Rubrik launched Agent Cloud (RAC) at Google Cloud Next 2026 on April 22, a governance layer for autonomous AI agents built on Google’s Gemini Enterprise Agent Platform. The integration gives enterprises visibility into what their agents are doing, the ability to enforce compliance policies in real time, and the option to reverse agent actions when they go wrong.

How It Works

RAC is powered by Rubrik’s Semantic AI Governance Engine, which replaces static policy rules with continuous, intent-driven oversight of agent behavior. According to SiliconANGLE, the integration includes three core capabilities:

Agent Inventory auto-discovers agents running on the Gemini platform, providing visibility into risk profiles, access permissions, and policy violations across all deployed agents.

Semantic governance enforces compliance rules based on what agents are actually doing, not just what they were configured to do. Rather than relying on predetermined rule sets, the governance engine continuously monitors agent behavior and applies policies dynamically.

Agent Rewind allows administrators to undo a destructive action taken by an autonomous agent. Rubrik claims this capability is unique to its platform.

“Enterprises want the speed of Google Cloud’s AI technologies but also require the safety of Rubrik’s cyber resilience,” Devvret Rishi, GM for AI at Rubrik, said. “Through this collaboration, we will remove the governance bottleneck for customers developing with Gemini Enterprise Agent Platform.”

Satish Thomas, VP of Applied AI and Platform Ecosystem at Google Cloud, described Rubrik’s unified control layer as “critical for customer success as enterprises move into the autonomous era.”

Database Protection Expansion

Alongside the agent governance launch, Rubrik also announced a Cloud SQL integration that adds cyber resilience for managed PostgreSQL databases on Google Cloud. The integration provides immutable, automated backups with air-gapped ransomware protection and cross-region recovery, according to SiliconANGLE. New Cloud SQL instances are automatically discovered and brought under global backup policies.

Rubrik Zero Labs research found that 90% of IT and security leaders reported cyberattacks in the previous year, with 35% citing securing data across varied ecosystems as their top challenge.

Market Context

Gartner predicts that 40% of enterprise applications will integrate task-specific AI agents by the end of 2026, according to Investing.com. That projection frames the scale of the governance problem Rubrik is targeting: as agent deployments move from pilot to production, enterprises need real-time visibility into what agents are accessing, whether they are complying with policies, and how to remediate when they cause damage.

Rubrik’s positioning mirrors a broader convergence at Cloud Next 2026, where Google’s own Agentic Data Cloud, Cyera’s Ryft acquisition, and Check Point’s AI Defense Plane integration all addressed the same underlying demand. The enterprise agent governance market is forming rapidly, and Rubrik is betting that data protection expertise translates into agent oversight credibility.