud, routing, any scope for accelerated growth? [ie, modelling revenue growth]
A: “(customers) more of them want to replace routers with routing. And frankly, legacy vendors are on the wrong side of this trend”. Over 150 FlexRoute licenses.

GD: FlexRoute licenses are key I think, monitor this going forward.

Q: Deferred revenue, what goes into that?
A: Deferred revenue is 2 buckets.

  1. Services, short and long-term, recognised when service is provided.
  2. Product, “acceptance clauses with customers and the delivery of new features… timing of that depends on when you ship and it depends on when the customers accept”.

Q: Margins
A: …" it’s becoming a 30% to 32% operating margin business, and I think we’re comfortable with that, certainly, for the near term"

Q: 100g port prices declining, but huge revenue. So… lots of growth. Whats driving it?
A: Trend->“where the LAN, the WAN, and the Layer 2 and the Layer 3 are all coming together to be a highly elastic Universal Spine.”

Q: FlexRoute for web titans, anyone using it?
A: Blah blah, close relationship etc., “… FlexRoute license and routing in general, absolutely, some of our early adopters have been the cloud titans right from the beginning in 2016 and they continue to be major deployers.”

Q: R&D down for 2 quarters in a row, are you under-investing? Also… “your guidance implies a massive expansion of OpEx on a quarter-over-quarter basis”…[GD: I don’t quite get the logic here]. How can you get to those expense levels? [particularly in regards to head count].
A: “…there’s no question in our mind that we dedicate a lot of engineering resources to legal…”. R&D headcount growing consistently, lumpiness around prototype and NRE expenses [GD: whats NRE? Non-recurring expenses]

Q: Delays to roadmap based on legal?
A: Some.

Q: Cloud, big picture [GD: trying to model industry growth rates], compared to other companies?
A: 100g … “Arista is less affected by what other companies are is, frankly speaking, we’re on the right side of these trends”

Q: Discounting on routing products? Margins compared to switching?
A: " Routing, in general, tends to be fewer boxes and fewer ports, but higher margins. And switching is connecting to the servers and storage, so it tends to have more pricing pressure."

Q: Cloud pushed to Q4, doesnt that suggest acceleration all other things equal?
A: “… didn’t do as well in the cloud titan in Q3 as we did in Q2”, but balanced across verticals. Hence, not much change in numbers.

Q: “what verticals are helping offset some of this cloud certification overhang?”
A: “service providers, high-tech enterprise and cloud specialty providers”. “legacy fatigue (from those vendors)” ← [GD: this is interesting…]

Q: Software-only revenue and how meaningful it’s becoming? You mentioned over 150 FlexRoute software licenses and you also have CloudVision.
A: “We not only exceeded 150 routing customers, but we have now also crossed 200 CloudVision customers.”…“because it’s a software subscription over multiple years, I expect in the next two years that this business can become approximately a 5% of our revenue.

Q: New features in routing expanding TAM?
A: Everyone gets same merchant silicon… “but somehow, we’ve been able to get more routes, more data paths, more convergence, more scale, more services out of it.”…“we’ve been doing across three generations, Petra, Arad and Jericho to really maximize the feed, speed, scale, and performance. And I cannot underscore that enough.”

“we complement that with our own co-processors, silicon drivers, et cetera, and a set of capability that the silicon may not have to enhance it. And again, this is very important, because in routing, there’s routing features, and then, there’s use cases that we go into. And so, we’ve been going into some very specific use cases in the content, media, cloud, routing aggregation, high-performance core, even some 5G wireless.”

Q: International sales strong, what use-cases?
A: Use cases and country adoption is broad-based.

Q4 2017

Guidance: Revenues of approximately $450 million to $468 million, gross margin of approximately 63% to 65%, operating margin of approximately 32%. Our effective tax rate is expected to be approximately 20%, with diluted shares of approximately $81 million.

Q: Glide path 2018 full year revenue growth, mid-20s. Enterprise? and in particular, routing enterprise?
A: Q4 unusual bump in enterprise, but normal mix going forward. Routing is across all verticals.

