Sundry Photography
A bargain too good to ignore
Earnings season is in the process of getting started. In the dim past-say a year ago, it was a time to look forward to and speculate as to which companies would report a blow-out and whether reporting a blowout quarter might levitate the share price of the company reporting strong results to a further high. These days earnings season is something else with many investors and analysts worrying as to whether their holdings might see a share price implosion if results were not considered satisfactory or if company’s guided down based on elongating sales cycles and more large deal scrutiny. Even companies reporting strong results and who increase guidance have seen their shares shredded as investors focus on macro headwinds rather than trying to analyze the prospects for a given company. In other words, a typical bear market paradigm.
It can be difficult for both contributors and readers to try to navigate through times such as these in which so much is at risk, and in which volatility is at a high point, and in which often what is good is considered bad by investors and commentators. Even what I might consider to be defensive companies, at least in the tech space, have seen their valuations halved or more. And for those of us who are stock pickers, recent data suggests that correlation is rising with the performance of individual companies having less than usual impact on valuations and share prices.
There seems to be little doubt now that a recession is impending although its severity and duration are still unknown. Are there vendors in the IT space whose performance will be relatively unaffected by macro headwinds. In my opinion, one such vendor is Arista Networks (NYSE:ANET). While not really a desirable condition, the fact is that Arista has a backlog of more than one year, and cannot fulfill the orders it continues to receive at the cadence its customers desire. Probably Arista has managed the supply chain headwind better than its rivals in the space, but nonetheless, according to some anecdotal checks, some prospective customers have received delivery quotes for some Arista products that extend into 2024. A recession is likely to ease supply chain woes without greatly impacting Arista’s projected revenue growth, and it should enable the company to improve its gross margins as the necessity for expedite fees begins to wane. That, notionally, at any rate, is a reason to consider Arista shares in this toxic environment for IT vendors with a recession likely impending.
It has been two years since I last wrote about Arista Networks for Seeking Alpha. Unlike many other IT shares over the past 2 years, Arista has appreciated by about 50% Of course to do so revenues have had to increase by more than 60% and non-GAAP net income had to increase by almost 80%. The company has gained substantial market share in its relevant markets over that time, and continues to do so.
Arista is set to report earnings on 10/31, a scary date for some, perhaps. But I believe the earnings to be reported at that time will once again be more treat than trick. Will that propel the shares in the weeks after earnings? With the shares down by 11% ( as of 10/19) just in the last 30 days, I can be pardoned, if I hope another strong earnings report will have some positive impact on the shares-but it has taken quite a bit to move shares positively in this environment. That said, however, this is not an article to recommend the shares for one quarterly earnings report.
The investment case for Arista in my mind is a function of its continuing market share gains, the success of various components of its growing product portfolio, and a business model that should resonate with investors in this extremely difficult environment. It isn’t just that the company’s EV/S is less than 6X but that its free cash flow margin is nearly 30% and that its Rule of 40% metric has been and will remain in the mid-high 40 range, with some upside possible. In my mind that makes the shares a significant bargain.
Reviewing Arista’s recent financials
The headline numbers for Arista last quarter included revenue growth of 49%, to a level that was about 8% above the mid-point of prior guidance. Gross margins were 62%, also at or above the upper end of prior expectations, mainly because of lower expedite fees than had been projected. Overall, non-GAAP operating margins for the quarter were more than 40% of revenue and the company’s effective tax rate was 21%. This all resulted in EPS of $1.08 which compared to a prior forecast of $.91. I usually like to focus on cash flow generation. Last quarter, the company generated $102 million in operating cash flow, a considerable decline sequentially and year on year which was a function of increases in inventory and prepayments as the company struggled with supply chain issues. DSO did drop to 51 days from 67 sequentially and inventory turns did increase, but overall inventories grew by more than 23% sequentially.
