From Data Warehouse to Agentic Control Plane
Snowflake’s planned acquisition of Natoma and its expanded AWS deal together mark a strategic turn from pure cloud database expansion toward becoming an “agentic control plane” for enterprise AI, where data, actions, and governance live in one place. This shift reflects a wider change in enterprise AI data management: organizations no longer want isolated copilots, but governed AI agents that can act safely across tools, workflows, and sensitive data. During Snowflake’s fiscal 2027 Q1 earnings call, CEO Sridhar Ramaswamy said the goal is to let AI agents send emails, open Jira tickets, and summarize Slack threads inside Snowflake’s AI interfaces while staying within enterprise security policies. The dual announcements signal that AI priorities are moving quickly from model experimentation toward operational control, with Snowflake positioning itself as the platform where intelligence and action meet under shared governance.
Natoma: Governance Layer for Rogue and Shadow AI Agents
Natoma gives Snowflake a gateway that governs how AI agents interact with external tools through the Model Context Protocol (MCP), a fast-emerging standard for connecting agents to SaaS apps, databases, and APIs. Natoma’s platform enforces identity checks, access policies, and audit trails on every tool call, answering the new question in enterprise AI: not only what an agent knows, but what it is allowed to do. This directly targets fears about rogue agents, shadow AI projects, and data leakage as companies move beyond simple chatbots. According to engineering.com, Snowflake cites internal research showing that 96% of organizations still face significant barriers scaling AI. By embedding Natoma into services such as Snowflake Intelligence and Cortex Code, Snowflake adds a native AI agent governance and identity layer that can extend across Slack, email, CRMs, cloud infrastructure, and internal APIs without sacrificing centralized policy control.
Sixth Acquisition and the New Enterprise AI Stack
Natoma is Snowflake’s sixth acquisition announcement since June 2025, signaling an aggressive Snowflake acquisition strategy aimed at building a full enterprise AI data management stack. Earlier deals brought in Crunchy Data for PostgreSQL, Datometry for database migration, Select Star for data discovery, Observe for AI-powered observability, and TensorStax for AI-driven data pipeline planning. Together, these moves extend Snowflake from analytics into observability, data movement, and now AI agent governance. The pattern shows Snowflake preparing for an era where data warehousing, application context, and operational telemetry must feed AI agents in near real time while staying governed. With Natoma, the company pushes its control ambitions beyond data and developer workflows into “where work actually happens” across business applications, so AI agents can coordinate workflows end to end while still operating under enterprise security, permissions, observability, and policy enforcement.
AWS Expansion: Scaling Infrastructure for Agentic AI
In parallel, Snowflake signed a five-year, USD 6 billion (approx. RM27.6 billion) AWS agreement centered on Graviton-powered compute and AI infrastructure, underscoring the sheer scale of AI workloads the company expects. The expanded collaboration focuses on deeper integrations around generative and agentic AI, closer alignment with AWS Marketplace sales, and joint programs for customer migration and deployment. 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.” The partnership also aligns with rising concern over where AI models run relative to governed data. Rather than exporting sensitive datasets into external systems, Snowflake and AWS promote patterns where models run near the core cloud database expansion layer, keeping policies, lineage, and access controls intact as AI agents query and act on production data.
What Snowflake’s Moves Reveal About AI Agent Governance Priorities
Taken together, the Natoma acquisition and AWS agreement reveal how enterprise AI priorities are pivoting from model choice toward AI agent governance, integration, and operational resilience. Early generative AI pilots focused on response quality; now buyers ask whether agents can respect permissions, support audits, and integrate with existing SaaS, cloud, and on-prem systems. Natoma attacks the governance gap by enforcing identity and policy at the tool-call level, while AWS provides the elastic infrastructure needed to run many agents close to governed data. For enterprises, the appeal is a single platform where AI agents can reason over Snowflake data, reach into systems like Slack and Jira, and then take action under consistent controls. As more organizations aim for an “agentic enterprise,” Snowflake’s strategy shows that owning the control plane around data, identity, and actions may matter more than owning the underlying models themselves.