Databricks Boosts Sales Forecast, Driving Valuation to $134 Billion
The Takeaway
- Databricks seeks $5 billion at $134 billion valuation, 32 times sales
- Databricks projects 55% sales growth, but gross margin has fallen to 74%
- CEO Ali Ghodsi has warned of an AI bubble despite Databricks’ growth from AI
Databricks CEO Ali Ghodsi has been more vocal than other Silicon Valley executives about the dangers of an AI bubble. His company’s new $5 billion round of fundraising, at a higher valuation than ever, epitomizes the risks and rewards of the boom: While sales are growing faster than forecast, AI development costs are squeezing gross profit margins.
Privately held Databricks began asking investors for cash two weeks ago in a fundraising round that values the company at $134 billion, which is roughly 32 times this year’s expected sales of about $4.1 billion, according to investor documents and a person familiar with the matter. Last year, Databricks raised cash at 24 times that year’s sales, and the year before, at 26 times.
Investors appear to be willing to give Databricks a higher multiple because it has been able to continue a streak of several years with a more than 50% growth rate, an unusually fast pace among mature software firms.
In fact, the company has increased its sales projections at least twice this year, according to the documents and the person. In September, it revised its sales projection from $3.8 billion to $4 billion, before revising it upward again slightly. It now expects sales to grow by 55% this year.
At the same time, the company has told investors its gross margin is falling faster than anticipated, to 74% compared to an earlier plan for 77%, due to increasing usage of its AI products.
Overall, Databricks is operating at roughly breakeven, anticipating about $10 million in free cash flow this year. That’s a big improvement from when the company was burning hundreds of millions in cash a year, as recently as 2023. But it’s far less profitable, by the measure of free cash flow margin, than similarly sized public companies like Palantir and Snowflake.
Databricks’ valuation puts it in the middle of the pack relative to some of its rivals. Snowflake and Datadog are valued at 21 times and 16 times expected sales, respectively. Palantir, a favorite stock among retail traders, is the only publicly traded software firm with a richer valuation, more than 90 times this year’s expected sales.
Databricks sells a data lake, a type of database for storing different types of corporate data. Recently, it has been trying to win more business by convincing customers they can use its AI agents to automate human resources and IT service management tasks, replacing more traditional software. Databricks has said it derives about a quarter of its revenue from AI products.
The company has close ties with OpenAI, one of its biggest customers. In September, Databricks said it would spend $100 million on OpenAI models over several years. It’s previously told investors its top 10 customers collectively make up less than 15% of its usage.
Bubble Warning
While touting how his own company has benefited from AI sales, CEO Ali Ghodsi, a co-founder, has warned investors that other AI executives have become overexuberant about the near-term capabilities of AI. In an interview with Goldman Sachs CEO David Solomon in September, Ghodsi said, “It’s early days. That’s why we’re in a bubble, right?”
He added that Databricks was benefiting from customers moving quickly to try to automate more business tasks. “The only thing I don’t know is if there is a big correction, that could set us back,” he added.
What concerns some investors more immediately is rising costs. Databricks’ gross margin fell to about 74% in the third quarter from 79% in the same period last year. The decline was due to increasing expenses tied to running a new cloud storage offering for customers and buying graphics processing units to run small AI models, the company told investors recently.
Publicly traded software firms with similar data infrastructure businesses, such as Snowflake, have also reported declining gross margins in recent quarters.
Ghodsi has been happy to raise money privately while pushing off an initial public offering that would be one of the most highly anticipated listings in tech in years. Databricks is the fifth most valuable private tech company in the U.S., after OpenAI, SpaceX, Anthropic and xAI. Its $10 billion funding round last December was among the largest ever by a private company.
Insight Partners, a tech investment firm that buys shares in public and private companies, is an existing investor in Databricks and is expected to lead the new round and invest at least $500 million. The company would use the capital from the new funding round to buy back shares from employees while covering taxes related to the share sales.