GD: worth watching routing progress

Q: Competition with routing, cost and performance equivalent to Arista? And whats FlexRoutes advantage?
A: 7500R and 7280R are gold standard. Across multiple use-cases, they’re occupying multiple positions and places in the cloud for that. So there’s not a like for like comparison.

Q: Microsoft 16% revenue, why better than expected, and 2018?
A: Multiple use-cases for Arista products, and MSFTs success. 10% customer this year.

Q: Domestical/international growth rates for enterprise? 2018 tax rate?
A: 19-21% tax rate. re: Enterprise, "…“lots of new customers - and it was almost neck to neck with Q3, record number of million dollar customers in both Q3 and Q4. CloudVision is becoming a very important piece of the Enterprise customer purchase.”. 60/40 international/domestic.

Q: Revenue growth rates? And 400g timing?
A: Q1 40% yoy, mid 20s for whole year. 100g very long tail. 400g doesnt take over from 100g, more of both.

Q: Cloud titan international versus domestic?
A: “one of the things we noticed as a trend in 2017 that we think will continue in 2018 with all the cloud titans, is since they began a lot of their expansion early in the US, a lot of the later expansion will be in international sites. So we expect at least a 50-50, maybe even a higher contribution from international with our cloud titan”

Q: First quarter forecast assumes cloud titan catchup?
A: Q1 seasonally slow. Normal guidance.

Q: …pros and cons of Arista participating in the campus switching market. I assume that’s something you’re not that interested in, but just wanted to understand strategically how you would frame that.
A: “And what our customers are really pushing us to do is look at campus as an extension of the cloud”… “no intentions of participating at this time in any kind of traditional campus markets, we have every intention of disrupting the traditional campus market as currently defined.”

Q: International contribution bouncing around.
A: Good growth, faster than US.

Q: International, and US telcos.
A: 60% of our customer acquisition in 2017 was international. Size of deals is smaller. Service providers (US telcos), some big wins, 2018 important year, service provider better in 2018 than 2017.

Q: Deferred revenue going forward?
A: Product deferred balance back to normal.

GD: Check Aristas deferred revenue. What is that?

Q: Mid 20s growth, why massive decelaration from 40%?
A: Visibility, good solid growth rate for now. 50% was unexpected. “Hope is not a strategy”.

Q: $1.5b cash.
A: Too early re: capital allocation.

Q: Deferred. Whats driving it down?
A: " One is we’re closing out on features, et cetera, that we needed to deliver and on acceptances that we had in contract. The R Series is over a year and in the market now. So we’re seeing a lot of those close out. So we’re not generating as much deferred revenue because of that. I think secondly, we were still qualifying some of these designer rounds in the fourth quarter and that did affect the linearity of invoicing and stuff. I think the underlying business is solid. It’s more the timing of shipments billing and how it’s long to the deferred.

… So I think as we go forward, deferred revenue will be much less of a topic from a product perspective, and hence you’ll see the services deferred revenue continue to grow."

Q1 2018

“Given some tough 2017 comparables, we believe that the current consensus for the balance of the year, which calls for year-over-year growth in the mid-20% range, remains relevant.
Guidance (non-GAAP), revenue: $500m-$514m;gross margin of approximately 62% to 64%; operating margin of approximately 32% to 34%.

Q: 945 certification deferred revenue done and dusted?
A: Pretty much. Small bit left.

Q: AI?
A: Very early, experimental. Partners with Nvidia and Pure.

Q: Privacy/Trade wars, any change in customers spend behaviour?
A: … um… not really. Some smaller more localised builds by cloud companies.

Q: You said “Mid-20% growth” but you grew 40% in Q1. So second half decelaration? And ASC 606, $19m deferred revenue impact. Did that impact revenue in the quarter?
A: “the mid-20s is a reasonable way to think about the growth rate” for remainder of year. ASC 606 just impacted balance sheet.

