Defining Snowflake’s Agentic Control Play
Snowflake’s planned acquisition of Natoma is an enterprise AI governance move in which the cloud data platform provider uses a specialized gateway to control how AI agents access applications, data, and workflows while keeping everything inside governed security and compliance boundaries. Snowflake describes this vision as an “agentic control plane,” where agents can send emails, summarize Slack threads, open Jira tickets, and query data from a single interface, but every action is checked against permissions, identity, and policy. The deal marks Snowflake’s sixth major acquisition since June 2025, signaling an accelerated Snowflake acquisition strategy centered on AI infrastructure rather than only analytics. Instead of treating AI as a sidecar, Snowflake is turning its platform into the main control layer for enterprise AI, aligning model access, tools, and data governance in one place.
Natoma and the Governance Gap for AI Agents
Natoma’s core product is a gateway for Model Context Protocol (MCP) servers that lets AI agents call external tools under strict identity, policy, and audit controls. Each tool call is tied to a specific user, their permissions, and a decision on whether an action should proceed—answering rising worries about shadow AI, data leakage, and rogue agents. Natoma’s founders say they built the company on the belief that AI agents would only reach production “if organizations could trust and control how those agents access data, use tools, and take action.” By bringing Natoma into its cloud data platform, Snowflake gives Cortex Code (Coco), Cortex Agents, and Snowflake Intelligence a shared governance and identity layer. This turns AI agents from experimental copilots into controllable, auditable actors embedded directly into day-to-day business applications.
Sixth Deal Underscores a Broader Snowflake Acquisition Strategy
Natoma is part of a pattern: since June 2025, Snowflake has announced six acquisitions spanning PostgreSQL (Crunchy Data), database migration (Datometry), data discovery (Select Star), observability (Observe), and AI pipelines (TensorStax). Together they show a Snowflake acquisition strategy that stretches beyond data warehousing toward a full cloud data platform for enterprise AI. Rather than building every piece from scratch, Snowflake is buying specialized capabilities that complete its vision of an agentic enterprise: governed data, migration, discovery, observability, and now agent control. For customers, this signals an intention to sell an end-to-end stack—from storage and compute to AI agents and monitoring—under a single governance model. It also raises competitive pressure on rivals that still treat AI governance, observability, and pipelines as separate products instead of integrated parts of the data platform.
AWS Partnership Expansion and AI Infrastructure Ambitions
The Natoma announcement landed the same day Snowflake revealed an expanded five-year, USD 6 billion (approx. RM27.6 billion) AWS partnership centered on Graviton-powered compute and AI infrastructure. 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 enlarged AWS partnership expansion focuses on deeper integration for generative and agentic AI, tighter marketplace collaboration, and joint migration programs. Keeping AI close to governed data is a shared theme: instead of exporting information to external systems, Snowflake and AWS want foundation models running near data already controlled in the cloud. This deepens Snowflake’s role as a foundational platform for enterprise AI infrastructure, anchoring its agentic control plane on a hyperscale cloud base that can handle growing compute demands.
Enterprise AI Governance as Competitive Differentiator
Snowflake cites internal research that 96% of organizations still face major barriers when scaling AI, shifting attention from model quality toward governance, permissions, and workflow integration. By combining Natoma’s MCP-based governance gateway with its own AI services and strengthened AWS infrastructure, Snowflake aims to turn enterprise AI governance from a problem into a product feature. The company’s platform narrative now centers on being the place where AI agents reason over trusted data, act inside existing SaaS and internal systems, and remain observable and policy-compliant from one console. This positions Snowflake as more than a data warehouse: it becomes a control system for AI-driven operations. If the strategy succeeds, Snowflake could set expectations that a modern cloud data platform must also manage AI agent identity, access, and actions—not only store and query data.