ANET Thoughts & Questions

This isn’t an analysis of the company’s earnings report. Just a few random thoughts and questions for the group.

Investing in ANET hasn’t been an easy decision for me. I wrote recently that I hate legal issues, but they can provide opportunities. So far, this has been an opportunity I seized. While still somewhat troubling, I’ve become more sanguine about the legal issues surrounding ANET. I anticipate the recent ITC ruling in January will have an affect on margins as ANET is forced to use more domestic manufacturing. The margin impact won’t be devastating by any means and the company is prepared. ANET has just under $900M in cash reserves as well. Here is what the company had to say on the legal front:

Arista will provide a legal update since 2016 marks the completion of two years since Cisco Systems filed its assault of lawsuits. Despite having to weather the many baseless accusations, we at Arista have demonstrated through our actions and our results. We will continue to uphold our dignity, integrity and adhere to the law the Arista way. Clearly, we’re dealing here with a powerful competitor with enormous clout, deep financial pockets and even political lobbying capabilities often misrepresenting our intentions.

So let me take a moment to quickly review the facts. To date, Arista has prevailed or obtained favorable Administrative Law Judge, ALJ, decisions on nine out of 14 asserted patents in ITC and district court. Four of the five remaining patents primarily cover commonly standard implemented networking features used across the switching industry such as private VLANs, access list and CoPP. On December 14, 2016, Arista received a favorable jury verdict on Cisco’s Command-Line Interface copyright claims. We consider this an important moral and legal victory not only for Arista but the broader industry to have that standard networking language of communication.

Regarding customs, late evening on Friday, January 13, 2017, we were surprised to receive a letter from Customs and Border Protection, CBP, that revoked its previously issued 177 ruling on November 18, 2016. It is important to emphasize that CBP has not ruled that Arista’s products infringe. CBP currently have no stated position on that issue. Instead, CBP has set up a process to obtain input from both parties before issuing a new ruling. We do look forward to cooperating with customs in forthcoming weeks. We are in inter-party process to resolve the matter. Arista is firmly resolute and steadfast in lawful supply of our products and servicing them without disruption to our customers. As a reminder, the ITC and customs orders do not prohibit us from selling non-infringing products manufactured in the domestic USA.

So, the legal issues are, of course, still a risk, but ANET appears to be handling them well. They have a specific portion of their IR website devoted to these matters if you want to check it out.

The company continues to grow in a straight line at about a 30+% clip - incredibly steady growth. They show a very compelling chart in their investor’s presentation this quarter (I didn’t look at past presentations). They show their top 25 customers on one axis and the last 12 quarters on the other axis. The chart depicts whether the customer made additional purchases each quarter. So there are 300 (12 x 25) customer-quarters on the chart. In 286 customer-quarters, the customer made additional purchases. In only 14 cases did the customer not buy more! That is an impressive trend.

Additionally, ANET brings on roughly one new customer per day.

I have a few questions:

First question is on guidance. ANET only provides Revenue ($325M), GM (62.5%), Op Margin (27%) and tax (29%) guidance. Is there anyone out there who can help me determine how this relates to EPS?

The 2nd question is on customer concentration and is 2 part. Microsoft is their only greater than 10% customer and increased from 12% to 16% this year. I’m not sure how I feel about that and am wondering if anyone has thoughts? Also, I presume Microsoft is using ANET for their Azure Cloud architecture (but that is a presumption based on zero tech knowledge) which brings to mind the following. I wonder who AMZN and IBM, for instance, are using?

Take care all,


First question is on guidance. ANET only provides Revenue ($325M), GM (62.5%), Op Margin (27%) and tax (29%) guidance. Is there anyone out there who can help me determine how this relates to EPS?

I’ll take a shot at this…

Honestly, it is best if you’re looking at an Income Statement as you read this. It almost doesn’t even have to be Arista’s, but theirs is easy enough to find online.

The top line on an income statement is revenue. Then you subtract cost of goods sold (or services rendered) which leaves you with gross profit. Gross profit divided by revenue is gross margin. They’ve explicitly told you revenue and gross margin, so you can predict gross profit by multiplying the revenue and gross margin guidance figures.

From gross profit, the next step on the income statement is to subtract all operating expenses, which leaves you with operating profit. Operating expenses are stuff like R&D, sales and marketing, depreciation, corporate overhead, etc. As was the case with gross margin/profit, you can figure out operating profit from revenue and operating margin.

Now we’re almost to the bottom line: net income and EPS. But first, there are a small handful of non-operating expenses to take care of. These are things like interest, taxes, and “other” (which is often foreign exchange gains or losses). Interest usually isn’t a large number - either as an expense or as income. Foreign exchange isn’t predictable. If they could predict it, they probably wouldn’t be selling communications equipment! So, that leaves taxes, and they’ve told you what the tax rate is on operating profit. So, as you can see, their guidance has taken you about 95%, maybe more, of the way to net income.

For EPS, you’ll have to try to figure out how many (diluted) shares will be outstanding. But that’s a number that typically doesn’t fluctuate wildly from quarter-to-quarter (although Arista’s seems to be growing a lot; that’s probably due to stock-based compensation, but I really haven’t checked to see if there were secondary offering(s) involved). You’re probably safe looking at the trend of the past few quarters and adjusting last quarter’s number based on that. You might want to check recent press releases to see if share buybacks or secondary offerings have been announced before just going with a trend-based number. Just as you played with recent numbers and trends for the diluted share count, you might want to do the same thing for “interest” and “other”, depending on how close you’re trying to get to a good net income number. Divide net income by the diluted number of shares that you’re guessing, and that’s it! You just predicted EPS. They pay Wall Street analysts how much?!?!?