Q: Cloud spenders reporting big beats in CapEx, so cap exp is growing in 2018. But you’re guiding lower for rest of 2018. Whats the gap?
A: Networking small part of their CapExp, and big spend a multi-year exercise, so hard to tell when the networking bit will hit.

Q: Change in mix of top-of-rack switches and (versus?) spine within web-scale customers?
A: Not much different.

Q: Verticals under versus over performing?
A: No.

Q: How about enterprise?
A: Clearly our number 2 vertical. We continue to do very very well.

Q: Where are we at in 100g cycle?
A: Not even in the first inning.

Q: Gross margin.
A: yeah, 62-65% for year.

Q: would you be able to share insight into opportunities, outlook and the current state of enterprise penetration for both routing and switching?
A: Switching better than routing. Would like to see us do better, but doing very well.

Q: Dumb question on cloud vertical margin
A: Cloud = lower margin

Q: Progress expanding customer base? International?
A: The M&E vertical is a particularly strong one for us…“Enterprises are risk averse and only in the last year, I would say, they have started to take Arista seriously”

GD: Whats M&E?

Q: Number of routing customers? And discuss backlog?
A: re: Backlog-> No. Year end FlexRoute > 200. Only going to give yearly numbers (but think we’re going to double).

Q: Mid-20s growth rate for whole year, or remainder?
A: Remainder.

Q: International expansion 80-120. Whats driving international?

GD: How was this calculated? 67% in US…and therefore?

A: Made lots of international investments 2015-2016. Starting to pay off. Mirrors top 5 verticals. Only Tier-2 service providers not so represented (more a US thing).

Q: Any change in pricing environment?
A: No.

Q: Incremental pressure from Cisco et al, or white boxes?
A: Normal aggression.

Q: You’re not seeing any incremental uptick of white box solutions by your customers?
A: I am not seeing any difference in competitive behavior due to white box. And white box adoption, or rather disaggregated EOS, I’m not seeing any shift or change.

Q: Winning replacement (ie, of competitors) router deals? Or new footprints?
A: …actual market has been shrinking… The new change is in the white area side, where more and more of the interfaces are moving to ethernet from traditional SONiC or T1 or ATM type of interfaces. In terms of new footprint versus existing, for us, it’s all new.

GD: Not sure I really understand this

Q: Campus?
A: Arista does not plan to participate in the traditional campus. As things evolve in the campus and they become more and more akin or aligned with Arista’s different value-add and differentiated capabilities, we’re open to that.

GD: Whats traditional campus given they announced entry into the campuse the next quarter]

Q: Cloud vertical, catch up spend in March? ie, any pent-up spend? ie, no ongoing pauses in spending?
A: Hurt by the certifications (legal stuff). But back to normal now.

Q: Cloud titan stickiness, loyalty? [GD: Dumb question]
A: Stickiness is with the software. “…tied to the operating system and their automation stack”

Q: “innovation you’re driving on the disaggregator software solutions with the hyperscalers” [GD: no idea what he’s talking about]. Road map of monetisation?
A: Not very many customers outside cloud titans.

Q: Number of 10% customers?
A: Only at end of year. But 4/5 verticals contributed to top 10 customers [GD: diversified]

Q: Days Sales Outstanding (DSO). Down to 39 days (all time lows), but cloud titans revenue up which suggests DSO should be up too. Whats the discrepancy?
A: Timing on billings. Mostly a linear business.

Q: Last year Q2, and this year Q1. Q3 and Q4 grow at 25% yoy, normal. But Q2 2017 was huge, what happened then (product related), and what now, slowdown into Q2 (lower than seasonal).