The company, dealing with continued supply chain issues, forecast very little sequential growth in revenues the balance of the year, and in turn, that has led to what would be minimal sequential growth in EPS. Last quarter, the company recognized $80 million of revenue from its balance sheet. The forecast that the company provided at the time of the conference call had no expectation of additional revenue from the balance sheet so essentially it suggests that actual shipments will rise by about 8% sequentially. If that cadence continues, it will obviously suggest significant upside to current consensus revenue and earnings expectations for the next several quarters.
The company enjoyed particular success with its cloud titan customers and with cloud specialty customers this past quarter. In particular, it called out Meta (META) and Microsoft (MSFT) as 10%+ customers and revenues from some of the Tier2/Tier 3 specialty internet vendors also achieved substantial growth. While relying on purchases from Meta, in particular, for growth might seem counter-intuitive, the migration of cloud users from 100gb to 400gb switches is inescapable at this point. It just isn’t possible to maintain reasonable network performance without migrating, even if the overall growth of FB has been challenging at this point and the other cloud titans also need to revamp their networks to achieve higher performance to allow for steady usage growth. In any event, given that lead times at Arista for new orders are 12 months or longer, even for cloud titans, the actual orders they might place at this point aren’t going to reflect in Arista’s business in the next several quarters.
Arista shares, after wavering a bit in the wake of its earnings report and its guidance, did manage to rally about 13% subsequent to earnings, as the overall tech bear market rally reached its zenith in August. But since that point, the shares have fallen by more than 20%, and are now just 17% above the low for the year set in the middle of June.
The company these days is not providing specific guidance beyond its current quarter, and its CFO, Ita Brennan, may not be the best communicator in the enterprise technology space. Ms. Brennan indicated that the guidance being provided was not explicit beyond the quarter most recently concluded. The issue isn’t demand, but managing the supply chain. The company, by all accounts continues to experience demand running ahead of supply, and we have heard through industry checks of lead times for some orders extending to more than 1 year. But other analysts have suggested that Arista was able to expedite some product shipments at the end of the quarter as it gains experience in juggling commitments from suppliers so that it can complete and ship partially finished switches.
Just exactly how these supply cross currents played out in the quarter to be reported at the end of October is not immediately apparent, at least to me. It seems to have been a strong quarter in terms of demand, and the company has been willing to make elevated levels of forward commitments to suppliers to obtain needed components. What I do think is apparent is that the company’s demand growth remains far above consensus expectations, and the company continues to grab share from its leading competitor, Cisco (CSCO). There has been a recent report that overall component shortages eased in September, but the key for Arista is that the easing has to be for the components necessary to complete the assembly of its switches, not just an overall less constrained component supply chain. Some components that Arista needs have had lead times of greater than one year, which in turn means that some proportion of Arista’s own portfolio cannot be shipped within a reasonable time frame.
Some thoughts on the components of demand growth
The world of data switching is probably not considered particularly glamorous by many readers/investors, but without high performance infrastructure, the internet as consumers and other users know it could not function. And over the years, when it comes to high performance switching for the internet, and for other workloads, Arista has emerged as the leading company in the space. Overall, Arista sells its products to 5 main groups of customers of whom the hyperscalers/cloud titans are the largest single category and most prominent in analyst minds. Other sizeable customer groups include the enterprise, financial services, smaller, more specialized internet providers, i.e. Tier 2/Tier 3 and internet service providers. A bit less than 40% of Arista’s business comes from the cloud titan segment, and an equivalent level comes from the enterprise and financial segment. The balance, just more than 25% comes from service providers and Tier 2/3 cloud.
At the moment, talking about demand might seem superfluous; the company’s largest customers, the cloud titans can’t obtain delivery for Arista switches in less than a year. Other customers such as those at large enterprise often receive delivery quotes out more than 18 months. For at least the next two quarters, and perhaps longer, revenue growth for Arista will be a function of the ability of the company to increase its deliveries against its large order backlog. As mentioned, company management hasn’t felt comfortable in forecasting any amelioration of its supply chain issues-not for the quarter that will be reported, and not thereafter, although the CEO talked about some relief in the 2nd half of 2023.
