Fastly helps internet-enabled enterprise to deliver faster, more powerful sites and applications on secure, programmable edge cloud platform.
Fastly Investment Research Summary
- Dozen of Mega-trends are enabling next generation of internet usage behaviour and have more than enough Edge Compute+CDN’s Total Addressable Market (TAM) of approximately $72 billion on the back of approximately $4 billion of revenue collectively captured by 3 CDN leaders – Akamai, Cloudflare & Fastly.
- Make no mistake, both CDN and edge computing is a commodity business. Let’s repeat, CDN and edge computing is a commodity and will still be a commodity down the road while all CDN provider’s USP will be diminished in the long run. This indicates that no one clear player in the market will have a sustainable moat. Fastly’s only economic moat is about building up switching costs for the clients. Fastly’s customer acquisition strategy is more sustainable due to the nature of high switching cost.
- Edge Computing is the tech buzzword now like how “Cloud Computing” is a buzzword in 2010. Edge Computing is definitely not a new technology but it’s buzzing due to driven demand of such need. It is also being buzzed due to the rise of IoT and 5G. CDN providers and public cloud vendors are also buzzing it now to capture the enterprise customers’ attention.
- 7 other risks to be priced-in which might derail the investment thesis while complementing 2 wild cards or opportunities.
- Valuation is suggesting a 64% upside over the next 24-month period.
Weak USP <=> Commodity offering
Both edge computing and CDN offerings are currently offered by Amazon’s Lambda@Edge, Cloudflare’s Workers, Akamai’s EdgeWorkers, Microsoft’s Azure Edge Zones, IBM’s IBM Edge Application Manager, Limelight’s Edge Compute……. (and the list goes on…). Just to summarize, ALL public cloud providers and ALL CDN providers had already provided or soon to provide edge computing services.
Compute@Edge is not unique to Fastly and any unique features in Compute@Edge would definitely not be unique for long until competitors catch up.
Fastly’s ONLY Moat: Dive-in with in-depth switching cost within the industry vertical
This chart was extracted from Cloudflare’s Investor Day presentation on 12 Feb 2021.
This chart was extracted from https://w3techs.com/technologies/details/cn-fastly on 25 February 2021.
Combining both charts, Fastly’s share of the Internet is more than Akamai and almost the same as Amazon CloudFront but nowhere near to Cloudflare. If you compare the revenue per share of the Internet, Cloudflare obviously has huge upsides to charge more to capitalize further. Akamai’s revenue is approximately 10 times more than Fastly but approximately 50% lesser in share of the internet which either Akamai charging 20 times more than it should or Fastly is underpricing which boths scenarios also indicate huge upsides for Fastly to get more revenue from re-pricing or more enterprise customers migrate away for cost-saving. W3tech is showing a very promising territory that Cloudflare is getting into which is mass-market approach and rising in rank in terms of number of high-traffic sites.
I hope this analogy would make sense to you: Would you like to land 5 big clients within the same vertical which own 100 subsidiaries each or land 10000 small clients across 50 different verticals with only 1-2 subsidiaries each? That’s how I look at Fastly versus the rest of the CDN providers or Edge Cloud providers. Go Niche. Go Deeply Integrated. Go Specific. Then cross verticals.
Across many Fastly’s content and earning calls, Fastly has always been saying “Built by developers, to empower developers”, “Giving back the control to the hands of developers”, “We support our customer’s needs and customer’s needs is our core priority”. Fastly’s customers also repeatedly champion for Fastly and love Fastly’s solution a lot. Some of Fastly’s customers also say “Fastly’s support team is great and they are a fun team to hang long to grow together”. This is no coincidence. It’s all about building switching costs as an economic moat that the customers won’t leave and it attracts more new customers to join.
Fastly’ CEO says: “We believe that we are unique in delivering the specific set of interconnected tools that our customers have shown us are the exact right combination required to build the future of the internet. “. It’s essentially saying Fastly’s tool is specifically designed for the industry vertical (think digital publishing and e-commerce).
Every industry vertical has different use cases and you just can’t use one generalized / commoditized feature to land every enterprise customer in different industry verticals. To land big clients, you would need to ensure you have the agility to customize features that no other CDN providers are willing to deploy development resources. Other CDN providers or edge computing providers are using a one-size-fits-all approach compared to Fastly’s agile approach which is willing to dedicate development resources to customize solutions on catering for big clients’ needs. In some way, Fastly is acting like a software house which builds customized solutions via API to one client. But let’s not forget that the client’s use case is not unique.
