Part 2 of 6 part series
In this five part series I analyze what I think are four really good cloud platform stocks within the investing theme of building blocks for digital transformation. I will post my evaluation of each one every few weeks, culminating in my selection of the best of the bunch. Follow me on Twitter @OwlWealthy to read my analysis when it's fresh off the press.
Throughout the history of the computer industry, computing architecture designs have gone through multiple cycles. These architecture debates are about where to compute and process data. From a centralized compute approach that launched the mainframe period, to a decentralized approach that rose to prominence during the PC era, and more recently back to centralization with the advent of cloud computing. We are now seeing the pendulum swing back towards decentralization with the advent of edge computing. And, Fastly (NYSE: FSLY), and edge computing platform provider is quickly emerging as a leader in this space.
What’s different with the current shift in computing paradigms this time around is that it’s not an either/or situation. Cloud computing and edge computing are complimentary, not mutually exclusive. More and more organizations are using both computing computing models for the best of both worlds. Fastly doesn't need the cloud titans (e.g., Amazon Web Services, Microsoft Azure, Google Cloud) to fail for them to succeed.
What is edge computing and why is it important?
Edge computing places content, data and processing closer to the applications, things and users that consume and interact with them; rather than on a centralized server or in the cloud. 91% of today’s data is created and processed in centralized data centers. By 2022 about 75% of all data will need analysis and action at the edge (Source: Gartner).
What is driving this data and compute migration to the edge? Primarily it is overcoming the kryptonite to digital transformation - latency. Latency is the time that passes between a user action and the resulting response from the app or website. Too much latency, users will get frustrated and move on to the next app or content provider.
What is the solution to latency? Edge computing. The benefits of edge computing are multifold and absolutely necessary for meeting today's digitally savvy customer expectations. Edge computing delivers speed, security, scalability and cost efficiencies (by reducing bandwidth requirements from local network points to the central cloud).
The massive amounts of data that today's digital experiences create cannot be processed by a single cloud or an on-premise location. Latency will always be a problem, especially for emerging scenarios that will become mainstream, with the help of edge computing. In virtual reality (VR) scenarios any data processing that takes over 7 milliseconds will cause motion sickness causing VR unusable. Autonomous driving cannot afford latency. Life and death decisions must be taken in real-time, the AI powering autonomous vehicles don't have the time to send data to the cloud. Edge computing is required to make autonomous driving feasible. These are emerging scenarios. Even for the consumers of today's dominant digital experiences, streaming a movie, playing a video game with someone on the other side of the planet, or watching the super bowl in real time, there is not patience for latency.
Let's dig into Fastly, using the evaluation criteria I outlined in my article about the cloud platform providers powering digital transformation and investing gains.
Today, many leading companies whose digital businesses depend on meeting their customers expectations of speed and an amazing experience use Fastly. Customers include Shopify, Spotify, Yelp, Pinterest, Slack, Twitter, and the list goes on and on. These are businesses who are building on Fastly to deliver user experiences that are driving their growth.
Fastly is a platform that enables world class digital businesses. It just so happens this platform is at the cloud edge, and its programmable, which provides vastly improved levels of performance, scalability and security. This boils down to differentiation for Fastly customers, allowing them to beat their competitors.
A good example of this is how the New York Times uses Fastly. If you open the New York Times app on your phone, that is a conversation you are having with a Fastly server. The New York Times has uploaded their code to the Fastly edge platform. Their paywall is on Fastly's edge platform so the New York Times has the ability to determine what you're entitled to view and what personalization will make your experience better. When you request to read an article from the New York Times app on your phone, you are not being routed to a data center on the opposite coast of where you are. Rather, because of Fastly's edge cloud platform, the processing of this request is taking place very close to where you are, providing a high performant and personalized experience.
Make no mistake, it's not just performance and personalization that Fastly's edge cloud platform is providing the New York Times. Security is a big driver as to why the New York Times chose Fastly. They rely on Fastly to secure their busiest time of the year, election night. Not only does the load spike, with the service receiving 100,000 requests per second, the level of cyber attacks spike during this crucial time. The New York Times is a natural target. Just listen to their CTO describe how much the New York Times relies on Fastly to be able to deliver on their biggest night of the year. It's also a good sign to see a passionate customers be your advocate.
There are two adjacent markets that make up Fastly's total addressable market (TAM) - the traditional Content Delivery Network (CDN) space and the emerging area of edge computing. Each contributing about half to the TAM, these markets combined are forecasted to be a $35.4B (TAM) by 2022 (Source: Markets & Markets). Fastly reached FY 2019 full year revenue of ~$200 million, representing 0.5% of the TAM. A promising sign that there is a lot of market left to claim.
The greatest green space for Fastly, and competitors, is in the nascent edge computing market. Sure, Fastly can show meaningful performance advantages over traditional CDN providers with their more heavily slanted software approach. But, what gets me most excited about Fastly is the fast growing edge computing market, and the technological edge Fastly appears to have earned over competitors.
Here is a good breakdown of how Fastly sees the market opportunity.
