Databricks completes $10 billion financing, Meta enters the game, founder: may go public in a year

Data analysis software startup Databricks announced on Wednesday that Meta has become one of its investors, participating in a recent funding round totaling $10 billion, one of the largest investments in venture capital history. Following this funding round, Databricks is now valued at $64 billion.

Databricks is headquartered in San Francisco and currently has 8,000 employees. The company's products have long been used for significant data science projects. For example, businesses use the platform to integrate and analyze large amounts of data from different systems. A retailer, for instance, may want to combine various datasets to determine which products sell best at what times of the year to forecast inventory needs.

Recently, Databricks has gained greater attention as a crucial component of artificial intelligence projects. Data is key to driving the AI revolution, and Databricks provides a unified platform for integrating and standardizing structured and unstructured data, which is essential for building and deploying machine learning models.

As the company behind training Llama, a popular open-source large language model (LLM), Meta plays an important role in the AI field, and Databricks is developed based on Llama.

Ali Ghodsi, co-founder and CEO of Databricks, stated in a media interview this week that Databricks maintains close collaboration with Meta's Llama team. He also mentioned that this partnership even extends to Meta co-founder and CEO Mark Zuckerberg.

"We have discussed open-source software in the past, and he is very concerned about open-source models and Llama-related content."

Compared to tech peers like Alphabet and Microsoft, Meta's startup investment activities are relatively few. However, Databricks is a rapidly growing company that is heading toward an initial public offering (IPO). To date, Databricks has raised a total of $14 billion in equity financing. With this new influx of funds, Databricks plans to invest in new AI products, strengthen its global market expansion efforts, and provide funding for new acquisitions.

Ghodsi told the media, "I wouldn't be particularly surprised if we go public in a year."

Last December, Ghodsi also stated that it would be foolish to conduct an IPO last year due to the U.S. elections, the new government taking office, and ongoing economic concerns, with the earliest theoretical possibility for an IPO being sometime in 2025. However, analysts believe that the company also indicated it would use part of the new funds to provide liquidity for "current and former employees," suggesting that the IPO may be delayed to a later time.

On Wednesday, Databricks also announced a $5.25 billion credit facility led by financial institutions such as JP Morgan, Barclays, Citigroup, Goldman Sachs, and Morgan Stanley. Ghodsi stated that even in a high-interest-rate environment, credit is preferable to diluting existing shareholders' equity through equity financing Media reports indicate that funding obtained from banks enabled Databricks to train its open-source LLM—DBRX—at a cost of approximately $10 million. At that time, DBRX outperformed Meta's Llama and other alternative models in some tests, but was quickly surpassed by other models thereafter.

Analysts believe this is one of the reasonable reasons for Databricks to ally with Meta. Meta has sufficient funds for capital expenditures to train models, while Databricks can allocate funds for other purposes, Ghodsi stated. However, he did not specify whether Meta is one of its clients.

In addition to Meta, the Qatar Sovereign Wealth Fund—Qatar Investment Authority—also participated in this $10 billion financing round. Ghodsi mentioned that Databricks is open to allowing its software to operate in the data centers of major operators in the Middle East. Currently, its services are only available through the cloud platforms of Amazon, Google, and Microsoft