For example, news publishers like The New York Times made a feature request for some new use cases, Fastly developed it for The New York Times to beta test and public launch then rolled out to the rest of Fastly’s news publisher. This gives a strong feedback loop and pulls in the rest of the news publishers who are not on Fastly yet and making the migration way from public cloud vendors to Fastly.
News publishing industry is competitive in nature and the “fear of missing out – FOMO” is helping Fastly to land more enterprise clients organically within the same verticals. When a CTO in a news publisher who uses Akamai or Cloudflare or Amazon, he/she starts to notice every other news publisher is on Fastly with features that their current cloud poriver don’t have, the CTO will first make a feature request to the current cloud providers. However, if such a feature request is so niche in the use case that only the news publishing industry will need it, good luck having that the current cloud providers will prioritize the feature request in the development pipeline. It’s just in a matter of time that the CTO will be inpatient and migrate to Fastly.
Fastly already conquered the news publishing’s cdn and edge computing market share. And they did the same with other verticals such as e-commerce – Shopify, Target, Etsy just to name a few. Their next move is on Video, Audio, SaaS, Travel Marketplaces and Fintech. My next guess is they are also trying to make their way into the game industry which they are late in the game as of now.
Page Load Time Matters and it Drives Demand for Edge Computing
Back in 1999, internet users were willing to wait up to 8 seconds for page load on desktop. You have to remember back then, the internet use case was just dominated with Internet 1.0 websites which without much web complexity. You were just plainly downloading the content from the websites and minimal uploading from your end to the websites.
Today, the web application is getting very complex to the point that the speed load is slower year over year (on mobile). This is due to the fact that you are simultaneously downloading a lot of contents from the web page and also uploading a lot more data back to the web page’s server. There are countless web cookies, trackers, interactive plugins, targeted ads, dynamic contents and live contents within the same page session or mobile app which adds up the application complexity while internet users are browsing it on mobile devices.
Why would enterprise customers want to add so many elements within the same page and what are you uploading to their server? Enterprises want it to be smarter with more data analysis which then resulted for better targeting with high relevance. Furthermore, the enterprise also wishes to achieve all of these heavy processing and return immediate responses within milliseconds while wanting dozens of the plugins to be able to feedback to compute and deliver results within the same milliseconds timeframe. As you type in the search bar for character by character, the algotherim is computing and predicting what you want and deciding what’s the best strategy to rank the result for you all within milliseconds. If e-commerce buyers can’t get what they want within the first result page, they will not click into the second result page. E-commerce operators would want to change the product results instantly and live within the first page itself when you land on the first page of the result page. How do they plan to do that? When you land on the result page, the enterprise would want to analyze how long you spend on reading/watching/scrolling/interacting the first 2-3 products while gauging what you are really looking for and quickly swap out/in the next 7 products without the consumer even realizing. You should watch The Social Dilemma on Netflix then you know what I am talking about. This is entirely based on the live interaction that you have provided and the compute engine processes it immediately and responds back to you immediately. Personalization element is also pushing the boundaries of what results should be given based on how you are interacting with the application. When you ask Apple’s Siri or Google Assistant or similar AI assistant a question with more than 30 words, it’s not likely to provide an accurate response at the moment. Even if the tech giants cracked the complex question with the most accurate answer, it’s not likely to deliver it back to you without buffering as all of the heavy computing is done at a central location instead of distributed.
Let’s take a look at amazon.com’s PageSpeed Insight at https://developers.google.com/speed/pagespeed/insights/?url=amazon.com
This PageSpeed snapshot was take around 23 February 2021
Looking at the PageSpeed snapshot shared above, you would notice that the Speed Index is 8.9 s while Time to Interactive is 12.6 s. Every web application would very much want every single element to be all loaded within 30ms while adding even more complex stuff along the way. This is where Edge Computing plays a role too. As more websites or mobile apps set a higher page load benchmark without compromising the data analytics capabilities, the demand for edge computing will explode.
In conclusion, the future of internet usage is about relevancy, data analytics, personalisation, real-time, zero buffering, high complexity and big computing done at the closest server located in the same town that you are using the devices. This change of behaviour drives the demand of edge computing.
Edge computing changes the way developers build applications. To simplify it, Amazon’s Edge Computing explained it 8 words: “Bringing data processing and analysis closer to end-points”. Instead of relying on one single core compute engine in a single centralized massively-scaled cloud server campus in San Francisco to run the logic processing for global users located all over the world of 195 countries, the developer will distribute more processing to be done on thousands of small-scaled distributed server campus which are located within 1000km (or nearer) of the user. Instead of relying on one master database in San Francisco to authenticate user access, developers can now distribute the authentication process by replicating the database to the distributed server which is nearer to the users. For telemetry use cases, instead of sending 10GB worth of imagery data all the way from Sydney to San Francisco for central processing, the developer would utilise distributed computing to process such data within Sydney’s distributed server which reduces the latency tremendously. However, there’s tremendous tailwinds on edge computing as it requires developers to change the existing compute, authentication, database storage and cache architecture. It involves huge effort to overhalve the legacy applications to the distributed architectures. CTOs are still busy adapting to big data, cybersecurity, machine learning and artificial intelligence.