In the absence of any guidance, most analysts have chosen to provide conservative estimates that show very little sequential revenue growth beyond the levels of Q2, and not much additional growth in 2023. And it appears that Arista is actually in better shape when it comes to supply chain than competitors such as Cisco.
I certainly have no insight as to when supply chain issues will sufficiently unsnarl-if that is a word-sufficiently such that Arista’s growth will be a function of real market demand and its competitive position, but I think in providing an analysis and an investment case, it is important to consider those factors as keys to its valuation.
Over the years, the company has broadened its product portfolio significantly, but at the end of the day, its competitive position currently is based on its offerings in the 400gb switch space-basically the company’s core. Currently, 60%+ of its revenues come from the switches that are used in data centers and cloud networks. Another 10-15% of revenues come from its campus switches and from its routing offerings-Arista describes these as adjacencies. And the balance of its revenues comes from software and services. Those percentages can and have varied somewhat depending on which group of customers is leading the pack with regards to demand.
I don’t want to claim specific domain expertise when it comes to evaluating networking offerings. This is a highly complex, technical field and what is important to some users is not important to others. Arista emerged as a substantial company by catering to the needs of cloud titans and that remains its single largest group of users. Its CEO is considered by most to be one of the visionaries in this space. And she has shaped a team that has significantly broadened its user base and today, enterprises and financial taken together are equivalent in importance for Arista to the cloud titan customers.
I have linked here to an evaluation published about a week earlier, comparing Arista and Cisco. Unlike many other 3rd party analysis, this one is pretty unambiguous. Arista is considered better because of ease of deployment, favorable total cost of ownership and somewhat better service and support. The Arista switches are said to provide visibility for many metrics that are not supported by others. Our of the box, no third party solutions are needed. Another link from Forbes that speaks to Arista market share gains is here.
Here is a link to Arista’s latest investor presentation. Of course much of it is a commercial. But I think the metric to note, which is from a 3rd party source shows market share trends. The graph shows that Arista currently has a market share of about 19% while Cisco’s share has fallen to 37%. It is a pretty stark representation of the market share gains that Arista has been able to achieve. It also depicts that the threat of what are called white box switches in the market has never been realized.
Of course, market share gains in the current environment, in which supply is constrained for all classes of products from all vendors, mean less than might otherwise the case. And Arista has an outsize market share in the cloud titan space. When cloud titan demand has been soft as was the case at the end of the last decade, it has seen its growth falter. As cloud titan demand has improved, Arista’s revenue growth has accelerated. The 49% revenue growth achieved last quarter was the strongest percentage showing for the company in some years.
As mentioned, about 10%-15% of Arista’s revenue comes from campus switches and routers, products that are just now becoming mature and are seeing significant acceptance in the enterprise. At one time, prior to the various impacts of the pandemic on the evolution of business in the networking space, and of severe component shortages that in turn has limited supply of switches and routers, Arista probably hoped to reach a greater run rate than $600 million/yr. for both products combined by this time. Of course, prospecting for new customers, when it is not possible to fulfill orders from existing users, is not the best use of the company’s sales resources. Looking at some of the commentary on Reddit and other 3rd party analysis, Arista certainly has a very competitive products in campus switching and internet routing, but if delivery takes more than a year, it really isn’t feasible to aggressively sell to new users.
What’s a reasonable long-term growth assumption for Arista?
Arista’s growth has seen highs and lows over the past several years. Growth troughed in a period in which the company’s largest customer class, cloud titans, constrained their capital spending. In fact, Arista revenues declined about 4% in 2020, which also showed a nadir for the stock price at the time of the Covid-19 inspired panic. As that changed, so did Arista’s growth which recently peaked at 49%, pretty impressive for a company with a revenue run rate of greater than $4 billion. One thing of note, the ethernet switch market is characterized by significant product cycles that drive growth. One factor that constrained Arista’s growth was the transition between 100gb and 400gb switches. Arista actually introduced its 400gb product in the fall of 2018. But market acceptance of that technology in high volume took until the spring of 2021. The next major technology will be the 800gb switch, but that product isn’t yet really in the market.
