Solutions  ·  2026-09-22

Fastly launches AI Firewall, AI Runtime Control, and API Security for AI agents (GA)

SolutionsHigh impactGlobal
On Sept 21, 2026 Fastly announced three now-available capabilities on its edge platform: AI Firewall (edge prompt-injection/LLM-attack blocking evaluated in the request path), AI Runtime Control (single-endpoint model routing with virtual keys, token-spend visibility, rate limits, budget controls, failover), and expanded API Security that enforces API contracts against agentic/agent-assisted traffic. All ship on Fastly's network (~5T requests/day).
This is a major CDN/edge vendor bringing AI-application and agent security into production infrastructure — first-class GA prompt-injection blocking plus runtime governance of model spend, credentials, and agent-to-API actions, addressing the 'agents calling enterprise APIs' gap at scale with no added compute vendor lock-in.
Enterprises running production AI apps, model providers, coding agents, or agent-to-API integrations on Fastly should pilot AI Firewall + API Security in observe mode as part of AI runtime governance now; GA.
Fastly press releaseFastly investor releasecitybiz coverage
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →