Differentiation of CDNs

Much has been written here since 6/23/20 when Beth Kindig said FSLY May not be worth current share price. I appreciate the examples of what FSLY can do that others can’t; but, I believe Peter Offringa at Softwareatackinvesting.com said it well back in November how FASLY is different:
I edited very little from The following. I recommend reading the entirety of his Blog.
-Edge Compute (Copute@edge)is the newest offering from Fastly and the area with the most potential. In the other five product areas, Fastly’s offering is equal to or incrementally better than that of competitors. These are still areas of commoditization, where competition around features and price will continue.
Where Fastly is making its big bet and primary technology investment going forward is in edge computing. This involves providing programmability for customers in multiple, distributed locations. The benefit for customers is similar to the driver behind CDN. By moving processing and delivery closer to end users, customers can reduce latency and bandwidth requirements associated with sending all requests to a central data center. Fastly achieves programmability by exposing a hosted runtime and development environment for customers on its POP servers.
This is important for investors to consider, as this capability hasn’t existed previously. A plethora of use cases have emerged in the last couple of years that elevate the priority of running code closer to the edge of the network.
Fastly’s intent to focus on edge computing was taken to the next level in November 2019, with the announcement of their Compute@Edge product. This offers a language-agnostic compute environment running on Fastly’s distributed network. It is designed to allow developers to build more advanced edge applications with “greater security, more robust logic, and new levels of performance”.
The platform is deployed in a “serverless” model, which means that Fastly is not continuously running servers with customer code waiting for user requests. Funny, “serverless” is a bit of a misnomer. A server still processes code in response to a user request. It’s just that this server is not started until the request comes in, versus the normal model of keeping a server running continuously with active threads waiting for requests.
In Fastly’s model, when a request comes to a POP, the edge compute platform spins up a web container to process that specific request. In the past, this server start-up involves noticeable latency, ranging from hundreds of milliseconds to even several seconds. If a user request has to wait more than 100-200 milliseconds for a server response (not including time to process and return content), then the lag becomes noticeable to humans. Fastly claims their servers can start up in 35 microseconds, which is extremely fast and pretty much obviates the core argument against serverless modes of operation in the past. Fastly claims this is 100x faster than competitive serverless solutions. This is a fair assessment, as AWS Lambda previously took several hundred milliseconds and now requires at least several milliseconds to boot.
The engine behind Fastly’s serverless infrastructure is based on WebAssembly, which was originally designed to allow browsers to run code locally, but has been extended to the server. WebAssembly is a great framework for this, as it is already designed for high performance, low memory footprint and security. Fastly built their own compiler and runtime, called Lucet, which leverages and extends WebAssembly for server-side execution requirements. You can read about Lucet in this blog post. It has been open sourced, so that developers understand how it works.

Me Here: I don’t try to make a habit of investing in the hope of a future TAM. I believe current share price is more than enough to explain current FSLY offerings. I don’t understand why FSLY’s margins are 30% lower than NET’s. It could be this next level R&D that FSLY has put together for actual computing at edge is the reason?

I’ve reduced my allocation In FSLY having more than doubled since their earning report from 14% to 5%. I put 5% back into TWLO after what I believed was better than FSLY’s. I still don’t like how the TWLO CEO characterized growth the prior two quarters (with SendGrid acquisition confusing the numbers) but Wow. They’re rev growth acceleration apples to apples (without SendGrid) was great and their margins are much better than FSLY.

I put the rest of what was in Fastly into WORK after their great quarter that wasn’t as good as Zm; but WORK’s 87% margins (twice that of FSLY) and currently higher rev growth than is projected for FSLY, I like this one better now. I understand how WORK is a communications platform (eg Zm integration) and only a little how it wouldn’t compete with Zm if when Zm expands its Ability to become a platform.

I continue also to hold NET at 5% due primarily to their fast pace of product development and 70% margins, higher than FSLY and TWLO.