Last year, according to IDC’s quarterly tracker, the market grew a bit less than 10%, with some growth acceleration during 2021. The market has grown a bit faster in 2022 according to the latest tracking data with growth accelerating to 20% according to the authoritative Dell’Oro group. According to Dell’Oro, Arista gained 400 bps of share last quarter, although its share is still just in the low 20% range, while Cisco’s share is in the high 30% range. On the other hand, other analysis linked here suggests, longer term market growth is expected to be more in the range of mid-high single digits. Dell’Oro has long been considered to be the gold standard for market research in the networking space, so I am most inclined to consider its view as most likely to be realized.
As mentioned earlier, while market share trends in an era of constrained supply are not necessarily indicative of long-term performance, Arista has been gaining share in its space for years and I expect it will continue to do so for the foreseeable future. Most analysts rate Arista hardware and particularly it EOS very highly and its user satisfaction has been at very elevated levels, mainly a function of its highly rated customer support. I am not going attempt to review the expanding Arista portfolio here, and try to compare it to what is offered by Cisco, or by the numerous other competitors in the space. For the most part, Arista has been announcing new products at a rapid pace; some of these have more significance than others. The company’s latest offering that it calls CUE (Cognitive Unified Edge) is aimed at commercial enterprises. In this space, particularly, there can be a long lag time between product introduction and widescale user deployment so I think it is prudent not to put too much emphasis on one product introduction or another.
Currently, the analyst consensus for Arista revenue growth in 2023, absent much in the way of guidance by the company, is for revenue growth of 15.6%. To a significant extent, growth will be contingent on solving supply chain issues. If Arista does better at that than some anticipate, its growth will be significantly higher. Demand, at least in the next 12 months is not going to be a significant factor in calibrating growth.
I use a 3 year growth rate in an attempt to smooth out some of the fluctuations due to product cycles, large customer behavior and at this point, supply chain issues. I expect a 3 year revenue CAGR in the low 20% range composed of market share gains in high performance switching, particularly in the enterprise and in Tier2/3 internet providers, continued growth in adjacent product areas, and strong performance of the company’s services offering. It is probably impossible at this point to determine just how much pent-up demand there is given the excessive lead times at this point.
The Arista business model
Arista is and has been a very profitable company. The current business model reflects both supply chain issues-which has resulted in elevated levels of expedite fee-and the very strong growth of the company’s cloud titan revenues which inevitably have lower gross margins than the other customers to whom Arista sells. Currently, non-GAAP gross margins are in the range of 62%. Historically, gross margins had been in the mid-60% range.
Given the background of the founders of this company, and its basic operating philosophy, it is not surprising to find that research and development spending is elevated compared to most other IT hardware vendors. In particular, last quarter non-GAAP research and development spending reached $148 million, or 14% of revenues. The comparable metric for Cisco last quarter was 11%.
The largest expense ratio divergence for Arista relates to sales and marketing expense which last quarter was just 6% of revenue, while the comparable number for Cisco was 17%. There are two reasons for this dramatic difference in sales and marketing expense ratios. One of these is that the concentration of Arista’s sales to cloud titans. The sales resources necessary to service cloud titans is proportionately much smaller than typical enterprise customers.
In addition, like some other IT companies such as Atlassian (TEAM) and Veeva (VEEV), the company essentially lets its product portfolio do its marketing. It has historically been able to sell based on word of mouth, and through expanding deployments at users, which enable the company to conserve sales and marketing expense. Of course, at the moment, with product deliveries constrained as they are, the need for sales and marketing spend is far less than usual, and last quarter, the company actually reduced sequential spend in that category. The company’s general and administrative cost are only, $15 million, or 1.4% of revenue, a remarkable level for a company of the size of Arista. General and administrative costs only grew by 10% year on year last quarter, compared to the 49% growth in revenues.
Overall, Arista’s non-GAAP operating margin last quarter was greater than 40%, compared to 38% in the prior quarter, and to 38% in the year earlier quarter. While gross margins may track higher depending on expedite fees and the mix of customers, I doubt that the company will be able to grow non-GAAP margins from this point.
