Updates: Cloudflare - The Three S-Curves (Pt.1) - How Will NET Navigate API Security Market Dynamics?

Summary

• It has been a while since we conducted a deep dive on Cloudflare. This has turned into a 15k word report that we're going to break down in four parts.
• As Matthew Prince has been referring to Cloudflare's journey in Act 1, 2, and 3, we're going to use the same framework for analyzing their business.
• Hence, this Part 1 will cover Act 1 - their mature application performance & security business - which is being influenced by industry dynamics related to API security.
• Part 2 will cover Act 2, Cloudflare's prospects in SASE. As per the previous Fortinet and Palo Alto recap reports, we will discuss the broader networking and NetSec landscape too, aka the middle and last miles.
• Part 3 will cover Act 3, Cloudflare's edge compute offering and we will review the competition.
• In Part 4 we'll discuss the latest quarterly results and provide our valuation analysis.

Intro

As a tech investor, Cloudflare (NET) needs little introduction. When NET was founded in 2009, its mission was to help make the internet faster, safer, and more reliable for users. This necessitated a rapid build-out of a globally dense network, which the company achieved via a three-pronged strategy: 1) colocation data centre arrangements (mostly with Equinix), 2) purchasing OEM COTS (Original Equipment Manufacturer; Commodity Off-the-Shelf) x86 servers, and 3) applying then-pioneering SDN (Software-Defined Networking) to operationalize the data centres and connect them together.

As you may already know, adopting a colocation data centre strategy is a great way for a vendor to expand their network fairly quickly (though not as quickly as using pre-existing PoPs, such as those owned by the Cloud Service Providers, or CSPs) while retaining full control of the infrastructure (which is not possible with CSPs). Moreover, it saves substantial costs because they are paying Equinix’s 40-50% gross margin rather than AWS’ 65-75% gross margin.

NET also saved significant costs by bypassing premium server brands, like Dell and HPE, and buying basic x86 servers directly from OEMs, such as Quanta and Lenovo, and leveraging the open-source Open Compute Project to design its own server for the OEMs to manufacture. By doing so, NET avoided the higher costs associated with the additional services, proprietary technologies, and brand premiums that companies like Dell and HPE include in their pricing. This allowed NET to efficiently scale its infrastructure while maintaining control over its capex, expediting its expansion.

However, the data centre and hardware strategy would not have been worthwhile if it were not for the SDN. NET leveraged SDN on their COTS hardware to innovate and deliver pioneering services to website owners and Internet users. By utilizing SDN, NET transformed basic x86 servers, equipped with Linux, into highly efficient routers and switches. This was achieved by configuring Linux routing tables, enabling these servers to manage network traffic dynamically and intelligently.