Fortinet adds inline AI guardrails to FortiGate firewalls
retained text
Fortinet announced the FortiGuard Inline AI Security Service for FortiGate customers. The company says it inspects prompts, model responses, MCP tool calls and agent actions, and can block or redact sensitive data. Its FortiAIGate product page says the gateway can also set model permissions, budgets and rate limits. There are two deployment paths. The new service is cloud-delivered through existing FortiGate infrastructure, while FortiAIGate can be self-hosted on Kubernetes with NVIDIA GPUs for teams that need more control over where prompts, data and logs are processed. This is a concrete control point for AI traffic, but the launch material is still vendor evidence. It does not include an independent benchmark or customer performance result, so latency and operational cost are still open questions. Sources: Fortinet announcement: [https://investor.fortinet.com/news-releases/news-release-details/fortinet-expands-ai-security-fortiguard-inline-ai-security](https://investor.fortinet.com/news-releases/news-release-details/fortinet-expands-ai-security-fortiguard-inline-ai-security) FortiAIGate product page: [https://www.fortinet.com/products/fortiaigate](https://www.fortinet.com/products/fortiaigate)
document fetched 2026-10-27 00:24 · canonical https://reddit.com/r/cybersecurity/comments/1xg6ulf/fortinet_adds_inline_ai_guardrails_to_fortigate