Looking at Google Trends to compare the search interest of “cloud computing”, “big data”, “edge computing”, “cloud gaming” and “machine learning”, you would notice the interest for edge computing is way lower than the rest as we are still in very early phase of developing the infrastructure of edge computing. The infrastructure is referring to both hardware and software which there’s a lack of mainstream open-source contribution to an open-source edge computing software stake. To escalate it, Fastly is championing WebAssembly as the standard to help developers overcome the problems.
9 Reasons to be Super-Bullish on Fastly
There are 9 growth factors which would sustain Fastly’s growth rate or even super-charge Fastly’s growth rate moving forward. This investment idea is suitable for value investors who already understand what is a Content Delivery Network (CDN) and what’s the USP of Fastly. If you do not know what Fastly does, you make check out other explanations first then come back to visit this write up. Edge computing is known to be the number one reason that all other investors are betting on Fastly or Cloudflare. Akamai was founded in 1998 and was and still the preferred choice of CDN provider for the majority of the CTO while it was deemed as a legacy CDN.
Cloudflare and Fastly are competing at different segments of the CDN market
The next-generation CDN providers like Cloudflare focus on small business or startup while Fastly goes niche to focus on large customers with millions of visits or web requests.
Fastly’s revenue is heavily concentrated with 89% from 324 enterprise customers who collectively contributed close to $74 million. Fastly’s average enterprise customer spending is approximately $782,000 on a trailing twelve-months basis.
Cloudflare’s full year 2020 revenue is at $431 million which is almost 50% more than Fastly’s full year 2020 revenue of $291 million. Cloudflare’s have 828 enterprise customers contributing 46% of FY2020 revenue or approximately $198 million which simple calculation indicates an average enterprise customer spend of $240,000 on a trailing twelve-months basis. Cloudflare’s Non-GAAP gross margin is much higher at 78% compared to Fastly’s 60.9%.
The bottom line here is if Fastly’s customer made the migration decision from Akamai to Fastly and spent way more than Cloudflare’s average enterprise customer spending, the decision-maker would make a considerable comparison that Fastly’s tech is the best match the needs compared to Cloudflare’s mass-market approach.
Take a look at the top 100 most visited websites in the United States of those using hosted on Fastly’s CDN. At least 30 of the most visited 100 websites are suspected to be using Fastly which migrated from legacy CDN.
I would like to provide more fundamental reasons to support more than just Edge computing as the core / fundamental investment thesis. Akamai’s team had shared that edge computing had been around for a long time. In fact, Akamai’s edge computing innovations represent approximately 20% of Akamai’s more than 400 technology patents, and edge computing solutions generated well over $2 billion in revenue over the twelve months ended June 30, 2020. Although many investors are bullish on Fastly’s deep tech but bear in mind that Akamai has a huge arsenal of engineers than Fastly which indicates Akamai can at any one point of time out-compete Fastly. Hence, it’s best to ride on mega-trends that drive tremendous demand for next-generation CDN that can benefit more than 1 player. So, let’s dive into the 9 other reasons outside of edge computing on why value investors should be bullish on Fastly.
Reason 1: 5G Adoption will accelerate CDN & Edge Computing demand
The recent filing revealed that Warren Buffett’s Berkshire Hathaway is betting on 5G mass adoption with an $8 billion investment in Verizon while Verizon’s 4Q2020 earning presentation shared that 5G target sites are expected to grow to at least 30,000 sites during FY2021.
Traditional CDN infrastructure is not capable of delivering content for 5G speed and capacity and Fastly’s is one of the few that has reached over 100 terabits per second (Tbps) of connected edge capacity. But do note that Akamai once reached 181 Tbps as disclosed in an earnings call dated Feb 10, 2021. Akamai CEO also shared, “Nobody in the marketplace comes anywhere close to our capacity to serve customers at the edge on a global scale.” which is a warning sign to monitor moving forward for Fastly if Fastly were able to scale the infrastructure and surpass a new milestone.
