Late last week, independent journalist Eric Newcomer reported that Databricks is raising new capital at a valuation of “about $27 billion.” A few days later, another publication chimed in, saying that they had heard the round could be worth $29 billion at a slightly higher valuation.
Well, well!
Last year, The Exchange covered Databricks’ financial progress as a private company. Databricks, as a refresher, provides its customers with analytics and data science tooling and crossed a $350 million run rate at the end of Q3 2020.
That figure was up from $200 million in the year-ago period. As we wrote at the time, Databricks was “an obvious IPO candidate” and a company with “broad private-market options.” Reports that it has raised more capital underscore our previous notes.
But we took it all one step further after news surfaced that Databricks could go public in the first half of 2021, noodling around with all the financial information we could scrape together for the company to come up with a valuation range. Our resulting figures were a bit low compared to recent news, which forms the crux of our work today: Can we come up with a set of numbers that help make sense of Databricks at $27 billion?
Databricks declined to comment. But that won’t stop us from having fun. So, let’s remind ourselves of what we know about Databricks’ growth history, economics and scale.
From there we will be able to check our estimates against its purported new valuation range and come up with some implied multiples. Then, we’ll contrast those with some high-flying public companies.
Do the numbers somewhat fit? Can we see Databricks making sense at more than $25 billion, more than four times its 2019-era private valuation of $6.2 billion? Let’s find out.
What’s it worth?
In our previous work, we ran a number of growth scenarios to come up with different estimates for Databricks’ current scale. Sparing you several hundred words, given how the company grew from $200 million in annual run rate to $350 million between Q3 2019 and Q3 2020, we estimated that the company would close Q1 2021 with between $425 million and $486.5 million in annualized revenue.
Looking at different data points from the Bessemer Cloud Index at the time while using some flexible-but-not-conservative estimates for Databricks’ revenue quality, we came up with market comps that would have given it a sales multiple of between 20x and 38x. At the high end of our revenue run rate guess for Q1 2021, and top multiple, you get a valuation of $18.5 billion.
That’s far less than $27 billion. So, something is wrong in our series of estimates. That could include:
- Databricks growing more quickly than we estimated, making it both larger and more attractive to investors.
- Databricks having more attractive revenue economics than we anticipated.
- Databricks having stronger SaaS metrics than we expected.
- The markets changing since we last wrote, repricing Databricks.
The first three options are obvious and impossible to parse. We took shots in the dark based on flashes of light. We cannot do much more sans a flashlight. But we can peek at today’s comps and see what we can learn.
Previously, we employed Okta and Slack to generate the sales multiple comps for Databricks, as they shared some characteristics that we felt formed a good zone of possibility regarding the private company’s possible public value. Looking today, those two companies are worth the following:
- Okta: 38.5x enterprise value/annualized revenue.
- Slack: 25.5x enterprise value/annualized revenue.
Slack’s number has gone up a bit thanks to Salesforce, but Okta is about where it was before. This means that the top end of our estimates stay under the $20 billion mark. And yet, $27 billion?
Perhaps investors agree with Newcomer’s perspective that Databricks “has a similar investor story to Snowflake.” Snowflake, of course, went public to great effect and has one of the richest revenue multiples in the world. Yahoo Finance lists it as 166.22x as of this morning, using a trailing revenue number instead of an annualized quarterly figure. So, the multiple would be a bit smaller if we used the same revenue-counting method used for Databricks, but still far higher than what other public software companies have managed.
Perhaps Databricks is just more attractive to investors than our selected comps would imply, or its growth has been rather impressive since Q3 2020. Either way, the prices that are bouncing around are indicative of a round that, if closed where it is reported to be circling, would imply something notable going on at the unicorn.
Returning to our original question: Can we come up with a set of numbers that help make sense of Databricks at $27 billion? No.
From where we are, the $27 billion number reads rather high. This is one S-1 that we very much want to read.