In a nutshell:
Companies are accelerating their digital transformation as a result of COVID-19.
Companies are building their own applications, rather than buying, therefore requiring the developers who are building to make decisions on how to differentiate their apps through services that companies like Fastly provide.
Developers are migrating quickly and en masse to developing on the edge because of its performance, security and scalability superiority. Central cloud platforms will continue to be used but more and more processing will be pushed to the edge.
Fastly is well positioned because companies want to use multiple central clouds to build negotiating power when dealing with the cloud titans (e.g., Amazon Web Service, Azure, Google Cloud, etc.), and to load balance. Fastly is viewed as an honest broker. They are cloud agnostic, you can build once and it will inter-operate across the multi-clouds you use. They can use Amazon, Azure and Google for central cloud workloads, and Fastly for edge and maintain one set of centralized logs, access and firewalls.
This is one of those unique market opportunities where, provided Fastly stays ahead of the competition in terms of product, they can continue to charge a premium. Sure premium pricing will attract competitors, and Fastly will have to continue to drive down cost of goods sold. They already have a huge cost advantage as their heavy software approach drives down operation costs. In contrast, traditional CDN providers just throw more capital intensive hardware at the problem of latency. But, customers will pay a premium for the differentiation Fastly can offer.
There are two main advantage Fastly has over competitors from a product perspective, programmability and better performance. I'll get into these advantages in more detail, share the product validation they are getting from some extremely technically discerning customers and discuss how their competitive edge is about to get a lot better.
Fastly's big bet and primary technology investment is in creating a programmable edge computing platform. Fastly is a pioneer by enabling programmability at the edge. While it's straightforward for legacy CDN providers to manage and deliver static content across a network of globally distributed servers, it is much more complicated to enable developers to run code on edge servers.
Programmability at the edge is extremely important for the very demanding and huge potential use cases that have emerged recently, including the Internet of Things (IoT), gaming, and Virtual Reality/Augmented Reality.
Fastly’s stated performance advantage blows away the competition. Fastly's technology can start up in 35 microseconds, which they claim is 100x faster than competitive solutions.
Fastly's technical superiority and better "mouse trap" for solving the latency problem is getting major endorsements from some extremely technical customers choosing Fastly.
There are several platform companies that build their SaaS and PaaS services on Fastly, including Mux, Pantheon and GitHub. Usage of Fastly by these platforms is significant, as these companies are among the most technically demanding in the world.
There are also a number of innovators that are building their businesses on Fastly, including Shopify, Pinterest and Khan Academy.
Even Amazon, which has a competitive product to Fastly, chose Fastly to run Amazon.com and IMDB at the edge because of their superior technology (source: https://www.streamingmediablog.com/2020/05/fastly-amazon-homepage.html).
Fastly's product lead over its competitors is about to get a whole lot bigger. Fastly will reach a step change in terms of their product capability with the General Availability (GA) launch of their Compute@Edge platform at the beginning of 2021. Described by some fans as a new and better Internet, it is designed to allow developers to build more advanced edge applications with “greater security, more robust logic, and new levels of performance”.
One of the biggest advancement coming with the Compute@Edge launch is the support of more programming languages. One of the limitations with Fastly's programmable edge today is that customers have to learn VCL (Varnish Configuration Language) to use it. VCL is not that popular with developers as it has a lot of limitations.
Here is how Fastly CEO Joshua Bixby explains the impact compute@edge will have on getting more customers onto their programmable edge:
The idea that we will be able to go in even more to developers and let them use the languages that they are comfortable with and where they have tooling and integrate with the tooling they have.
Right now, it’s doable but it requires more effort.
So, talking to a customer, we can be in a situation where unless they try it, they will be skeptical about a lot and it’s harder for them to try because a lot more changes have to happen.
So [with compute@edge] it will be an easier conversation with the customer or a prospect and getting them to do a test.
Fastly was built by developers for developers
- Fastly CEO, Joshua Bixby
Fastly is rooted in the developer mindset. The founder of the company came up with the idea for Fastly as a developer looking for a Content Development Network he could could write code for. Frustrated with the offerings available at the time, Artur Bergman built his own solution that was more performant, less expensive and easier to manage.
Fastly is extremely developer friendly. They have extensive documentation for developers, tutorials, reusable code components for many common edge computing use cases; all making it as easy as possible for developers to integrate their code with Fastly.
They make it very easy for developers to transact with them. They don't have to go thru a long drawn out procurement cycle led by a Fastly sales rep. If a developer gets inspired at midnight working on their application, within 30 seconds they can be up and running with Fastly.
Developers are going to be very happy when they discover the new Fastly Compute@Edge supports the popular Rust programming language. For five years running, Rust has taken the top spot in the annual Stack Overflow developer rating as the most loved programming language, by a wide margin. Enabling developers to code in their favored Rust language will go a long way in attracting developers to the Fastly platform. Developers get the indisputable benefits of the edge, without having to sacrifice their preferred coding language.
