Study the Legends: Infrastructure IPOs
For any early-stage VC, it’s important to study past success patterns as you evaluate new opportunities. At Bain Capital Ventures, I focus exclusively on infrastructure software, because I’m genuinely curious to learn about these types of products, and I think that a number of factors make infrastructure the most exciting area for investment in the world of software. I took a little time this week to “study the legends,” as they say, and learn from some of the best infra companies in the business.
To start, it’s useful to establish a definition of what exactly we mean when we say infrastructure as each infra investor seems to have his or her own unique definition. Some say it’s any software which is relatively complex to architect and build. Others say it’s anything with a technical buyer or technical user. My general criteria is any software that enables the development, delivery, or security of other software. This encompasses everything from horizontal API products to endpoint security to databases.
To start, I curated a list of all infrastructure companies which have gone public on U.S. exchanges since 2015. In a future analysis, I’d love to look at infra companies which have gone public over the last twenty years, but in attempting this, I quickly reached a point at which I couldn’t graphically represent everything that I wanted to.
The first graphic plots companies with the year in which they initially listed against their valuation at the time of listing. As recently as seven years ago, the upper echelons of infrastructure exits were pegged around $1B. Companies like Atlassian, Twilio, and Pure Storage were the high flyers of the day with initial valuations ranging from $3-5B. In today’s world, the hottest infra companies like Confluent, Gitlab, Hashicorp, and of course Snowflake list with valuations north of $10B.
Of course, a large part of this shift is explained by companies’ decision to delay IPO, leading to a higher market cap when listed, but part of the trend is due to growth in markets for infrastructure software. For example, prior to Snowflake, the enterprise data warehouse was ubiquitous, but the need to manage warehouse hardware and over-provision compute acted as an entry barrier for companies with fewer resources. Snowflake’s flexible architecture improved the status quo, but also expanded the market.
While the data warehouse market sits at $21B today, it is projected to grow to $51B by 2028 as Snowflake and other cloud-first products expand the breadth of possibilities. It’s worth noting that the chart above is not perfectly to scale because Snowflake’s valuation at IPO is such an outlier. But in any case, a similar dynamic exists with companies like HashiCorp, GitLab, and Confluent. The infrastructure-as-code paradigm made popular by HashiCorp’s products was an incremental improvement beyond standard cloud hardware provisioning methods of the day, but it also made the cloud more accessible to companies with fewer resources and expanded the size of the market, leading to a higher IPO outcome.
Breaking these infrastructure IPOs down into category (very basic categories, I should add), it’s interesting to see the balance in value creation across infra. If you control for Snowflake’s $33B market cap at IPO, the dev tools and ops section would more clearly lead the way with $63B in value creation.
Notably, only a couple companies in the AI/ML space have gone public since 2015. I think this shows both the untapped potential of this area of investment in private markets, but also the difficulty of productionizing ML applications. The complexities involved in taking working ML models into a deployed, customer-impacting environment are non-trivial, and this is an area where my team spends a lot of time.
For the fun of it, I placed the highest valued infra IPOs from the last few years on a bar chart to create a kind of stack ranking. It’s left as an exercise to the reader to mentally impose an updated version of this chart with today’s current market caps, but a few companies (e.g. MongoDB, which didn’t even make this list) have sustained high growth rates and would sit at much higher positions in the rankings today.
Though I spend time with companies at the earliest stages in their journey, I plan to continue studying the legends of my space, and I’m excited to see what this similar charts will look like in a few more years!