Snowflake is acquiring Natoma, a startup focused on controlling how AI agents access enterprise systems, in a purchase that addresses one of the biggest challenges of enterprise AI: agents are now taking actions across business applications, but many companies lack the controls to fully govern them.
The acquisition gives Snowflake technology to manage the permissions, identities and security policies that determine what AI agents can do inside an organization. As enterprises move beyond AI chatbots and begin deploying autonomous agents that can update records, access applications and execute workflows, concern over governance remains a stumbling block slowing adoption of agentic AI.
Founded in 2024, Natoma developed a platform built around the Model Context Protocol (MCP), a commonly used framework that allows AI systems to connect to enterprise software, databases and business tools. While MCP is helping advance the adoption of agentic AI, it is also creating new security concerns because agents can potentially access sensitive information or trigger actions across multiple systems.
Snowflake plans to integrate Natoma’s technology into products including Snowflake Intelligence, Cortex Agents and Cortex Code, creating a centralized layer for managing how AI agents interact with enterprise environments.
“Snowflake’s acquisition of Natoma is one of the clearest indicators yet that the AI battleground is shifting from models to governance, orchestration, and trusted enterprise execution,” Phil Fersht, Chief Analyst at HFS Research, told Techstrong.ai.
“The challenge is that governance alone does not guarantee enterprise adoption. Snowflake must prove it can simplify deployment while balancing security, usability, and interoperability across increasingly complex enterprise environments. The risk is that MCP ecosystems become fragmented before standards mature, creating another layer of enterprise complexity instead of reducing it.”
Governance Gateway
Natoma’s platform acts as a governance gateway between AI agents and enterprise applications. The technology verifies identity, manages authorization, enforces policies and records activity, enabling organizations to monitor and control AI-driven actions. The company offers a library of more than 100 MCP server integrations to help simplify secure connections across enterprise environments.
Enterprise managers have become deeply concerned about the rapid growth of unofficial AI deployments inside organizations. Employees are connecting AI tools to Slack, Salesforce, Jira, GitHub and other business platforms, often without centralized governance. These fragmented deployments create blind spots that expose sensitive data or allow unintended actions.
The more autonomy that agents gain, the greater the potential problem. Without clear authorization controls, agents may access systems or data that were never intended to be part of a workflow.
This is an identity and authorization problem rather than simply an AI problem. While MCP standardizes connectivity between agents and enterprise systems, it does not provide governance on its own. Companies still need mechanisms to determine which agents can access specific resources, under what conditions and with what level of oversight.
Snowflake believes that capability will become a foundational requirement as enterprises expand their use of agentic AI. The company recently reported that 96% of organizations continue to encounter significant challenges when attempting to scale AI initiatives, with governance and operational controls among the major obstacles.
Natoma also adds expertise in identity governance and privileged access management. The startup raised $7 million in seed funding in May 2025 and was led by CEO Pratyus Patnaik, a former Okta executive.
Big picture, the Natoma acquisition demonstrates that Snowflake is continuing its ongoing investment in building the governance and orchestration layers needed to support enterprise AI agents. Financial terms of the transaction were not disclosed.
Since 2025, Snowflake has acquired six firms, including companies focused on observability, data pipelines, migration technology and AI operations. The Natoma announcement coincided with a separate five-year, $6 billion agreement between Snowflake and AWS focused on AI infrastructure.