5G will and has unlocked many new business use-cases that were never made available prior. Imagine a massive multiplayer game that streams within a mobile browser powered by 5G and edge computing. Check out Intel’s case study on “How to leverage 5G with Cloud Gaming” at https://www.intel.com/content/dam/www/public/us/en/documents/case-studies/playgiga-cloud-gaming.pdf
To monitor the strength of this investment thesis, one should measure the 5G infrastructure deployment status (supply) and also the 5G adoption rate (demand).
Reason 2: Live streaming event’s demand will accelerate demand for low latency network via CDN
Marketer had been making use of Zoom to conduct various live streaming sessions to reach out to prospects as studies find that live video has a higher engagement and conversion rate. One of Fastly’s customers – Twitch, had cultivated rising game streaming demand globally. During the year 2020, the majority of the world population had been used to enjoying the comfort of work-from-home to watch various business or leisure events via live streaming. In fact, 2020 is the year forcing many inaugural events to be hosted virtually and live for the very first time and it will not stop there. In a post-COVID-19 world, more physical events will be held and event organizers would recognize the demands of online viewers to sell live viewing access which pushes demand for the live streaming event further. The boom of short-video streamings and OTT streamings are another enabler at play currently. Approximate 40% of Fastly’s revenue comes from the video streaming industry at one point in time. Fastly’s enterprise customer – Vimeo’s recently filed S-4 spin-off exercises from IAC, it shared that “While the COVID-19 pandemic has accelerated the demand for video from businesses, we believe this shift is secular and will endure given the fundamental need for businesses to communicate online. For example, we believe that the businesses streaming their town halls, classes and events today will continue to value the global reach of live streaming along with their physical events in the future. We see countless examples of the power of this reach in our user base today”. Another of Fastly’s customers – Clubhouse also accelerating the demand for live events. TED Conferences would be another great example that Fastly powered the live stream. To monitor the strength of this investment thesis, one should measure the continuous trending momentum of the live streaming events.
Reason 3: eSports will accelerate demand for low latency network via CDN
Fastly’s 4Q 2020 letter shared that “Our efforts in this rapidly-growing vertical saw significant expansion with a large gaming company, who is leveraging our platform to deliver new titles and support the launch of the next-generation consoles. We also added a leading interactive gaming company to our customer base, supporting their storefront for games available via the web.”. In the same letter, it also shared an internal prototype use case, “edge-native, multiplayer version of the popular video game Doom”. This means that the computation of the entire game’s interaction was done in the distributed point instead of at the central point which could unlock more possibilities of gaming experience to game developers.
Fastly’s Q2 2020 letter also shared “As online gaming becomes a popular choice for entertainment at home, a developer of massively multiplayer online (MMO) games chose Fastly to optimize and accelerate in-game assets and interactivity to over 75 million of their browser and mobile casual games customers.”
Online games have risen in both desktop and mobile while the rise of the live massive multiplayer mobile game is on the rise too. Do note that the keyword here is “live”, “massive”, “multiplayer” and “mobile” which makes a big difference. A non-live mobile game like Candy Crush doesn’t require low latency. A non-massive live mobile game like a turn-based game such as Crash Royale or chess is a 1-on-1 play that usually has plenty of waiting time for the player to act before the times up. A massively multiplayer game is coined as Battle Royale which has 50-100 players within the game sessions like PUBG, Mobile Legends, Marvel Super War and Free Fire. The recent rise of battle royale was made possible thanks to Unreal Engine and Unity, while more game developers are tapping into the engine to develop a more competitive gaming experience. Lastly, desktop gaming on battle royale games was the mainstream compared to the mobile version of battle royale games due to the mobile device capabilities issue. My next reason will change the landscape and further boost the rise of mobile battle royale gaming.
Reason 4: Cloud gaming will accelerate demand for low latency network via CDN
Cloud gaming will enable device-agnostic gaming which allows gamers to have a level playing field. The current gamers are forced to purchase expensive smartphones or better desktop gaming gears to compete with other gamers. Cloud gaming built on the promise of zero install, zero download, zero update and low latency at any devices to play all games, To gauge the momentum of cloud gaming, one should monitor the number of games on these platforms: Nvidia’s Geforce Now, Google’s Stadia, Microsoft’s Xbox Project xCloud, Sony’s PlayStation Now, Amazon’s Luna.
Reason 5: Demand for reducing reliance on Amazon or Marketplace will accelerate demand for e-Commerce traffic outside of Marketplace
We are not talking about a boring topic of brick and mortar going to e-Commerce but we are talking about merchants having increasing concerns about fully relying on Amazon for their e-Commerce contribution. Shopify is enabling e-Commerce sellers to reduce their reliance on Amazon and level up the playing field. The growth of Shopify will accelerate growth for Fastly as Shopify is using Fastly. Approximately 60% of Fastly’s revenue comes from the e-Commerce industry at one point in time. And why e-Commerce will give rise to Fastly CDN adoption? Refer to Reason 8 for cybersecurity concerns. As e-Commerce merchants can’t afford to suffer downtime during the shopping period and it’s a wise investment decision to turn on Fastly to be ready for DDoS attacks.
