Databricks has made Unity AI Gateway generally available, positioning it as a central control layer for AI costs, security and access across enterprise models, agents, tools, skills and MCP integrations.
Databricks has made Unity AI Gateway generally available, describing the product as a unified governance layer for enterprise AI systems. In a company blog post, Databricks said the gateway is designed to govern spending, security and access across AI agents, models, Model Context Protocol (MCP) integrations, skills and tools.
The release addresses operational challenges that can emerge as organizations expand AI deployments across teams and applications. Databricks identifies three related concerns: unpredictable consumption costs, security and intellectual-property risks, and the need to preserve developer flexibility across models and tool providers.
According to Databricks, AI costs can be difficult to manage because many services charge by token consumption rather than by fixed user seats. That pricing model can make spending less predictable as usage rises, particularly when organizations operate multiple AI applications and providers.
Databricks says Unity AI Gateway provides end-to-end observability, cost controls and routing capabilities. The company presents those features as a way to give enterprises a broader view of AI usage and to direct requests in accordance with cost and performance policies.
Rather than requiring organizations to manage controls independently for every model or AI application, the gateway is intended to apply a common governance approach across different components. Databricks says that includes models, agents, tools, skills and MCPs, which are increasingly used to connect AI systems with external data sources and services.
Databricks also focuses on the security implications of AI systems that can retrieve sensitive information, call software tools or take actions on behalf of users. In its announcement, the company cited risks involving prompt injection, supply-chain attacks, data retention and AI memory systems that can create additional copies of sensitive information.
Unity AI Gateway applies runtime guardrails and policies to AI interactions, according to Databricks. The company says the product is designed to preserve developer choice across models and tools while providing organizations with a central mechanism for applying access and security controls.
This distinction is important for agentic systems, where governance can extend beyond a model request. Enterprises may also need to control which data an agent can access, which tools it can invoke, and what actions it is permitted to perform. Databricks positions Unity AI Gateway as infrastructure for managing those connected decisions at runtime.
The generally available designation indicates that Databricks is offering Unity AI Gateway for production use. Separate recaps from Databricks partners Kanerika and Qubika also list Unity AI Gateway among capabilities described as generally available for enterprise deployments.
The product does not replace an organization's own policies for access, acceptable use or AI spending. Instead, Databricks presents the gateway as a central layer for observing AI activity and enforcing the policies an organization chooses across an expanding mix of AI services.
Databricks expands its AI governance tooling Databricks has made Unity AI Gateway generally available, describing the product as a unified governance layer for enterprise AI systems.
In a company blog post, Databricks said the gateway is designed to govern spending, security and access across AI agents, models, Model Context Protocol (MCP) integrations, skills and tools.
The release addresses operational challenges that can emerge as organizations expand AI deployments across teams and applications.
Continue reading
