The investor love affair with “developer friendly” enterprise SaaS is still in full bloom. The last entry into the market, Zoom ($ZM), had a stunning debut. Our next contestant is Fastly ($FSLY) which is scheduled to price on Thursday night.
We’ve been covering software infrastructure for 30 years. During the past ten (or so) it’s become commonplace for companies providing complex and scalable software infrastructure (like Stripe, New Relic ($NEWR), Twilio ($TWLO), Atlassian ($TEAM), Slack ($SK), MongoDB ($MDB) and Elastic ($ESTC) to offer developers the opportunity to incorporate their technology into their new applications with little or no interaction with those dreaded “sales types.”
Costs become real when new applications are deployed and they typically start quite small but scale up rapidly based on usage. At some point they become large enough to be considered an “enterprise sale” by institutional standards. But this happens much latter in the process and because it remains in the SaaS model the revenue is recurring and can continue to grow. Line of business applications have been going to these early “self service” model as well led by companies like HubSpot ($HUBS) and Zendesk ($ZEN).
Content Delivery Networks (CDNs) have been around for over 20 years. Akamai ($AKAM) invented the category and dominated the market during the first stages of the internet. The Victoria’s Secret fashion show couldn’t happen without them. Akamai has built up a huge infrastructure of servers in datacenters and follows what you might call a “we do it for you” model which works well for certain types of content, especially when the are static and not interactive, like video streams and large files.
This first generation of CDN technology wasn’t designed for the high level of interactivity common on mobile devices and the fine-grained control that is required to effectively service thousands of different client types at the same time. During the same period security has become more of a challenge and companies need their infrastructure to help them be effective while providing high levels of security.
This is where “next generation” CDN vendors like Fastly and Cloudflare come into the marketplace. Both of these companies offer comparable solutions with Cloudflare being the larger and more established of the two. Cloudflare grew out of the security space and that heritage is manifest in some aspects of their solution (like DNS control) whereas Fastly is based on an open-source software project called Varnish which grew out of the media world. There is even a competitor in the form of Varnish Software which offers a commercial version of the open-source product.
What is Fastly?
Although described as a CDN Fastly is really more like software-defined networking in the cloud. Rather than relying on specialized networking hardware (like routers) this software enables general purpose hardware to perform strong feats of network-oriented processing.
First imagine high performance general purpose server a node in the network. Here’s a crude picture of one with the underlying subcomponents. Think of the boxes below the Arista switches as virtual servers.
This gives you a highly scalable way to deliver processing and content management. Put things get more interesting when you deploy them into a cloud and add layers of control software that can work together. That picture looks more like this one:
In this case everything is in the cloud but the control plane and management tools can still provide very fast granular control all the way to the “edge” of the cloud where a device or application interfaces.
Because the vast majority of this runs on general purpose hardware it is much more programmable than traditional networking hardware.
Competition & Market Dynamics
The big dog in this space is Akamai ($AKAM) which has built a $2.7B business as a CDN and has a $12.5B market capitalization. Akamai is characterized as a kind of “legacy CDN” which is hotly contested by some. They have made some adaptations and improved their r offerings to be more competitive with upstarts like Fastly but their entire base of software was designed and built in a different era.
One casual example illustrates this well. Being able to purge content in the network quickly is an important feature of any CDN. Akamai worked hard to improve this aspect of their system and got it down to a matter of seconds (in most cases.) But a developer who uses the Fastly platform wanted to purge their cache thousands of times per second. Order of magnitude differences like this can’t be bridged by patches to an existing system.
Akamai will nonetheless be a competitor, especially in large corporate accounts. They have strong sales teams and an installed base that won’t be given up easily. Akamai does not have a developer-friendly sales model however so there is little they can do to slow down Fastly in terms of onboarding developers and DevOps teams.
There are other startups in the space including Varnish Software and the well-funded Cloudflare that are developer-friendly and offer solutions that are much closer to what Fastly provides. The outcome will depend on how fast these companies can deliver new software and new paying customers.
This section wouldn’t be complete without acknowledging a meme in this market which is the threat of Amazon (AWS) taking the same open source (in this case Varnish) and rolling out their own service. Investors have heard this with respect to companies like MongoDB ($MDB) and Elastic ($ESTC) and are definitely worried about it.
It’s hard to estimate the TAM for this market but it’s clear that it’s measured in the tens of billions per year – especially given the opportunity to substitute cloud-based general purpose hardware for specialized networking gear.
Valuation and Stock Conclusion
Investors are already bidding up the value of $FSLY even prior to pricing so we will have to see where it prices and trades. These stocks have been flying past our valuation estimates based on PFV in this market – sometimes on the open. This is one we would like to own but it’s going to depend on the price.
If one applies a P/S multiple based on how some of these names are trading now the shares could change hands at $35.