Databricks Launches First Multi-AI Indemnity to Protect Enterprise AI Innovation

Following our recent announcement of Agent Bricks, which allows users to auto-optimize an agent on their own data by applying our agentic research techniques such as prompt optimization or reinforcement learning, we are introducing the industry’s first Multi-AI Indemnity. Thousands of enterprises use Databricks every day to build and deploy agents into production, often using multiple models and tools orchestrated together. But it’s not always about tokens and tools. We know that our customers also need certainty. Available today, the Multi-AI Indemnity in Databricks Model Serving provides IP protection across all major frontier AI models. Together with our strategic partnerships with Anthropic, OpenAI, Gemini, and Meta, we will soon have all frontier models natively available on Databricks, with a Multi-AI Indemnity to unlock new use cases and capabilities. This indemnity applies to Databricks Foundational Models API and joins Databricks’s existing indemnity covering the Databricks Platform’s ready-to-use tools such as Genie and the Databricks Assistant.

Databricks gives customers choice for all leading AI models

The frontier model providers constantly release new models and differentiated capabilities. In response to so much change, our customers don’t want to be locked into a single provider or model family. That’s why we forged partnerships to offer all of the most popular AI models, including OpenAI GPT, Anthropic Claude, Meta Llama, and Google Gemini natively in the Databricks Platform. With Databricks, customers have the most flexibility to build sophisticated AI applications and agents. 

After deploying in Databricks Model Serving, these models can then be empowered by tools and governed MCP servers, and run directly in SQL on Databricks to be close to your enterprise data. Databricks can also help you determine if the latest model releases improve your agent. Track how your AI agents make decisions in real time and generate custom evaluations for your task and your data. 

Multi-AI Indemnity protects your ability to innovate

Of course, model flexibility means nothing without confidence. As enterprises increasingly deploy AI solutions, one concern consistently emerges: protection against third-party copyright claims related to AI-generated content. Until now, customers have had to navigate this uncertainty themselves or rely on very limited protections from each model vendor.

All of these existing AI indemnities are too narrow in scope, covering only the provider's own models and services. This leaves customers exposed, since most organizations use multiple models in their agent systems today.

To fill this gap, Databricks Multi-AI Indemnity protects customers across all of the leading frontier models from OpenAI, Anthropic, Meta, Google, as well as Databricks’ own first-party models. Once a customer implements a few required industry-standard mitigations, the indemnity automatically applies. The protection safeguards customers from IP claims, including judgments, settlements, and legal fees, when using supported models from those providers through Databricks Model Serving.  

Enabling innovation on Databricks

Databricks already has a comprehensive platform for building, evaluating, and monitoring agents deployed by thousands of our enterprise customers. By providing comprehensive protection across all major AI models, we're helping establish a new standard for how AI platforms should support their customers. We believe this will unlock new use cases in enterprises and let our customers’ AI scientists innovate with confidence. To learn more about how this industry-first protection can accelerate your AI initiatives, contact your Databricks representative.