The Convergence of Cloud Secrets & AI Risk
SentinelOne, Wednesday, May 13th, 2026
AI-specific credentials surged 140% in 2025, creating new attack vectors that compound traditional cloud security risks.
SentinelOne's 2026 report reveals that AI adoption has become the primary driver of cloud risk, with AI-related secrets like OpenAI API Keys increasing 140% year-over-year. The proliferation of unmanaged AI credentials and "shadow AI" usage creates unique vulnerabilities that expose sensitive data and enable prompt injection attacks, while traditional cloud secrets remain prevalent across diverse platforms including AWS, Azure, GCP, and payment processors.
Legacy vulnerabilities continue to serve as reliable initial access points for attackers, who exploit these weaknesses to pivot into cloud and AI services to harvest corporate data. Security leaders must implement continuous surface monitoring, DevSecOps automation, and centralized governance of AI credentials to address the interconnected risks posed by the convergence of cloud and AI infrastructures.