Zscaler report finds 91% surge in AI traffic is outpacing enterprise security

Zscaler published the ThreatLabz 2026 AI Security Report on January 27, 2026. It analyzes 989.3 billion AI/ML transactions across roughly 9,000 organizations on its Zero Trust Exchange during 2025. The report found a 91% year-over-year rise in AI/ML activity across more than 3,400 applications, a fourfold increase. Data transfers to these apps rose 93% to 18,033 terabytes. Finance and insurance accounted for 23% of traffic. Technology and education saw 202% and 184% growth, respectively. ChatGPT processed 115 billion transactions, moved 2,021 terabytes of data, and triggered 410 million data loss prevention violations. Red team testing found critical vulnerabilities in 100% of enterprise AI systems examined. The median time to first critical failure was 16 minutes, and 90% of systems were compromised within 90 minutes. Embedded AI in SaaS and supply chain weaknesses increase risk by enabling agentic AI to automate attacks. Many organizations lack inventories of AI models, widening oversight gaps. Zscaler recommends AI-native zero-trust defenses and launched the Zscaler AI Security Suite to manage assets, enforce secure access, and provide runtime protections that inspect traffic, classify prompts, and align with the NIST AI Risk Management Framework.

Why It Matters

• Inventory AI assets, including embedded models in SaaS and supply chain components, to restore visibility into data flows and reduce shadow AI risk.
• Deploy AI-native zero-trust access with inline inspection to block exfiltration from AI apps amid a 93% surge in data transfers.
• Implement runtime guardrails and automated red teaming to detect and mitigate breaches that can appear in a median 16 minutes.
• Elevate AI governance and board-level reporting to address the universal vulnerabilities found in 100% of tested enterprise systems.