Please let me know what you :thinking:



Hi Jason, I read your post and realised that I also didn’t know why FSLYs gross margin was so much lower than NETs.

That seems like something important to know! Note that Cost of Revenue doesn’t include R&D, but the costs that keep the service running for their customers.

From the 10-Q:


"Cost of revenue consists primarily of fees paid for bandwidth, peering, and colocation. Cost of revenue also includes personnel costs, such as salaries, benefits, bonuses, and stock-based compensation for our customer support and infrastructure employees, and non-personnel costs, such as amortization of
capitalized internal-use software development costs and depreciation of our network equipment. Our arrangements with network service providers require us to pay fees based on bandwidth use, in some cases subject to minimum commitments, which may be underutilized.


“Cost of revenue consists primarily of expenses that are directly related to providing our service to our paying customers. These expenses include expenses related to operating in co-location facilities, network and bandwidth costs, depreciation of our equipment located in co-location facilities, certificate authority services costs for paying customers, related overhead costs, the amortization of our capitalized internal-use software, and the amortization of acquired developed technologies. Cost of revenue also includes employee-related costs, including salaries, bonuses, benefits, and stock-based compensation for employees whose primary responsibilities relate to supporting our paying customers and delivering paid customer support. Other costs included in cost of revenue include credit card fees related to processing customer transactions and allocated overhead costs.”

So not super-clear why the difference, although you might figure that FSLY is spending more due to increased usage and building out server infrastructure.

Heres the GAAP smackdown as at last quarter:

| TICKER                 | FSLY    | NET     |
| CLOSE                  | $86.5   | $35.5   |
| REVENUE                | $62.9M  | $91.3M  |
| REVENUE TTM            | $218M   | $249M   |
| REV. TTM 𝝳 (YOY)       | 38.20%  | 48.91%  |
| EV                     | $8.24B  | $10.7B  |
| EV/S                   | 37.8    | 42.9    |
| GROSS MARGIN           | 56.70%  | 77.20%  |
| GM $ TTM               | $122M   | $194M   |
| EV/GM $                | 67.5    | 55.1    |
| GM $ 𝝳 (YOY)%          | 38.98%  | 50.71%  |
| RPS                    | $2.28   | $841m   |
| FCF                    | -$18.8M | -$30.6M |
| CASH                   | $92.6M  | $118M   |
| SHORT-TERM INVESTMENTS | $94.1M  | $479M   |
| DEBT                   | $31.3M  | $50.8M  |
| DEFERRED REV           | $0.00   | $38.1M  |
| EBIT MARGIN            | -9.80%  | -24.60% |
| STOCK BASED COMP       | $6.33M  | $12.9M  |
| DILUTED SHARES         | 95.4M   | 296M    |
| SHARESWADIL_GROWTH     | 277.23% | 248.01% |

So I have FSLY at 57% gross margin, and NET at 77%. So 2000 basis points difference. FSLY is still cheaper by EV by 24%.

From the last conference call a bunch of analysts talked about gross margin, an example response:

“Yeah, it’s Adriel here. With respect to what we’re reflecting in operating leverage for the rest of the year, I’ll start first on the gross margin side. So for Q2, I do think we’re going to get some incremental gross margin improvement in Q2 relative to Q1. A big reason there is we only had a couple of weeks’ worth of sort of that revenue and I think just given what we’re seeing at the – sort of what we saw in March, I think we should see a good outcome in Q2 relative to gross margin.”

Not sure that added much!



Thanks Sarksnz for your most thorough reply.

For some reason I had FSLY at 43% margins. Not sure where I got that. 57% is better but NET at 77% and the difference still unknown - I’ll just assume NET’s Freemium model of customer acquisition is more efficient? And if that’s the case then their giving away their new Teams security feature until September might work out well for them in a couple quarters (watch out Zs)?


