Inside Databricks’ Contrarian Playbook: Burn $1.5 Billion to Buy Big Growth

The company is predicting the cash burn will be rewarded with a revenue growth rate of 50% or higher for each of the next three years, as it looks to get more companies hooked on its tools that help data scientists make sense of large pools of data. The company expects to pull in about $1.6 billion in revenue this fiscal year, which ends in January.

San Francisco-based Databricks also expects to get closer in size to publicly traded rival Snowflake, which had nearly twice the revenue as Databricks last year. Databricks forecasted that its sales will be 80% of Snowflake’s by the 2026 fiscal year, according to Databricks’ own projections and FactSet’s forecast of Snowflake. That would put Databricks’ revenue at about $3.7 billion.

Jenna Kozel King, a Databricks spokesperson, said the company couldn’t comment on private financials.

The company is anticipating a continued surge in corporate interest in harnessing large-language models made popular by ChatGPT. The company’s revenue forecasts could fall flat if interest subsides. But companies, especially before a potential initial public offering, tend to keep forecasts somewhat conservative to make sure they live up to expectations with large institutional investors who are writing checks. A potential Databricks IPO is closely watched among bankers hoping for more high-growth tech firms to increase investor appetite for public listings. Ali Ghodsi, Databricks’ chief executive, said recently that the firm would eventually go public but wasn’t in a rush.

Databricks has positioned itself as a kind of anti-OpenAI in the way it sells software to help companies make their own AI apps from scratch. The company makes money through subscriptions to its set of tools that essentially gives data scientists and engineers instructions on how to use vast amounts of data they collect about customers and other parts of their business. It also charges customers based on how much they use those tools. Microsoft, OpenAI’s largest investor, also plans to start selling a new version of Databricks’ software that helps customers make AI apps for their businesses, The Information reported last month.

To be sure, Databricks isn’t in danger of running out of cash after raising its last round at just the right time. The company raised $1.6 billion from Morgan Stanley and other investors in August 2021, about seven months before the Federal Reserve started hiking interest rates, which increased capital costs for startups. It had about $2 billion in cash on its balance sheet at the end of last year.

But the company is about to reload. T. Rowe Price, the mutual fund giant that had already invested in Databricks in 2021, last month agreed to invest more than $250 million in the company, one of the people said. It’s unclear who else plans to invest in the new round. The $43 billion valuation would make Databricks the third-most valuable U.S. private tech company, after rocket company SpaceX and payments firm Stripe. The Information first reported the funding talks, and Bloomberg first reported its potential valuation.

The funding round won’t increase the value of existing Databricks’ shareholders’ stakes in the business, however. The company expects to raise funding at about the same per-share price as last round, people familiar with the matter said.

Databricks’ losses have accumulated because of its high operating expenses—including high-priced engineers—which exceed how much revenue Databricks generates, even without accounting for stock-based compensation expenses. Another big investment by the company has been in salespeople. The company increased its number of account executives by 80% last year and expects to grow that headcount by half again this year, the people said. Overall, Databricks’ headcount grew by 65% last year, during a time when most tech firms were trying to reduce their workforces.

Databricks, meanwhile, has also been one of the most aggressive private tech firms on the mergers and acquisitions market. In June, it bought AI startup MosaicML for $1.3 billion in a mostly stock deal.

Overall, the company planned to slow its operating expense growth this year, from about 80% last year to about 37% this year, the forecasts show. One factor that may embolden Databricks is that it has high gross profits—which exclude operating expenses as well as stock-based compensation. Its gross profits are nearly 80% of revenue, on par with buzzy publicly traded software firms like DataDog and higher than other well-known software firms like Salesforce and MongoDB.

Databricks’  forecasts show that investors putting money in at a $43 billion valuation would be giving the company a significant valuation premium over Snowflake. At that level, investors would put about $25 of value on each dollar of revenue Databricks expects to pull in this year, compared to about $17 for Snowflake, which is publicly traded with an enterprise value of about $50 billion. Snowflake is profitable, larger and generates cash, but is growing more slowly than Databricks and has slimmer gross margins.

The valuation gap is even greater when looking at the company's projected operating income in 2026. Investors in Databricks would value each dollar of its operating income at $800, about 10 times where Snowflake trades.

The two companies compete head to head for some customers. Databricks’ focus expanded in recent years to data warehousing, a type of database software designed for speedy data retrieval. Databricks’ data warehousing product, which competes with Snowflake’s core product, was generating more than $130 million in annual recurring revenue, the company said a few months ago.