Reason 6: Better channel partners with top 3 public cloud vendors such as Azure, AWS and Google Cloud will accelerate the migration from legacy CDN to next-generation CDN
The general-purpose CDN providers like Amazon Cloudfront / Google Cloud CDN / Microsoft Azure CDN / Cloudflare CDN are a one-size-fits-all approach that doesn’t meet the needs of enterprise customers who demand low-latency. However, it’s very hard for Fastly to win over enterprise clients who were using legacy CDN. The decision-maker for such an important migration to next-generation CDN like Fastly is made by the coders/back-end engineers / CTO who know the technicality. The typical procurement process in big companies is troublesome for tech guys as they need to jump through the many hoops for rallying support on such a decision or red tape if it’s a new vendor and if there’s a contract to commit for the long term. The reason being is big companies will negotiate a bulk-purchase rate with Fastly given that the company who wants to migrate can commit a full migration or a large bandwidth commit, which then involves a 12-months or 24-months lock-down contract. If the company that wants to migrate is already using Azure or AWS or Google Cloud, they just need to activate Fastly selection as an “add-on” without going through the lengthy procurement process as it’s all billed under the single billing account arrangement within the cloud vendor. Hence the recent partnership with the cloud vendors is expecting an acceleration of more enterprise companies to migrate away from legacy CDN. Google Cloud’s partnership with Fastly was announced in September 2020, while the partnership with Microsoft Azure was announced in January 2019. However, I had yet to verify whether the single billing account arrangement was in Azure or AWS. The bottom line is that the great relationship with the cloud vendor indeed accelerates the transition.
Reason 7: Hyper-growth in real-time data analytics to make a real-time decision at the edge will accelerate demand for low-latency CDN
Again, we are not talking about Big Data but real-time data analytics. The increasing inter-connected data points across multiple IoT devices in various locations of the world give rise to the needs for edge computing to enable enterprise customers to make faster decisions. This real-time data analytics trend was enabled by Snowflake, Palantir, MongoDB, Cloudera, Teradata and all major cloud providers. The real-time logging powered by Fastly and the massive millions of event logs delivery per second are the key considerations for delivering real-time data analytic solutions. This is a game-changer as the nature of a content delivery network is meant to be distributed across multiple places and it will take hours or days to fully receive all the delivery logs across all of the distributed networks.
Reason 8: Spike in cyber-attackers’ capacity and capabilities in DDoS Attack will demand a CDN with stronger cybersecurity
As shared by Cloudflare’s blog post dated 18 Nov 2020, the attack size, frequency and sophistication of DDoS attacks are concerning the enterprise customers which then accelerate the migration to next-generation CDN which can provide massive logs in real-time to the cybersecurity monitoring tools.
Reason 9: Internet penetration and hyper-growth of OTT activities in emerging markets accelerate the acquisition of new enterprise customers
CDN provider selection for enterprise customers is not an all-in decision as there’s a multi-CDN approach via a smart selection or dynamic selection. Hence, don’t be surprised on why some big enterprise customers were using Akamai or Limelight and also concurrently using Fastly which was due to Fastly’s capacity availability in the markets which the enterprise customers wish to tap into that the legacy CDN does not have sufficient capacity. Another reason for the multi-CDN approach is also to prevent “single point of failure” and hence introduce redundancy to the architecture as in black swan events that one CDN provider does fail, the enterprise company can switch to another CDN provider dynamically. As of 4Q 2020 disclosure, Fastly is already in 56 markets and 26 countries while 30% of the revenue contributes outside of the United States. Riding on Fastly’s existing customer’s needs, Fastly is expanding to more new markets for capturing more revenue streams. For example, Fastly’s customer – Taboola is serving ads on Singapore Press Holdings (SPH)’s digital media assets and hence having PoP in Singapore. As per Taboola demanding for more presence within South Asia, Fastly could leverage on the demand and build more new markets.
Here are a couple of potential risks that might derail the investment thesis:
Risk 1: Patent infringement
Google Patent Search tool is showing that Akamai has 575 results (https://patents.google.com/?assignee=Akamai&num=100&oq=assignee:(Akamai)&sort=new) while Fastly have only 56 results (https://patents.google.com/?assignee=fastly&num=100&oq=assignee:(fastly)&sort=new). As of December 31, 2019, Fastly had 40 issued patents, 47 patent applications pending for examination, as well as 10 pending provisional applications. In the same period, Cloudflare had over 110 issued patents and over 70 pending patent applications. However, this doesn’t mean that there’s potential infringement but such high-value tech would face it soon or later. Hence, this is one risk that investors should price-in too.