I’m not technically savvy like Greg, but I do address margins here: https://discussion.fool.com/why-i-bought-fsly-34514084.aspx

I suggest you listen to the company’s conference call on Q1 that was released around May 6. You can always learn a ton from these. I’ve gleaned a little about what I think the future will look like:

  • much improved margins
  • 60%+ revenue growth

Please share your opinions once you’ve dug into the call.



NET vs FSLY CC notes

The following is from my notes listening to the calls back to back today.

They both use the word Platform; but, my take is that They are both currently providing a breadth of offerings. The difference is: FSLY will move Compute@edge (huge focus on the call: this being real Ap logic and compute at the edge giving them ‘their highest margin product and making others ‘not a credible choice’) from Beta to Limited Release to GA not perhaps until 2022 -and- NET already eating their own dog food, Teams (what sounds like a Zs killer), with 1000 companies using it now and likely paying for it in September.

Fastly will become a true platform given what I copied from Peter Offringa’s blog in the initial post of this thread, but not for ‘a while’.

I’m likely to sell some of my 16% position in CRWD and add to my 5% NET position for the above reasons. But also because on the the NET CC when talking about their 50+% international including BIDU for 5 years now and ‘augmenting’ that with a fifty city roll out with JD Cloud in November; that they’re seeing ‘No decrease in the surge of rev growth at the end of the period, just being extremely prudent’ regarding guidance.

If anyone seeing the problem in this logic please let me know,



I found a couple posts from February on the thread Muji started with his Deep Dive into NET when he initiated an 8% position to replace Zs. Phoolio explains why NET has better margins and 12x explains why NET doesn’t consider themselves a CDN.


From NET S-1:

Instead of including the costs as COGS, certificate authority services costs for free customers and bandwidth and co-location costs for free customers that would hit gross margins, they include it as S&M expenses which would impact Operating Margins. In my opinion it is six of one and half of the other.

Operating margins would be the ultimate factor to consider.
Finally, I have no idea to what degree the costs above impact gross margins.

Just wanted to point this distinction out, and I’m appreciative of the person who informed me of it.



most CDNs charge per data usage whereas Cloudflare charges a fixed cost. This is because they consider themselves first and foremost a security company and legacy security companies that Cloudflare competes with such as Barracuda, F5, and anyone else out there offering a WAF appliance sells them at fixed costs. So they felt their service and therefore CDN should be fixed prices regardless of data volume. This is disruptive to the CDN industry and it is also disruptive to the WAF appliance industry because cloudflare becomes 100% software SaaS based.

Hope that helps,



I posit that the statement in my last post by 12x,’most CDNs charge per data usage whereas Cloudflare charges a fixed cost’, and ‘This being disruptive to the CDN providers’, is an under-statement. This would be the ultimate commoditization of CDN’s no?

My reason for selling FSLY is that the CC was nearly entirely about Compute@edgde and I do believe that could be a game changer at some point.

But, although Smorgasbord pointed out that FSLY has some current advantages Muji’s Deep Dive into NET in February pointed out NETs superiority to FSLY in multiple areas of service, at length. Which, with faster rev growth and 20% better margins currently, explains the 24% higher PS NET sports currently.

I’m betting that Cloudflare’s Teams Product, with all the capabilities of Zs, already in GA and given away to over 1000 companies ‘at the expense of current Gross Margins(stated repeatedly in their CC)’ and Cloudflare’s total commoditization of the CDN space will lead to outsized returns relative to FSLY over the next 12 months (FSLY CC stated only Limited Release of Compute@edge until perhaps 2022).

I say this knowing now that FSLY is usage based and NET is not and this and some other specifications of the FSLY product may, for a period of time, grow FSLY revenues faster than NET. And, this may keep the share price moving higher until Compute@edge picks up the decline due to the comoditization of FSLY’s other CDN services.

Hmm, did that make sense?



(FSLY CC stated only Limited Release of Compute@edge until perhaps 2022)

Here’s a quote from Joshua Bixby, CEO, during the conference call:
We have said in the past and continue to maintain that we’re moving more aggressively through that beta and into the LA phase, limited availability phase, this year, and this will have a revenue impact next year in a way that we’ll be able to quantify and speak to.



