Databricks has announced customers will be able to access Mistral AI large language models (LLMs) through a new partnership between the firms, marking the latest major industry deal for the French startup.
Under the terms of the partnership, which also saw Databricks participate in Mistral AI’s Series A funding round, the firm’s Mistral 7B and Mixtral 8x7B models will be integrated within the Databricks Data Intelligence Platform.
Databricks said customers will be able to interact with the models via the Mosaic AI Playground tool, customize them based on their own internal datasets, and use them as “optimized model endpoints through Mosaic AI Model Serving”.
“With these out-of-the-box integrations, today’s enterprises can rapidly leverage Mistral AI’s models for their generative AI applications, without compromising on security, data privacy, and governance that are core to the Databricks platform,” the company said in a statement.
The firm said that, since the beginning of 2024, it has observed more than 1,000 enterprises using Mistral models on Databricks. The partnership announcement therefore represents a “natural next step” to further bolster customer offerings.
Arthur Mensch, Founder and CEO of Mistral AI, said the deal marks a major milestone for the company and its goal of creating a more democratized AI industry landscape.
"We are delighted to forge this strategic alliance with Databricks, reaffirming our shared commitment to the portability, openness and accessibility of generative artificial intelligence for all,” he said.
“By seamlessly integrating our models into Databricks' data intelligence platform, we are advancing our shared mission of democratizing AI. This integration marks an important step in extending our innovative solutions to Databricks' vast customer base and continues to drive innovation and significant advances in AI.”
Mistral AI is the talk of the town
The partnership announcement marks the latest in a flurry of major deals for Mistral AI, which has experienced a rapid rise to prominence.
Mistral AI first burst onto the global AI scene in early 2023 and quickly secured $113 million in seed funding round in June. This was quickly followed by another $415 million as part of its Series A in which Databricks participated.
Since then, commercial ties with industry heavyweights have been rolling in.
In December, Google Cloud confirmed it had struck a deal to distribute Mistral models across its infrastructure, and more recently, Microsoft invested in the Parisian startup to bring its newest AI model, Mistral Large, to Azure.
The latter of the deals was a significant move that highlighted both the popularity of the firm’s offerings, but also Microsoft’s potential willingness to move away from a monolithic approach to AI collaboration through its OpenAI relationship.
Mistral has been keen to frame itself as a potential long-term counterweight to the dominance of US-based firms such as Anthropic and OpenAI. Speaking in the wake of the Series A funding round last year, Mensch specifically mentioned the company’s goal is to create a “European champion” in the burgeoning global AI industry.
The company’s approach to AI development appears to be one that resonates with hyperscalers both from a technical and perception standpoint.
Mensch previously said Mistral’s objectives aim to create an “open, responsible, and decentralized approach” to generative AI.
As a primarily European-based company, Mistral represents an ideal regional partner for US hyperscalers, all of which are contending with heightened scrutiny from EU regulators.