Did that help? Are there questions?

Thanks and best wishes,
TMFDatabaseBob (no positions in companies mentioned, drat!, but Arista has long been on my watch list; it is primarily the legal issues that kept me on the sidelines)
See my holdings here:
Peace on Earth


Google lists ANET’s EPS at $2.29 (…)

Feb 17 - Close
Range 110.31 - 119.45
52 week 55.00 - 119.45
Open 111.00
Vol / Avg. 5.91M/733,344.00
Mkt cap 8.32B
P/E 52.02
Div/yield -
EPS 2.29
Shares 70.13M
Beta -
Inst. own 64%

Arista has long been on my watch list; it is primarily the legal issues that kept me on the sidelines)

I had a starter position in ANET, but when CISCO sued, I bought a bunch more. And then when the ITC reversal came and the price plummeted, I sold some Puts. All obviously doing really well now.

Big companies like Cisco have a patent portfolio as a defensive strategy, so for them to actually go offensive with it meant that they knew they were losing. Arista can lose the lawsuit battle but they will still win the market share war.

Someone somewhere in Fooldom posted his concern about commodity hardware being a threat to Arista’s price moat. But, software is still the key for Arista. It can be worth it for companies like Google and Facebook to develop their own solutions since they have the scale and every last bit of custom optimization saves them money, but that’s OK - there are many more companies big and medium for whom trying to reinvent what Arista has would be a big mistake.


Thank you A.J. for taking the trouble to post that very interesting summary of the current legal state of affairs.

No trouble at all. While my posts have a small level of altruism, I can assure you the motivation is primarily ulterior. They help me stay up on the companies I follow and allow for obtaining input on questions I have.

I reference back to the posts I write from time to time as well which helps as I don’t have the best memory.

Take care,


Did that help? Are there questions?

It seems I do still have some questions. ANET gives non-GAAP Gross and Operating Margins guidance. But we need to use GAAP income to determine the effect of taxes. So, I want to get to GAAP income using their guidance to take out the effect of taxes and then add back in stock based compensation and litigation expenses, though I’m considering not treating the litigation expenses as one time since these may be on going for many more quarters (and since I don’t know how long, I’d rather take a conservative view).

I guess I could model income from operation on a GAAP basis using mid-points from historical data and then add back in what I need to. However, I’m wondering if I’m missing something that doesn’t require the “guesswork” on my end. Is there a more precise way I’m missing here?


I was a bit surprised that you’d want the kind of tutorial I gave. I’m sorry for not giving the answer you needed. I hope someone found it useful. Your follow-up gets to the heart of the matter, I think.

I don’t think you’re missing anything. Assuming they’re not guiding the stock-based comp amount, that would be an additional piece of guesswork on your part.

Sorry again.

Thanks and best wishes,
TMFDatabaseBob (no position in companies mentioned)
See my holdings here:
Peace on Earth


I appreciate the time you spent answering the question. I wanted to make sure I wasn’t missing something. In that case, I’ll spend the time giving everyone my estimates. You gave me exactly what I was looking for. Just validation there wasn’t a simpler way of doing things.

I’ll be back in touch with thoughts on the guidance. My nature is to take a conservative approach on the outlook and hope for better outcomes…however, I’m not sure this has served me well in the past. Stand by…


Is there a more precise way I’m missing here? – A.J.

For a rapidly growing company like ANET, why is “precision” important? I don’t see how anyone can predict future revenue growth rates with any precision, so any “precision” downstream of that factor would be kinda delusional, wouldn’t you think?

He is no fool who gives what he cannot keep to gain what he cannot lose.

1 Like

For a rapidly growing company like ANET, why is “precision” important?

That’s a good question and one I’ve thought about. I bet it is the weight given to analysts opinion and the “beat and raise” game or whatever you want to call it.

I know the guidance I’d give if I were a company. In fact, it is the same guidance I give my boss every single month - conservative as all get out. I only give the orders I know for a fact are coming in the door that month. Everything else is gravy.

But, I like to track how the companies I follow are doing relative to their forecasts. And I like to see what kind of trends develop.

For instance, in my notes recently I wrote down that TWLO would do $81.5M based on the forecast they gave of $73M. It was simple math based upon the percentage they beat the last quarter. TWLO did $82M in the quarter. I’m not suggesting I will be able to predict quarterly revenues, and yes, this was an outlier, but it just helps me keep track of things.

Appreciate your feedback though and you are probably correct in that precision is not warranted.

Take care,


But, I like to track how the companies I follow are doing relative to their forecasts. And I like to see what kind of trends develop. – A.J.


And I like to compare that data to the analyst projections. See if analysts are out to lunch or management is sandbagging. Maybe both. LOL


Okay. For the record, here is how I’m modeling the next quarter for ANET.

My revenue assumption is $337M which is around a 2% beat above the top end of guidance.
I’m using a GM of 62% which is around 250 basis points lower than the most recent quarters. I’m doing this as an assumed impact on the ITC ruling in Jan.

I took out R&D expense which of $72.5M which is only up about $1M from last quarter, but that seems to be the trend. I used 11.5% and 6.8% for S/M and G/A expenses which seem to be a good representation of percentages over the past few quarter.

I added back in stock based compensation at 5% and legal at $12M (3.6%) minus “Income tax effect on non-GAAP exclusions.” This brings me to $74.27M on 74.9M shares outstanding.

EPS - $0.96/share

This would be growth of 41.1% which is pretty impressive.

The above number may be heavy on revenue but are generally conservative elsewhere.

Take care,