Databricks agrees to acquire MosaicML.

Image Credit: Databricks

Data lakehouse vendor Databricks today announced it has signed a definitive agreement to acquire MosaicML, an artificial intelligence (AI) startup based in San Francisco. The deal is estimated to be valued at $1.3 billion. 

In a press release, Databricks said it plans to bring MosaicML’s entire team and technology under its umbrella, providing enterprises with a unified platform to manage data assets and build secure generative AI models. The move comes as enterprises across sectors continue to look for ways to leverage large language models (LLMs) and target different use cases.

“Every organization should be able to benefit from the AI revolution with more control over how their data is used. Databricks and MosaicML have an incredible opportunity to democratize AI and make the lakehouse the best place to build generative AI and LLMs,” Ali Ghodsi, cofounder and CEO of Databricks, said in the release.

MosaicML helps with building LLMs

Modern enterprises want to build generative AI models but are held back by the challenge of feeding their data into these systems. MosaicML gives these companies a platform to build, train and deploy their state-of-the-art models using their proprietary data. 

The company, which works with the likes of Replit and Allen Institute for AI, also offers its own commercially usable, open-source MPT series of models, which organizations can fine-tune on their data to quickly and easily deploy their private LLMs. In both cases, MosaicML enables developers to maintain full control over the models they build with model ownership and data privacy built into the platform’s design.

>>Follow VentureBeat’s ongoing generative AI coverage<<

Now, with this deal, all these offerings from MosaicML will come under the umbrella of Databricks, which provides enterprises with a platform to store structured, unstructured and semi-structured data — an element critical to training AI models.

As the companies explained, once the transaction closes, the entire MosaicML team, including its AI research department, will join Databricks and the platform’s training and inference tools will be integrated into the lakehouse. This will create a unified offering, giving enterprises both easy access to data and the tools needed to build, train and deploy their own private generative AI models.

The combined offering is also expected to bring down the cost of training and using LLMs from millions of dollars to thousands. 

Availability of MosaicML integration

Databricks noted that MosaicML’s platform will be supported, scaled and integrated over time to offer customers a seamless unified experience. However, the company has not shared an exact timeline for when the transaction will close or the integration go live. 

In this space, the data giant competes with Snowflake, which recently made its own generative AI push with the acquisition of Neeva.