FSLY appears to be following up with their more aggressive release of ComputeEdge. See news today.

Are they pushing it out faster or is there more demand pulling it?

From todays pr: "Customers are testing use cases that range from content transformation at the edge, to identity enforcement, to enterprise-wide data loss prevention and data protection. RVU, the UK’s leading resource for market comparison sites and apps such as Uswitch, Money, and Bankrate, is encouraged by the observability built into Compute@Edge’s development experience. “Fastly’s new Compute@Edge offering aligns with RVU’s vision for the future and our team is excited to leverage the data and deeper insights into performance that it provides,” said Tom Booth, Head of Infrastructure and Security at RVU. “In addition to the broad benefits of externalizing business logic to the edge, having more visibility and scalability are critical to our workflows and business processes, and from our experience so far, we see promise that these capabilities are finally attainable through serverless technology.”


FSLY appears to be following up with their more aggressive release of ComputeEdge. See news today.

Here’s a link to the latest blog post on it: https://www.fastly.com/blog/observability-compute-edge

Some snippets:

Today, we’re announcing critical new observability features for Compute@Edge — logging, tracing capabilities, and granular, real-time metrics. These features, which are now available in beta, bring observability to the forefront of the serverless compute environment.


You can correlate and stitch together the lifetime of a request in a third-party visualization system, like Splunk or Datadog, after for further analysis. It’s easy to set up and lets you trace function performance after deployment.


When you’re developing, a println on Compute@Edge works the same as a println on your laptop.

If you’re a developer, these are big deals. Sounds like debugging in this serverless environment is pretty easy. As an ex-developer myself, having “print” work on a just-started-up-for-me-then-killed-right-afterwards process is a godsend. And having all sorts of data available to be easily pumped into Splunk or DataDog (!) will give you great overall metrics.

For me, it’s a sign of a mature company when even a new offering isn’t just about features or performance, but about how usable it is. That Fastly is doing Compute@Edge in a Beta is, for me anyway, a sign that when they finally open it up it’ll be a winner. Note that they’ll be going from Beta to Limited Availability this year. When Fastly releases it to what may be called GA (General Availability) it’ll be purchasable, like all of Fastly, via the web - no need to call a salesperson unless you want some advice.


Not sure if I missed this in previous posts but I actually never knew that there was a strategic partnership in place with DataDog. Here is the blurb within that news release. Really pleased to now know that.

“Serverless architecture has changed the game for operationalizing and managing product workloads,” said Ilan Rabinovitch, Vice President, Product & Community at Datadog, a monitoring and security platform for cloud applications, and a Fastly integration partner for data storage and analysis. “As an integrated logging and data visualization partner, we’re thrilled to be part of this advanced serverless offering that will enable a new generation of low latency applications. Our partnership will provide Compute@Edge users real-time visibility and insight into infrastructure and performance from edge to origin.”


Excellent insight and discussion here, thanks to everyone on this thread.

For a non-SW person like me, at a 10k ft level, my take is following… would be curious to see what others think…

  1. NET is really a secure access provider to cloud, just as ZS is … except that ZS is focused on large enterprise customers where NET is in mid-market to SMB space…

And both NET and ZS taking away $$ from not only PANW and FEYE etc… but also to some extent networking gear companies like CSCO… as they offer a replacement of old enterprise access networks

BTW - this also means both NET and ZS benefit tremendously with WFH… (I would have thought NET to accelerate revenue but this clarification that they offered free access to 1000 customers means they will see revenue build up next Q)

  1. FSLY on the other hand… is almost like building a faster, smaller, nimbler version of AWS or Azure… focused on edge… not competing with them but complementing… so they become premium, high speed version of big cloud…

if this is anywhere near good enough explanation, I see them both separate investment opportunities in their own right… and not much comparable… just like you would not compare CRWD to DDOG… very different function / purpose even though they are both cloud companies…