The Deal
Databricks acquired MosaicML for $1.3B in June 2023. MosaicML had developed efficient training techniques and released MPT models that demonstrated frontier-competitive performance at lower cost. The company had raised only $64M before the acquisition, making this a significant premium.
Strategic Rationale
Databricks wanted to offer customers the ability to train custom AI models on their own data within the Databricks lakehouse platform. MosaicML training infrastructure and research team provided this capability immediately, rather than Databricks needing to build it from scratch.
Impact
The acquisition signaled that AI model training was becoming a feature of data platforms rather than a standalone business. It accelerated Databricks positioning as an end-to-end AI platform and put pressure on Snowflake and other data companies to develop similar capabilities.
What Happened After
MosaicML team integrated into Databricks as their AI research division. They launched DBRX (an open model) and Mosaic AI training products that allow enterprise customers to fine-tune models on proprietary data. Databricks valuation grew to $43B, partially driven by the AI capabilities MosaicML enabled.