Risk 2: Key-man risk on innovation
As the Google Patent Search tool indicates that the founder Artur Bergman represents as the major patent inventor for 46.4% of the Fastly patent search results, the next-generation product vision and innovation is heavily concentrated on one single person which is the current Chief Architect and Founder, Artur Bergman. As of December 31, 2019, Fastly had 191 employees in Fastly’s research and development group which the research and development expenses were $46.5 million in the year ended December 31, 2019. The key-man risk should be reduced gradually as the R&D department increases as domain knowledge is being compounded and transferred across. Akamai’s payroll and related costs in research and development expenses for the years ended December 31, 2019, were $382 million which is 8 times more than Fastly. Furthermore, Akamai’s net cash provided by operating activities for the year ended December 31, 2020, were $1.2 billion with a significant war chest to accelerate innovation ahead of Fastly. This risk was arised due to Artur Bergman selling close to $200 million worth of Class A common stock after he exercised employee stock options between Nov 2019 till Feb 2021. I would hope I’m right that he’s using his personal money to fund a stealth team outside of Fastly for accelerating innovation. Or else, this is not a good sign.
Risk 3: Increasing OPEX or burn-rate and doesn’t meet the rule of 40
As a final risk, Fastly’s incoming earnings quarters might be recording increasing CAPEX (expect to spend 11%-13% of revenue on CAPEX) or increasing researching and development expenditure or possible many acqui-hire related activities which might cause many sell-side analysts giving bearish call due to declining free cash flows. Due to aggressive bandwidth purchase which indicates the risk of increasing unutilized bandwidth over the quarter, the gross margin might be volatile as it doesn’t achieve the economics of scales during growth in capacity expansion. Fastly is an emerging investment idea that gearing up for the long-term instead of the near-term. Hence, if you are not ready for short term losses (<24 months), you should not hold this stock and should move to other investment ideas. This investment idea is not matured yet and bets on multi enablers to accelerate the revenue growth.
Risk 4: Market correction on Price-to-sales ratio
Fastly is currently trading at 30-35 times price-to-sales ratio and revenue grow at merely 30%-40% compounded annual growth rate (CAGR) which there’s no one SaaS company in history with such growth rate to trade at such a high price-to-sales ratio (Please correct me if I’m wrong but I guess it would only be a handful that ever trades at such high ratio for a long period of time). It’s understandable that such 30-35 times of price-to-sales is justifiable if the company is growing at 80%-160% or more but Fastly isn’t. Currently, Investors are too bullish to price-in Dollar-Based Net Expansion Rate (DBNER) which is indeed growing at a range of 136%-147% for sustaining the 3X price-to-sales ratio.
Risk 5: A sudden and major security breach to Fastly’s CDN
As Fastly championed for security capabilities, it would be a fool to not price-in a possibility of corporate espionage that attempted to score a successful breach in Fastly’s CDN. This risk was highlighted due to this disclosure: “On July 29th at 00:00 UTC, Fastly was notified by a customer (customer X) that a single log line intended for a different customer (customer Y) was received by customer X’s log system.” https://www.fastly.com/security-advisories/incorrect-delivery-of-partial-log
Risk 6: A predatory acquisition at play by the public cloud vendors
This might be too much of a conspiracy theory but it makes sense for either Microsoft, Google, Amazon or IBM to acquire Fastly as analysts are continuously bearish on Fastly to a point that the stock is trading at 20 times the price-to-sales ratio and the cloud major would able to justify the valuation to make an acquisition offer and the trade position would be unable to realise its maximum return. Amazon’s Twitch, Amazon.com and IMDB are using Fastly. Fastly is currently traded at around $9 billion market capitalization. The risk of such acquisition is high if the market capitalization goes below $5 billion at any one point in time. This risk was surfaced due to rumour that Cisco is planning to acquire Fastly (Source: https://www.msn.com/en-us/money/topstocks/heres-why-fastly-stock-is-soaring-today/ar-BB1bDFqz)
Risk 7: weaker growth in DBNER and enterprise customer acquisition
|Sales & Marketing Expenses ‘000||Cloudflare||Fastly|
Historically, Fastly is not a sales-driven SaaS company with a mega team of sales personnel or sales engineers as they are busy catering to inbound / organic requests. Since IPO, Fastly has gear up the sales and marketing expenses from $50 million a year to $100 million a year. However, the conversion to number of new enterprise customers added is low compared to Cloudflare’s size of enterprise sales team and the number of new large customers added. This risk would be mitigated by the new Chief Revenue Officer – Brett Shirk, who has almost 25 years of experience in SaaS enterprise sales to transform Fastly into both an innovation-driven and sales-driven organization.