While the company uses SBC, it is at modest levels, and has actually declined a bit as a percent of revenues. Last quarter, SBC was 5% of revenues, flat sequentially, and up less than 15% year on year.
The company’s free cash flow last quarter was just $100 million, or a margin of 10%. The issue was mainly related to inventory which rose by $160 million in the quarter as well as a very elevated level of prepayments the company incurred in order to secure future supply of components. Inventories rose as the company couldn’t get all of the components necessary to complete assemblies of finished switches.
Quarterly cash flows are notoriously variable, and not usually representative. I imagine quarterly cash flow performance in the balance of the year and beyond will trend much closer to historical conversion rates. I have estimated a full year free cash flow margin of 29%.
Anet has a current cash balance of $2.9 billion. It has been buying back some shares; the outstanding share count has fallen by a bit less than 1% over the past year. It has also bought a couple of small companies, and I expect it will continue to make tuck-in acquisitions which will bolster its CAGR and enhance its competitive position. As the company is generating about $1 billion/year in free cash flow, it will either accelerate its share buyback plan, find additional tuck-in acquisitions or some combination of the two.
Valuation/wrapping up
This remains a difficult period in which to invest in the IT space or into shares of growth businesses. Fed speakers continue to roil markets with comments about peak inflation and strength in the job market. The knee jerk reaction to these pronouncements, whether or not based on substance, has been to sell IT shares, and particularly shares in growth companies. At some point, of course, the reality of demand destruction, such as that exhibited in Wednesday’s report on housing starts and permits will intrude into the discussion.
One company not currently worried about demand destruction is Arista Networks. While it doesn’t explicitly publish backlog data, at this point, it has scheduled more than a year of deliveries, and even its best customers are unable to get deliveries for new orders in less than 12 months, almost regardless of circumstances. While that is really not desirable from a business standpoint, it certainly mitigates any fears about a recessionary environment upending revenue growth assumptions.
Arista shares have been pummeled along with just about every other company in the tech space. While last quarter was a substantial upside, and the shares did ultimately participate in this past summer’s bear market rally, they have given up all of that gain, and more. The shares now trade at an EV/S of less than 6X based on forward revenues, with a Rule of 40 metric that should be in the mid-high 40s, depending on whatever amelioration the company can engineer in its supply chain problems. The company continues to gain share-according to the latest data; it took 400 bps of share, mainly from Cisco last quarter. The company has a very attractive business model with non-GAAP operating margins of greater than 40% last quarter, despite dealing with substantial expedite fees that have constrained gross margins.
While never easy to dissect, the competitive strength of the company’s product portfolio-both in terms of core switching, but in routing and campus switching as well has never been more pronounced. While currently benefitting from a significant upgrade cycle from the company’s cloud titan users, the company has also seen conspicuous strength from customers in both the enterprise and in Tier 2 and Tier 3 internet providers.
By all accounts, Q3 which will be reported at the end of this month is likely to have exceeded what was cautious guidance based on supply chain woes. Just how much additional the company was able to ship is not knowable, but it seems probable that the upside will be noticeable, and should create some additional operating leverage. That said, I am sure that guidance will be cautious, with little of specifics discussed about the outlook for 2023. While I don’t suggest that IBM (IBM) and Arista are close analogs, there might be some read through by the comments that the former company management made about the state of demand for IT solutions during its Wednesday conference call. A recession is obviously impending, but not all IT companies, and that includes Arista, are going to see demand issues as 2023 progresses.
Arista shares are obviously not going to perform outside of the context of the current environment for IT shares. They didn’t do so after the company achieved a noticeable upside compared to expectations in August, and I can’t really forecast that an upside when results are reported on 10/31 will necessarily propel the shares despite the lack of any significant speculation with regards to the quarter, at least to this point. That said, at some point, the realities of Arista’s positioning in what is still a healthy market will be realized. I think there can be considerable positive alpha to be found in these shares over the coming 12 months.
Be the first to comment