Xage Fabric prevents unauthorized access and sensitive data exposure

Xage Security has released zero trust platform designed to secure AI environments. Built on the same proven zero trust principles Xage uses to protect critical infrastructure, the platform delivers control over AI data access, tool usage, and multi-agent workflows, eliminating jailbreak risks and ending AI adoption anxiety.

As the race to adopt AI continues, so too does the fear of unintended consequences, like rogue AI behavior and sensitive data leaks. Organizations want the competitive edge of AI, yet they need to be hypervigilant about protecting against the mounting risks of AI implementation. Today’s stopgap measures, such as existing LLM firewalls/guardrails and siloed LLM deployment, are costly, clumsy, and vulnerable to jailbreaks.

Xage’s dynamic zero trust approach brings previously unknown rigor and certainty to the security of AI implementations, empowering enterprises to unlock AI’s full potential safely.

“AI is being embraced at a pace that rivals the early days of internet adoption, only faster, deeper, and across every industry,” said Mark Gudiksen, Managing Partner at Piva Capital. “But with that momentum comes risk. We’ve already seen examples in the news of what happens when AI systems operate without the right safeguards. The long-term success of AI depends not just on innovation, but on the rigorous controls needed to govern it. Control isn’t optional—it’s the foundation for safe, scalable AI and the enabler for universal AI adoption.”

A new paradigm for securing AI

AI introduces constantly shifting, many-to-many connections between users, agents, LLMs, APIs, and data sources. Without tight, identity-first controls, this web of interactions can lead to unauthorized access, data exposure, and unmanageable risk.

Digital infrastructure grows more complex by the day, outpacing what conventional security can handle. Spanning compute, storage, networking, environmental controls, and hybrid or multi-cloud deployments, infrastructure needs embedded fortifications that guide AI agents and mitigate risks as they gain agency and enhanced permissions.

“Identity must be reimagined for AI. Agents should have cryptographically verifiable identities, scoped permissions, and clear delegation chains. They should be subject to the same principles of least privilege, credential rotation, and behavioral monitoring that govern human access. In short, it’s about knowing who (or what) is acting on your behalf, and ensuring they’re authorized to do so,” said Frank Dickson, Global VP of Security and Trust at IDC. “Applying zero trust principles to AI provides organizations with the ability to safeguard their AI initiatives while maintaining compliance and governance across complex, distributed environments.”

Xage’s identity-first zero trust architecture solves these security and trust challenges by enforcing real-time, context-aware controls across every layer, ensuring only the right people and applications have the right access, every time.

The Xage Fabric Platform delivers unified zero trust protection across the entire AI and data center stack—end-to-end, edge-to-core, and across any environment. Purpose-built for today’s most demanding environments, it offers:

• Full-stack security: Safeguards every layer, from physical infrastructure to digital workloads and sensitive data.
• Identity-centric defense: Granular identity verification protects sessions, tokens, and credentials, blocking lateral movement and limiting attack spread.
• Resilience by design: Delivers always-on, tamperproof and quantum-safe protection—even in air-gapped or sovereign deployments.
• Granular, jailbreak-proof data security: Data access control enforced at the network-level to block AI data leakage, leveraging Model Context Protocol (MCP).
• Secure MCP and A2A: Hardened, identity- and entitlement-aware MCP servers, MCP proxies, and AI-agent access shields to enforce zero trust for AI components and data.
• Rogue AI containment: Least-privilege restrictions rigorously enforced to prevent AI agents from carrying out harmful or unauthorized actions.
• Worry-free AI deployment: Organizations can design the AI workflows they want—for example, connecting an AI chatbot, such as Copilot or Claude, to their sensitive data—while knowing that the security risks are taken care of.

“Generative AI has opened incredible opportunities, but it also introduces threats that can’t be left to chance,” said Duncan Greatwood, CEO of Xage Security. “Too often, teams find themselves reacting to threats piecemeal instead of blocking them outright. The Xage Fabric Platform flips that script. We’re moving from an overreliance on prompt filters, which are vulnerable to jailbreaking, to true zero trust enforcement at the network protocol level. That means no more just hoping AI will behave as intended. Now, enterprises can be certain that they have unbreakable protection against internal or external data leakage and against the risk of rogue AI behavior.”