Opportunity 1: Revenue Expansion from existing enterprise customers
Sell-side analysts and the investment community who are bearish due to too much focus on a single blue-chip customer – Tiktok while should also be reminded that Fastly had even more blue-chip customers which is way bigger than Tiktok. At the moment, Fastly already has 324 enterprise customers. Some of Fastly’s enterprise customers are mega in nature which have many new business units opening at nearly every quarter while also acquiring rivals or new verticals every year. Take a look at Gannett who currently owns 383 websites which collectively attracted 160 million unique visitors a month. Another of Fastly’s enterprise customers – Reddit had raised $367 million and aiming to raise up to $500 million for expanding the ever-growing community size and expecting reddits time-spent to grow to new high. Last example – Shopify is growing at 94% year-on-year in the top line as well which Shopify’s merchants are indirectly fueling the usage of Fastly’s edge computing solution. These are the key under-currents that are fueling the future growth of Fastly’s revenue.
There are some public sources indicating who are the customers but I had gauged the customers by using danetsoft website to identify the origin server was hosted with which CDN provider. However, there’s no real evidence that danetsoft might be accurate in knowing the company’s architecture is using Fastly or not while it might not be a mutually-exclusive customer due to the fact the top websites tend to deploy multi-CDN selection. Here’s the link: http://data.danetsoft.com/.
Suspected list of Fastly’s current enterprise customers:
Video Content: TikTok, Twitch, Vimeo, fuboTV, Super Bowl, Atresmedia, MediaPro, Fox TV & Sports, Viacom CBS & CBS Sports, A&E
Sound Content: Spotify, Clubhouse, Shazam, 7digital
E-commerce and related SaaS: Amazon.com, eBay, Target, Etsy, Shopify, 1800flowers.com, Ticketmaster, Wayfair, Grubhub, Deliveroo, KAYAK, Alaska Airlines, Airbnb, HomeAway, Hotel Tonight, JetBlue, Tripadvisor, Eurostar, Holidayextras, Aer Lingus, Love Holidays, Dollar Shave Club, Big Cartel, Dunelm, Yottaa, Adobe Magento, CommentSold, Wanelo, Kickstarter, La Redoute, Daniel Wellington, Boots UK, Catch Digital, 1stdibs, Adorama, Casper, Overstock, Rakuten
Software / SaaS: Slack, Firebase, Taboola, Wix, New Relic, Pantheon, Jimdo, LaunchDarkly, imgur, Sonatype, Opera, Drupal Association, HashiCorp, Speedtest, Hoodoo, Cloudera, Chartboost, Firebase
Publisher: Fandom, CNN, CNET, Gamepedia, IGN, Buzzfeed, BBC, Fast Company, Forbes, The New York Times, Financial Times, Guardian, Gannett (USA Today), Nikkei, Vox Media, WIRED, Business Insider, Condé Nast, Hearst, PRISA, RCS MediaGroup, ZEIT ONLINE, Lonely Planet, Dictionary.com, Urban Dictionary, Wikia, Wenner Media
Investment Content: The Motley Fool, Seeking Alpha, Investopedia, Bloomberg
Financial / FinTech: PayPal, Stripe, Affirm, Wealthfront, Chartbeat, Virgin Money, RVU, Simple, Marqwta, wepay, Bankrate, GoFundMe, ConsenSys
Social Network / Community: Twitter, Reddit, Quora, GitHub, Stack Overflow, Stack Exchange, Pinterest, Foursquare, Yelp, Meetup, Disqus
Education: Blackboard, UC Davis, MIT, Khan Academy, TED, Skillsoft, AAMC, Education.com, Kahoot, Inspired eLearning, Coolmath.com, Nearpod
Game: Paradox Interactive
Opportunity 2: WebAssembly or WASM
Fastly built Compute@Edge based on WebAssembly. Fastly (along with Google, Microsoft, Amazon, Cloudflare engineers) been a supporter and contributor to the emerging open-source software development framework – WebAssembly or known as WASM. WASM is built on the promise that allows any internet user to run the web app at near-native speed. Layman explains it allows apps to run faster due to it being compiled in lightweight to run on your device / browser instead of the server’s end as it skipped the latency issue. However, WASM is still in its infancy stage while tackling many problems on it’s way. Taking a look at the number of questions tagged to WASM on stackoverflow, it’s not bad at all with 1,520+ questions but the developers crowd is still relatively smaller than other emerging technology. (https://stackoverflow.com/questions/tagged/webassembly)
This opportunity has arised due to Fastly sponsor-hiring the Mozilla’s Wasmtime and WASI team to work on outside-the-browser projects. Ex-Mozilla & now-Fastly’s Aaron Turon also tweeted that they are expanding the team around Aug 2020 which is just 2 months before the hiring announcement of the Mozilla team. The WebAssembly Fastly’s customer – Shopify leverages WASM to enable third-party apps to be run with WASM on Shopify merchant’s site.
