Databricks and OpenAI launch $100M Agent brick platform for enterprise AI

Databricks, a data intelligence platform company, has partnered with OpenAI in a $100 million deal to launch “Agent brick,” an enterprise AI platform that integrates advanced models like GPT-5 into business-ready agent systems. The collaboration aims to help companies build AI agents that can analyze data, automate workflows, and maintain compliance with regulations and cybersecurity standards, positioning both companies to capture growing enterprise demand for frontier AI capabilities.

What you should know: The Agent brick platform runs on Databricks’ data intelligence infrastructure, allowing enterprises to deploy large language models at scale while maintaining governance and security standards.

• Companies can build AI applications and agents directly on their existing data without compromising security or moving sensitive information to external platforms.
• The system is designed to handle context-aware operations, including regulatory compliance and cybersecurity requirements that are critical for enterprise deployments.

What they’re saying: Both companies emphasized the overwhelming enterprise demand driving this partnership.

• “We’re seeing overwhelming demand from enterprise customers looking to build AI apps and agents on their data, tailored to their unique business needs,” said Databricks Co-founder and CEO Ali Ghodsi.
• “Enterprise demand for frontier AI is accelerating,” added OpenAI COO Brad Lightcap. “Our partnership with Databricks brings our most advanced models to where secure enterprise data already lives, making it easier for businesses to experiment, deploy, and scale AI agents with real impact.”

The bigger picture: Databricks executives discussed the current state of AI development and adoption during a recent Stanford event, revealing mixed perspectives on industry hype and progress.

• CTO Matei Zaharia noted that “some areas are overhyped, some areas are probably underhyped,” but acknowledged that practical AI applications are “landing every day in production that are working well.”
• Chairman Ion Stoica highlighted the rapid timeline of AI development, pointing out that “ChatGPT was released less than three years ago” and emphasizing how quickly the technology has moved from research labs to mainstream business discussions.

Technical challenges remain: Despite significant progress, Databricks leaders identified key areas where AI models still need improvement.

• Stoica noted that current models are “very knowledgeable, but also imprecise, and hallucinating,” indicating that accuracy and reliability improvements are still needed.
• Zaharia observed that while large language model capabilities may be “slowing down a little bit,” the cost of running these models is “going down dramatically,” potentially enabling edge deployment and broader accessibility.

Why this matters: The partnership represents a strategic move to democratize enterprise AI by embedding advanced capabilities within existing data infrastructure, similar to no-code and low-code approaches that have made other technologies more accessible to business users.