After $43B valuation, Databricks acquires data replication startup Arcion for $100M

Databricks has remained a hot startup at a time when interest from investors has cooled across the ecosystem. Just last month the company raised $500 million at an eye-popping $43 billion valuation. That would be big money anytime, but especially in the current fundraising climate.

With that kind of dough in its coffers, the company went shopping this morning and bought data replication startup Arcion for $100 million.

Data replication for a data lakehouse like Databricks provides a way to move data in a consistent way between sources. “This acquisition will enable Databricks to natively provide a scalable, easy-to-use, and cost-effective solution to ingest data from various enterprise data sources,” the company stated.

Prior to this, customers would need to use a third party tool like Informatica, Qlik or Alteryx to get the data into Databricks. Arcion will give the company a tool of its own.

“Today, organizations have to pay and configure multiple tools to get data into Databricks. Arcion will provide them with native capabilities to get that data seamlessly into Databricks. That removes a lot of friction that today exists in the Data + AI journey that organizations have to go through,” Databricks co-founder and CEO Ali Ghodsi told TechCrunch.

Arcion, which raised $18 million since its inception in 2016, provides Databricks with more than 20 connectors to enterprise databases and data sources, including Oracle, PostgreSQL, Redis, SAP, Salesforce and Snowflake, giving the company its very own ingestion engine to operate on its platform to bring data in from these and other popular sources. It’s worth noting that the company also had built connectors to Databricks as well.

The idea behind having a data storage solution like Databricks is about putting that data to work to do things like building dashboards, developing data applications and feeding machine learning models for AI, but to do that, companies need to move the underlying data into something like Databricks.

Recently, the company has been focusing on AI, as generative AI has altered the software landscape this year. In March, it announced an open source large language model called Dolly, just months after OpenAI released ChatGPT. In May, the company continued its AI emphasis, acquiring AI-focused data governance platform Okera.

In June, the company announced that it was buying OpenAI competitor MosaicML for $1.3 billion, indicating that it was serious about being at the center of the growing AI market. While Arcion is a fraction of the price, and is less directly focused on AI, it gives the company an essential ingredient in data ingestion, and the fact that the companies have been partners in the past, could help as they come together.

According to PitchBook, the company was valued at $65 million as of February 2022. Assuming that was accurate, Arcion’s sale price represents a $35 million premium on its last private mark.

The $100 million figure works out to a roughly 54% mark-up for Arcion’s backers. Is that a strong outcome? On one hand, certainly. Every investor likes to put capital into a startup and see it quickly appreciate in value, especially when the value accrued is in the form of something that can be returned to the venture capitalists own backers.

At the same, venture capital is a hits game. Therefore, a 54% gain on the company’s Series A value is likely less than its backers at the time were hoping for. But because the company last raised in early 2022, when the venture market was still coming down from its then-recent peak, we can infer that Arcion likely picked up a richer valuation at the time than it might have today, given similar traction and results. So, the entry price for the deal might have been a little bit pricey, making the implied return of the sale to Databricks a bit more attractive than it looks even on paper.

For the investors in Arcion’s seed round, who bought in at a roughly $12 million valuation per PitchBook, the deal is a nice 8x return in less than three years. That’s real venture math.