Fortinet has brought its AI security gateway, FortiAIGate, to the Equinix Distributed AI Hub, giving enterprises a validated way to secure and govern artificial intelligence that no longer lives in one place. Announced on November 5, 2026, the move embeds security and policy enforcement directly into AI infrastructure, so organizations can scale AI across clouds, data centers and edge environments without losing control of their data or their costs.
Why Distributed AI Needs Security Built In
Enterprise AI is quickly moving beyond isolated applications. Models, agents, data and infrastructure now span clouds and geographies, and that distribution creates new challenges around governance, data protection, regional requirements and performance. Fortinet’s Anthony James, Executive Vice President of Marketing, put it plainly in the announcement: security needs to be built into that architecture from the start, not bolted on after AI is already in production. The Equinix Distributed AI Hub connects AI agents, models, data, compute and cloud providers through private, high-bandwidth, low-latency interconnection, and it is exactly this kind of connected environment that needs guardrails at the point where AI interactions happen. The same shift toward managed, always-on AI systems is visible across the industry, from managed agent runtimes to self-tuning inference platforms that route work to the right model automatically.
What FortiAIGate Does on the Equinix Distributed AI Hub
FortiAIGate is now available on the Equinix Distributed AI Hub as part of a new validated reference architecture. In this setup, FortiAIGate has been validated as an AI security guardrail, applying security closer to where AI workloads and data reside while maintaining consistent protection as deployments scale. It inspects AI traffic, enforces security policies, protects sensitive data and provides visibility into AI usage within the hub environment. Equinix’s Maryam Zand, Vice President of Partnerships and Ecosystem Development, said enterprises need to connect models, data and clouds while retaining control over their architecture and security choices, and the validated architecture gives customers a way to apply FortiAIGate security and governance to AI interactions across that connected environment. Equinix says the hub itself spans 282 AI data centers across 77 metros in 36 countries, putting AI infrastructure within 10 milliseconds of the majority of internet users worldwide.
Governing AI Interactions With Low-Latency Guardrails
The core of the announcement is governance. FortiAIGate’s low-latency guardrails inspect prompts and responses in both directions and apply centralized security policies across sanctioned large language model and MCP traffic, giving organizations full control without slowing teams down. That two-way inspection matters because risks flow in both directions: a prompt can carry sensitive data out, and a response can carry manipulated or unsafe content back in. Centralized policy also means the same rules follow AI interactions wherever they run, rather than being re-implemented for each model or cloud. This is the same defensive thinking behind open-source AI models built for defensive cybersecurity, which aim to give security teams tools that work with their existing stack rather than against it.
Protecting Sensitive Data and Supporting AI Sovereignty
Data protection is the second pillar. FortiAIGate keeps control, management and inspection within customer infrastructure, with no data leaving the environment, which supports data residency and region-specific governance requirements. That matters for regulated industries and governments that must prove where data is processed, not just promise it. The approach fits a broader pattern of sovereign and on-premises AI deployments, where organizations want the power of modern models without surrendering control of their data. Fortinet frames this as part of its AI-Native Security Fabric, which unifies AI security, secure networking and infrastructure protection, and the two companies also collaborate on network security through FortiGate VM on Equinix Network Edge.
Controlling AI Costs and Gaining Visibility Into Shadow AI
Security is not the only concern the integration addresses. FortiAIGate tracks per-token large language model costs and lets organizations set budgets with limits and alerts, giving them greater control over AI consumption as adoption grows. That cost visibility is increasingly important as AI coding agents and other autonomous tools multiply the number of model calls an organization makes every day. Visibility extends to unsanctioned use as well: FortiAIGate governs approved AI interactions, while FortiGate provides visibility into unsanctioned AI usage across the network, so security teams can see shadow AI instead of guessing at it. And because the inspection leverages GPUs, including from NVIDIA, Fortinet says the network layer can handle security inspection with zero performance impact on AI workloads.
What This Means for Enterprises Scaling AI
The Fortinet and Equinix announcement is less about a single product and more about a pattern: AI infrastructure is becoming distributed by default, and security has to be distributed with it. A validated reference architecture gives IT and security teams a tested starting point instead of a do-it-yourself integration, which shortens the path from AI pilot to production. For enterprises, the practical checklist is straightforward: inspect AI traffic in both directions, keep sensitive data inside your own infrastructure, track what AI actually costs, and watch for unsanctioned use. Organizations that build those controls in now will scale AI with far less rework than those that add them after a breach or a budget surprise forces the issue.