Here are the price target given by these analysts with average price target from $49 to $110 with average price target at $84:
- $110 by Timothy Haron @ Oppenheimer
- $105 by Rishi Jaluria by DA Davidson
- $95 by Robert Majek @ Raymond James Financial, Inc.
- $90 by Brad Zelnick @ Credit Suisse
- $90 by William Power @ Robert W. Baird
- $85 by Jeff Van Rhee @ Craig-Hallum
- $77 by Brad Reback @ Stifel
- $75 by Tal Liani @ Bank of America
- $65 by James Fish @ Piper Sandler
- $49 by Walter Pritchard @ Citi
Fastly’s weighted average basic shares outstanding is expected to be 118.4 million for the full year 2021. Fastly’s last traded share price was $75.76 on 25 February 2021 with a price-to-sales ratio of 31 times. At the high-range of FY2021 revenue guidance at $385 million, it implies YoY revenue growth rate of approximately 30%. FY21 forward revenue per share is approximately $3.25 per share, implying a reasonable price-to-sales ratio at 23 times.
Cloudflare’s latest weighted-average shares outstanding is at 303.8913 million. Cloudflare’s last trade share price was $76.43 on 25 February 2021 with a price-to-sales ratio of 54 times. At the high-range of FY2021 revenue guidance at $593 million, it implies YoY revenue growth rate of approximately 38%. FY21 forward revenue per share averaged is approximately $1.95 per share, implying a high price-to-sales ratio at 39 times on the back of a mere 38% growth rate. This is the main reason why even Cloudflare is an amazing company but it’s still too expensive for investors to even consider at the moment.
I would value Fastly with a 2-years Revenue CAGR of 30% and 30 times price-to-sales ratio, which then derive a target price of $125 which represents approximately 64% upside over the course of 2-years.
Bonus: Warning signals to be monitored over the earning release:
- Declining annual revenue retention rate (Now at 99%)
- Declining Dollar-Based Net Expansion Rate (DBNER) (Now at 133%)
- Stagnant count of enterprise customers (Now at 324)
- Revenue YoY growth rate declined consecutively (Now at 40%)
- Compute@Edge is yet to surface as a major revenue contributor or the count of customers using Compute@Edge is low (Unknown or Nil at the moment, to be verified). However, there are 60 customers participating in the beta program of Compute@Edge.
Websites and insights that helped to support the above investment thesis and huge credits to the content owner:
- Fastly’s customer empowerment conference – Altitude’s video: https://czarviz.com/fastlyaltitude/
- Testimonials and Educational Content by Fastly on Vimeo Channel: https://vimeo.com/fastly
- Fastly Investor Relations: https://investors.fastly.com/
- Vimeo’s S-4 Filing: https://www.sec.gov/Archives/edgar/data/0001837686/000110465920139169/tm2037917-1_s4.htm
- Fastly’s S-1 Filing: https://www.sec.gov/Archives/edgar/data/1517413/000119312519111675/d702138ds1.htm
- Jaguar Analytis’s research: https://www.jaguaranalytics.com/wp-content/uploads/2020/02/FSLY.pdf
- Christopher Seifel’s research: https://twitter.com/SeifelCapital/status/1295843014275993602
- Peter Offringa’s research: https://softwarestackinvesting.com/fastly-fsly-q3-recap/
- Edge Computing Use Case: https://medium.com/@mfcaulfield/edge-computing-9-killer-use-cases-for-now-the-future-ed2083ff4e6c
- Edge Computing, the devil is in the details: https://www.fiercewireless.com/tech/editor-s-corner-edge-computing-devil-details
- The Edge Completes the Cloud: A Gartner Trend Insight Report: https://www.gartner.com/en/doc/3889058-the-edge-completes-the-cloud-a-gartner-trend-insight-report
One last thing, a bonus reason to consider is Cathie Wood’s ARK ETF also bullish on Fastly https://cathiesark.com/ark-combined-holdings-of-fsly, but do your own due diligence and don’t be a herd.
Happy to be corrected or debated! Please leave your comments below.