# Zscaler, Inc. partners with Anthropic and Mistral on AI security

URL: https://decisioninsights.ai/zscaler-inc.-partners-with-anthropic-and-mistral-on-ai-security/

Zscaler, Inc. is working with frontier AI model providers Anthropic and Mistral to help enterprises adopt AI securely and defend against attacks driven by AI. The collaborations address two areas: stronger defenses against increasingly sophisticated AI-driven attacks, and secure enterprise use of AI applications and autonomous AI agents. The company states that AI now speeds up vulnerability discovery and exploitation, leaving traditional firewalls and VPNs unable to adequately protect the enterprise.Zscaler's Zero Trust architecture connects users, workloads, devices, and AI agents directly to the applications and tools each is authorized to use, which eliminates internet-facing attack surfaces and prevents lateral movement. Its Security for AI solution is combined with advanced capabilities from frontier AI models to give enterprises greater visibility and control over AI agents, surface emerging threats and attack paths, and protect sensitive data.Through Anthropic's Cyber Verification Program, Zscaler is working to advance its own vulnerability discovery and security research. Separately, Zscaler integrated Claude Mythos 5.1 into its Endpoint AI Security offering to power Exploit Paths. That capability correlates security evidence across complex attack sequences, helps identify potential threats, and shows defenders where to break the attack chain. A limited preview is planned.Zscaler's Security for AI solution, delivered through its sovereign Zero Trust Exchange in Europe, integrates with Mistral. The integration is intended to help European enterprises adopt AI while meeting their security, data protection, and sovereignty requirements.“AI is fundamentally changing the threat landscape, making Zero Trust more critical than ever,” said Jay Chaudhry, CEO, chairman, and founder of Zscaler. “We must eliminate attack surfaces, prevent lateral movement, and ensure AI agents access only the tools needed for their tasks, never the entire corporate network. By combining our Zero Trust architecture with innovations from leading frontier AI companies, we’re helping customers safely embrace AI and defend against AI-driven attacks.”The company plans to unveil additional innovations for securing the agentic enterprise at a launch scheduled for December 9, 2026.
