SIAgentNews.com
← All news

Zscaler partners with Anthropic and Mistral to secure enterprise AI

The cybersecurity firm is integrating frontier AI models into its Zero Trust platform to protect sensitive data and manage autonomous AI agents.

· First reported by mykxlg.com

Ink and watercolour: picture for the story 'Zscaler partners with Anthropic and Mistral to secure enterprise AI'
AI-made ink and watercolour for this story. No real people or products are shown.
200+public service locations globally

The short version

  • Zscaler is integrating Anthropic's Claude Mythos 5.1 to power 'Exploit Paths' for identifying complex attack sequences.
  • The company is working with Anthropic's Cyber Verification Program to advance vulnerability discovery research.
  • Mistral integration helps European enterprises meet specific data protection and sovereignty requirements.
  • The partnership aims to secure autonomous AI agents by limiting their access to only necessary tools.

Why it matters

As companies move toward using autonomous AI agents, they need a way to ensure these tools don't have unrestricted access to internal networks. This partnership provides a framework for businesses to use advanced AI without exposing their entire corporate infrastructure to new risks.

Securing the AI frontier

Zscaler has announced new partnerships with prominent frontier AI model providers, specifically Anthropic and Mistral. These collaborations are designed to help large enterprises adopt artificial intelligence while simultaneously defending against increasingly complex AI-driven cyberattacks. The initiative combines Zscaler's security expertise with the latest innovations in frontier AI models.

The primary goal of these partnerships is to provide visibility and control over how organizations use AI applications and autonomous agents. As companies scale their AI adoption, they face the risk of sensitive data being exposed or exploited by malicious actors. Zscaler aims to address these concerns by integrating advanced AI capabilities into its existing security infrastructure.

By working directly with the companies building the models, Zscaler hopes to stay ahead of emerging threats. The collaboration focuses on two critical areas: strengthening defenses against sophisticated attacks and enabling the secure deployment of AI tools. This approach is intended to give customers a clearer understanding of their attack paths and a better way to protect their data.

Technical integrations with Anthropic

One of the specific technical integrations involves Anthropic's Claude Mythos 5.1 model. Zscaler is incorporating this model into its Endpoint AI Security to power a feature called Exploit Paths. This capability is designed to correlate security evidence across various complex attack sequences to help defenders identify potential threats.

The Exploit Paths feature is intended to show security teams exactly where they can break an attack chain. By visualizing these sequences, defenders can more effectively stop a breach before it reaches its final objective. Zscaler plans to offer a limited preview of this specific capability to its customers in the near future.

Additionally, Zscaler is participating in Anthropic's Cyber Verification Program. This program allows the cybersecurity firm to advance its own research into vulnerability discovery. By collaborating on the research side, Zscaler can better understand how AI might be used to find weaknesses in software and systems.

Sovereignty and Mistral in Europe

The partnership with Mistral focuses on a different set of requirements, particularly for the European market. Zscaler's Security for AI solution integrates with Mistral through its sovereign Zero Trust Exchange. This setup is specifically designed to help European enterprises meet strict data protection and sovereignty requirements.

Data sovereignty refers to the idea that data should be subject to the laws and governance of the nation where it is collected. Many European companies require that their AI interactions remain within specific geographic boundaries or under specific legal jurisdictions. The Mistral integration ensures these companies can use AI without violating these regulations.

This collaboration allows businesses to adopt AI tools while maintaining a high level of security and compliance. It addresses the unique needs of organizations that must balance the benefits of AI with the necessity of local data control. Zscaler positions this as a way to provide a secure environment for AI deployment across different regions.

The role of Zero Trust architecture

A core component of these partnerships is Zscaler's Zero Trust architecture. Zero Trust is a security model that requires constant verification of every user, device, and application. Unlike traditional firewalls or VPNs, which often allow broad access once a user is inside the network, Zero Trust connects entities directly to only the specific tools they need.

Jay Chaudhry, CEO of Zscaler, stated, "We must eliminate attack surfaces." This means that if an AI agent is compromised, it cannot move freely across the entire corporate network to access unrelated sensitive information. The goal is to ensure that agents only access the specific tools required for their assigned tasks.

This architecture is considered essential as AI accelerates the speed at which vulnerabilities are discovered and exploited. By eliminating internet-facing attack surfaces, Zscaler helps organizations reduce their overall risk profile. This allows for the deployment of autonomous agents in a way that limits their reach to only authorized applications.

What comes next for AI security

Zscaler is preparing to unveil more innovations specifically designed to secure the "agentic enterprise." This refers to a business environment where autonomous AI agents perform many of the tasks traditionally handled by humans. The company plans to launch these new capabilities on December 9, 2026, at 1:00 PM ET.

The upcoming launch will showcase a unified Zero Trust Platform with new tools to protect AI agents and sensitive data. These tools are intended to work wherever the agents operate, whether they are on-premises or in the cloud. The goal is to provide a comprehensive way to manage the security of AI as it becomes more integrated into daily business operations.

As the technology evolves, the focus will likely shift toward managing the permissions of these autonomous agents. Organizations will need to decide which data these agents can see and which actions they can perform independently. Zscaler's partnerships with Anthropic and Mistral provide the foundational security layers needed to make that transition possible.

What to watch

  • How the Exploit Paths feature performs in real-world threat detection.
  • The adoption rate of sovereign AI solutions in the European market.
  • New security standards for autonomous AI agent permissions.

Try it

Look for Zscaler's upcoming launch on December 9, 2026, to see their new tools for protecting autonomous AI agents.

Sources and links

Read the original at mykxlg.com

First reported by mykxlg.com. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news