Even Google Says You Cannot Do AI Security On One Platform
Security Boulevard, Thursday, May 28th, 2026
Google Cloud's COO states that effective AI security requires platform-independent solutions across multiple clouds and SaaS applications.
According to Francis de Souza, COO of Google Cloud, enterprises cannot rely on single-platform AI security solutions because the attack surface now includes models, prompts, agents, training pipelines, and MCP servers spanning multiple clouds and SaaS environments.
AI agents discover and expose previously unknown data assets that security teams weren't aware of, expanding the blast radius significantly. De Souza emphasized that security posture must extend across all platforms where agents operate, not just within a single cloud provider.
Salt Security argues that platform-independent, vendor-neutral security tools built specifically for agentic environments are essential because each cloud-native tool only sees its own platform, creating incomplete threat visibility.
CISOs cannot wait for platform providers to fully implement their own security advice and must establish independent security layers to map their entire agentic infrastructure in real time.