In a strategic move that reflects the growing importance of governance in enterprise artificial intelligence, Snowflake has announced the acquisition of Natoma, a startup specialized in the MCP (Model Context Protocol) protocol. This operation aims to provide companies with the necessary tools to securely and controllably manage AI agents operating in heterogeneous environments, a crucial step to move from pilots to real production.

What does Natoma bring to Snowflake?

Natoma offers a platform that provides access to MCP-based tools, along with governance and observability capabilities. By integrating into Snowflake, it will allow companies to securely connect Cortex Agents, Snowflake Intelligence, Cortex Code, and other AI platforms with enterprise systems including SaaS applications, cloud environments, VPCs, and on-premises infrastructures. Essentially, Natoma acts as the control and governance fabric for these connections, an element that is becoming critical as real-time autonomous agents connect via MCP.

MCP Governance: A Key Challenge for CIOs

According to Phil Fersht, CEO of HFS Research, "MCP is becoming the connective fabric for enterprise agents, but without identity, policies, privileged access controls, and auditability, it can quickly become a shadow AI risk." CIOs must now govern not only who queries a table, but also what agents can see, which systems they can touch, and what actions they can trigger. As Robert Kramer, Managing Director of KramerERP, points out, "MCP is a protocol, not a governance model in itself. It can standardize connections, but it can also standardize risk if access is too broad."

The acquisition of Natoma aims precisely to fill that gap, providing a governed MCP with verified servers, identity-aware authorizations, policy enforcement, auditability, and gateway control. However, as Fersht warns, most companies are not yet ready for mass MCP adoption: "They want the productivity and context benefits, but their governance, identity, data classification, and access control models are still catching up."

Implications for Business Strategy

For Michael Ni, Principal Analyst at Constellation Research, "Data platforms won the analytics era. Whoever governs agents, context, and autonomous actions will win the agentic era." Snowflake is not alone in this race: SaaS providers like Salesforce, ServiceNow, and Workday are integrating agentic orchestration capabilities, while hyperscalers like Microsoft, AWS, and Google consolidate their development tools. In this context, the acquisition of Natoma positions Snowflake as a key player in the AI control plane.

The real test will be the frictionless integration of Natoma's capabilities into Snowflake's offering, avoiding adding another layer of complexity. CIOs will need to manage permissions, policies, and controls at scale, a challenge we have already seen in other areas such as the identity layer for autonomous agents or the lessons from vendor neutrality in OpenTelemetry.

Recommendations for CIOs

Analysts agree that CIOs should avoid treating MCP as a plug-and-play miracle. Critical points to watch include identity-based permissions, least-privilege access, audit logs, human oversight for high-risk actions, data leakage controls, and clear ownership when an agent makes a bad decision. In an environment where digital sovereignty and network security are priorities, AI agent governance is shaping up to be the next big challenge.

Snowflake has not disclosed the financial terms or the closing date of the transaction, but the move is already generating expectations about how AI governance will evolve in the enterprise. As we have seen in cases like digital transformation with Microsoft 365, the key will be integration and gradual adoption.

Original source: ComputerWorld. Analysis and adaptation by ForgeNEX.