A lot of Fastly's optionality potential resides with its new Compute@Edge platform that is currently in beta. Its customers are innovating on the new platform (in Beta) are demonstrating some new use cases that have a lot of potential to accelerate Fastly's growth.
Here are some examples of the sort of use cases that are being developed on Compute@Edge by early adopters of the Beta version:
Personalization: In the past it’s felt like a trade off that developers have had to make between performance and personalization. Anytime developers have wanted to personalize a website or app, they've had to roundtrip the data to a central cloud location. With edge computing, there is no longer a trade-off. Developers are now building highly personalized experiences on Fastly's edge platform, without having to compromise performance.
IoT and Machine learning inference: New scenarios are requiring the application of machine learning models that empower Internet of Things (IoT) devices (machinery, cars, etc.) to make decisions throughout the course of their operation. This model is called machine learning inference. A trained algorithm gets deployed to an IoT device to make predictions, based on data generated at this endpoint, to guide decision logic applied by the device. A good use case example of where machine learning inference will be required is autonomous cars. Because self-driving cars need to make split second life or death decisions based on data the car is collecting in real time (e.g., the behavior of other cars on the road, hazards on the road, etc.), self-driving cars need the brains to make these decisions. They need an algorithm that's been trained extensively across millions of scenarios a self-driving car could face. The level of compute and data required to train this self-driving algorithm makes the central cloud the only feasible location today for this processing. But once the algorithm is trained to a suitable level, it can be used by autonomous vehicles to quickly infer and make the sort of local decisions that are necessary to keep their passengers safe. The real time nature of decision making required to make self-driving cars a reality requires an edge platform so car don't waste milliseconds trying to communicate with a central cloud. While this specific self-driving use case has not been disclosed by Fastly as a real scenario it's handling today (it very well could be), they have disclosed they are seeing customers use their new compute@edge platform for IoT scenarios where machine learning inference is being done.
Security: Customers are testing new types of edge compute security scenarios on Fastly's compute@edge platform. This includes identity enforcement, enterprise-wide data loss prevention and data protection. Read how Fastly CEO Joshua Bixby sees the opportunity:
"And then one area that I think is very interesting, one that I think is underappreciated, is compute edge not only being a toolbox for web and application and mobile developers, but security developers. And we are seeing more and more security use cases. We're talking to a senior security person who said, listen, digital loss prevention. So think of this as I want to make sure that when 50,000 credit cards leave my network, I know. I want to centralize and be able to do digital loss prevention. So, there are some security use cases that I think really broaden out the total addressable market for us and are very interesting. These are all new and we're still in the exploration stage."
Fastly is an $8.7 billion company, with close to half of that valuation realized over the last few months. While it's likely to be a bumpy ride I can realistically see Fastly double in the next year or two thanks to the acceleration of digital transformation alone. This stock has high risk but if their edge@compute launch is successful, they are small enough that big gains will be realized by investors who get in well ahead of the new platforms launch.
The company has benefitted from the amazing technical leadership of its founder Artur Bergman who recently stepped down as CEO of the company. As part of this move, Bergman maintained his product leadership role by becoming the company's chief architect and executive chairperson. Joshua Bixby was promoted from President to CEO and has been impressive since becoming the chief executive.
This appears to be a good move for shareholders as Bergman recognized the role where he could help the company grow the most. "I like being in the trenches" Bergman shared at the time of his role change, citing his "true strengths and passions" as a developer instead of company leader.
It is too early in his tenure to review Glassdoor ratings for Joshua Bixby (as I usually like to do as a means of evaluating the CEO by their toughest critics, employees), but I will be checking back soon.
My bottom line
I am a current shareholder of Fastly (NYSE: FSLY) and as of the publication of this article (July 6th, 2020) I added to my position at the price of $83.05. The theme of applications, content and data processing moving further out to the edge is too good to ignore. Plus, this mega trend is being accelerated as a result of COVID-19 triggering companies to move faster with their own digital transformation.
I believe Fastly is extremely well positioned to capture an outsized share of this fast growing market. Their soon to be edge@compute platform launch will attract more developers due to its support of the most popular programming language today, and will spur new use cases as their customers innovate at the edge.
At a 40 Price to Sales (P/S) ratio, this is an expensive stock. Their net expansion rate of 133% is among the highest for cloud companies, but their gross margin is lower than ideal because of the capital expenditures required to scale their platform.
Fastly is a high growth and high risk company. Expect a lot of volatility. I recommend, if you are interested in the stock, easing into a position and adding over time as you learn more about the company.
As with any investment decision you make it’s important to complete your own due diligence and assess if the investment matches your risk profile. This stock, while it has a lot of potential to continue its rocket ship like rise, has a lot of risk.
-------------------------------------------------------------------------------------------------------------------------- This stock idea falls into the category of "taking intelligent risk". I don't do this often but I believe with the right investment hypothesis, research and buying what you know; allocating a small portion (i.e., 5% of your portfolio) of your total investment portfolio to taking measured risk can be worth it for the potential of hitting the odd home run that super charges your returns. Before making investment decisions please do your own research and/or seek the advice of a professional adviser.
Comments