Defining Snowflake’s New Bet: An Agentic Control Plane
Snowflake’s acquisition of Natoma marks a strategic move toward an “agentic control plane,” a governed environment where enterprise AI agents can act across business systems while staying within security and compliance policies, signaling a shift from passive analytics to active, autonomous workflows. Announced during Snowflake’s fiscal 2027 first-quarter earnings call, the Natoma deal is the sixth Snowflake acquisition since June 2025, highlighting an aggressive Snowflake acquisition strategy aimed at enterprise AI agents rather than only cloud database expansion. CEO Sridhar Ramaswamy framed the goal as turning AI from insight into action across everyday tools like email, calendars, Slack, and Jira, all from Snowflake’s AI products such as Snowflake Intelligence and Cortex Code (Coco). This approach positions Snowflake as more than a cloud data warehouse, steering it toward being a central execution layer for governed, agentic AI inside large organizations.
Natoma’s Role: Containing Rogue Agents and Shadow AI
Natoma brings a governance gateway built around the Model Context Protocol (MCP), a standard for connecting AI systems to external tools, APIs, and SaaS applications. Its platform enforces identity checks, permission policies, and audit trails at the level of each tool call, recording who requested what, with which rights, and whether the action was allowed. This design aims to contain “rogue agents” and reduce risks such as unauthorized data access, data leakage, and shadow AI deployments running outside official oversight. According to Snowflake’s earnings call, Natoma extends Snowflake’s agentic control plane beyond data and development workflows into the applications where daily work occurs. For enterprise AI agents, the implication is clear: production deployment will depend less on model accuracy and more on whether organizations can show governed behavior, traceable actions, and controlled access across their existing software stack.
From Analytics to Agentic Workloads Across the Enterprise Stack
Snowflake’s purchase of Natoma caps a rapid sequence of deals—Crunchy Data, Datometry, Select Star, Observe, TensorStax, and now Natoma—that together reshape its role in enterprise data. Earlier acquisitions strengthened databases, migration, observability, and data pipeline planning; Natoma moves the focus up the stack to how AI agents act on that data. The company’s narrative has shifted from cloud database expansion and classic analytics to operational AI workloads that coordinate workflows and trigger changes in external systems. Natoma’s MCP gateway is key here, because it pulls context from SaaS apps, internal APIs, and infrastructure into Snowflake’s own AI layer. Snowflake says this enables deep research-style queries that span Snowflake data, the web, Google Docs, and Slack while also allowing immediate actions such as sending emails or opening Jira tickets, all within governed boundaries.
AWS Partnership Growth Fuels Infrastructure for Agentic AI
Announced alongside the Natoma deal, Snowflake’s expanded partnership with AWS underlines the infrastructure demands of agentic AI. The companies revealed a five-year, USD 6 billion (approx. RM27,600,000,000) commitment centered on Graviton-powered compute and AI infrastructure to support Snowflake’s growing agentic AI ambitions. According to AWS CEO Matt Garman, enterprises are “rapidly moving from experimenting with AI to putting intelligent agents to work that drive real business outcomes.” The collaboration targets deeper integration around generative and agentic AI, broader AWS Marketplace sales motions, and joint programs to speed customer migration and deployment. Architecturally, both firms promote keeping AI systems close to governed data rather than exporting sensitive information to external tools. For Snowflake, this AWS partnership growth strengthens its position as a foundational platform where enterprise AI agents run near the data they need, under tight governance.
Enterprise AI Priorities: Governance, Not Just Models
Both the Natoma acquisition and the AWS expansion highlight a broader change in enterprise AI priorities. Early generative AI interest focused on model performance and access to the latest large models. Now, buyers are more concerned with identity, permissions, auditability, and integration with existing workflows. Snowflake cited internal research indicating that 96% of organizations still face major obstacles in scaling AI across the enterprise, underscoring that technology gaps lie as much in governance and control as in modeling capabilities. Natoma’s secure connectivity layer, combined with Snowflake’s Cortex Agents, Snowflake Intelligence, and Cortex Code, is aimed at closing that gap by providing a single, policy-aware interface to systems ranging from CRM tools to cloud infrastructure. As enterprises move toward agentic AI, platforms that combine data, governance, and execution are likely to define the next phase of AI adoption.
