What Snowflake’s Natoma Move Says About Enterprise AI Now
Snowflake’s planned acquisition of Natoma is a strategic move in enterprise AI infrastructure, showing how data platforms are evolving into governed control planes for action-taking AI agents across business systems. The deal targets a growing gap between experimental AI pilots and production deployments, where enterprises need agents that can act on email, calendars, tickets, and SaaS tools without breaking security rules. Snowflake CEO Sridhar Ramaswamy describes this as building an “agentic control plane,” in which Cortex Code (Coco) and Snowflake Intelligence become a single, governed interface for everyday work. Natoma’s gateway for Model Context Protocol (MCP) connects AI agents to external software with identity checks, policy enforcement, and detailed audit trails on each tool call. This combination aligns with the shift from model-centric innovation toward controlled, action-oriented AI grounded in existing cloud data consolidation strategies.
Rogue Agents and the Need for an Agentic Control Plane
Natoma addresses a new risk in enterprise AI: “rogue” agents acting across systems with little visibility or governance. Its MCP-based gateway lets enterprises decide, per tool call, who an AI agent acts on behalf of, what permissions apply, and whether an action should proceed. That means sending emails, summarizing Slack threads, or opening Jira tickets can all flow through a governed environment with enterprise security, observability, and policy enforcement. Natoma’s founders say they built the company on the belief that AI agents would transform work only if organizations could trust and control how those agents access data and tools. Their vision mirrors Snowflake’s view that enterprises need a trusted control plane for the agentic era, with AI grounded in their own data and governed by their own policies across complex technology stacks.
An Aggressive Snowflake Acquisition Strategy Comes Into Focus
Natoma is Snowflake’s sixth acquisition since June 2025, signaling a clear Snowflake acquisition strategy aimed at building an end-to-end enterprise AI infrastructure stack. Earlier purchases ranged from PostgreSQL provider Crunchy Data to Datometry for database migration, Select Star for data discovery, Observe for AI-powered observability, and TensorStax for data pipeline planning. Rather than isolated bets, these targets form a pattern: consolidate core data, migration, observability, and now agent governance into one platform. According to The Register, this latest move pushes Snowflake’s agentic control plane “beyond data and development workflows into everyday applications where work actually happens.” For buyers, it suggests that AI capabilities, governance, and cloud data consolidation will increasingly be bundled, reducing the need to stitch together multiple vendors for pipelines, monitoring, and secure agent operations.
AWS Partnership Expansion: Infrastructure for the Agentic Enterprise
Snowflake announced the Natoma deal alongside a five-year, USD 6 billion (approx. RM27.6 billion) AWS partnership expansion focused on Graviton-powered compute and AI infrastructure. The timing shows that governance and scale are now inseparable priorities in enterprise AI infrastructure. As workloads shift from simple copilots to action-taking agents, Snowflake needs massive, closely integrated cloud capacity to keep AI systems near governed enterprise data rather than pushing sensitive data out to external tools. The expanded collaboration includes deeper integrations for generative and agentic AI, more activity in the AWS Marketplace, and joint migration and deployment efforts. Matt Garman, CEO of AWS, said enterprises are “rapidly moving from experimenting with AI to putting intelligent agents to work that drive real business outcomes,” underscoring why infrastructure commitments now sit alongside governance-focused acquisitions.
From Experimentation to Integrated AI and Data Platforms
Together, the Natoma acquisition and AWS partnership expansion highlight a rapid pivot in enterprise priorities: from model selection toward governed, integrated AI and data platforms. Internal Snowflake research cited in Engineering.com notes that 96% of organizations still face barriers scaling AI, with governance, permissions, auditability, and workflow integration now front of mind. Natoma’s control plane for MCP tools, combined with Snowflake’s data cloud and AWS’s scale, aims to lower those hurdles by bringing intelligence, action, and policy into one environment. This pattern mirrors a broader industry shift: AI vendors are racing to bundle data management, security, observability, and agent governance into unified enterprise AI infrastructure. For customers, the promise is fewer “shadow AI” deployments and more consistent control as AI agents move from pilots to core operational roles.