Revenue looks like:

Quarter Revenue TTM 𝝳 (q-1) 𝝳 (YoY) 𝝳 (TTM)
2016Q3 290.26m 1.047b 8.0% 33.4% 36.7%
2016Q4 327.97m 1.129b 13.0% 33.6% 34.8%
2017Q1 335.48m 1.222b 2.3% 38.5% 35.7%
2017Q2 405.21m 1.359b 20.8% 50.8% 39.5%
2017Q3 437.63m 1.506b 8.0% 50.8% 43.9%
2017Q4 467.87m 1.646b 6.9% 42.7% 45.8%
2018Q1 472.49m 1.783b 1.0% 40.8% 45.9%
2018Q2 519.85m 1.898b 10.0% 28.3% 39.7%

A: “… confluence of 7500, 100-gig and routing (and good cloud spend). And then enterprise kicked in”.

Q: M+A, inorganic strategies?
A: Being careful…

Q2 2018

Mojo Networks acquisition
Entry into campus market - “Customers have been asking us for some time”.
Days Sales Outstanding - 46 days up from 39.
Guidance 3rd Quarter:

  1. $540-$552 Revenue.
  2. Gross margin 63-65%
  3. Operating margin: 32-34%
  4. Tax rate: 21.5%
  5. Diluted shares: 81.1m

Q: Campus? And in particular, enterprise interest?
A: Lots of interest, want the common architecture across data center and campus. Material impact next year.

Q: Hyperscale demand? white box etc
A: Good and healthy, top in Q1 and Q2 and expect same in 2nd half. Nothing new.

Q: Mojo customers, revenue synergies?
A: Good alignment from technical architecture.

Q: Operating margins, campus build up will drive margins down?
A: Long term see 32-34% [GD: as guided]

Q: Peer stating cloud deals got pushed out? Arista too?
A: No

Q: 100g. And metrics?
A: 35-45% market share. 100g very strong, 400g later (2020…), but in tandem with 100g. Metrics depends on the vertical, cloud is all 100g, and enterprise early adopters. Arista very strong in 10,25,40, but 100g is much stronger.

Q: Wiring closet for campus? will see wiring closet switches? [GD: No idea what this is]
A: First approach Cognitive Management Plane and Spine. Phase 2, wireless endpoints or edges if required.

Q: Router business, not much mentioned.
A: Fully expect to double FlexRoute licenses [200 end of last year]. Service providers not doing so well with routing.

Q: Beat in an impressive way, but guidance the same. So softness for next half?
A: No particular softness. Deferred revenue is stabilised now. Happy with guidance.

Q: Transition cloud titan spend, new data centers, or expanding existing footprint?
A: Both, more 100g in both examples.

Q: Pricing?
A: No real changes. Competitive market.

Q: Mojo $100m revenue target.
A: Sounds very high!

Q: Seasonality shifted to first half of year (cloud titans have inventory stored?). Normally see more revenue at back end of year.
A: Q1 grew 40% yoy, but easier compare. Don’t think anything different in seasonality, so guidance stands. And we’re seeing healthy demand.

Quarter Revenue TTM 𝝳 (q-1) 𝝳 (YoY) 𝝳 (TTM)
2017Q1 335.48m 1.222b 2.3% 38.5% 35.7%
2018Q1 472.49m 1.783b 1.0% 40.8% 45.9%

Q: 400g position? 100g you were a long way out so got lots of market.
A: Jayshree - I thought we’d have sedate gains in 100g, but we did really well “competition did not respond to 100g as well as we did”. 400g - multiyear, making sure our architecture is 400g capable. Usually first to market, expect to be with 400g too.

Q: Operating margin guidance 400bp decline?
A: 32-34% operating margin is what we’ll grow into [GD: Around 30% now? TODO: remove $400m settlement from Q2-2018 numbers]

Q: Building and scaling campus channel?

Q: Normalising of 100g after strong sales in 2017?
A: Not normalising, lots of demand.


Just quickly to say, what an interesting and helpful post on the comments in the CC Sarksnz, thanks.

Wow, Greg! What a great Conference Call review. Getting all those Q&A’s is very very useful. Makes me think I should take another look at Arista. Just that I have too many positions right now.
Thanks very much for the post,

1 Like

As fyi … NRE is non-recurring engineering.

1 Like

I was liquidating my ANET position as the holdings became long term. I don’t have a lot of shares remaining, but I might just hold them longer than planned.